Correlation between aggregated molecular cancer subtypes and selected clinical features

Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma (Primary solid tumor)

17 October 2014 | analyses__2014_10_17

Maintainer Information

Citation Information

Maintained by TCGA GDAC Team (Broad Institute/MD Anderson Cancer Center/Harvard Medical School)

Cite as Broad Institute TCGA Genome Data Analysis Center (2014): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1W37V5D

Overview

Introduction

This pipeline computes the correlation between cancer subtypes identified by different molecular patterns and selected clinical features.

Summary

Testing the association between subtypes identified by 8 different clustering approaches and 45 clinical features across 250 patients, 16 significant findings detected with P value < 0.05 and Q value < 0.25.

3 subtypes identified in current cancer cohort by 'Copy Number Ratio CNMF subtypes'. These subtypes do not correlate to any clinical features.
6 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'HISTOLOGICAL.TYPE'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'HISTOLOGICAL.TYPE'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'HISTOLOGICAL.TYPE'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'HISTOLOGICAL.TYPE'.
5 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'AGE', 'HISTOLOGICAL.TYPE', 'AGE_AT_DIAGNOSIS', and 'STAGE_EVENT.CLINICAL_STAGE'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'HISTOLOGICAL.TYPE', 'RADIATIONS.RADIATION.REGIMENINDICATION', 'RADIATION_THERAPY_TYPE', and 'INITIAL_PATHOLOGIC_DX_YEAR'.
4 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'HISTOLOGICAL.TYPE', 'RADIATIONS.RADIATION.REGIMENINDICATION', 'RADIATION_THERAPY_TYPE', and 'INITIAL_PATHOLOGIC_DX_YEAR'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 8 different clustering approaches and 45 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 16 significant findings detected.


Clinical Features	Statistical Tests	Copy Number Ratio CNMF subtypes	METHLYATION CNMF	RNAseq CNMF subtypes	RNAseq cHierClus subtypes	MIRSEQ CNMF	MIRSEQ CHIERARCHICAL	MIRseq Mature CNMF subtypes	MIRseq Mature cHierClus subtypes
Time to Death	logrank test	0.364 (1.00)	0.0878 (1.00)	0.152 (1.00)	0.00602 (1.00)	0.436 (1.00)	0.519 (1.00)	0.765 (1.00)	0.981 (1.00)
AGE	Kruskal-Wallis (anova)	0.032 (1.00)	0.00769 (1.00)	0.0018 (0.616)	0.0161 (1.00)	0.171 (1.00)	8.5e-06 (0.00304)	0.415 (1.00)	0.546 (1.00)
PATHOLOGY T STAGE	Fisher's exact test	0.166 (1.00)	0.995 (1.00)	0.777 (1.00)	0.946 (1.00)	0.932 (1.00)	0.00149 (0.511)	0.892 (1.00)	0.825 (1.00)
PATHOLOGY N STAGE	Fisher's exact test	0.246 (1.00)	0.415 (1.00)	0.0774 (1.00)	0.387 (1.00)	0.832 (1.00)	0.354 (1.00)	0.339 (1.00)	0.102 (1.00)
PATHOLOGY M STAGE	Fisher's exact test	0.0537 (1.00)	0.571 (1.00)	0.197 (1.00)	0.0499 (1.00)	0.267 (1.00)	0.00764 (1.00)	0.124 (1.00)	0.22 (1.00)
HISTOLOGICAL TYPE	Fisher's exact test	0.16 (1.00)	1e-05 (0.00357)	1e-05 (0.00357)	1e-05 (0.00357)	1e-05 (0.00357)	1e-05 (0.00357)	8e-05 (0.0277)	1e-05 (0.00357)
RADIATIONS RADIATION REGIMENINDICATION	Fisher's exact test	0.878 (1.00)	0.171 (1.00)	0.438 (1.00)	0.255 (1.00)	0.251 (1.00)	0.0434 (1.00)	1e-05 (0.00357)	1e-05 (0.00357)
NUMBERPACKYEARSSMOKED	Kruskal-Wallis (anova)	0.536 (1.00)	0.635 (1.00)	0.461 (1.00)	0.628 (1.00)	0.147 (1.00)	0.324 (1.00)	0.408 (1.00)	0.238 (1.00)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.248 (1.00)	0.601 (1.00)	0.0389 (1.00)	0.0689 (1.00)	0.397 (1.00)	0.228 (1.00)	0.446 (1.00)	0.118 (1.00)
RACE	Fisher's exact test	0.0975 (1.00)	0.32 (1.00)	0.153 (1.00)	0.469 (1.00)	0.339 (1.00)	0.59 (1.00)	0.099 (1.00)	0.499 (1.00)
ETHNICITY	Fisher's exact test	0.839 (1.00)	0.31 (1.00)	0.496 (1.00)	0.518 (1.00)	0.26 (1.00)	0.178 (1.00)	0.0475 (1.00)	0.294 (1.00)
WEIGHT KG AT DIAGNOSIS	Kruskal-Wallis (anova)	0.462 (1.00)	0.799 (1.00)	0.332 (1.00)	0.481 (1.00)	0.476 (1.00)	0.765 (1.00)	0.889 (1.00)	0.679 (1.00)
TUMOR STATUS	Fisher's exact test	0.609 (1.00)	0.84 (1.00)	0.74 (1.00)	0.606 (1.00)	0.362 (1.00)	0.88 (1.00)	0.566 (1.00)	0.399 (1.00)
TUMOR SAMPLE PROCUREMENT COUNTRY	Fisher's exact test	0.381 (1.00)	0.497 (1.00)	0.766 (1.00)	0.453 (1.00)	0.0295 (1.00)	0.239 (1.00)	0.0764 (1.00)	0.0207 (1.00)
NEOPLASMHISTOLOGICGRADE	Fisher's exact test	0.123 (1.00)	0.839 (1.00)	0.00964 (1.00)	0.0353 (1.00)	0.0292 (1.00)	0.0646 (1.00)	0.0342 (1.00)	0.0459 (1.00)
TOBACCO SMOKING YEAR STOPPED	Kruskal-Wallis (anova)	0.866 (1.00)	0.801 (1.00)	0.975 (1.00)	0.478 (1.00)	0.12 (1.00)	0.0569 (1.00)	0.317 (1.00)	0.0617 (1.00)
TOBACCO SMOKING PACK YEARS SMOKED	Kruskal-Wallis (anova)	0.536 (1.00)	0.635 (1.00)	0.461 (1.00)	0.628 (1.00)	0.147 (1.00)	0.324 (1.00)	0.408 (1.00)	0.238 (1.00)
TOBACCO SMOKING HISTORY	Fisher's exact test	0.738 (1.00)	0.584 (1.00)	0.0998 (1.00)	0.355 (1.00)	0.238 (1.00)	0.49 (1.00)	0.124 (1.00)	0.459 (1.00)
PATIENT AGEBEGANSMOKINGINYEARS	Kruskal-Wallis (anova)	0.204 (1.00)	0.93 (1.00)	0.489 (1.00)	0.368 (1.00)	0.796 (1.00)	0.705 (1.00)	0.936 (1.00)	0.998 (1.00)
RADIATION TOTAL DOSE	Kruskal-Wallis (anova)	0.82 (1.00)	0.974 (1.00)	0.311 (1.00)	0.897 (1.00)	0.68 (1.00)	0.415 (1.00)	0.0236 (1.00)	0.297 (1.00)
RADIATION THERAPY TYPE	Fisher's exact test	0.929 (1.00)	0.193 (1.00)	0.556 (1.00)	0.672 (1.00)	0.12 (1.00)	0.243 (1.00)	2e-05 (0.00696)	1e-05 (0.00357)
RADIATION THERAPY STATUS	Fisher's exact test	0.262 (1.00)	0.886 (1.00)	0.454 (1.00)	0.658 (1.00)	0.396 (1.00)	0.86 (1.00)	1 (1.00)	0.458 (1.00)
RADIATION THERAPY SITE	Fisher's exact test	0.0858 (1.00)	0.0606 (1.00)	0.0669 (1.00)	0.117 (1.00)	0.0861 (1.00)	0.151 (1.00)	0.797 (1.00)	0.504 (1.00)
RADIATION ADJUVANT UNITS	Fisher's exact test	1 (1.00)	0.928 (1.00)	0.74 (1.00)	1 (1.00)	0.413 (1.00)	0.881 (1.00)	1 (1.00)	0.761 (1.00)
PREGNANCIES COUNT TOTAL	Kruskal-Wallis (anova)	0.479 (1.00)	0.291 (1.00)	0.0112 (1.00)	0.0973 (1.00)	0.461 (1.00)	0.106 (1.00)	0.436 (1.00)	0.115 (1.00)
PREGNANCIES COUNT STILLBIRTH	Kruskal-Wallis (anova)	0.246 (1.00)	0.00484 (1.00)	0.153 (1.00)	0.578 (1.00)	0.848 (1.00)	0.477 (1.00)	0.218 (1.00)	0.216 (1.00)
PATIENT PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT	Kruskal-Wallis (anova)	0.735 (1.00)	0.423 (1.00)	0.933 (1.00)	0.0932 (1.00)	0.00877 (1.00)	0.307 (1.00)	0.279 (1.00)	0.357 (1.00)
PREGNANCIES COUNT LIVE BIRTH	Kruskal-Wallis (anova)	0.619 (1.00)	0.114 (1.00)	0.0455 (1.00)	0.0515 (1.00)	0.326 (1.00)	0.026 (1.00)	0.341 (1.00)	0.036 (1.00)
PATIENT PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT	Kruskal-Wallis (anova)	0.896 (1.00)	0.151 (1.00)	0.849 (1.00)	0.801 (1.00)	0.396 (1.00)	0.915 (1.00)	0.555 (1.00)	0.64 (1.00)
PREGNANCIES COUNT ECTOPIC	Kruskal-Wallis (anova)	0.162 (1.00)	0.769 (1.00)	0.975 (1.00)	0.984 (1.00)	0.686 (1.00)	0.794 (1.00)	0.849 (1.00)	0.658 (1.00)
POS LYMPH NODE LOCATION	Fisher's exact test	0.862 (1.00)	0.888 (1.00)	0.988 (1.00)	0.993 (1.00)	0.809 (1.00)	0.899 (1.00)	0.646 (1.00)	0.7 (1.00)
MENOPAUSE STATUS	Fisher's exact test	0.0142 (1.00)	0.0477 (1.00)	0.0193 (1.00)	0.026 (1.00)	0.407 (1.00)	0.00307 (1.00)	0.414 (1.00)	0.396 (1.00)
LYMPHOVASCULAR INVOLVEMENT	Fisher's exact test	0.164 (1.00)	0.592 (1.00)	0.482 (1.00)	0.596 (1.00)	0.666 (1.00)	0.445 (1.00)	0.715 (1.00)	0.0762 (1.00)
LYMPH NODES EXAMINED HE COUNT	Kruskal-Wallis (anova)	0.248 (1.00)	0.601 (1.00)	0.0389 (1.00)	0.0689 (1.00)	0.397 (1.00)	0.228 (1.00)	0.446 (1.00)	0.118 (1.00)
LYMPH NODES EXAMINED	Kruskal-Wallis (anova)	0.193 (1.00)	0.588 (1.00)	0.987 (1.00)	0.595 (1.00)	0.644 (1.00)	0.533 (1.00)	0.18 (1.00)	0.641 (1.00)
KERATINIZATION SQUAMOUS CELL	Fisher's exact test	0.838 (1.00)	0.00981 (1.00)	0.0619 (1.00)	0.0717 (1.00)	0.0298 (1.00)	0.0905 (1.00)	0.0458 (1.00)	0.0192 (1.00)
INITIAL PATHOLOGIC DX YEAR	Kruskal-Wallis (anova)	0.656 (1.00)	0.386 (1.00)	0.227 (1.00)	0.154 (1.00)	0.0327 (1.00)	0.127 (1.00)	2.69e-05 (0.00933)	1.22e-06 (0.000438)
HISTORY HORMONAL CONTRACEPTIVES USE	Fisher's exact test	0.617 (1.00)	0.759 (1.00)	0.0854 (1.00)	0.03 (1.00)	0.679 (1.00)	0.416 (1.00)	0.9 (1.00)	0.395 (1.00)
HEIGHT CM AT DIAGNOSIS	Kruskal-Wallis (anova)	0.0625 (1.00)	0.275 (1.00)	0.657 (1.00)	0.418 (1.00)	0.407 (1.00)	0.685 (1.00)	0.533 (1.00)	0.753 (1.00)
CORPUS INVOLVEMENT	Fisher's exact test	0.352 (1.00)	0.694 (1.00)	0.0605 (1.00)	0.274 (1.00)	0.803 (1.00)	0.318 (1.00)	0.755 (1.00)	0.486 (1.00)
CHEMO CONCURRENT TYPE	Fisher's exact test	0.538 (1.00)	0.21 (1.00)	1 (1.00)	0.256 (1.00)	1 (1.00)	0.442 (1.00)	0.206 (1.00)	0.14 (1.00)
CERVIX SUV RESULTS	Kruskal-Wallis (anova)	0.0166 (1.00)	0.0821 (1.00)	0.0298 (1.00)	0.0961 (1.00)	0.508 (1.00)	0.207 (1.00)	0.348 (1.00)	0.302 (1.00)
AJCC TUMOR PATHOLOGIC PT	Fisher's exact test	0.15 (1.00)	0.444 (1.00)	0.222 (1.00)	0.114 (1.00)	0.392 (1.00)	0.00142 (0.488)	0.235 (1.00)	0.829 (1.00)
AGE AT DIAGNOSIS	Kruskal-Wallis (anova)	0.0189 (1.00)	0.0102 (1.00)	0.00227 (0.773)	0.0125 (1.00)	0.21 (1.00)	6.47e-06 (0.00232)	0.37 (1.00)	0.611 (1.00)
STAGE EVENT CLINICAL STAGE	Fisher's exact test	0.0865 (1.00)	0.105 (1.00)	0.00183 (0.624)	0.273 (1.00)	0.0469 (1.00)	0.00057 (0.197)	0.00658 (1.00)	0.193 (1.00)

Clustering Approach #1: 'Copy Number Ratio CNMF subtypes'

Table S1. Description of clustering approach #1: 'Copy Number Ratio CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	105	57	76

'Copy Number Ratio CNMF subtypes' versus 'Time to Death'

P value = 0.364 (logrank test), Q value = 1

Table S2. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	233	50	0.0 - 195.8 (15.9)
subtype1	105	19	0.0 - 195.8 (14.5)
subtype2	54	16	0.1 - 147.3 (17.1)
subtype3	74	15	0.0 - 173.3 (17.8)

Figure S1. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.032 (Kruskal-Wallis (anova)), Q value = 1

Table S3. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	234	47.5 (13.5)
subtype1	105	46.5 (14.4)
subtype2	56	45.6 (13.5)
subtype3	73	50.2 (11.9)

Figure S2. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #2: 'AGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.166 (Fisher's exact test), Q value = 1

Table S4. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	121	50	9	2
subtype1	56	26	2	2
subtype2	23	14	3	0
subtype3	42	10	4	0

Figure S3. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.246 (Fisher's exact test), Q value = 1

Table S5. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	111	47
subtype1	58	18
subtype2	21	13
subtype3	32	16

Figure S4. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.0537 (Fisher's exact test), Q value = 1

Table S6. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	91	8	93
subtype1	36	4	50
subtype2	18	3	22
subtype3	37	1	21

Figure S5. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.16 (Fisher's exact test), Q value = 1

Table S7. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	5	197	5	22	3	6
subtype1	3	78	3	14	2	5
subtype2	2	49	1	4	1	0
subtype3	0	70	1	4	0	1

Figure S6. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'Copy Number Ratio CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.878 (Fisher's exact test), Q value = 1

Table S8. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	30	208
subtype1	12	93
subtype2	8	49
subtype3	10	66

Figure S7. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'Copy Number Ratio CNMF subtypes' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.536 (Kruskal-Wallis (anova)), Q value = 1

Table S9. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	73	19.0 (14.0)
subtype1	34	16.8 (11.5)
subtype2	16	21.0 (12.9)
subtype3	23	20.8 (17.7)

Figure S8. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'Copy Number Ratio CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.248 (Kruskal-Wallis (anova)), Q value = 1

Table S10. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	135	0.9 (2.1)
subtype1	67	1.0 (2.6)
subtype2	26	1.0 (1.8)
subtype3	42	0.8 (1.3)

Figure S9. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.0975 (Fisher's exact test), Q value = 1

Table S11. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	8	18	24	1	169
subtype1	6	7	9	1	74
subtype2	0	1	8	0	45
subtype3	2	10	7	0	50

Figure S10. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #10: 'RACE'

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

P value = 0.839 (Fisher's exact test), Q value = 1

Table S12. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	16	149
subtype1	8	67
subtype2	4	34
subtype3	4	48

Figure S11. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'ETHNICITY'

