Correlation between aggregated molecular cancer subtypes and selected clinical features

Thyroid Adenocarcinoma (Primary solid tumor)

21 August 2015 | analyses__2015_08_21

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 (2015): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1ST7P5N

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 10 different clustering approaches and 17 clinical features across 503 patients, 79 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 correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', and 'TUMOR_SIZE'.
3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', and 'NUMBER_OF_LYMPH_NODES'.
CNMF clustering analysis on RPPA data identified 4 subtypes that correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', and 'EXTRATHYROIDAL_EXTENSION'.
Consensus hierarchical clustering analysis on RPPA data identified 7 subtypes that correlate to 'Time to Death', 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', and 'TUMOR_SIZE'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 4 subtypes that correlate to 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', 'NUMBER_OF_LYMPH_NODES', 'TUMOR_SIZE', and 'ETHNICITY'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 5 subtypes that correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', 'RESIDUAL_TUMOR', 'NUMBER_OF_LYMPH_NODES', 'TUMOR_SIZE', 'RACE', and 'ETHNICITY'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'Time to Death', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', 'NUMBER_OF_LYMPH_NODES', and 'TUMOR_SIZE'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes correlate to 'Time to Death', 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', 'NUMBER_OF_LYMPH_NODES', and 'TUMOR_SIZE'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'Time to Death', 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', 'RESIDUAL_TUMOR', 'NUMBER_OF_LYMPH_NODES', and 'RACE'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'Time to Death', 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'HISTOLOGICAL_TYPE', 'EXTRATHYROIDAL_EXTENSION', 'RESIDUAL_TUMOR', 'NUMBER_OF_LYMPH_NODES', and 'TUMOR_SIZE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 10 different clustering approaches and 17 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 79 significant findings detected.


Clinical Features	Statistical Tests	Copy Number Ratio CNMF subtypes	METHLYATION CNMF	RPPA CNMF subtypes	RPPA cHierClus subtypes	RNAseq CNMF subtypes	RNAseq cHierClus subtypes	MIRSEQ CNMF	MIRSEQ CHIERARCHICAL	MIRseq Mature CNMF subtypes	MIRseq Mature cHierClus subtypes
Time to Death	logrank test	0.338 (0.495)	0.223 (0.367)	0.262 (0.405)	0.000213 (0.000771)	0.591 (0.767)	0.411 (0.578)	0.0336 (0.0762)	0.00883 (0.0235)	0.0299 (0.0705)	0.00422 (0.0121)
YEARS TO BIRTH	Kruskal-Wallis (anova)	1.45e-07 (3.53e-06)	0.156 (0.289)	0.00137 (0.00427)	0.000996 (0.00326)	0.141 (0.266)	0.00239 (0.00709)	0.056 (0.119)	0.00138 (0.00427)	0.0242 (0.0597)	0.000229 (0.000811)
PATHOLOGIC STAGE	Fisher's exact test	3e-05 (0.000121)	1e-05 (4.25e-05)	0.00025 (0.000867)	0.00084 (0.0028)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)
PATHOLOGY T STAGE	Fisher's exact test	0.177 (0.312)	0.00242 (0.00709)	0.0117 (0.0302)	0.00017 (0.000628)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	5e-05 (0.000189)	5e-05 (0.000189)	4e-05 (0.000158)	3e-05 (0.000121)
PATHOLOGY N STAGE	Fisher's exact test	1 (1.00)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	0.0084 (0.0227)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	0.00064 (0.00218)	1e-05 (4.25e-05)
PATHOLOGY M STAGE	Fisher's exact test	0.584 (0.764)	0.636 (0.787)	0.67 (0.802)	0.756 (0.87)	0.758 (0.87)	0.403 (0.576)	0.466 (0.619)	0.336 (0.495)	0.436 (0.587)	0.372 (0.535)
GENDER	Fisher's exact test	0.438 (0.587)	0.91 (0.973)	0.434 (0.587)	0.942 (0.98)	0.875 (0.953)	0.174 (0.311)	0.841 (0.935)	0.636 (0.787)	0.948 (0.98)	0.213 (0.356)
RADIATION THERAPY	Fisher's exact test	0.102 (0.201)	0.857 (0.94)	0.0982 (0.196)	0.743 (0.865)	0.436 (0.587)	0.212 (0.356)	1 (1.00)	0.711 (0.838)	0.286 (0.434)	0.934 (0.98)
HISTOLOGICAL TYPE	Fisher's exact test	0.665 (0.802)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)
RADIATION EXPOSURE	Fisher's exact test	0.621 (0.784)	0.647 (0.792)	0.974 (0.993)	0.235 (0.381)	0.984 (0.996)	0.253 (0.398)	0.951 (0.98)	0.905 (0.973)	0.411 (0.578)	0.852 (0.94)
EXTRATHYROIDAL EXTENSION	Fisher's exact test	0.131 (0.253)	1e-05 (4.25e-05)	0.00585 (0.0163)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)	1e-05 (4.25e-05)
RESIDUAL TUMOR	Fisher's exact test	0.124 (0.242)	0.233 (0.381)	0.134 (0.256)	0.0936 (0.192)	0.151 (0.283)	0.0198 (0.0503)	0.178 (0.312)	0.193 (0.331)	0.029 (0.0696)	0.0474 (0.102)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.585 (0.764)	2.03e-08 (6.91e-07)	0.0617 (0.13)	0.433 (0.587)	8.35e-09 (3.77e-07)	8.87e-09 (3.77e-07)	8.43e-09 (3.77e-07)	4.47e-09 (3.77e-07)	0.0113 (0.0295)	1.43e-07 (3.53e-06)
MULTIFOCALITY	Fisher's exact test	0.0968 (0.196)	0.943 (0.98)	0.158 (0.289)	0.18 (0.313)	0.27 (0.414)	0.244 (0.391)	0.817 (0.92)	0.603 (0.777)	0.841 (0.935)	0.639 (0.787)
TUMOR SIZE	Kruskal-Wallis (anova)	0.00426 (0.0121)	0.622 (0.784)	0.768 (0.876)	0.0203 (0.0506)	0.0428 (0.0944)	0.00195 (0.00593)	0.032 (0.0735)	0.0378 (0.0846)	0.161 (0.291)	0.028 (0.068)
RACE	Fisher's exact test	0.256 (0.4)	0.654 (0.794)	0.301 (0.449)	0.295 (0.444)	0.71 (0.838)	0.00106 (0.0034)	0.361 (0.525)	0.202 (0.343)	0.0461 (0.101)	0.087 (0.18)
ETHNICITY	Fisher's exact test	0.885 (0.958)	0.433 (0.587)	0.609 (0.778)	0.936 (0.98)	0.00695 (0.0191)	0.0314 (0.0731)	0.715 (0.838)	0.797 (0.904)	0.975 (0.993)	0.251 (0.398)

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	56	360	83

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

P value = 0.338 (logrank test), Q value = 0.49

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	497	16	0.2 - 169.3 (30.6)
subtype1	55	3	1.0 - 136.0 (30.6)
subtype2	359	12	0.2 - 169.3 (30.0)
subtype3	83	1	0.2 - 139.9 (32.5)

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 'YEARS_TO_BIRTH'

P value = 1.45e-07 (Kruskal-Wallis (anova)), Q value = 3.5e-06

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

	nPatients	Mean (Std.Dev)
ALL	499	47.3 (15.8)
subtype1	56	58.7 (14.7)
subtype2	360	45.7 (15.7)
subtype3	83	46.4 (13.5)

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 3e-05 (Fisher's exact test), Q value = 0.00012

Table S4. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	282	50	111	2	46	6
subtype1	14	14	16	1	10	1
subtype2	219	31	78	0	26	4
subtype3	49	5	17	1	10	1

Figure S3. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.177 (Fisher's exact test), Q value = 0.31

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

nPatients	T1	T2	T3	T4
ALL	142	165	167	23
subtype1	8	21	22	5
subtype2	109	115	119	15
subtype3	25	29	26	3

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

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

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

nPatients	0	1
ALL	225	224
subtype1	24	23
subtype2	162	162
subtype3	39	39

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

'Copy Number Ratio CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.584 (Fisher's exact test), Q value = 0.76

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

nPatients	0	1
ALL	278	9
subtype1	24	1
subtype2	209	6
subtype3	45	2

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

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.438 (Fisher's exact test), Q value = 0.59

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

nPatients	FEMALE	MALE
ALL	365	134
subtype1	37	19
subtype2	267	93
subtype3	61	22

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

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.102 (Fisher's exact test), Q value = 0.2

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

nPatients	NO	YES
ALL	185	290
subtype1	21	30
subtype2	142	204
subtype3	22	56

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

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 0.665 (Fisher's exact test), Q value = 0.8