'Copy Number Ratio CNMF subtypes' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.462 (Kruskal-Wallis (anova)), Q value = 1

Table S13. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	216	75.1 (22.4)
subtype1	90	75.8 (19.0)
subtype2	54	71.8 (16.9)
subtype3	72	76.7 (29.1)

Figure S12. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'Copy Number Ratio CNMF subtypes' versus 'TUMOR_STATUS'

P value = 0.609 (Fisher's exact test), Q value = 1

Table S14. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	71	25
subtype1	32	9
subtype2	16	8
subtype3	23	8

Figure S13. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

'Copy Number Ratio CNMF subtypes' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.381 (Fisher's exact test), Q value = 1

Table S15. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	23	5	1	10	7	178	14
subtype1	7	3	0	3	4	83	5
subtype2	8	1	1	2	1	43	1
subtype3	8	1	0	5	2	52	8

Figure S14. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'Copy Number Ratio CNMF subtypes' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.123 (Fisher's exact test), Q value = 1

Table S16. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	17	106	97	1	14
subtype1	9	49	40	0	6
subtype2	2	17	31	1	4
subtype3	6	40	26	0	4

Figure S15. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'Copy Number Ratio CNMF subtypes' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.866 (Kruskal-Wallis (anova)), Q value = 1

Table S17. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	34	2000.4 (11.2)
subtype1	14	2000.6 (11.7)
subtype2	10	2001.2 (11.7)
subtype3	10	1999.4 (11.2)

Figure S16. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'Copy Number Ratio CNMF subtypes' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.536 (Kruskal-Wallis (anova)), Q value = 1

Table S18. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	73	19.0 (14.0)
subtype1	34	16.8 (11.5)
subtype2	16	21.0 (12.9)
subtype3	23	20.8 (17.7)

Figure S17. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'Copy Number Ratio CNMF subtypes' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.738 (Fisher's exact test), Q value = 1

Table S19. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	31	8	2	52	112
subtype1	13	3	0	24	56
subtype2	9	1	1	11	21
subtype3	9	4	1	17	35

Figure S18. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'Copy Number Ratio CNMF subtypes' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.204 (Kruskal-Wallis (anova)), Q value = 1

Table S20. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	67	21.5 (7.8)
subtype1	30	20.1 (6.2)
subtype2	16	20.7 (9.0)
subtype3	21	24.2 (8.7)

Figure S19. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_TOTAL_DOSE'

P value = 0.82 (Kruskal-Wallis (anova)), Q value = 1

Table S21. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	103	3861.7 (1648.6)
subtype1	41	3987.6 (1645.6)
subtype2	28	3543.8 (1858.6)
subtype3	34	3971.7 (1474.4)

Figure S20. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY_TYPE'

P value = 0.929 (Fisher's exact test), Q value = 1

Table S22. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	INTERNAL
ALL	19	75	12	12
subtype1	7	33	5	4
subtype2	4	19	4	4
subtype3	8	23	3	4

Figure S21. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY_STATUS'

P value = 0.262 (Fisher's exact test), Q value = 1

Table S23. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	25	3
subtype1	9	1
subtype2	6	2
subtype3	10	0

Figure S22. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY_SITE'

P value = 0.0858 (Fisher's exact test), Q value = 1

Table S24. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	28	16
subtype1	2	0	12	3
subtype2	0	2	9	5
subtype3	0	0	7	8

Figure S23. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_ADJUVANT_UNITS'

P value = 1 (Fisher's exact test), Q value = 1

Table S25. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	34	3
subtype1	11	1
subtype2	13	1
subtype3	10	1

Figure S24. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'Copy Number Ratio CNMF subtypes' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.479 (Kruskal-Wallis (anova)), Q value = 1

Table S26. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	207	3.5 (2.4)
subtype1	92	3.4 (2.1)
subtype2	49	3.4 (2.6)
subtype3	66	3.8 (2.7)

Figure S25. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'Copy Number Ratio CNMF subtypes' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.246 (Kruskal-Wallis (anova)), Q value = 1

Table S27. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	103	0.1 (0.4)
subtype1	52	0.1 (0.3)
subtype2	18	0.0 (0.0)
subtype3	33	0.1 (0.5)

Figure S26. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'Copy Number Ratio CNMF subtypes' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.735 (Kruskal-Wallis (anova)), Q value = 1

Table S28. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	128	0.5 (0.8)
subtype1	63	0.4 (0.8)
subtype2	24	0.5 (0.7)
subtype3	41	0.5 (1.0)

Figure S27. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'Copy Number Ratio CNMF subtypes' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.619 (Kruskal-Wallis (anova)), Q value = 1

Table S29. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	209	2.6 (1.9)
subtype1	94	2.5 (1.6)
subtype2	50	2.8 (2.3)
subtype3	65	2.7 (1.8)

Figure S28. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'Copy Number Ratio CNMF subtypes' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.896 (Kruskal-Wallis (anova)), Q value = 1

Table S30. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	113	0.8 (1.7)
subtype1	57	0.7 (1.1)
subtype2	22	0.7 (1.0)
subtype3	34	1.2 (2.6)

Figure S29. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'Copy Number Ratio CNMF subtypes' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.162 (Kruskal-Wallis (anova)), Q value = 1

Table S31. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	105	0.1 (0.3)
subtype1	51	0.0 (0.2)
subtype2	19	0.2 (0.4)
subtype3	35	0.2 (0.5)

Figure S30. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'Copy Number Ratio CNMF subtypes' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.862 (Fisher's exact test), Q value = 1

Table S32. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	2	6	29	1	7
subtype1	1	1	12	1	3
subtype2	0	3	7	0	1
subtype3	1	2	10	0	3

Figure S31. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'Copy Number Ratio CNMF subtypes' versus 'MENOPAUSE_STATUS'

P value = 0.0142 (Fisher's exact test), Q value = 1

Table S33. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	16	68	105
subtype1	0	4	29	53
subtype2	0	2	12	27
subtype3	2	10	27	25

Figure S32. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'Copy Number Ratio CNMF subtypes' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.164 (Fisher's exact test), Q value = 1

Table S34. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	68	70
subtype1	37	30
subtype2	9	18
subtype3	22	22

Figure S33. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'Copy Number Ratio CNMF subtypes' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.248 (Kruskal-Wallis (anova)), Q value = 1

Table S35. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	135	0.9 (2.1)
subtype1	67	1.0 (2.6)
subtype2	26	1.0 (1.8)
subtype3	42	0.8 (1.3)

Figure S34. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'Copy Number Ratio CNMF subtypes' versus 'LYMPH_NODES_EXAMINED'

P value = 0.193 (Kruskal-Wallis (anova)), Q value = 1

Table S36. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	156	21.7 (12.6)
subtype1	75	20.9 (12.4)
subtype2	33	25.1 (12.7)
subtype3	48	20.7 (12.9)

Figure S35. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'Copy Number Ratio CNMF subtypes' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.838 (Fisher's exact test), Q value = 1

Table S37. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	43	94
subtype1	19	42
subtype2	8	21
subtype3	16	31

Figure S36. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'Copy Number Ratio CNMF subtypes' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 0.656 (Kruskal-Wallis (anova)), Q value = 1

Table S38. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	236	2007.9 (4.9)
subtype1	103	2008.1 (5.0)
subtype2	57	2007.8 (4.6)
subtype3	76	2007.7 (5.0)

Figure S37. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'Copy Number Ratio CNMF subtypes' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.617 (Fisher's exact test), Q value = 1

Table S39. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	49	57
subtype1	5	23	23
subtype2	3	12	11
subtype3	2	14	23

Figure S38. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'Copy Number Ratio CNMF subtypes' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.0625 (Kruskal-Wallis (anova)), Q value = 1

Table S40. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	207	161.8 (6.5)
subtype1	89	162.3 (7.4)
subtype2	49	163.0 (6.3)
subtype3	69	160.4 (5.3)

Figure S39. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'Copy Number Ratio CNMF subtypes' versus 'CORPUS_INVOLVEMENT'

P value = 0.352 (Fisher's exact test), Q value = 1

Table S41. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	89	16
subtype1	44	6
subtype2	14	5
subtype3	31	5

Figure S40. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'Copy Number Ratio CNMF subtypes' versus 'CHEMO_CONCURRENT_TYPE'

P value = 0.538 (Fisher's exact test), Q value = 1

Table S42. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	1	15	1
subtype1	1	4	1
subtype2	0	7	0
subtype3	0	4	0

Figure S41. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'Copy Number Ratio CNMF subtypes' versus 'CERVIX_SUV_RESULTS'

P value = 0.0166 (Kruskal-Wallis (anova)), Q value = 1

Table S43. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	13	12.4 (5.5)
subtype1	7	9.1 (2.9)
subtype2	3	20.1 (5.2)
subtype3	3	12.5 (1.7)

Figure S42. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'Copy Number Ratio CNMF subtypes' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.15 (Fisher's exact test), Q value = 1

Table S44. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1A1	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3	T3B	T4	TIS	TX
ALL	1	1	30	65	24	4	8	7	9	22	1	8	2	1	14
subtype1	0	1	13	29	13	2	5	3	7	9	1	1	2	0	6
subtype2	0	0	7	13	3	0	0	4	2	8	0	3	0	1	2
subtype3	1	0	10	23	8	2	3	0	0	5	0	4	0	0	6

Figure S43. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'Copy Number Ratio CNMF subtypes' versus 'AGE_AT_DIAGNOSIS'

P value = 0.0189 (Kruskal-Wallis (anova)), Q value = 1

Table S45. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	238	47.6 (13.5)
subtype1	105	46.5 (14.4)
subtype2	57	45.7 (13.3)
subtype3	76	50.5 (11.7)

Figure S44. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'Copy Number Ratio CNMF subtypes' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.0865 (Fisher's exact test), Q value = 1

Table S46. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA1	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE III	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	1	31	69	32	3	6	5	7	27	1	34	3	7
subtype1	1	0	1	0	12	35	15	0	4	2	5	9	1	14	1	5
subtype2	1	0	0	1	7	13	7	0	2	3	2	12	0	8	0	0
subtype3	2	2	0	0	12	21	10	3	0	0	0	6	0	12	2	2

Figure S45. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #2: 'METHLYATION CNMF'

Table S47. Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3	4	5	6
Number of samples	56	41	44	51	37	14

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.0878 (logrank test), Q value = 1

Table S48. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	238	53	0.0 - 195.8 (15.9)
subtype1	55	12	0.1 - 137.2 (15.0)
subtype2	41	6	0.1 - 195.8 (15.6)
subtype3	44	10	0.0 - 182.9 (19.7)
subtype4	49	19	0.0 - 154.3 (17.2)
subtype5	35	5	0.0 - 134.3 (20.9)
subtype6	14	1	0.1 - 53.2 (5.5)

Figure S46. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'AGE'

P value = 0.00769 (Kruskal-Wallis (anova)), Q value = 1

Table S49. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	239	47.4 (13.6)
subtype1	54	45.3 (11.6)
subtype2	40	50.5 (14.3)
subtype3	44	46.7 (13.9)
subtype4	50	42.6 (12.9)
subtype5	37	53.0 (13.1)
subtype6	14	50.9 (14.8)

Figure S47. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

'METHLYATION CNMF' versus 'PATHOLOGY.T.STAGE'

P value = 0.995 (Fisher's exact test), Q value = 1

Table S50. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	125	48	8	4
subtype1	33	13	1	1
subtype2	20	7	2	1
subtype3	24	10	1	1
subtype4	21	10	2	1
subtype5	20	5	2	0
subtype6	7	3	0	0

Figure S48. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY.N.STAGE'

P value = 0.415 (Fisher's exact test), Q value = 1

Table S51. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	115	50
subtype1	30	13
subtype2	15	12
subtype3	22	11
subtype4	23	5
subtype5	18	6
subtype6	7	3

Figure S49. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY.M.STAGE'

P value = 0.571 (Fisher's exact test), Q value = 1

Table S52. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	99	8	89
subtype1	23	5	21
subtype2	17	0	16
subtype3	21	1	16
subtype4	19	2	16
subtype5	16	0	13
subtype6	3	0	7

Figure S50. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'METHLYATION CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S53. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	4	204	5	22	2	6
subtype1	4	20	4	20	2	6
subtype2	0	39	0	2	0	0
subtype3	0	44	0	0	0	0
subtype4	0	51	0	0	0	0
subtype5	0	37	0	0	0	0
subtype6	0	13	1	0	0	0

Figure S51. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'METHLYATION CNMF' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.171 (Fisher's exact test), Q value = 1

Table S54. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	35	208
subtype1	4	52
subtype2	3	38
subtype3	8	36
subtype4	11	40
subtype5	7	30
subtype6	2	12

Figure S52. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'METHLYATION CNMF' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.635 (Kruskal-Wallis (anova)), Q value = 1

Table S55. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	75	18.4 (13.4)
subtype1	15	15.1 (11.4)
subtype2	14	20.2 (16.7)
subtype3	14	15.1 (8.0)
subtype4	15	17.8 (14.0)
subtype5	13	22.5 (12.8)
subtype6	4	24.5 (22.3)

Figure S53. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'METHLYATION CNMF' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.601 (Kruskal-Wallis (anova)), Q value = 1

Table S56. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	142	1.0 (2.4)
subtype1	37	1.1 (2.9)
subtype2	23	1.3 (2.0)
subtype3	24	1.2 (2.2)
subtype4	24	0.6 (1.4)
subtype5	25	1.2 (3.3)
subtype6	9	0.6 (0.9)

Figure S54. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'METHLYATION CNMF' versus 'RACE'

P value = 0.32 (Fisher's exact test), Q value = 1

Table S57. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	8	19	26	1	172
subtype1	1	4	3	0	43
subtype2	1	4	4	0	29
subtype3	1	6	6	0	28
subtype4	4	3	8	1	34
subtype5	1	0	2	0	29
subtype6	0	2	3	0	9

Figure S55. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.31 (Fisher's exact test), Q value = 1

Table S58. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	155
subtype1	6	35
subtype2	4	25
subtype3	1	33
subtype4	5	29
subtype5	1	22
subtype6	0	11

Figure S56. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'ETHNICITY'

'METHLYATION CNMF' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.799 (Kruskal-Wallis (anova)), Q value = 1

Table S59. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	221	75.0 (22.6)
subtype1	50	75.6 (15.1)
subtype2	41	75.4 (23.1)
subtype3	38	71.0 (21.0)
subtype4	44	74.9 (20.5)
subtype5	35	76.8 (29.4)
subtype6	13	77.9 (35.4)

Figure S57. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'METHLYATION CNMF' versus 'TUMOR_STATUS'

P value = 0.84 (Fisher's exact test), Q value = 1

Table S60. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	75	28
subtype1	18	7
subtype2	10	5
subtype3	18	4
subtype4	15	8
subtype5	9	3
subtype6	5	1

Figure S58. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'TUMOR_STATUS'

'METHLYATION CNMF' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.497 (Fisher's exact test), Q value = 1

Table S61. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	20	5	1	11	7	185	14
subtype1	4	3	1	2	1	43	2
subtype2	4	0	0	2	2	30	3
subtype3	3	1	0	3	0	33	4
subtype4	7	0	0	0	3	38	3
subtype5	2	1	0	4	1	29	0
subtype6	0	0	0	0	0	12	2

Figure S59. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'METHLYATION CNMF' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.839 (Fisher's exact test), Q value = 1

Table S62. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	15	112	98	1	14
subtype1	6	24	23	0	3
subtype2	1	21	18	0	1
subtype3	2	22	17	0	3
subtype4	1	22	22	1	3
subtype5	4	18	11	0	3
subtype6	1	5	7	0	1

Figure S60. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'METHLYATION CNMF' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.801 (Kruskal-Wallis (anova)), Q value = 1

Table S63. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	34	2000.4 (11.2)
subtype1	6	1999.0 (14.1)
subtype2	8	1998.8 (14.6)
subtype3	6	1999.8 (5.8)
subtype4	4	2002.0 (5.6)
subtype5	6	1999.2 (14.0)
subtype6	4	2007.0 (8.3)

Figure S61. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'METHLYATION CNMF' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.635 (Kruskal-Wallis (anova)), Q value = 1

Table S64. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	75	18.4 (13.4)
subtype1	15	15.1 (11.4)
subtype2	14	20.2 (16.7)
subtype3	14	15.1 (8.0)
subtype4	15	17.8 (14.0)
subtype5	13	22.5 (12.8)
subtype6	4	24.5 (22.3)

Figure S62. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'METHLYATION CNMF' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.584 (Fisher's exact test), Q value = 1

Table S65. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	33	8	3	53	108
subtype1	6	1	0	10	31
subtype2	7	2	1	8	18
subtype3	6	2	0	9	22
subtype4	5	0	1	14	21
subtype5	5	3	1	9	11
subtype6	4	0	0	3	5

Figure S63. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'METHLYATION CNMF' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.93 (Kruskal-Wallis (anova)), Q value = 1