Table S10. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	355	100	36
subtype1	0	40	12	4
subtype2	7	257	67	29
subtype3	1	58	21	3

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

'Copy Number Ratio CNMF subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.621 (Fisher's exact test), Q value = 0.78

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

nPatients	NO	YES
ALL	420	17
subtype1	49	3
subtype2	299	12
subtype3	72	2

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

'Copy Number Ratio CNMF subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 0.131 (Fisher's exact test), Q value = 0.25

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

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	132	18	330	1
subtype1	15	4	34	1
subtype2	99	13	237	0
subtype3	18	1	59	0

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

'Copy Number Ratio CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.124 (Fisher's exact test), Q value = 0.24

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

nPatients	R0	R1	R2	RX
ALL	383	51	4	30
subtype1	38	9	0	6
subtype2	284	35	3	16
subtype3	61	7	1	8

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

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.585 (Kruskal-Wallis (anova)), Q value = 0.76

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

	nPatients	Mean (Std.Dev)
ALL	385	3.6 (6.0)
subtype1	42	2.8 (4.9)
subtype2	275	3.6 (6.2)
subtype3	68	3.7 (5.6)

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

'Copy Number Ratio CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.0968 (Fisher's exact test), Q value = 0.2

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

nPatients	MULTIFOCAL	UNIFOCAL
ALL	225	264
subtype1	21	35
subtype2	159	193
subtype3	45	36

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

'Copy Number Ratio CNMF subtypes' versus 'TUMOR_SIZE'

P value = 0.00426 (Kruskal-Wallis (anova)), Q value = 0.012

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

	nPatients	Mean (Std.Dev)
ALL	401	3.0 (1.6)
subtype1	50	3.6 (1.6)
subtype2	285	2.9 (1.6)
subtype3	66	2.9 (1.5)

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

'Copy Number Ratio CNMF subtypes' versus 'RACE'

P value = 0.256 (Fisher's exact test), Q value = 0.4

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

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	52	27	328
subtype1	0	1	4	39
subtype2	1	41	18	237
subtype3	0	10	5	52

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

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

P value = 0.885 (Fisher's exact test), Q value = 0.96

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

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	38	359
subtype1	4	37
subtype2	27	265
subtype3	7	57

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

Clustering Approach #2: 'METHLYATION CNMF'

Table S19. Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3
Number of samples	277	156	70

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.223 (logrank test), Q value = 0.37

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	501	16	0.2 - 169.3 (30.6)
subtype1	277	7	0.2 - 169.3 (32.0)
subtype2	154	5	0.2 - 166.6 (23.4)
subtype3	70	4	0.3 - 147.4 (31.4)

Figure S18. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.156 (Kruskal-Wallis (anova)), Q value = 0.29

Table S21. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	503	47.3 (15.8)
subtype1	277	47.1 (15.5)
subtype2	156	48.7 (15.2)
subtype3	70	45.0 (17.9)

Figure S19. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'METHLYATION CNMF' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	284	51	111	2	47	6
subtype1	148	19	68	0	37	4
subtype2	90	30	28	1	4	2
subtype3	46	2	15	1	6	0

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.00242 (Fisher's exact test), Q value = 0.0071

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	143	166	169	23
subtype1	65	88	103	20
subtype2	53	58	44	1
subtype3	25	20	22	2

Figure S21. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	227	226
subtype1	96	160
subtype2	101	29
subtype3	30	37

Figure S22. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.636 (Fisher's exact test), Q value = 0.79

Table S25. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	280	9
subtype1	165	6
subtype2	74	3
subtype3	41	0

Figure S23. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.91 (Fisher's exact test), Q value = 0.97

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	368	135
subtype1	204	73
subtype2	112	44
subtype3	52	18

Figure S24. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

'METHLYATION CNMF' versus 'RADIATION_THERAPY'

P value = 0.857 (Fisher's exact test), Q value = 0.94

Table S27. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	186	293
subtype1	102	168
subtype2	57	86
subtype3	27	39

Figure S25. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

'METHLYATION CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S28. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	357	102	36
subtype1	4	226	18	29
subtype2	2	75	79	0
subtype3	2	56	5	7

Figure S26. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'METHLYATION CNMF' versus 'RADIATION_EXPOSURE'

P value = 0.647 (Fisher's exact test), Q value = 0.79

Table S29. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	423	17
subtype1	239	8
subtype2	125	6
subtype3	59	3