Table S66. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	67	21.5 (7.9)
subtype1	14	19.3 (4.4)
subtype2	15	21.9 (10.1)
subtype3	11	20.1 (5.8)
subtype4	13	23.1 (9.5)
subtype5	10	23.5 (8.8)
subtype6	4	21.0 (6.2)

Figure S64. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'METHLYATION CNMF' versus 'RADIATION_TOTAL_DOSE'

P value = 0.974 (Kruskal-Wallis (anova)), Q value = 1

Table S67. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	104	3933.6 (1618.2)
subtype1	22	3930.1 (1721.5)
subtype2	15	3355.2 (2028.8)
subtype3	18	4321.4 (1175.0)
subtype4	28	3842.7 (1787.4)
subtype5	16	4331.2 (851.0)
subtype6	5	3525.0 (2117.0)

Figure S65. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'METHLYATION CNMF' versus 'RADIATION_THERAPY_TYPE'

P value = 0.193 (Fisher's exact test), Q value = 1

Table S68. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	IMPLANTS	INTERNAL
ALL	20	72	15	1	11
subtype1	3	20	1	0	1
subtype2	3	11	0	1	2
subtype3	5	13	3	0	0
subtype4	6	16	5	0	4
subtype5	2	8	5	0	4
subtype6	1	4	1	0	0

Figure S66. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'METHLYATION CNMF' versus 'RADIATION_THERAPY_STATUS'

P value = 0.886 (Fisher's exact test), Q value = 1

Table S69. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	29	3
subtype1	4	0
subtype2	2	0
subtype3	7	1
subtype4	8	2
subtype5	7	0
subtype6	1	0

Figure S67. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'METHLYATION CNMF' versus 'RADIATION_THERAPY_SITE'

P value = 0.0606 (Fisher's exact test), Q value = 1

Table S70. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	28	13
subtype1	1	0	3	1
subtype2	0	0	1	5
subtype3	0	1	7	0
subtype4	1	1	9	4
subtype5	0	0	7	3
subtype6	0	0	1	0

Figure S68. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'METHLYATION CNMF' versus 'RADIATION_ADJUVANT_UNITS'

P value = 0.928 (Fisher's exact test), Q value = 1

Table S71. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	31	3
subtype1	5	0
subtype2	4	0
subtype3	5	1
subtype4	10	2
subtype5	6	0
subtype6	1	0

Figure S69. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'METHLYATION CNMF' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.291 (Kruskal-Wallis (anova)), Q value = 1

Table S72. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	212	3.6 (2.5)
subtype1	51	2.9 (2.1)
subtype2	35	3.8 (2.5)
subtype3	39	3.8 (2.5)
subtype4	43	3.8 (2.9)
subtype5	33	3.9 (2.5)
subtype6	11	2.8 (2.1)

Figure S70. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'METHLYATION CNMF' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.00484 (Kruskal-Wallis (anova)), Q value = 1

Table S73. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	107	0.1 (0.4)
subtype1	24	0.0 (0.2)
subtype2	21	0.0 (0.0)
subtype3	21	0.3 (0.7)
subtype4	21	0.0 (0.0)
subtype5	16	0.0 (0.0)
subtype6	4	0.0 (0.0)

Figure S71. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'METHLYATION CNMF' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.423 (Kruskal-Wallis (anova)), Q value = 1

Table S74. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	133	0.5 (0.8)
subtype1	26	0.5 (0.6)
subtype2	24	0.4 (1.2)
subtype3	29	0.5 (0.7)
subtype4	28	0.6 (1.0)
subtype5	21	0.3 (0.5)
subtype6	5	0.4 (0.5)

Figure S72. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'METHLYATION CNMF' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.114 (Kruskal-Wallis (anova)), Q value = 1

Table S75. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	214	2.6 (1.8)
subtype1	50	2.1 (1.8)
subtype2	36	3.3 (2.2)
subtype3	41	2.5 (1.6)
subtype4	42	2.8 (1.9)
subtype5	34	2.4 (1.3)
subtype6	11	2.5 (1.6)

Figure S73. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'METHLYATION CNMF' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.151 (Kruskal-Wallis (anova)), Q value = 1

Table S76. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	117	0.9 (1.8)
subtype1	26	0.7 (1.0)
subtype2	22	0.3 (0.8)
subtype3	23	1.1 (2.3)
subtype4	23	0.8 (1.6)
subtype5	18	1.6 (3.1)
subtype6	5	0.2 (0.4)

Figure S74. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'METHLYATION CNMF' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.769 (Kruskal-Wallis (anova)), Q value = 1

Table S77. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	109	0.1 (0.3)
subtype1	23	0.1 (0.3)
subtype2	22	0.0 (0.2)
subtype3	21	0.0 (0.2)
subtype4	21	0.1 (0.4)
subtype5	17	0.2 (0.5)
subtype6	5	0.2 (0.4)

Figure S75. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'METHLYATION CNMF' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.888 (Fisher's exact test), Q value = 1

Table S78. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	2	7	33	1	9
subtype1	1	1	7	0	2
subtype2	0	1	3	0	2
subtype3	0	0	8	0	2
subtype4	0	2	7	1	1
subtype5	1	3	5	0	2
subtype6	0	0	3	0	0

Figure S76. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'METHLYATION CNMF' versus 'MENOPAUSE_STATUS'

P value = 0.0477 (Fisher's exact test), Q value = 1

Table S79. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	15	68	105
subtype1	0	4	14	29
subtype2	1	3	15	14
subtype3	0	3	14	20
subtype4	0	2	4	26
subtype5	1	2	15	11
subtype6	0	1	6	5

Figure S77. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'METHLYATION CNMF' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.592 (Fisher's exact test), Q value = 1

Table S80. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	68	74
subtype1	20	20
subtype2	8	14
subtype3	13	13
subtype4	12	9
subtype5	13	12
subtype6	2	6

Figure S78. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'METHLYATION CNMF' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.601 (Kruskal-Wallis (anova)), Q value = 1

Table S81. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	142	1.0 (2.4)
subtype1	37	1.1 (2.9)
subtype2	23	1.3 (2.0)
subtype3	24	1.2 (2.2)
subtype4	24	0.6 (1.4)
subtype5	25	1.2 (3.3)
subtype6	9	0.6 (0.9)

Figure S79. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'METHLYATION CNMF' versus 'LYMPH_NODES_EXAMINED'

P value = 0.588 (Kruskal-Wallis (anova)), Q value = 1

Table S82. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	163	21.7 (12.3)
subtype1	42	20.8 (9.0)
subtype2	25	24.4 (15.8)
subtype3	29	23.8 (13.8)
subtype4	29	20.5 (11.7)
subtype5	27	21.0 (11.0)
subtype6	11	17.6 (15.5)

Figure S80. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'METHLYATION CNMF' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.00981 (Fisher's exact test), Q value = 1

Table S83. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	49	96
subtype1	1	15
subtype2	6	21
subtype3	10	22
subtype4	15	18
subtype5	15	13
subtype6	2	7

Figure S81. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'METHLYATION CNMF' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 0.386 (Kruskal-Wallis (anova)), Q value = 1

Table S84. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	241	2007.7 (5.0)
subtype1	55	2008.6 (4.1)
subtype2	41	2007.3 (5.8)
subtype3	44	2006.8 (5.4)
subtype4	51	2007.0 (5.3)
subtype5	36	2007.9 (5.0)
subtype6	14	2009.6 (3.5)

Figure S82. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'METHLYATION CNMF' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.759 (Fisher's exact test), Q value = 1

Table S85. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	47	58
subtype1	5	15	11
subtype2	1	8	11
subtype3	1	7	12
subtype4	2	10	10
subtype5	1	4	10
subtype6	0	3	4

Figure S83. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'METHLYATION CNMF' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.275 (Kruskal-Wallis (anova)), Q value = 1

Table S86. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	213	161.7 (7.0)
subtype1	49	162.9 (6.7)
subtype2	39	162.1 (6.8)
subtype3	38	159.3 (7.8)
subtype4	42	161.6 (6.8)
subtype5	32	161.6 (6.9)
subtype6	13	163.6 (6.1)

Figure S84. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'METHLYATION CNMF' versus 'CORPUS_INVOLVEMENT'

P value = 0.694 (Fisher's exact test), Q value = 1

Table S87. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	95	17
subtype1	29	5
subtype2	15	2
subtype3	15	4
subtype4	15	1
subtype5	16	5
subtype6	5	0

Figure S85. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'METHLYATION CNMF' versus 'CHEMO_CONCURRENT_TYPE'

P value = 0.21 (Fisher's exact test), Q value = 1

Table S88. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	3	18	1
subtype1	0	2	0
subtype2	1	1	0
subtype3	1	5	1
subtype4	0	8	0
subtype5	1	2	0
subtype6	0	0	0

Figure S86. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'METHLYATION CNMF' versus 'CERVIX_SUV_RESULTS'

P value = 0.0821 (Kruskal-Wallis (anova)), Q value = 1

Table S89. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	13	12.4 (5.5)
subtype1	1	7.1 (NA)
subtype2	1	11.1 (NA)
subtype3	3	8.4 (1.8)
subtype4	4	18.0 (6.0)
subtype5	3	10.8 (4.4)
subtype6	1	13.8 (NA)

Figure S87. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'METHLYATION CNMF' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.444 (Fisher's exact test), Q value = 1

Table S90. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1A1	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3	T3B	T4	TIS	TX
ALL	1	1	32	66	25	4	8	7	10	19	1	7	4	1	15
subtype1	0	0	3	23	7	0	2	3	3	5	0	1	1	0	1
subtype2	1	0	4	10	5	2	1	0	2	2	1	1	1	0	5
subtype3	0	0	8	14	2	2	1	2	1	4	0	1	1	0	2
subtype4	0	0	11	7	3	0	1	1	2	6	0	2	1	1	4
subtype5	0	1	6	8	5	0	2	1	1	1	0	2	0	0	2
subtype6	0	0	0	4	3	0	1	0	1	1	0	0	0	0	1

Figure S88. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'METHLYATION CNMF' versus 'AGE_AT_DIAGNOSIS'

P value = 0.0102 (Kruskal-Wallis (anova)), Q value = 1

Table S91. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	243	47.5 (13.5)
subtype1	56	45.5 (11.4)
subtype2	41	50.7 (14.2)
subtype3	44	46.7 (13.9)
subtype4	51	43.0 (13.1)
subtype5	37	53.0 (13.1)
subtype6	14	50.9 (14.8)

Figure S89. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'METHLYATION CNMF' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.105 (Fisher's exact test), Q value = 1

Table S92. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA1	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE III	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	1	35	68	34	4	7	5	7	23	1	33	5	7
subtype1	0	0	0	0	3	25	10	0	1	1	2	5	0	5	0	4
subtype2	1	1	0	0	6	8	5	2	0	0	1	2	1	7	2	1
subtype3	0	1	0	0	7	14	3	1	1	2	0	3	0	10	1	1
subtype4	1	0	0	0	13	9	5	0	4	1	1	8	0	6	1	1
subtype5	2	0	1	1	5	9	7	1	1	1	1	4	0	3	1	0
subtype6	0	0	0	0	1	3	4	0	0	0	2	1	0	2	0	0

Figure S90. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #3: 'RNAseq CNMF subtypes'

Table S93. Description of clustering approach #3: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	77	100	62

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.152 (logrank test), Q value = 1

Table S94. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	234	51	0.0 - 195.8 (15.9)
subtype1	75	14	0.1 - 147.4 (15.0)
subtype2	98	18	0.0 - 195.8 (17.3)
subtype3	61	19	0.0 - 154.3 (15.6)

Figure S91. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.0018 (Kruskal-Wallis (anova)), Q value = 0.62

Table S95. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	235	47.2 (13.4)
subtype1	75	46.3 (12.0)
subtype2	99	50.7 (14.0)
subtype3	61	42.5 (12.6)

Figure S92. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.777 (Fisher's exact test), Q value = 1

Table S96. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	123	47	8	3
subtype1	46	16	1	1
subtype2	48	21	5	2
subtype3	29	10	2	0

Figure S93. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.0774 (Fisher's exact test), Q value = 1

Table S97. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	112	49
subtype1	45	14
subtype2	40	27
subtype3	27	8

Figure S94. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.197 (Fisher's exact test), Q value = 1

Table S98. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	96	8	88
subtype1	35	5	25
subtype2	37	1	43
subtype3	24	2	20

Figure S95. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S99. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	4	200	5	22	2	6
subtype1	4	38	5	22	2	6
subtype2	0	100	0	0	0	0
subtype3	0	62	0	0	0	0

Figure S96. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'RNAseq CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.438 (Fisher's exact test), Q value = 1

Table S100. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	33	206
subtype1	8	69
subtype2	14	86
subtype3	11	51

Figure S97. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq CNMF subtypes' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.461 (Kruskal-Wallis (anova)), Q value = 1

Table S101. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	75	18.4 (13.4)
subtype1	21	15.9 (12.0)
subtype2	40	20.6 (15.1)
subtype3	14	15.7 (9.0)

Figure S98. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'RNAseq CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0389 (Kruskal-Wallis (anova)), Q value = 1

Table S102. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	139	1.0 (2.4)
subtype1	52	0.8 (2.5)
subtype2	60	1.2 (1.9)
subtype3	27	0.9 (3.1)

Figure S99. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.153 (Fisher's exact test), Q value = 1

Table S103. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	7	18	26	1	170
subtype1	1	7	4	0	59
subtype2	2	7	11	0	70
subtype3	4	4	11	1	41

Figure S100. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.496 (Fisher's exact test), Q value = 1

Table S104. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	151
subtype1	6	50
subtype2	5	63
subtype3	6	38

Figure S101. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'ETHNICITY'

'RNAseq CNMF subtypes' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.332 (Kruskal-Wallis (anova)), Q value = 1

Table S105. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	217	74.9 (22.2)
subtype1	71	77.3 (21.1)
subtype2	92	74.2 (25.2)
subtype3	54	72.9 (17.9)

Figure S102. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'RNAseq CNMF subtypes' versus 'TUMOR_STATUS'

P value = 0.74 (Fisher's exact test), Q value = 1

Table S106. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	74	26
subtype1	26	8
subtype2	29	9
subtype3	19	9

Figure S103. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

'RNAseq CNMF subtypes' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.766 (Fisher's exact test), Q value = 1

Table S107. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	20	5	1	11	7	181	14
subtype1	4	3	1	3	2	59	5
subtype2	10	2	0	7	3	72	6
subtype3	6	0	0	1	2	50	3

Figure S104. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'RNAseq CNMF subtypes' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.00964 (Fisher's exact test), Q value = 1

Table S108. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	15	110	96	1	14
subtype1	6	30	35	0	5
subtype2	7	60	29	0	4
subtype3	2	20	32	1	5

Figure S105. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'RNAseq CNMF subtypes' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.975 (Kruskal-Wallis (anova)), Q value = 1

Table S109. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	34	2000.4 (11.2)
subtype1	11	1999.7 (12.2)
subtype2	19	2000.5 (11.9)
subtype3	4	2002.0 (5.6)

Figure S106. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'RNAseq CNMF subtypes' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.461 (Kruskal-Wallis (anova)), Q value = 1

Table S110. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	75	18.4 (13.4)
subtype1	21	15.9 (12.0)
subtype2	40	20.6 (15.1)
subtype3	14	15.7 (9.0)

Figure S107. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'RNAseq CNMF subtypes' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.0998 (Fisher's exact test), Q value = 1

Table S111. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	33	8	3	52	106
subtype1	12	1	0	14	40
subtype2	16	7	2	25	37
subtype3	5	0	1	13	29

Figure S108. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'RNAseq CNMF subtypes' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.489 (Kruskal-Wallis (anova)), Q value = 1

Table S112. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	67	21.5 (7.9)
subtype1	19	19.3 (3.9)
subtype2	37	23.0 (8.9)
subtype3	11	20.3 (8.9)

Figure S109. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'RNAseq CNMF subtypes' versus 'RADIATION_TOTAL_DOSE'

P value = 0.311 (Kruskal-Wallis (anova)), Q value = 1

Table S113. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	102	3950.7 (1603.4)
subtype1	30	4026.2 (1702.4)
subtype2	39	3933.7 (1509.6)
subtype3	33	3902.2 (1665.4)

Figure S110. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY_TYPE'

P value = 0.556 (Fisher's exact test), Q value = 1

Table S114. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	INTERNAL
ALL	19	72	15	11
subtype1	6	24	2	1
subtype2	8	27	7	6
subtype3	5	21	6	4

Figure S111. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY_STATUS'

P value = 0.454 (Fisher's exact test), Q value = 1

Table S115. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	28	3
subtype1	8	0
subtype2	12	1
subtype3	8	2

Figure S112. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY_SITE'

P value = 0.0669 (Fisher's exact test), Q value = 1

Table S116. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	28	13
subtype1	1	0	4	1
subtype2	0	1	10	10
subtype3	1	1	14	2

Figure S113. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'RNAseq CNMF subtypes' versus 'RADIATION_ADJUVANT_UNITS'

P value = 0.74 (Fisher's exact test), Q value = 1

Table S117. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	31	3
subtype1	5	0
subtype2	15	1
subtype3	11	2

Figure S114. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'RNAseq CNMF subtypes' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.0112 (Kruskal-Wallis (anova)), Q value = 1

Table S118. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	208	3.5 (2.5)
subtype1	67	2.9 (1.9)
subtype2	89	4.0 (2.7)
subtype3	52	3.7 (2.6)

Figure S115. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'RNAseq CNMF subtypes' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.153 (Kruskal-Wallis (anova)), Q value = 1

Table S119. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	104	0.1 (0.4)
subtype1	33	0.0 (0.2)
subtype2	48	0.1 (0.5)
subtype3	23	0.0 (0.0)

Figure S116. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'RNAseq CNMF subtypes' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.933 (Kruskal-Wallis (anova)), Q value = 1

Table S120. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	130	0.5 (0.8)
subtype1	35	0.4 (0.6)
subtype2	62	0.5 (0.9)
subtype3	33	0.5 (0.9)

Figure S117. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'RNAseq CNMF subtypes' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.0455 (Kruskal-Wallis (anova)), Q value = 1

Table S121. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	210	2.6 (1.8)
subtype1	66	2.2 (1.7)
subtype2	91	2.8 (1.8)
subtype3	53	2.8 (2.0)

Figure S118. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'RNAseq CNMF subtypes' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.849 (Kruskal-Wallis (anova)), Q value = 1

Table S122. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	114	0.9 (1.9)
subtype1	36	0.6 (1.0)
subtype2	53	1.2 (2.5)
subtype3	25	0.6 (1.1)

Figure S119. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'RNAseq CNMF subtypes' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.975 (Kruskal-Wallis (anova)), Q value = 1

Table S123. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	106	0.1 (0.3)
subtype1	33	0.1 (0.3)
subtype2	50	0.1 (0.4)
subtype3	23	0.1 (0.3)

Figure S120. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'RNAseq CNMF subtypes' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.988 (Fisher's exact test), Q value = 1

Table S124. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	2	6	32	1	8
subtype1	1	2	10	0	3
subtype2	1	2	13	1	4
subtype3	0	2	9	0	1

Figure S121. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'RNAseq CNMF subtypes' versus 'MENOPAUSE_STATUS'

P value = 0.0193 (Fisher's exact test), Q value = 1

Table S125. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	15	65	105
subtype1	0	5	22	36
subtype2	2	8	36	37
subtype3	0	2	7	32

Figure S122. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'RNAseq CNMF subtypes' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.482 (Fisher's exact test), Q value = 1

Table S126. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	67	71
subtype1	25	29
subtype2	27	32
subtype3	15	10

Figure S123. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'RNAseq CNMF subtypes' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.0389 (Kruskal-Wallis (anova)), Q value = 1

Table S127. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	139	1.0 (2.4)
subtype1	52	0.8 (2.5)
subtype2	60	1.2 (1.9)
subtype3	27	0.9 (3.1)

Figure S124. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'RNAseq CNMF subtypes' versus 'LYMPH_NODES_EXAMINED'

P value = 0.987 (Kruskal-Wallis (anova)), Q value = 1

Table S128. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	159	21.8 (12.4)
subtype1	58	21.3 (11.0)
subtype2	67	22.7 (14.5)
subtype3	34	20.7 (10.0)

Figure S125. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'RNAseq CNMF subtypes' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.0619 (Fisher's exact test), Q value = 1

Table S129. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	48	94
subtype1	4	23
subtype2	28	44
subtype3	16	27

Figure S126. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'RNAseq CNMF subtypes' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 0.227 (Kruskal-Wallis (anova)), Q value = 1

Table S130. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	237	2007.7 (5.0)
subtype1	76	2008.6 (4.1)
subtype2	99	2007.5 (5.6)
subtype3	62	2007.1 (5.1)

Figure S127. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'RNAseq CNMF subtypes' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.0854 (Fisher's exact test), Q value = 1

Table S131. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	47	56
subtype1	6	20	15
subtype2	1	16	28
subtype3	3	11	13

Figure S128. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'RNAseq CNMF subtypes' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.657 (Kruskal-Wallis (anova)), Q value = 1

Table S132. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	209	161.8 (6.7)
subtype1	67	162.6 (6.6)
subtype2	90	161.4 (6.8)
subtype3	52	161.4 (6.5)

Figure S129. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'RNAseq CNMF subtypes' versus 'CORPUS_INVOLVEMENT'

P value = 0.0605 (Fisher's exact test), Q value = 1

Table S133. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	93	16
subtype1	39	7
subtype2	33	9
subtype3	21	0

Figure S130. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'RNAseq CNMF subtypes' versus 'CHEMO_CONCURRENT_TYPE'

P value = 1 (Fisher's exact test), Q value = 1

Table S134. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	2	17	1
subtype1	0	4	0
subtype2	1	6	0
subtype3	1	7	1

Figure S131. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'RNAseq CNMF subtypes' versus 'CERVIX_SUV_RESULTS'

P value = 0.0298 (Kruskal-Wallis (anova)), Q value = 1

Table S135. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	13	12.4 (5.5)
subtype1	2	10.8 (5.2)
subtype2	7	9.7 (2.9)
subtype3	4	18.0 (6.0)

Figure S132. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'RNAseq CNMF subtypes' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.222 (Fisher's exact test), Q value = 1

Table S136. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1A1	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3	T3B	T4	TIS	TX
ALL	1	1	32	64	25	4	8	7	10	18	1	7	3	1	15
subtype1	0	0	4	32	10	2	3	3	3	5	0	1	1	0	2
subtype2	1	1	17	20	9	1	4	2	5	9	1	4	2	0	7
subtype3	0	0	11	12	6	1	1	2	2	4	0	2	0	1	6

Figure S133. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'RNAseq CNMF subtypes' versus 'AGE_AT_DIAGNOSIS'

P value = 0.00227 (Kruskal-Wallis (anova)), Q value = 0.77

Table S137. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	239	47.3 (13.3)
subtype1	77	46.4 (11.9)
subtype2	100	50.8 (13.9)
subtype3	62	42.9 (12.7)

Figure S134. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'RNAseq CNMF subtypes' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.00183 (Fisher's exact test), Q value = 0.62

Table S138. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA1	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE III	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	1	34	67	34	4	6	5	7	23	1	33	4	7
subtype1	1	0	0	0	4	37	14	1	1	1	2	6	0	6	0	4
subtype2	2	2	1	1	15	19	10	3	2	2	4	9	1	19	4	2
subtype3	1	0	0	0	15	11	10	0	3	2	1	8	0	8	0	1

Figure S135. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #4: 'RNAseq cHierClus subtypes'

Table S139. Description of clustering approach #4: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	55	150	34

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.00602 (logrank test), Q value = 1

Table S140. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	234	51	0.0 - 195.8 (15.9)
subtype1	54	11	0.1 - 137.2 (14.9)
subtype2	147	28	0.0 - 195.8 (17.9)
subtype3	33	12	0.0 - 99.9 (11.4)

Figure S136. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.0161 (Kruskal-Wallis (anova)), Q value = 1

Table S141. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	235	47.2 (13.4)
subtype1	53	46.4 (11.3)
subtype2	148	48.9 (13.8)
subtype3	34	41.1 (12.8)

Figure S137. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.946 (Fisher's exact test), Q value = 1

Table S142. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	123	47	8	3
subtype1	32	13	1	1
subtype2	77	27	6	2
subtype3	14	7	1	0

Figure S138. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.387 (Fisher's exact test), Q value = 1

Table S143. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	112	49
subtype1	31	11
subtype2	65	34
subtype3	16	4

Figure S139. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.0499 (Fisher's exact test), Q value = 1

Table S144. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	96	8	88
subtype1	23	5	21
subtype2	61	1	56
subtype3	12	2	11

Figure S140. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S145. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	4	200	5	22	2	6
subtype1	3	19	4	22	2	5
subtype2	1	147	1	0	0	1
subtype3	0	34	0	0	0	0

Figure S141. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'RNAseq cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.255 (Fisher's exact test), Q value = 1

Table S146. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	33	206
subtype1	4	51
subtype2	24	126
subtype3	5	29

Figure S142. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq cHierClus subtypes' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.628 (Kruskal-Wallis (anova)), Q value = 1

Table S147. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	75	18.4 (13.4)
subtype1	14	15.8 (11.6)
subtype2	51	19.7 (14.3)
subtype3	10	15.3 (10.0)

Figure S143. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'RNAseq cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0689 (Kruskal-Wallis (anova)), Q value = 1

Table S148. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	139	1.0 (2.4)
subtype1	37	0.7 (2.4)
subtype2	86	1.3 (2.5)
subtype3	16	0.2 (0.6)

Figure S144. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.469 (Fisher's exact test), Q value = 1

Table S149. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	7	18	26	1	170
subtype1	1	5	3	0	41
subtype2	4	12	16	1	105
subtype3	2	1	7	0	24

Figure S145. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.518 (Fisher's exact test), Q value = 1

Table S150. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	151
subtype1	6	34
subtype2	9	97
subtype3	2	20

Figure S146. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'ETHNICITY'

'RNAseq cHierClus subtypes' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.481 (Kruskal-Wallis (anova)), Q value = 1

Table S151. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	217	74.9 (22.2)
subtype1	51	75.7 (15.4)
subtype2	138	75.2 (25.2)
subtype3	28	72.1 (16.9)

Figure S147. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'RNAseq cHierClus subtypes' versus 'TUMOR_STATUS'

P value = 0.606 (Fisher's exact test), Q value = 1

Table S152. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	74	26
subtype1	19	6
subtype2	49	16
subtype3	6	4

Figure S148. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

'RNAseq cHierClus subtypes' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.453 (Fisher's exact test), Q value = 1

Table S153. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	20	5	1	11	7	181	14
subtype1	4	3	1	2	2	40	3
subtype2	12	2	0	9	3	114	10
subtype3	4	0	0	0	2	27	1

Figure S149. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'RNAseq cHierClus subtypes' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.0353 (Fisher's exact test), Q value = 1

Table S154. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	15	110	96	1	14
subtype1	6	24	21	0	4
subtype2	8	77	54	0	8
subtype3	1	9	21	1	2

Figure S150. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'RNAseq cHierClus subtypes' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.478 (Kruskal-Wallis (anova)), Q value = 1

Table S155. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	34	2000.4 (11.2)
subtype1	6	1996.2 (12.6)
subtype2	25	2001.4 (11.5)
subtype3	3	2000.3 (5.5)

Figure S151. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'RNAseq cHierClus subtypes' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.628 (Kruskal-Wallis (anova)), Q value = 1

Table S156. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	75	18.4 (13.4)
subtype1	14	15.8 (11.6)
subtype2	51	19.7 (14.3)
subtype3	10	15.3 (10.0)

Figure S152. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'RNAseq cHierClus subtypes' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.355 (Fisher's exact test), Q value = 1

Table S157. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	33	8	3	52	106
subtype1	6	1	0	9	32
subtype2	24	7	2	34	60
subtype3	3	0	1	9	14

Figure S153. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'RNAseq cHierClus subtypes' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.368 (Kruskal-Wallis (anova)), Q value = 1

Table S158. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	67	21.5 (7.9)
subtype1	13	19.6 (4.5)
subtype2	44	22.4 (8.3)
subtype3	10	19.7 (9.2)

Figure S154. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'RNAseq cHierClus subtypes' versus 'RADIATION_TOTAL_DOSE'

P value = 0.897 (Kruskal-Wallis (anova)), Q value = 1

Table S159. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	102	3950.7 (1603.4)
subtype1	21	3877.2 (1745.6)
subtype2	63	4032.5 (1518.2)
subtype3	18	3750.5 (1791.5)

Figure S155. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY_TYPE'

P value = 0.672 (Fisher's exact test), Q value = 1

Table S160. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	INTERNAL
ALL	19	72	15	11
subtype1	3	17	1	1
subtype2	13	41	12	7
subtype3	3	14	2	3

Figure S156. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY_STATUS'

P value = 0.658 (Fisher's exact test), Q value = 1

Table S161. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	28	3
subtype1	4	0
subtype2	20	2
subtype3	4	1

Figure S157. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY_SITE'

P value = 0.117 (Fisher's exact test), Q value = 1

Table S162. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	28	13
subtype1	1	0	3	1
subtype2	0	1	16	11
subtype3	1	1	9	1

Figure S158. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'RNAseq cHierClus subtypes' versus 'RADIATION_ADJUVANT_UNITS'

P value = 1 (Fisher's exact test), Q value = 1

Table S163. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	31	3
subtype1	5	0
subtype2	18	2
subtype3	8	1

Figure S159. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'RNAseq cHierClus subtypes' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.0973 (Kruskal-Wallis (anova)), Q value = 1

Table S164. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	208	3.5 (2.5)
subtype1	48	2.9 (2.1)
subtype2	131	3.7 (2.6)
subtype3	29	3.8 (2.6)

Figure S160. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'RNAseq cHierClus subtypes' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.578 (Kruskal-Wallis (anova)), Q value = 1

Table S165. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	104	0.1 (0.4)
subtype1	23	0.0 (0.2)
subtype2	69	0.1 (0.4)
subtype3	12	0.0 (0.0)

Figure S161. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'RNAseq cHierClus subtypes' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.0932 (Kruskal-Wallis (anova)), Q value = 1

Table S166. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	130	0.5 (0.8)
subtype1	25	0.5 (0.6)
subtype2	86	0.4 (0.8)
subtype3	19	0.7 (1.0)

Figure S162. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'RNAseq cHierClus subtypes' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.0515 (Kruskal-Wallis (anova)), Q value = 1

Table S167. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	210	2.6 (1.8)
subtype1	48	2.1 (1.8)
subtype2	133	2.7 (1.8)
subtype3	29	2.9 (2.1)

Figure S163. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'RNAseq cHierClus subtypes' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.801 (Kruskal-Wallis (anova)), Q value = 1

Table S168. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	114	0.9 (1.9)
subtype1	25	0.7 (1.0)
subtype2	75	1.0 (2.2)
subtype3	14	0.7 (1.0)

Figure S164. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'RNAseq cHierClus subtypes' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.984 (Kruskal-Wallis (anova)), Q value = 1

Table S169. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	106	0.1 (0.3)
subtype1	22	0.1 (0.3)
subtype2	72	0.1 (0.4)
subtype3	12	0.1 (0.3)

Figure S165. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'RNAseq cHierClus subtypes' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.993 (Fisher's exact test), Q value = 1

Table S170. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	2	6	32	1	8
subtype1	1	1	7	0	2
subtype2	1	4	20	1	5
subtype3	0	1	5	0	1

Figure S166. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'RNAseq cHierClus subtypes' versus 'MENOPAUSE_STATUS'

P value = 0.026 (Fisher's exact test), Q value = 1

Table S171. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	15	65	105
subtype1	0	4	16	27
subtype2	2	10	48	60
subtype3	0	1	1	18

Figure S167. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'RNAseq cHierClus subtypes' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.596 (Fisher's exact test), Q value = 1

Table S172. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	67	71
subtype1	21	19
subtype2	37	45
subtype3	9	7

Figure S168. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'RNAseq cHierClus subtypes' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.0689 (Kruskal-Wallis (anova)), Q value = 1

Table S173. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	139	1.0 (2.4)
subtype1	37	0.7 (2.4)
subtype2	86	1.3 (2.5)
subtype3	16	0.2 (0.6)

Figure S169. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'RNAseq cHierClus subtypes' versus 'LYMPH_NODES_EXAMINED'

P value = 0.595 (Kruskal-Wallis (anova)), Q value = 1

Table S174. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	159	21.8 (12.4)
subtype1	41	20.3 (9.5)
subtype2	99	22.0 (13.5)
subtype3	19	23.8 (11.9)

Figure S170. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'RNAseq cHierClus subtypes' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.0717 (Fisher's exact test), Q value = 1

Table S175. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	48	94
subtype1	1	13
subtype2	40	67
subtype3	7	14

Figure S171. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'RNAseq cHierClus subtypes' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 0.154 (Kruskal-Wallis (anova)), Q value = 1

Table S176. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	237	2007.7 (5.0)
subtype1	54	2008.9 (3.8)
subtype2	149	2007.7 (5.1)
subtype3	34	2006.3 (6.0)

Figure S172. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'RNAseq cHierClus subtypes' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.03 (Fisher's exact test), Q value = 1

Table S177. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	47	56
subtype1	5	15	12
subtype2	2	29	36
subtype3	3	3	8

Figure S173. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'RNAseq cHierClus subtypes' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.418 (Kruskal-Wallis (anova)), Q value = 1

Table S178. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	209	161.8 (6.7)
subtype1	50	162.3 (6.6)
subtype2	133	161.3 (6.7)
subtype3	26	163.0 (6.5)