Figure S27. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'METHLYATION CNMF' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S30. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	133	18	333	1
subtype1	93	17	161	0
subtype2	25	0	122	0
subtype3	15	1	50	1

Figure S28. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'METHLYATION CNMF' versus 'RESIDUAL_TUMOR'

P value = 0.233 (Fisher's exact test), Q value = 0.38

Table S31. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	385	52	4	30
subtype1	208	36	4	18
subtype2	124	9	0	8
subtype3	53	7	0	4

Figure S29. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'METHLYATION CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 2.03e-08 (Kruskal-Wallis (anova)), Q value = 6.9e-07

Table S32. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	388	3.7 (6.2)
subtype1	229	4.4 (6.6)
subtype2	101	1.7 (4.3)
subtype3	58	3.9 (6.7)

Figure S30. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'METHLYATION CNMF' versus 'MULTIFOCALITY'

P value = 0.943 (Fisher's exact test), Q value = 0.98

Table S33. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	227	266
subtype1	126	146
subtype2	71	82
subtype3	30	38

Figure S31. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

'METHLYATION CNMF' versus 'TUMOR_SIZE'

P value = 0.622 (Kruskal-Wallis (anova)), Q value = 0.78

Table S34. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	403	3.0 (1.6)
subtype1	224	3.0 (1.6)
subtype2	124	3.1 (1.7)
subtype3	55	2.8 (1.4)

Figure S32. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'TUMOR_SIZE'

'METHLYATION CNMF' versus 'RACE'

P value = 0.654 (Fisher's exact test), Q value = 0.79

Table S35. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	52	27	330
subtype1	1	30	17	191
subtype2	0	14	9	86
subtype3	0	8	1	53

Figure S33. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #16: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.433 (Fisher's exact test), Q value = 0.59

Table S36. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	38	361
subtype1	24	208
subtype2	7	100
subtype3	7	53

Figure S34. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #3: 'RPPA CNMF subtypes'

Table S37. Description of clustering approach #3: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3	4
Number of samples	64	64	43	51

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.262 (logrank test), Q value = 0.4

Table S38. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	222	14	1.2 - 166.6 (35.5)
subtype1	64	6	1.2 - 158.8 (33.4)
subtype2	64	2	4.0 - 147.4 (40.6)
subtype3	43	2	3.0 - 166.6 (32.5)
subtype4	51	4	2.7 - 147.8 (37.5)

Figure S35. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.00137 (Kruskal-Wallis (anova)), Q value = 0.0043

Table S39. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	222	48.1 (16.8)
subtype1	64	51.8 (14.9)
subtype2	64	51.5 (15.1)
subtype3	43	43.6 (18.0)
subtype4	51	42.8 (18.1)

Figure S36. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.00025 (Fisher's exact test), Q value = 0.00087

Table S40. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	117	33	46	20	4
subtype1	32	12	13	3	2
subtype2	22	11	19	12	0
subtype3	30	8	3	2	0
subtype4	33	2	11	3	2

Figure S37. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.0117 (Fisher's exact test), Q value = 0.03

Table S41. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	51	83	76	11
subtype1	23	23	16	1
subtype2	12	19	27	6
subtype3	9	22	12	0
subtype4	7	19	21	4

Figure S38. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S42. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	98	95
subtype1	37	19
subtype2	25	31
subtype3	25	11
subtype4	11	34

Figure S39. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RPPA CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.67 (Fisher's exact test), Q value = 0.8

Table S43. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	118	5
subtype1	38	2
subtype2	35	1
subtype3	21	0
subtype4	24	2

Figure S40. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.434 (Fisher's exact test), Q value = 0.59

Table S44. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	154	68
subtype1	44	20
subtype2	47	17
subtype3	32	11
subtype4	31	20

Figure S41. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.0982 (Fisher's exact test), Q value = 0.2

Table S45. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	92	125
subtype1	32	30
subtype2	25	39
subtype3	20	21
subtype4	15	35

Figure S42. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RPPA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S46. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	2	158	52	10
subtype1	0	36	27	1
subtype2	0	44	13	7
subtype3	1	32	10	0
subtype4	1	46	2	2

Figure S43. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RPPA CNMF subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.974 (Fisher's exact test), Q value = 0.99

Table S47. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	184	11
subtype1	55	4
subtype2	55	3
subtype3	29	2
subtype4	45	2

Figure S44. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'RPPA CNMF subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 0.00585 (Fisher's exact test), Q value = 0.016