Figure S174. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'RNAseq cHierClus subtypes' versus 'CORPUS_INVOLVEMENT'

P value = 0.274 (Fisher's exact test), Q value = 1

Table S179. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	93	16
subtype1	31	4
subtype2	52	12
subtype3	10	0

Figure S175. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'RNAseq cHierClus subtypes' versus 'CHEMO_CONCURRENT_TYPE'

P value = 0.256 (Fisher's exact test), Q value = 1

Table S180. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	2	17	1
subtype1	0	2	0
subtype2	2	12	0
subtype3	0	3	1

Figure S176. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'RNAseq cHierClus subtypes' versus 'CERVIX_SUV_RESULTS'

P value = 0.0961 (Kruskal-Wallis (anova)), Q value = 1

Table S181. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	13	12.4 (5.5)
subtype1	1	7.1 (NA)
subtype2	9	11.0 (3.7)
subtype3	3	18.4 (7.3)

Figure S177. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'RNAseq cHierClus subtypes' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.114 (Fisher's exact test), Q value = 1

Table S182. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1A1	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3	T3B	T4	TIS	TX
ALL	1	1	32	64	25	4	8	7	10	18	1	7	3	1	15
subtype1	0	0	3	23	6	1	2	2	3	5	0	1	1	0	2
subtype2	1	1	20	39	16	3	5	2	6	11	1	5	2	0	10
subtype3	0	0	9	2	3	0	1	3	1	2	0	1	0	1	3

Figure S178. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'RNAseq cHierClus subtypes' versus 'AGE_AT_DIAGNOSIS'

P value = 0.0125 (Kruskal-Wallis (anova)), Q value = 1

Table S183. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	239	47.3 (13.3)
subtype1	55	46.6 (11.2)
subtype2	150	49.0 (13.8)
subtype3	34	41.1 (12.8)

Figure S179. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'RNAseq cHierClus subtypes' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.273 (Fisher's exact test), Q value = 1

Table S184. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA1	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE III	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	1	34	67	34	4	6	5	7	23	1	33	4	7
subtype1	1	0	0	0	3	24	8	1	1	1	2	5	0	5	0	4
subtype2	3	2	1	1	23	37	18	3	3	2	4	15	1	25	4	3
subtype3	0	0	0	0	8	6	8	0	2	2	1	3	0	3	0	0

Figure S180. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #5: 'MIRSEQ CNMF'

Table S185. Description of clustering approach #5: 'MIRSEQ CNMF'

Cluster Labels	1	2	3
Number of samples	79	68	103

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.436 (logrank test), Q value = 1

Table S186. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	245	54	0.0 - 195.8 (16.1)
subtype1	79	16	0.1 - 160.4 (15.6)
subtype2	65	11	0.0 - 182.9 (15.0)
subtype3	101	27	0.0 - 195.8 (17.6)

Figure S181. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'AGE'

P value = 0.171 (Kruskal-Wallis (anova)), Q value = 1

Table S187. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	246	47.6 (13.5)
subtype1	77	46.3 (11.6)
subtype2	68	50.0 (13.2)
subtype3	101	46.9 (14.9)

Figure S182. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

'MIRSEQ CNMF' versus 'PATHOLOGY.T.STAGE'

P value = 0.932 (Fisher's exact test), Q value = 1

Table S188. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	127	52	9	4
subtype1	45	20	4	1
subtype2	38	14	1	1
subtype3	44	18	4	2

Figure S183. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY.N.STAGE'

P value = 0.832 (Fisher's exact test), Q value = 1

Table S189. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	117	51
subtype1	42	17
subtype2	35	18
subtype3	40	16

Figure S184. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY.M.STAGE'

P value = 0.267 (Fisher's exact test), Q value = 1

Table S190. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	99	8	96
subtype1	34	6	33
subtype2	30	1	25
subtype3	35	1	38

Figure S185. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'MIRSEQ CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S191. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	5	209	5	22	3	6
subtype1	4	47	4	17	3	4
subtype2	1	60	1	5	0	1
subtype3	0	102	0	0	0	1

Figure S186. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'MIRSEQ CNMF' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.251 (Fisher's exact test), Q value = 1

Table S192. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	35	215
subtype1	8	71
subtype2	8	60
subtype3	19	84

Figure S187. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CNMF' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.147 (Kruskal-Wallis (anova)), Q value = 1

Table S193. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	76	18.8 (13.8)
subtype1	18	14.8 (10.4)
subtype2	24	23.2 (14.8)
subtype3	34	17.8 (14.2)

Figure S188. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'MIRSEQ CNMF' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.397 (Kruskal-Wallis (anova)), Q value = 1

Table S194. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	144	1.1 (2.5)
subtype1	51	0.8 (2.3)
subtype2	48	1.5 (2.6)
subtype3	45	0.9 (2.5)

Figure S189. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.339 (Fisher's exact test), Q value = 1

Table S195. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	8	19	26	1	178
subtype1	1	8	7	0	54
subtype2	1	7	6	0	48
subtype3	6	4	13	1	76

Figure S190. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #10: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.26 (Fisher's exact test), Q value = 1

Table S196. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	157
subtype1	7	49
subtype2	2	47
subtype3	8	61

Figure S191. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #11: 'ETHNICITY'

'MIRSEQ CNMF' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.476 (Kruskal-Wallis (anova)), Q value = 1

Table S197. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	228	74.8 (22.4)
subtype1	75	76.5 (20.2)
subtype2	59	75.4 (29.1)
subtype3	94	73.0 (19.1)

Figure S192. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'MIRSEQ CNMF' versus 'TUMOR_STATUS'

P value = 0.362 (Fisher's exact test), Q value = 1

Table S198. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	75	28
subtype1	26	9
subtype2	22	5
subtype3	27	14

Figure S193. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #13: 'TUMOR_STATUS'

'MIRSEQ CNMF' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.0295 (Fisher's exact test), Q value = 1

Table S199. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	24	5	1	11	7	188	14
subtype1	11	3	1	2	3	54	5
subtype2	1	2	0	5	2	52	6
subtype3	12	0	0	4	2	82	3

Figure S194. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'MIRSEQ CNMF' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.0292 (Fisher's exact test), Q value = 1

Table S200. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	17	113	102	1	14
subtype1	7	30	35	0	7
subtype2	8	34	23	1	1
subtype3	2	49	44	0	6

Figure S195. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'MIRSEQ CNMF' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.12 (Kruskal-Wallis (anova)), Q value = 1

Table S201. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	34	2000.4 (11.2)
subtype1	8	1999.9 (12.7)
subtype2	9	1995.1 (11.5)
subtype3	17	2003.5 (9.9)

Figure S196. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'MIRSEQ CNMF' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.147 (Kruskal-Wallis (anova)), Q value = 1

Table S202. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	76	18.8 (13.8)
subtype1	18	14.8 (10.4)
subtype2	24	23.2 (14.8)
subtype3	34	17.8 (14.2)

Figure S197. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'MIRSEQ CNMF' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.238 (Fisher's exact test), Q value = 1

Table S203. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	33	8	3	54	114
subtype1	9	1	0	13	46
subtype2	8	3	2	18	29
subtype3	16	4	1	23	39

Figure S198. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'MIRSEQ CNMF' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.796 (Kruskal-Wallis (anova)), Q value = 1

Table S204. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	68	21.4 (7.8)
subtype1	16	20.4 (7.1)
subtype2	21	21.5 (6.4)
subtype3	31	21.9 (9.1)

Figure S199. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'MIRSEQ CNMF' versus 'RADIATION_TOTAL_DOSE'

P value = 0.68 (Kruskal-Wallis (anova)), Q value = 1

Table S205. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	110	3890.9 (1636.4)
subtype1	36	3934.4 (1684.4)
subtype2	24	3955.9 (1661.7)
subtype3	50	3828.3 (1620.7)

Figure S200. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY_TYPE'

P value = 0.12 (Fisher's exact test), Q value = 1

Table S206. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	IMPLANTS	INTERNAL
ALL	20	77	15	1	12
subtype1	6	30	2	0	2
subtype2	2	18	6	1	3
subtype3	12	29	7	0	7

Figure S201. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY_STATUS'

P value = 0.396 (Fisher's exact test), Q value = 1

Table S207. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	29	3
subtype1	8	0
subtype2	6	0
subtype3	15	3

Figure S202. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY_SITE'

P value = 0.0861 (Fisher's exact test), Q value = 1

Table S208. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	30	16
subtype1	1	0	3	7
subtype2	0	1	8	3
subtype3	1	1	19	6

Figure S203. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'MIRSEQ CNMF' versus 'RADIATION_ADJUVANT_UNITS'

P value = 0.413 (Fisher's exact test), Q value = 1

Table S209. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	36	3
subtype1	11	0
subtype2	8	0
subtype3	17	3

Figure S204. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'MIRSEQ CNMF' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.461 (Kruskal-Wallis (anova)), Q value = 1

Table S210. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	219	3.6 (2.5)
subtype1	69	3.4 (2.7)
subtype2	62	3.6 (2.4)
subtype3	88	3.6 (2.5)

Figure S205. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'MIRSEQ CNMF' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.848 (Kruskal-Wallis (anova)), Q value = 1

Table S211. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	109	0.1 (0.4)
subtype1	29	0.0 (0.2)
subtype2	32	0.1 (0.6)
subtype3	48	0.1 (0.2)

Figure S206. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'MIRSEQ CNMF' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.00877 (Kruskal-Wallis (anova)), Q value = 1

Table S212. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	135	0.5 (0.8)
subtype1	36	0.8 (1.1)
subtype2	40	0.3 (0.5)
subtype3	59	0.4 (0.8)

Figure S207. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'MIRSEQ CNMF' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.326 (Kruskal-Wallis (anova)), Q value = 1

Table S213. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	220	2.6 (1.9)
subtype1	67	2.6 (2.3)
subtype2	62	2.7 (1.7)
subtype3	91	2.7 (1.7)

Figure S208. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'MIRSEQ CNMF' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.396 (Kruskal-Wallis (anova)), Q value = 1

Table S214. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	119	0.9 (1.8)
subtype1	32	0.8 (1.5)
subtype2	36	1.1 (2.3)
subtype3	51	0.7 (1.6)

Figure S209. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'MIRSEQ CNMF' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.686 (Kruskal-Wallis (anova)), Q value = 1

Table S215. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	111	0.1 (0.3)
subtype1	29	0.1 (0.3)
subtype2	32	0.2 (0.4)
subtype3	50	0.1 (0.3)

Figure S210. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'MIRSEQ CNMF' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.809 (Fisher's exact test), Q value = 1

Table S216. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	2	7	33	1	9
subtype1	1	2	8	0	2
subtype2	1	2	11	1	5
subtype3	0	3	14	0	2

Figure S211. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'MIRSEQ CNMF' versus 'MENOPAUSE_STATUS'

P value = 0.407 (Fisher's exact test), Q value = 1

Table S217. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	16	70	109
subtype1	1	8	19	37
subtype2	1	3	23	26
subtype3	0	5	28	46

Figure S212. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'MIRSEQ CNMF' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.666 (Fisher's exact test), Q value = 1

Table S218. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	69	76
subtype1	24	29
subtype2	21	26
subtype3	24	21

Figure S213. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'MIRSEQ CNMF' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.397 (Kruskal-Wallis (anova)), Q value = 1

Table S219. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	144	1.1 (2.5)
subtype1	51	0.8 (2.3)
subtype2	48	1.5 (2.6)
subtype3	45	0.9 (2.5)

Figure S214. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'MIRSEQ CNMF' versus 'LYMPH_NODES_EXAMINED'

P value = 0.644 (Kruskal-Wallis (anova)), Q value = 1

Table S220. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	166	21.7 (12.4)
subtype1	56	20.3 (10.8)
subtype2	55	24.0 (14.4)
subtype3	55	21.0 (11.6)

Figure S215. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'MIRSEQ CNMF' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.0298 (Fisher's exact test), Q value = 1

Table S221. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	49	97
subtype1	4	25
subtype2	18	25
subtype3	27	47

Figure S216. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'MIRSEQ CNMF' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 0.0327 (Kruskal-Wallis (anova)), Q value = 1

Table S222. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	248	2007.7 (5.0)
subtype1	78	2008.8 (4.7)
subtype2	67	2007.5 (4.9)
subtype3	103	2007.1 (5.2)

Figure S217. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'MIRSEQ CNMF' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.679 (Fisher's exact test), Q value = 1

Table S223. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	49	62
subtype1	5	18	20
subtype2	2	15	15
subtype3	3	16	27

Figure S218. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'MIRSEQ CNMF' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.407 (Kruskal-Wallis (anova)), Q value = 1

Table S224. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	218	161.6 (7.1)
subtype1	70	162.1 (6.9)
subtype2	55	162.5 (7.5)
subtype3	93	160.7 (6.9)

Figure S219. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'MIRSEQ CNMF' versus 'CORPUS_INVOLVEMENT'

P value = 0.803 (Fisher's exact test), Q value = 1

Table S225. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	96	17
subtype1	34	7
subtype2	31	6
subtype3	31	4

Figure S220. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'MIRSEQ CNMF' versus 'CHEMO_CONCURRENT_TYPE'

P value = 1 (Fisher's exact test), Q value = 1

Table S226. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	3	18	1
subtype1	0	3	0
subtype2	1	3	0
subtype3	2	12	1

Figure S221. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'MIRSEQ CNMF' versus 'CERVIX_SUV_RESULTS'

P value = 0.508 (Kruskal-Wallis (anova)), Q value = 1

Table S227. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	13	12.4 (5.5)
subtype1	2	10.8 (5.2)
subtype2	4	10.2 (3.3)
subtype3	7	14.1 (6.6)

Figure S222. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'MIRSEQ CNMF' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.392 (Fisher's exact test), Q value = 1

Table S228. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1A1	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3	T3B	T4	TIS	TX
ALL	1	1	32	68	25	4	8	7	10	23	1	8	4	1	15
subtype1	0	0	8	29	8	2	3	3	6	6	0	4	1	0	3
subtype2	0	1	6	23	8	1	3	2	2	6	0	1	1	0	3
subtype3	1	0	18	16	9	1	2	2	2	11	1	3	2	1	9

Figure S223. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'MIRSEQ CNMF' versus 'AGE_AT_DIAGNOSIS'

P value = 0.21 (Kruskal-Wallis (anova)), Q value = 1

Table S229. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	250	47.7 (13.5)
subtype1	79	46.4 (11.5)
subtype2	68	50.0 (13.2)
subtype3	103	47.1 (14.9)

Figure S224. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'MIRSEQ CNMF' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.0469 (Fisher's exact test), Q value = 1

Table S230. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA1	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE III	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	1	35	70	34	4	7	5	7	27	1	34	5	7
subtype1	2	0	0	0	7	28	13	1	1	1	5	5	0	11	1	4
subtype2	0	1	1	1	7	24	9	1	3	2	2	5	0	7	2	1
subtype3	2	1	0	0	21	18	12	2	3	2	0	17	1	16	2	2

Figure S225. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #6: 'MIRSEQ CHIERARCHICAL'

Table S231. Description of clustering approach #6: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3	4	5
Number of samples	48	79	64	20	39

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.519 (logrank test), Q value = 1

Table S232. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	245	54	0.0 - 195.8 (16.1)
subtype1	47	9	0.1 - 137.2 (14.9)
subtype2	79	16	0.0 - 160.4 (16.1)
subtype3	63	13	0.0 - 182.9 (21.7)
subtype4	18	5	0.1 - 195.8 (13.7)
subtype5	38	11	0.1 - 144.2 (14.3)

Figure S226. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 8.5e-06 (Kruskal-Wallis (anova)), Q value = 0.003

Table S233. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	246	47.6 (13.5)
subtype1	46	44.1 (11.8)
subtype2	78	49.9 (13.7)
subtype3	64	51.2 (14.2)
subtype4	19	51.7 (9.1)
subtype5	39	39.1 (11.5)

Figure S227. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY.T.STAGE'

P value = 0.00149 (Fisher's exact test), Q value = 0.51

Table S234. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	127	52	9	4
subtype1	31	8	0	0
subtype2	28	23	4	2
subtype3	44	8	0	1
subtype4	8	5	2	1
subtype5	16	8	3	0

Figure S228. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY.N.STAGE'

P value = 0.354 (Fisher's exact test), Q value = 1

Table S235. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	117	51
subtype1	28	9
subtype2	27	18
subtype3	34	17
subtype4	10	2
subtype5	18	5

Figure S229. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY.M.STAGE'

P value = 0.00764 (Fisher's exact test), Q value = 1

Table S236. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	99	8	96
subtype1	17	2	22
subtype2	21	1	39
subtype3	33	1	21
subtype4	12	1	4
subtype5	16	3	10

Figure S230. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S237. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	5	209	5	22	3	6
subtype1	4	11	4	20	3	6
subtype2	1	76	1	1	0	0
subtype3	0	64	0	0	0	0
subtype4	0	20	0	0	0	0
subtype5	0	38	0	1	0	0