Table S48. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	55	10	149
subtype1	12	1	51
subtype2	22	6	34
subtype3	5	0	32
subtype4	16	3	32

Figure S45. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'RPPA CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.134 (Fisher's exact test), Q value = 0.26

Table S49. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	168	23	2	14
subtype1	50	6	1	6
subtype2	45	12	0	4
subtype3	34	2	0	0
subtype4	39	3	1	4

Figure S46. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RPPA CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0617 (Kruskal-Wallis (anova)), Q value = 0.13

Table S50. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	167	3.5 (5.8)
subtype1	47	3.7 (7.2)
subtype2	49	3.0 (4.5)
subtype3	26	2.7 (4.4)
subtype4	45	4.4 (6.3)

Figure S47. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RPPA CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.158 (Fisher's exact test), Q value = 0.29

Table S51. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	100	115
subtype1	34	27
subtype2	22	40
subtype3	20	22
subtype4	24	26

Figure S48. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'RPPA CNMF subtypes' versus 'TUMOR_SIZE'

P value = 0.768 (Kruskal-Wallis (anova)), Q value = 0.88

Table S52. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	191	3.3 (1.6)
subtype1	55	3.2 (1.6)
subtype2	52	3.4 (1.6)
subtype3	36	3.2 (1.3)
subtype4	48	3.4 (1.5)

Figure S49. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.301 (Fisher's exact test), Q value = 0.45

Table S53. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #16: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	16	13	149
subtype1	3	7	36
subtype2	3	3	48
subtype3	4	1	28
subtype4	6	2	37

Figure S50. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #16: 'RACE'

'RPPA CNMF subtypes' versus 'ETHNICITY'

P value = 0.609 (Fisher's exact test), Q value = 0.78

Table S54. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	16	168
subtype1	3	45
subtype2	5	49
subtype3	5	31
subtype4	3	43

Figure S51. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #4: 'RPPA cHierClus subtypes'

Table S55. Description of clustering approach #4: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3	4	5	6	7
Number of samples	47	28	33	29	49	27	9

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.000213 (logrank test), Q value = 0.00077

Table S56. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	222	14	1.2 - 166.6 (35.5)
subtype1	47	5	1.2 - 166.6 (33.5)
subtype2	28	1	6.0 - 95.3 (33.3)
subtype3	33	0	4.0 - 139.0 (47.4)
subtype4	29	2	3.0 - 155.5 (22.5)
subtype5	49	1	2.7 - 150.5 (37.7)
subtype6	27	1	6.0 - 112.8 (34.4)
subtype7	9	4	17.5 - 147.4 (42.1)

Figure S52. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.000996 (Kruskal-Wallis (anova)), Q value = 0.0033

Table S57. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	222	48.1 (16.8)
subtype1	47	52.6 (16.2)
subtype2	28	49.6 (15.4)
subtype3	33	47.6 (15.3)
subtype4	29	45.3 (19.2)
subtype5	49	40.1 (14.7)
subtype6	27	52.2 (16.7)
subtype7	9	61.8 (14.9)

Figure S53. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RPPA cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 0.00084 (Fisher's exact test), Q value = 0.0028

Table S58. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	117	33	46	20	4
subtype1	26	11	6	1	1
subtype2	16	3	7	2	0
subtype3	13	4	11	5	0
subtype4	17	4	7	0	1
subtype5	35	3	6	4	1
subtype6	9	7	7	4	0
subtype7	1	1	2	4	1

Figure S54. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 0.00017 (Fisher's exact test), Q value = 0.00063

Table S59. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	51	83	76	11
subtype1	16	22	9	0
subtype2	11	9	8	0
subtype3	3	11	17	2
subtype4	4	11	12	1
subtype5	11	21	16	1
subtype6	5	9	11	2
subtype7	1	0	3	5

Figure S55. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.0084 (Fisher's exact test), Q value = 0.023

Table S60. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	98	95
subtype1	28	11
subtype2	13	13
subtype3	10	22
subtype4	17	9
subtype5	16	26
subtype6	11	10
subtype7	3	4

Figure S56. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RPPA cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.756 (Fisher's exact test), Q value = 0.87

Table S61. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	118	5
subtype1	21	1
subtype2	19	0
subtype3	21	1
subtype4	13	1
subtype5	26	1
subtype6	11	0
subtype7	7	1

Figure S57. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.942 (Fisher's exact test), Q value = 0.98