Figure S231. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.0434 (Fisher's exact test), Q value = 1

Table S238. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	35	215
subtype1	3	45
subtype2	7	72
subtype3	12	52
subtype4	3	17
subtype5	10	29

Figure S232. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CHIERARCHICAL' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.324 (Kruskal-Wallis (anova)), Q value = 1

Table S239. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	76	18.8 (13.8)
subtype1	15	15.1 (11.4)
subtype2	29	21.6 (17.6)
subtype3	19	20.6 (10.9)
subtype4	2	16.0 (5.7)
subtype5	11	13.8 (9.5)

Figure S233. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.228 (Kruskal-Wallis (anova)), Q value = 1

Table S240. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	144	1.1 (2.5)
subtype1	35	0.7 (2.0)
subtype2	37	1.8 (3.2)
subtype3	45	1.0 (1.8)
subtype4	8	0.4 (0.7)
subtype5	19	0.9 (3.2)

Figure S234. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.59 (Fisher's exact test), Q value = 1

Table S241. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	8	19	26	1	178
subtype1	0	4	2	0	37
subtype2	3	5	11	0	53
subtype3	1	6	6	0	46
subtype4	1	2	3	0	14
subtype5	3	2	4	1	28

Figure S235. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.178 (Fisher's exact test), Q value = 1

Table S242. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	157
subtype1	4	30
subtype2	7	46
subtype3	1	48
subtype4	2	14
subtype5	3	19

Figure S236. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'ETHNICITY'

'MIRSEQ CHIERARCHICAL' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.765 (Kruskal-Wallis (anova)), Q value = 1

Table S243. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	228	74.8 (22.4)
subtype1	43	75.1 (14.8)
subtype2	75	74.5 (21.5)
subtype3	58	76.4 (29.8)
subtype4	18	68.3 (17.5)
subtype5	34	75.5 (20.3)

Figure S237. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'MIRSEQ CHIERARCHICAL' versus 'TUMOR_STATUS'

P value = 0.88 (Fisher's exact test), Q value = 1

Table S244. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	75	28
subtype1	17	5
subtype2	15	8
subtype3	26	8
subtype4	5	2
subtype5	12	5

Figure S238. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'TUMOR_STATUS'

'MIRSEQ CHIERARCHICAL' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.239 (Fisher's exact test), Q value = 1

Table S245. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	24	5	1	11	7	188	14
subtype1	3	3	0	1	2	37	2
subtype2	12	0	0	4	2	57	4
subtype3	2	2	0	5	1	50	4
subtype4	2	0	1	1	0	14	2
subtype5	5	0	0	0	2	30	2

Figure S239. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.0646 (Fisher's exact test), Q value = 1

Table S246. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	17	113	102	1	14
subtype1	6	21	17	0	4
subtype2	3	40	30	0	4
subtype3	6	31	26	0	1
subtype4	0	12	7	0	1
subtype5	2	9	22	1	4

Figure S240. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'MIRSEQ CHIERARCHICAL' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.0569 (Kruskal-Wallis (anova)), Q value = 1

Table S247. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	34	2000.4 (11.2)
subtype1	6	1999.0 (14.1)
subtype2	12	2005.6 (10.3)
subtype3	7	1993.4 (11.4)
subtype4	2	1992.5 (6.4)
subtype5	7	2002.0 (7.7)

Figure S241. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'MIRSEQ CHIERARCHICAL' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.324 (Kruskal-Wallis (anova)), Q value = 1

Table S248. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	76	18.8 (13.8)
subtype1	15	15.1 (11.4)
subtype2	29	21.6 (17.6)
subtype3	19	20.6 (10.9)
subtype4	2	16.0 (5.7)
subtype5	11	13.8 (9.5)

Figure S242. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'MIRSEQ CHIERARCHICAL' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.49 (Fisher's exact test), Q value = 1

Table S249. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	33	8	3	54	114
subtype1	6	1	0	10	26
subtype2	12	2	1	18	31
subtype3	4	5	1	16	31
subtype4	3	0	0	2	11
subtype5	8	0	1	8	15

Figure S243. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'MIRSEQ CHIERARCHICAL' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.705 (Kruskal-Wallis (anova)), Q value = 1

Table S250. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	68	21.4 (7.8)
subtype1	14	19.3 (4.4)
subtype2	25	23.1 (9.5)
subtype3	14	22.1 (7.2)
subtype4	2	20.0 (5.7)
subtype5	13	20.0 (8.2)

Figure S244. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_TOTAL_DOSE'

P value = 0.415 (Kruskal-Wallis (anova)), Q value = 1

Table S251. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	110	3890.9 (1636.4)
subtype1	18	4214.9 (1570.1)
subtype2	37	3608.0 (1701.3)
subtype3	23	4340.9 (1163.1)
subtype4	9	3488.2 (1690.4)
subtype5	23	3799.9 (1930.5)

Figure S245. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY_TYPE'

P value = 0.243 (Fisher's exact test), Q value = 1

Table S252. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	IMPLANTS	INTERNAL
ALL	20	77	15	1	12
subtype1	3	16	0	0	1
subtype2	4	27	3	0	7
subtype3	7	15	5	1	1
subtype4	1	5	2	0	1
subtype5	5	14	5	0	2

Figure S246. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY_STATUS'

P value = 0.86 (Fisher's exact test), Q value = 1

Table S253. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	29	3
subtype1	3	0
subtype2	7	0
subtype3	9	1
subtype4	2	0
subtype5	8	2

Figure S247. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY_SITE'

P value = 0.151 (Fisher's exact test), Q value = 1

Table S254. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	30	16
subtype1	1	0	3	1
subtype2	0	0	13	11
subtype3	0	1	6	1
subtype4	0	0	2	2
subtype5	1	1	6	1

Figure S248. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_ADJUVANT_UNITS'

P value = 0.881 (Fisher's exact test), Q value = 1

Table S255. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	36	3
subtype1	5	0
subtype2	16	1
subtype3	5	1
subtype4	3	0
subtype5	7	1

Figure S249. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'MIRSEQ CHIERARCHICAL' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.106 (Kruskal-Wallis (anova)), Q value = 1

Table S256. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	219	3.6 (2.5)
subtype1	42	2.8 (1.9)
subtype2	69	4.1 (3.1)
subtype3	57	3.7 (2.0)
subtype4	15	3.7 (2.5)
subtype5	36	3.2 (2.4)

Figure S250. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'MIRSEQ CHIERARCHICAL' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.477 (Kruskal-Wallis (anova)), Q value = 1

Table S257. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	109	0.1 (0.4)
subtype1	19	0.1 (0.2)
subtype2	26	0.0 (0.2)
subtype3	38	0.2 (0.5)
subtype4	9	0.0 (0.0)
subtype5	17	0.0 (0.0)

Figure S251. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.307 (Kruskal-Wallis (anova)), Q value = 1

Table S258. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	135	0.5 (0.8)
subtype1	20	0.6 (0.6)
subtype2	35	0.5 (1.1)
subtype3	45	0.4 (0.7)
subtype4	12	0.3 (0.7)
subtype5	23	0.5 (0.7)

Figure S252. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'MIRSEQ CHIERARCHICAL' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.026 (Kruskal-Wallis (anova)), Q value = 1

Table S259. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	220	2.6 (1.9)
subtype1	42	2.0 (1.6)
subtype2	66	3.0 (2.2)
subtype3	60	2.7 (1.6)
subtype4	16	3.1 (2.1)
subtype5	36	2.4 (1.8)

Figure S253. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.915 (Kruskal-Wallis (anova)), Q value = 1

Table S260. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	119	0.9 (1.8)
subtype1	21	0.6 (0.7)
subtype2	30	1.6 (3.1)
subtype3	40	0.7 (1.1)
subtype4	9	0.7 (1.4)
subtype5	19	0.6 (0.9)

Figure S254. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'MIRSEQ CHIERARCHICAL' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.794 (Kruskal-Wallis (anova)), Q value = 1

Table S261. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	111	0.1 (0.3)
subtype1	18	0.1 (0.3)
subtype2	29	0.1 (0.4)
subtype3	38	0.1 (0.4)
subtype4	9	0.1 (0.3)
subtype5	17	0.1 (0.2)

Figure S255. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'MIRSEQ CHIERARCHICAL' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.899 (Fisher's exact test), Q value = 1

Table S262. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	2	7	33	1	9
subtype1	1	0	5	0	2
subtype2	0	2	7	1	3
subtype3	1	3	12	0	3
subtype4	0	0	2	0	0
subtype5	0	2	7	0	1

Figure S256. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'MIRSEQ CHIERARCHICAL' versus 'MENOPAUSE_STATUS'

P value = 0.00307 (Fisher's exact test), Q value = 1

Table S263. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	16	70	109
subtype1	0	2	13	25
subtype2	1	7	23	31
subtype3	1	4	28	21
subtype4	0	2	4	7
subtype5	0	1	2	25

Figure S257. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'MIRSEQ CHIERARCHICAL' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.445 (Fisher's exact test), Q value = 1

Table S264. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	69	76
subtype1	20	15
subtype2	17	22
subtype3	18	28
subtype4	5	3
subtype5	9	8

Figure S258. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'MIRSEQ CHIERARCHICAL' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.228 (Kruskal-Wallis (anova)), Q value = 1

Table S265. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	144	1.1 (2.5)
subtype1	35	0.7 (2.0)
subtype2	37	1.8 (3.2)
subtype3	45	1.0 (1.8)
subtype4	8	0.4 (0.7)
subtype5	19	0.9 (3.2)

Figure S259. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'MIRSEQ CHIERARCHICAL' versus 'LYMPH_NODES_EXAMINED'

P value = 0.533 (Kruskal-Wallis (anova)), Q value = 1

Table S266. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	166	21.7 (12.4)
subtype1	38	21.4 (10.6)
subtype2	45	20.7 (12.3)
subtype3	51	23.8 (14.5)
subtype4	10	15.9 (8.3)
subtype5	22	22.4 (11.7)

Figure S260. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'MIRSEQ CHIERARCHICAL' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.0905 (Fisher's exact test), Q value = 1

Table S267. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	49	97
subtype1	0	10
subtype2	18	34
subtype3	19	27
subtype4	3	11
subtype5	9	15

Figure S261. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'MIRSEQ CHIERARCHICAL' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 0.127 (Kruskal-Wallis (anova)), Q value = 1

Table S268. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	248	2007.7 (5.0)
subtype1	47	2008.8 (4.2)
subtype2	79	2008.3 (4.6)
subtype3	63	2006.7 (5.2)
subtype4	20	2007.2 (6.5)
subtype5	39	2007.1 (5.4)

Figure S262. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'MIRSEQ CHIERARCHICAL' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.416 (Fisher's exact test), Q value = 1

Table S269. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	49	62
subtype1	4	13	10
subtype2	2	15	22
subtype3	1	13	17
subtype4	0	4	4
subtype5	3	4	9

Figure S263. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'MIRSEQ CHIERARCHICAL' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.685 (Kruskal-Wallis (anova)), Q value = 1

Table S270. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	218	161.6 (7.1)
subtype1	41	162.5 (6.8)
subtype2	72	161.0 (7.0)
subtype3	56	161.2 (7.5)
subtype4	17	160.7 (8.0)
subtype5	32	162.8 (6.2)

Figure S264. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'MIRSEQ CHIERARCHICAL' versus 'CORPUS_INVOLVEMENT'

P value = 0.318 (Fisher's exact test), Q value = 1

Table S271. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	96	17
subtype1	27	3
subtype2	21	5
subtype3	31	7
subtype4	5	2
subtype5	12	0

Figure S265. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'MIRSEQ CHIERARCHICAL' versus 'CHEMO_CONCURRENT_TYPE'

P value = 0.442 (Fisher's exact test), Q value = 1

Table S272. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	3	18	1
subtype1	0	1	0
subtype2	1	2	0
subtype3	2	7	0
subtype4	0	1	0
subtype5	0	7	1

Figure S266. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'CERVIX_SUV_RESULTS'

P value = 0.207 (Kruskal-Wallis (anova)), Q value = 1

Table S273. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	13	12.4 (5.5)
subtype1	1	7.1 (NA)
subtype2	3	10.5 (4.3)
subtype3	6	11.3 (3.7)
subtype5	3	18.4 (7.3)

Figure S267. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'MIRSEQ CHIERARCHICAL' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.00142 (Fisher's exact test), Q value = 0.49

Table S274. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1A1	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3	T3B	T4	TIS	TX
ALL	1	1	32	68	25	4	8	7	10	23	1	8	4	1	15
subtype1	0	0	3	23	5	1	2	3	1	1	0	0	0	0	2
subtype2	1	0	6	13	8	2	2	0	6	13	0	4	2	0	7
subtype3	0	1	12	24	7	0	3	1	0	4	0	0	1	0	3
subtype4	0	0	3	3	2	0	0	1	2	2	1	1	1	0	1
subtype5	0	0	8	5	3	1	1	2	1	3	0	3	0	1	2

Figure S268. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'MIRSEQ CHIERARCHICAL' versus 'AGE_AT_DIAGNOSIS'

P value = 6.47e-06 (Kruskal-Wallis (anova)), Q value = 0.0023

Table S275. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	250	47.7 (13.5)
subtype1	48	44.4 (11.6)
subtype2	79	50.0 (13.7)
subtype3	64	51.2 (14.2)
subtype4	20	52.3 (9.2)
subtype5	39	39.1 (11.5)

Figure S269. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'MIRSEQ CHIERARCHICAL' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.00057 (Fisher's exact test), Q value = 0.2

Table S276. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA1	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE III	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	1	35	70	34	4	7	5	7	27	1	34	5	7
subtype1	1	0	0	0	2	25	8	1	1	1	0	2	0	5	0	2
subtype2	2	1	0	0	7	15	10	2	1	0	3	13	0	15	3	2
subtype3	0	1	1	1	13	21	8	1	2	1	1	2	0	8	2	1
subtype4	1	0	0	0	6	1	1	0	0	1	2	2	1	3	0	2
subtype5	0	0	0	0	7	8	7	0	3	2	1	8	0	3	0	0

Figure S270. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #7: 'MIRseq Mature CNMF subtypes'

Table S277. Description of clustering approach #7: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	99	68	72

'MIRseq Mature CNMF subtypes' versus 'Time to Death'

P value = 0.765 (logrank test), Q value = 1

Table S278. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	235	52	0.0 - 182.9 (16.6)
subtype1	98	16	0.0 - 182.9 (15.6)
subtype2	65	13	0.1 - 147.4 (14.1)
subtype3	72	23	0.1 - 177.0 (28.6)

Figure S271. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'AGE'

P value = 0.415 (Kruskal-Wallis (anova)), Q value = 1

Table S279. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	236	47.7 (13.6)
subtype1	97	47.6 (12.8)
subtype2	67	49.3 (14.0)
subtype3	72	46.3 (14.3)

Figure S272. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.892 (Fisher's exact test), Q value = 1

Table S280. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	122	49	7	4
subtype1	46	23	4	2
subtype2	34	13	1	1
subtype3	42	13	2	1

Figure S273. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.339 (Fisher's exact test), Q value = 1

Table S281. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	112	48
subtype1	43	15
subtype2	36	13
subtype3	33	20

Figure S274. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.124 (Fisher's exact test), Q value = 1

Table S282. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	96	7	90
subtype1	35	6	41
subtype2	27	1	24
subtype3	34	0	25

Figure S275. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 8e-05 (Fisher's exact test), Q value = 0.028

Table S283. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	5	199	5	21	3	6
subtype1	4	70	3	14	3	5
subtype2	1	57	2	7	0	1
subtype3	0	72	0	0	0	0

Figure S276. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'MIRseq Mature CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S284. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	33	206
subtype1	3	96
subtype2	6	62
subtype3	24	48

Figure S277. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq Mature CNMF subtypes' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.408 (Kruskal-Wallis (anova)), Q value = 1

Table S285. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	71	18.7 (13.6)
subtype1	22	16.4 (15.2)
subtype2	23	19.6 (12.2)
subtype3	26	19.7 (13.8)

Figure S278. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'MIRseq Mature CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.446 (Kruskal-Wallis (anova)), Q value = 1

Table S286. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	140	1.0 (2.4)
subtype1	46	0.7 (2.2)
subtype2	43	1.2 (2.3)
subtype3	51	1.2 (2.7)

Figure S279. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.099 (Fisher's exact test), Q value = 1

Table S287. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	7	18	23	1	172
subtype1	5	9	13	0	64
subtype2	1	6	2	0	54
subtype3	1	3	8	1	54

Figure S280. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.0475 (Fisher's exact test), Q value = 1

Table S288. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	148
subtype1	12	58
subtype2	2	46
subtype3	3	44

Figure S281. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'ETHNICITY'

'MIRseq Mature CNMF subtypes' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.889 (Kruskal-Wallis (anova)), Q value = 1

Table S289. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	217	74.6 (22.3)
subtype1	91	73.8 (19.8)
subtype2	62	76.0 (28.6)
subtype3	64	74.4 (18.8)

Figure S282. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'MIRseq Mature CNMF subtypes' versus 'TUMOR_STATUS'

P value = 0.566 (Fisher's exact test), Q value = 1

Table S290. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	72	27
subtype1	25	9
subtype2	17	4
subtype3	30	14

Figure S283. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

'MIRseq Mature CNMF subtypes' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.0764 (Fisher's exact test), Q value = 1

Table S291. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	24	5	1	11	6	179	13
subtype1	15	3	1	2	4	68	6
subtype2	3	1	0	5	2	51	6
subtype3	6	1	0	4	0	60	1

Figure S284. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'MIRseq Mature CNMF subtypes' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.0342 (Fisher's exact test), Q value = 1

Table S292. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	15	106	100	1	14
subtype1	6	36	49	0	8
subtype2	8	35	22	0	2
subtype3	1	35	29	1	4

Figure S285. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'MIRseq Mature CNMF subtypes' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.317 (Kruskal-Wallis (anova)), Q value = 1

Table S293. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	32	2000.8 (11.4)
subtype1	11	2005.9 (5.9)
subtype2	9	1996.2 (16.2)
subtype3	12	1999.6 (10.1)

Figure S286. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'MIRseq Mature CNMF subtypes' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.408 (Kruskal-Wallis (anova)), Q value = 1

Table S294. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	71	18.7 (13.6)
subtype1	22	16.4 (15.2)
subtype2	23	19.6 (12.2)
subtype3	26	19.7 (13.8)

Figure S287. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'MIRseq Mature CNMF subtypes' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.124 (Fisher's exact test), Q value = 1

Table S295. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	32	7	3	51	112
subtype1	12	1	1	14	53
subtype2	7	3	2	18	29
subtype3	13	3	0	19	30

Figure S288. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'MIRseq Mature CNMF subtypes' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.936 (Kruskal-Wallis (anova)), Q value = 1

Table S296. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	63	21.1 (7.7)
subtype1	22	21.6 (7.7)
subtype2	19	20.7 (7.2)
subtype3	22	21.0 (8.3)

Figure S289. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_TOTAL_DOSE'

P value = 0.0236 (Kruskal-Wallis (anova)), Q value = 1

Table S297. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	107	3911.4 (1611.9)
subtype1	44	3495.4 (1847.3)
subtype2	25	3885.5 (1601.3)
subtype3	38	4410.1 (1162.8)

Figure S290. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY_TYPE'

P value = 2e-05 (Fisher's exact test), Q value = 0.007

Table S298. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	IMPLANTS	INTERNAL
ALL	19	74	15	1	12
subtype1	2	39	1	0	6
subtype2	2	18	5	0	4
subtype3	15	17	9	1	2

Figure S291. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY_STATUS'

P value = 1 (Fisher's exact test), Q value = 1

Table S299. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	27	3
subtype1	3	0
subtype2	5	0
subtype3	19	3

Figure S292. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY_SITE'

P value = 0.797 (Fisher's exact test), Q value = 1

Table S300. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	29	16
subtype1	1	0	12	8
subtype2	1	1	8	5
subtype3	0	1	9	3

Figure S293. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_ADJUVANT_UNITS'

P value = 1 (Fisher's exact test), Q value = 1

Table S301. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	35	3
subtype1	17	1
subtype2	9	1
subtype3	9	1

Figure S294. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'MIRseq Mature CNMF subtypes' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.436 (Kruskal-Wallis (anova)), Q value = 1

Table S302. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	210	3.6 (2.5)
subtype1	83	3.4 (2.5)
subtype2	62	3.7 (2.3)
subtype3	65	3.6 (2.6)

Figure S295. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'MIRseq Mature CNMF subtypes' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.218 (Kruskal-Wallis (anova)), Q value = 1

Table S303. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	104	0.1 (0.3)
subtype1	31	0.0 (0.2)
subtype2	27	0.0 (0.0)
subtype3	46	0.1 (0.5)

Figure S296. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'MIRseq Mature CNMF subtypes' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.279 (Kruskal-Wallis (anova)), Q value = 1

Table S304. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	131	0.5 (0.8)
subtype1	44	0.7 (1.1)
subtype2	37	0.4 (0.5)
subtype3	50	0.4 (0.7)

Figure S297. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'MIRseq Mature CNMF subtypes' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.341 (Kruskal-Wallis (anova)), Q value = 1

Table S305. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	211	2.7 (1.9)
subtype1	83	2.7 (2.1)
subtype2	62	2.8 (1.7)
subtype3	66	2.4 (1.8)

Figure S298. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'MIRseq Mature CNMF subtypes' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.555 (Kruskal-Wallis (anova)), Q value = 1

Table S306. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	113	0.8 (1.8)
subtype1	34	0.5 (0.9)
subtype2	32	1.2 (2.5)
subtype3	47	0.8 (1.7)

Figure S299. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'MIRseq Mature CNMF subtypes' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.849 (Kruskal-Wallis (anova)), Q value = 1

Table S307. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	107	0.1 (0.3)
subtype1	34	0.1 (0.3)
subtype2	27	0.1 (0.4)
subtype3	46	0.1 (0.3)

Figure S300. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'MIRseq Mature CNMF subtypes' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.646 (Fisher's exact test), Q value = 1

Table S308. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	1	6	32	1	9
subtype1	0	1	7	0	1
subtype2	1	1	8	1	4
subtype3	0	4	17	0	4

Figure S301. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'MIRseq Mature CNMF subtypes' versus 'MENOPAUSE_STATUS'

P value = 0.414 (Fisher's exact test), Q value = 1

Table S309. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	16	65	104
subtype1	0	9	23	44
subtype2	0	4	21	27
subtype3	2	3	21	33

Figure S302. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'MIRseq Mature CNMF subtypes' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.715 (Fisher's exact test), Q value = 1

Table S310. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	66	70
subtype1	23	22
subtype2	22	21
subtype3	21	27

Figure S303. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'MIRseq Mature CNMF subtypes' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.446 (Kruskal-Wallis (anova)), Q value = 1

Table S311. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	140	1.0 (2.4)
subtype1	46	0.7 (2.2)
subtype2	43	1.2 (2.3)
subtype3	51	1.2 (2.7)

Figure S304. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'MIRseq Mature CNMF subtypes' versus 'LYMPH_NODES_EXAMINED'

P value = 0.18 (Kruskal-Wallis (anova)), Q value = 1

Table S312. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	158	22.2 (12.5)
subtype1	55	20.8 (12.0)
subtype2	49	25.0 (13.0)
subtype3	54	21.0 (12.4)

Figure S305. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'MIRseq Mature CNMF subtypes' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.0458 (Fisher's exact test), Q value = 1

Table S313. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	46	92
subtype1	8	35
subtype2	17	25
subtype3	21	32

Figure S306. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'MIRseq Mature CNMF subtypes' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 2.69e-05 (Kruskal-Wallis (anova)), Q value = 0.0093

Table S314. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	237	2007.7 (5.0)
subtype1	98	2009.0 (4.5)
subtype2	67	2007.9 (4.8)
subtype3	72	2005.7 (5.3)

Figure S307. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'MIRseq Mature CNMF subtypes' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.9 (Fisher's exact test), Q value = 1

Table S315. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	49	58
subtype1	5	19	26
subtype2	3	15	14
subtype3	2	15	18

Figure S308. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'MIRseq Mature CNMF subtypes' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.533 (Kruskal-Wallis (anova)), Q value = 1

Table S316. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	208	161.4 (7.0)
subtype1	88	161.1 (6.3)
subtype2	58	162.0 (7.0)
subtype3	62	161.3 (8.0)

Figure S309. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'MIRseq Mature CNMF subtypes' versus 'CORPUS_INVOLVEMENT'

P value = 0.755 (Fisher's exact test), Q value = 1

Table S317. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	92	17
subtype1	29	4
subtype2	28	5
subtype3	35	8

Figure S310. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'MIRseq Mature CNMF subtypes' versus 'CHEMO_CONCURRENT_TYPE'

P value = 0.206 (Fisher's exact test), Q value = 1

Table S318. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	3	17	1
subtype1	0	2	0
subtype2	0	2	1
subtype3	3	13	0

Figure S311. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'MIRseq Mature CNMF subtypes' versus 'CERVIX_SUV_RESULTS'

P value = 0.348 (Kruskal-Wallis (anova)), Q value = 1

Table S319. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	12	12.0 (5.6)
subtype1	3	14.8 (10.0)
subtype2	5	9.4 (3.4)
subtype3	4	13.2 (3.1)

Figure S312. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'MIRseq Mature CNMF subtypes' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.235 (Fisher's exact test), Q value = 1

Table S320. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3B	T4	TIS	TX
ALL	1	31	66	24	4	8	7	9	21	7	4	1	15
subtype1	1	9	26	10	2	4	2	5	10	4	2	1	8
subtype2	0	4	21	9	0	4	3	2	4	1	1	0	5
subtype3	0	18	19	5	2	0	2	2	7	2	1	0	2

Figure S313. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'MIRseq Mature CNMF subtypes' versus 'AGE_AT_DIAGNOSIS'

P value = 0.37 (Kruskal-Wallis (anova)), Q value = 1

Table S321. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	239	47.7 (13.5)
subtype1	99	47.6 (12.7)
subtype2	68	49.5 (13.9)
subtype3	72	46.3 (14.3)

Figure S314. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'MIRseq Mature CNMF subtypes' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.00658 (Fisher's exact test), Q value = 1

Table S322. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	34	68	32	4	6	5	7	26	33	5	6
subtype1	3	1	0	8	32	14	1	2	0	4	13	13	1	4
subtype2	0	1	1	4	22	11	2	3	3	2	6	6	3	1
subtype3	1	0	0	22	14	7	1	1	2	1	7	14	1	1

Figure S315. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Clustering Approach #8: 'MIRseq Mature cHierClus subtypes'

Table S323. Description of clustering approach #8: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	50	42	31	116

'MIRseq Mature cHierClus subtypes' versus 'Time to Death'

P value = 0.981 (logrank test), Q value = 1

Table S324. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	235	52	0.0 - 182.9 (16.6)
subtype1	49	9	0.1 - 137.2 (15.0)
subtype2	40	7	0.1 - 147.4 (14.5)
subtype3	31	12	1.2 - 177.0 (35.6)
subtype4	115	24	0.0 - 182.9 (16.1)

Figure S316. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'AGE'

P value = 0.546 (Kruskal-Wallis (anova)), Q value = 1

Table S325. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	236	47.7 (13.6)
subtype1	48	45.8 (11.8)
subtype2	42	49.9 (14.3)
subtype3	31	48.9 (14.4)
subtype4	115	47.3 (13.8)

Figure S317. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY.T.STAGE'

P value = 0.825 (Fisher's exact test), Q value = 1

Table S326. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

nPatients	T1	T2	T3	T4
ALL	122	49	7	4
subtype1	28	13	0	1
subtype2	22	7	1	1
subtype3	20	5	1	0
subtype4	52	24	5	2

Figure S318. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGY.T.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY.N.STAGE'

P value = 0.102 (Fisher's exact test), Q value = 1

Table S327. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

nPatients	0	1
ALL	112	48
subtype1	30	8
subtype2	21	8
subtype3	13	13
subtype4	48	19

Figure S319. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.N.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY.M.STAGE'

P value = 0.22 (Fisher's exact test), Q value = 1

Table S328. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

nPatients	M0	M1	MX
ALL	96	7	90
subtype1	20	4	20
subtype2	18	1	13
subtype3	17	0	9
subtype4	41	2	48

Figure S320. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.M.STAGE'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S329. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

nPatients	ADENOSQUAMOUS	CERVICAL SQUAMOUS CELL CARCINOMA	ENDOCERVICAL ADENOCARCINOMA OF THE USUAL TYPE	ENDOCERVICAL TYPE OF ADENOCARCINOMA	ENDOMETRIOID ADENOCARCINOMA OF ENDOCERVIX	MUCINOUS ADENOCARCINOMA OF ENDOCERVICAL TYPE
ALL	5	199	5	21	3	6
subtype1	3	14	4	20	3	6
subtype2	1	41	0	0	0	0
subtype3	0	30	0	1	0	0
subtype4	1	114	1	0	0	0

Figure S321. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'HISTOLOGICAL.TYPE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S330. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	33	206
subtype1	1	49
subtype2	3	39
subtype3	19	12
subtype4	10	106

Figure S322. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq Mature cHierClus subtypes' versus 'NUMBERPACKYEARSSMOKED'

P value = 0.238 (Kruskal-Wallis (anova)), Q value = 1

Table S331. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

	nPatients	Mean (Std.Dev)
ALL	71	18.7 (13.6)
subtype1	13	14.2 (10.6)
subtype2	12	23.2 (11.6)
subtype3	10	19.5 (9.9)
subtype4	36	18.5 (15.9)

Figure S323. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'NUMBERPACKYEARSSMOKED'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.118 (Kruskal-Wallis (anova)), Q value = 1

Table S332. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	140	1.0 (2.4)
subtype1	34	0.6 (2.0)
subtype2	24	1.1 (1.8)
subtype3	25	2.4 (4.5)
subtype4	57	0.6 (1.1)

Figure S324. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.499 (Fisher's exact test), Q value = 1

Table S333. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	NATIVE HAWAIIAN OR OTHER PACIFIC ISLANDER	WHITE
ALL	7	18	23	1	172
subtype1	1	5	3	0	36
subtype2	1	5	1	0	31
subtype3	0	2	3	0	26
subtype4	5	6	16	1	79

Figure S325. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.294 (Fisher's exact test), Q value = 1

Table S334. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	17	148
subtype1	6	31
subtype2	1	29
subtype3	1	22
subtype4	9	66

Figure S326. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'ETHNICITY'

'MIRseq Mature cHierClus subtypes' versus 'WEIGHT_KG_AT_DIAGNOSIS'

P value = 0.679 (Kruskal-Wallis (anova)), Q value = 1

Table S335. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	217	74.6 (22.3)
subtype1	45	75.4 (14.9)
subtype2	39	78.0 (34.3)
subtype3	28	71.8 (17.7)
subtype4	105	73.8 (20.6)

Figure S327. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'WEIGHT_KG_AT_DIAGNOSIS'

'MIRseq Mature cHierClus subtypes' versus 'TUMOR_STATUS'

P value = 0.399 (Fisher's exact test), Q value = 1

Table S336. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

nPatients	TUMOR FREE	WITH TUMOR
ALL	72	27
subtype1	17	4
subtype2	11	2
subtype3	16	10
subtype4	28	11

Figure S328. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'TUMOR_STATUS'

'MIRseq Mature cHierClus subtypes' versus 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

P value = 0.0207 (Fisher's exact test), Q value = 1

Table S337. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

nPatients	BRAZIL	CANADA	NIGERIA	RUSSIA	UKRAINE	UNITED STATES	VIETNAM
ALL	24	5	1	11	6	179	13
subtype1	4	3	1	2	1	36	3
subtype2	3	1	0	5	2	26	5
subtype3	0	0	0	1	0	30	0
subtype4	17	1	0	3	3	87	5

Figure S329. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'TUMOR_SAMPLE_PROCUREMENT_COUNTRY'

'MIRseq Mature cHierClus subtypes' versus 'NEOPLASMHISTOLOGICGRADE'

P value = 0.0459 (Fisher's exact test), Q value = 1

Table S338. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

nPatients	G1	G2	G3	G4	GX
ALL	15	106	100	1	14
subtype1	6	22	18	0	4
subtype2	5	23	11	0	2
subtype3	0	17	14	0	0
subtype4	4	44	57	1	8

Figure S330. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'NEOPLASMHISTOLOGICGRADE'

'MIRseq Mature cHierClus subtypes' versus 'TOBACCO_SMOKING_YEAR_STOPPED'

P value = 0.0617 (Kruskal-Wallis (anova)), Q value = 1

Table S339. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

	nPatients	Mean (Std.Dev)
ALL	32	2000.8 (11.4)
subtype1	5	1996.4 (14.1)
subtype2	5	1995.4 (15.9)
subtype3	5	1995.6 (5.4)
subtype4	17	2005.2 (9.4)

Figure S331. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #16: 'TOBACCO_SMOKING_YEAR_STOPPED'

'MIRseq Mature cHierClus subtypes' versus 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

P value = 0.238 (Kruskal-Wallis (anova)), Q value = 1

Table S340. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	71	18.7 (13.6)
subtype1	13	14.2 (10.6)
subtype2	12	23.2 (11.6)
subtype3	10	19.5 (9.9)
subtype4	36	18.5 (15.9)

Figure S332. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #17: 'TOBACCO_SMOKING_PACK_YEARS_SMOKED'

'MIRseq Mature cHierClus subtypes' versus 'TOBACCO_SMOKING_HISTORY'

P value = 0.459 (Fisher's exact test), Q value = 1

Table S341. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

nPatients	CURRENT REFORMED SMOKER FOR < OR = 15 YEARS	CURRENT REFORMED SMOKER FOR > 15 YEARS	CURRENT REFORMED SMOKER, DURATION NOT SPECIFIED	CURRENT SMOKER	LIFELONG NON-SMOKER
ALL	32	7	3	51	112
subtype1	5	1	0	9	29
subtype2	3	2	1	10	19
subtype3	8	2	0	5	14
subtype4	16	2	2	27	50

Figure S333. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #18: 'TOBACCO_SMOKING_HISTORY'

'MIRseq Mature cHierClus subtypes' versus 'PATIENT.AGEBEGANSMOKINGINYEARS'

P value = 0.998 (Kruskal-Wallis (anova)), Q value = 1

Table S342. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

	nPatients	Mean (Std.Dev)
ALL	63	21.1 (7.7)
subtype1	12	19.7 (4.6)
subtype2	9	20.1 (5.3)
subtype3	8	19.9 (5.2)
subtype4	34	22.2 (9.4)

Figure S334. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #19: 'PATIENT.AGEBEGANSMOKINGINYEARS'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_TOTAL_DOSE'

P value = 0.297 (Kruskal-Wallis (anova)), Q value = 1

Table S343. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

	nPatients	Mean (Std.Dev)
ALL	107	3911.4 (1611.9)
subtype1	18	3763.4 (1871.6)
subtype2	13	3685.2 (1875.1)
subtype3	19	4549.5 (914.6)
subtype4	57	3797.0 (1631.5)

Figure S335. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #20: 'RADIATION_TOTAL_DOSE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 0.0036

Table S344. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

nPatients	COMBINATION	EXTERNAL	EXTERNAL BEAM	IMPLANTS	INTERNAL
ALL	19	74	15	1	12
subtype1	1	17	0	0	1
subtype2	0	11	3	0	3
subtype3	13	1	7	1	0
subtype4	5	45	5	0	8

Figure S336. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #21: 'RADIATION_THERAPY_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY_STATUS'

P value = 0.458 (Fisher's exact test), Q value = 1

Table S345. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

nPatients	COMPLETED AS PLANNED	TREATMENT NOT COMPLETED
ALL	27	3
subtype1	1	0
subtype2	2	0
subtype3	17	1
subtype4	7	2

Figure S337. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #22: 'RADIATION_THERAPY_STATUS'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY_SITE'

P value = 0.504 (Fisher's exact test), Q value = 1

Table S346. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

nPatients	DISTANT RECURRENCE	LOCAL RECURRENCE	PRIMARY TUMOR FIELD	REGIONAL SITE
ALL	2	2	29	16
subtype1	1	0	3	1
subtype2	0	1	5	4
subtype3	0	0	2	0
subtype4	1	1	19	11

Figure S338. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #23: 'RADIATION_THERAPY_SITE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_ADJUVANT_UNITS'

P value = 0.761 (Fisher's exact test), Q value = 1

Table S347. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

nPatients	CGY	GY
ALL	35	3
subtype1	5	0
subtype2	6	1
subtype3	2	0
subtype4	22	2

Figure S339. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #24: 'RADIATION_ADJUVANT_UNITS'

'MIRseq Mature cHierClus subtypes' versus 'PREGNANCIES_COUNT_TOTAL'

P value = 0.115 (Kruskal-Wallis (anova)), Q value = 1

Table S348. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

	nPatients	Mean (Std.Dev)
ALL	210	3.6 (2.5)
subtype1	44	2.9 (2.1)
subtype2	37	3.5 (2.5)
subtype3	30	3.3 (1.9)
subtype4	99	4.0 (2.7)

Figure S340. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #25: 'PREGNANCIES_COUNT_TOTAL'

'MIRseq Mature cHierClus subtypes' versus 'PREGNANCIES_COUNT_STILLBIRTH'

P value = 0.216 (Kruskal-Wallis (anova)), Q value = 1

Table S349. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

	nPatients	Mean (Std.Dev)
ALL	104	0.1 (0.3)
subtype1	20	0.0 (0.0)
subtype2	17	0.0 (0.0)
subtype3	26	0.2 (0.6)
subtype4	41	0.0 (0.2)

Figure S341. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #26: 'PREGNANCIES_COUNT_STILLBIRTH'

'MIRseq Mature cHierClus subtypes' versus 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

P value = 0.357 (Kruskal-Wallis (anova)), Q value = 1

Table S350. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	131	0.5 (0.8)
subtype1	23	0.5 (0.5)
subtype2	23	0.3 (0.4)
subtype3	27	0.4 (0.7)
subtype4	58	0.6 (1.1)

Figure S342. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #27: 'PATIENT.PATIENTPREGNANCYSPONTANEOUSABORTIONCOUNT'

'MIRseq Mature cHierClus subtypes' versus 'PREGNANCIES_COUNT_LIVE_BIRTH'

P value = 0.036 (Kruskal-Wallis (anova)), Q value = 1

Table S351. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	211	2.7 (1.9)
subtype1	43	2.2 (1.9)
subtype2	37	2.6 (1.3)
subtype3	31	2.2 (1.6)
subtype4	100	3.0 (2.2)

Figure S343. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #28: 'PREGNANCIES_COUNT_LIVE_BIRTH'

'MIRseq Mature cHierClus subtypes' versus 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

P value = 0.64 (Kruskal-Wallis (anova)), Q value = 1

Table S352. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

	nPatients	Mean (Std.Dev)
ALL	113	0.8 (1.8)
subtype1	22	0.7 (1.0)
subtype2	21	1.5 (2.9)
subtype3	25	0.6 (1.0)
subtype4	45	0.7 (1.7)

Figure S344. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #29: 'PATIENT.PATIENTPREGNANCYTHERAPEUTICABORTIONCOUNT'

'MIRseq Mature cHierClus subtypes' versus 'PREGNANCIES_COUNT_ECTOPIC'

P value = 0.658 (Kruskal-Wallis (anova)), Q value = 1

Table S353. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

	nPatients	Mean (Std.Dev)
ALL	107	0.1 (0.3)
subtype1	20	0.1 (0.2)
subtype2	17	0.1 (0.5)
subtype3	26	0.1 (0.3)
subtype4	44	0.1 (0.3)

Figure S345. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #30: 'PREGNANCIES_COUNT_ECTOPIC'

'MIRseq Mature cHierClus subtypes' versus 'POS_LYMPH_NODE_LOCATION'

P value = 0.7 (Fisher's exact test), Q value = 1

Table S354. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

nPatients	MACROSCOPIC PARAMETRIAL INVOLVEMENT	MICROSCOPIC PARAMETRIAL INVOLVEMENT	OTHER LOCATION, SPECIFY	POSITIVE BLADDER MARGIN	POSITIVE VAGINAL MARGIN
ALL	1	6	32	1	9
subtype1	0	1	5	0	2
subtype2	1	1	4	1	1
subtype3	0	2	13	0	2
subtype4	0	2	10	0	4

Figure S346. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #31: 'POS_LYMPH_NODE_LOCATION'

'MIRseq Mature cHierClus subtypes' versus 'MENOPAUSE_STATUS'

P value = 0.396 (Fisher's exact test), Q value = 1

Table S355. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

nPatients	INDETERMINATE (NEITHER PRE OR POSTMENOPAUSAL)	PERI (6-12 MONTHS SINCE LAST MENSTRUAL PERIOD)	POST (PRIOR BILATERAL OVARIECTOMY OR >12 MO SINCE LMP WITH NO PRIOR HYSTERECTOMY)	PRE (<6 MONTHS SINCE LMP AND NO PRIOR BILATERAL OVARIECTOMY AND NOT ON ESTROGEN REPLACEMENT)
ALL	2	16	65	104
subtype1	0	3	14	25
subtype2	0	3	13	16
subtype3	2	1	12	14
subtype4	0	9	26	49

Figure S347. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #32: 'MENOPAUSE_STATUS'

'MIRseq Mature cHierClus subtypes' versus 'LYMPHOVASCULAR_INVOLVEMENT'

P value = 0.0762 (Fisher's exact test), Q value = 1

Table S356. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	66	70
subtype1	20	15
subtype2	13	13
subtype3	6	18
subtype4	27	24

Figure S348. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #33: 'LYMPHOVASCULAR_INVOLVEMENT'

'MIRseq Mature cHierClus subtypes' versus 'LYMPH_NODES_EXAMINED_HE_COUNT'

P value = 0.118 (Kruskal-Wallis (anova)), Q value = 1

Table S357. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

	nPatients	Mean (Std.Dev)
ALL	140	1.0 (2.4)
subtype1	34	0.6 (2.0)
subtype2	24	1.1 (1.8)
subtype3	25	2.4 (4.5)
subtype4	57	0.6 (1.1)

Figure S349. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #34: 'LYMPH_NODES_EXAMINED_HE_COUNT'

'MIRseq Mature cHierClus subtypes' versus 'LYMPH_NODES_EXAMINED'

P value = 0.641 (Kruskal-Wallis (anova)), Q value = 1

Table S358. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

	nPatients	Mean (Std.Dev)
ALL	158	22.2 (12.5)
subtype1	38	20.8 (10.5)
subtype2	29	21.2 (12.5)
subtype3	25	26.1 (15.2)
subtype4	66	21.9 (12.4)

Figure S350. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #35: 'LYMPH_NODES_EXAMINED'

'MIRseq Mature cHierClus subtypes' versus 'KERATINIZATION_SQUAMOUS_CELL'

P value = 0.0192 (Fisher's exact test), Q value = 1

Table S359. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

nPatients	KERATINIZING SQUAMOUS CELL CARCINOMA	NON-KERATINIZING SQUAMOUS CELL CARCINOMA
ALL	46	92
subtype1	0	11
subtype2	13	15
subtype3	9	12
subtype4	24	54

Figure S351. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #36: 'KERATINIZATION_SQUAMOUS_CELL'

'MIRseq Mature cHierClus subtypes' versus 'INITIAL_PATHOLOGIC_DX_YEAR'

P value = 1.22e-06 (Kruskal-Wallis (anova)), Q value = 0.00044

Table S360. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

	nPatients	Mean (Std.Dev)
ALL	237	2007.7 (5.0)
subtype1	49	2009.0 (3.8)
subtype2	41	2008.9 (4.1)
subtype3	31	2003.0 (5.1)
subtype4	116	2007.9 (5.1)

Figure S352. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #37: 'INITIAL_PATHOLOGIC_DX_YEAR'

'MIRseq Mature cHierClus subtypes' versus 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

P value = 0.395 (Fisher's exact test), Q value = 1

Table S361. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

nPatients	CURRENT USER	FORMER USER	NEVER USED
ALL	10	49	58
subtype1	5	13	12
subtype2	0	9	9
subtype3	0	5	10
subtype4	5	22	27

Figure S353. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #38: 'HISTORY_HORMONAL_CONTRACEPTIVES_USE'

'MIRseq Mature cHierClus subtypes' versus 'HEIGHT_CM_AT_DIAGNOSIS'

P value = 0.753 (Kruskal-Wallis (anova)), Q value = 1

Table S362. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	208	161.4 (7.0)
subtype1	44	162.2 (6.9)
subtype2	37	162.5 (7.8)
subtype3	25	160.8 (8.0)
subtype4	102	160.9 (6.6)

Figure S354. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #39: 'HEIGHT_CM_AT_DIAGNOSIS'

'MIRseq Mature cHierClus subtypes' versus 'CORPUS_INVOLVEMENT'

P value = 0.486 (Fisher's exact test), Q value = 1

Table S363. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

nPatients	ABSENT	PRESENT
ALL	92	17
subtype1	27	4
subtype2	16	3
subtype3	17	6
subtype4	32	4

Figure S355. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #40: 'CORPUS_INVOLVEMENT'

'MIRseq Mature cHierClus subtypes' versus 'CHEMO_CONCURRENT_TYPE'

P value = 0.14 (Fisher's exact test), Q value = 1

Table S364. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

nPatients	CARBOPLATIN	CISPLATIN	OTHER
ALL	3	17	1
subtype1	0	0	0
subtype2	0	1	0
subtype3	3	9	0
subtype4	0	7	1

Figure S356. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #41: 'CHEMO_CONCURRENT_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'CERVIX_SUV_RESULTS'

P value = 0.302 (Kruskal-Wallis (anova)), Q value = 1

Table S365. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

	nPatients	Mean (Std.Dev)
ALL	12	12.0 (5.6)
subtype1	1	7.1 (NA)
subtype2	4	9.6 (3.9)
subtype3	3	11.0 (3.0)
subtype4	4	16.4 (7.0)

Figure S357. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #42: 'CERVIX_SUV_RESULTS'

'MIRseq Mature cHierClus subtypes' versus 'AJCC_TUMOR_PATHOLOGIC_PT'

P value = 0.829 (Fisher's exact test), Q value = 1

Table S366. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

nPatients	T1A	T1B	T1B1	T1B2	T2	T2A	T2A1	T2A2	T2B	T3B	T4	TIS	TX
ALL	1	31	66	24	4	8	7	9	21	7	4	1	15
subtype1	0	3	20	5	1	2	2	3	5	0	1	0	2
subtype2	0	6	12	4	0	3	1	1	2	1	1	0	2
subtype3	0	8	9	3	1	0	1	0	3	1	0	0	0
subtype4	1	14	25	12	2	3	3	5	11	5	2	1	11

Figure S358. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #43: 'AJCC_TUMOR_PATHOLOGIC_PT'

'MIRseq Mature cHierClus subtypes' versus 'AGE_AT_DIAGNOSIS'

P value = 0.611 (Kruskal-Wallis (anova)), Q value = 1

Table S367. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

	nPatients	Mean (Std.Dev)
ALL	239	47.7 (13.5)
subtype1	50	46.0 (11.6)
subtype2	42	49.9 (14.3)
subtype3	31	48.9 (14.4)
subtype4	116	47.4 (13.8)

Figure S359. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #44: 'AGE_AT_DIAGNOSIS'

'MIRseq Mature cHierClus subtypes' versus 'STAGE_EVENT.CLINICAL_STAGE'

P value = 0.193 (Fisher's exact test), Q value = 1

Table S368. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IA2	STAGE IB	STAGE IB1	STAGE IB2	STAGE II	STAGE IIA	STAGE IIA1	STAGE IIA2	STAGE IIB	STAGE IIIB	STAGE IVA	STAGE IVB
ALL	4	2	1	34	68	32	4	6	5	7	26	33	5	6
subtype1	1	0	0	3	22	8	1	1	1	2	4	4	0	3
subtype2	0	1	1	4	14	4	1	2	1	1	5	4	2	1
subtype3	0	0	0	12	5	4	0	1	1	0	2	6	0	0
subtype4	3	1	0	15	27	16	2	2	2	4	15	19	3	2

Figure S360. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #45: 'STAGE_EVENT.CLINICAL_STAGE'

Methods & Data

Input

Cluster data file = CESC-TP.mergedcluster.txt
Clinical data file = CESC-TP.merged_data.txt
Number of patients = 250
Number of clustering approaches = 8
Number of selected clinical features = 45
Exclude small clusters that include fewer than K patients, K = 3

Clustering approaches

CNMF clustering

consensus non-negative matrix factorization clustering approach (Brunet et al. 2004)

Consensus hierarchical clustering

Resampling-based clustering method (Monti et al. 2003)

Survival analysis

For survival clinical features, the Kaplan-Meier survival curves of tumors with and without gene mutations were plotted and the statistical significance P values were estimated by logrank test (Bland and Altman 2004) using the 'survdiff' function in R

Fisher's exact test

For binary clinical features, two-tailed Fisher's exact tests (Fisher 1922) were used to estimate the P values using the 'fisher.test' function in R

Q value calculation

For multiple hypothesis correction, Q value is the False Discovery Rate (FDR) analogue of the P value (Benjamini and Hochberg 1995), defined as the minimum FDR at which the test may be called significant. We used the 'Benjamini and Hochberg' method of 'p.adjust' function in R to convert P values into Q values.

Download Results

In addition to the links below, the full results of the analysis summarized in this report can also be downloaded programmatically using firehose_get, or interactively from either the Broad GDAC website or TCGA Data Coordination Center Portal.

References

[1] Brunet et al., Metagenes and molecular pattern discovery using matrix factorization, PNAS 101(12):4164-9 (2004)

[2] Monti et al., Consensus Clustering: A Resampling-Based Method for Class Discovery and Visualization of Gene Expression Microarray Data, Machine Learning 52(1):91-118 (2003)

[3] Bland and Altman, Statistics notes: The logrank test, BMJ 328(7447):1073 (2004)

[4] Fisher, R.A., On the interpretation of chi-square from contingency tables, and the calculation of P, Journal of the Royal Statistical Society 85(1):87-94 (1922)

[5] Benjamini and Hochberg, Controlling the false discovery rate: a practical and powerful approach to multiple testing, Journal of the Royal Statistical Society Series B 59:289-300 (1995)

Made with Nozzle