Table S62. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	154	68
subtype1	31	16
subtype2	18	10
subtype3	25	8
subtype4	21	8
subtype5	33	16
subtype6	20	7
subtype7	6	3

Figure S58. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.743 (Fisher's exact test), Q value = 0.87

Table S63. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	92	125
subtype1	22	23
subtype2	14	14
subtype3	14	19
subtype4	11	16
subtype5	20	28
subtype6	8	19
subtype7	3	6

Figure S59. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S64. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	2	158	52	10
subtype1	0	22	25	0
subtype2	0	22	5	1
subtype3	0	25	0	8
subtype4	0	22	7	0
subtype5	1	46	2	0
subtype6	1	12	13	1
subtype7	0	9	0	0

Figure S60. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RPPA cHierClus subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.235 (Fisher's exact test), Q value = 0.38

Table S65. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	184	11
subtype1	39	3
subtype2	26	0
subtype3	29	0
subtype4	22	2
subtype5	39	2
subtype6	23	3
subtype7	6	1

Figure S61. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'RPPA cHierClus subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S66. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	55	10	149
subtype1	4	0	42
subtype2	9	0	19
subtype3	15	3	14
subtype4	6	1	18
subtype5	12	0	36
subtype6	6	2	19
subtype7	3	4	1

Figure S62. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'RPPA cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.0936 (Fisher's exact test), Q value = 0.19

Table S67. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	168	23	2	14
subtype1	37	5	0	3
subtype2	22	3	0	2
subtype3	24	7	0	0
subtype4	19	2	2	1
subtype5	39	3	0	3
subtype6	23	1	0	3
subtype7	4	2	0	2

Figure S63. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RPPA cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.433 (Kruskal-Wallis (anova)), Q value = 0.59

Table S68. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	167	3.5 (5.8)
subtype1	30	3.0 (6.7)
subtype2	25	3.9 (7.0)
subtype3	26	3.3 (3.9)
subtype4	22	2.8 (4.2)
subtype5	38	4.0 (6.6)
subtype6	19	3.7 (5.9)
subtype7	7	4.0 (5.3)

Figure S64. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RPPA cHierClus subtypes' versus 'MULTIFOCALITY'

P value = 0.18 (Fisher's exact test), Q value = 0.31

Table S69. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	100	115
subtype1	27	17
subtype2	12	16
subtype3	12	20
subtype4	12	17
subtype5	25	22
subtype6	10	16
subtype7	2	7

Figure S65. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'RPPA cHierClus subtypes' versus 'TUMOR_SIZE'

P value = 0.0203 (Kruskal-Wallis (anova)), Q value = 0.051

Table S70. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	191	3.3 (1.6)
subtype1	38	3.4 (1.5)
subtype2	27	2.6 (1.6)
subtype3	29	3.3 (1.3)
subtype4	24	3.6 (1.5)
subtype5	45	3.1 (1.5)
subtype6	22	3.3 (1.6)
subtype7	6	5.2 (1.8)

Figure S66. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 0.295 (Fisher's exact test), Q value = 0.44

Table S71. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #16: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	16	13	149
subtype1	1	6	25
subtype2	2	1	19
subtype3	3	0	28
subtype4	2	1	20
subtype5	7	2	33
subtype6	1	2	16
subtype7	0	1	8

Figure S67. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #16: 'RACE'

'RPPA cHierClus subtypes' versus 'ETHNICITY'

P value = 0.936 (Fisher's exact test), Q value = 0.98

Table S72. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	16	168
subtype1	2	32
subtype2	1	23
subtype3	3	28
subtype4	2	22
subtype5	5	38
subtype6	2	17
subtype7	1	8

Figure S68. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #5: 'RNAseq CNMF subtypes'

Table S73. Description of clustering approach #5: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3	4
Number of samples	168	157	59	117

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.591 (logrank test), Q value = 0.77

Table S74. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	499	16	0.2 - 169.3 (30.5)
subtype1	168	7	0.2 - 155.5 (29.8)
subtype2	156	5	0.2 - 166.6 (23.7)
subtype3	59	2	0.3 - 157.2 (31.0)
subtype4	116	2	0.9 - 169.3 (33.2)

Figure S69. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.141 (Kruskal-Wallis (anova)), Q value = 0.27

Table S75. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	501	47.3 (15.8)
subtype1	168	48.3 (15.8)
subtype2	157	48.7 (15.3)
subtype3	59	45.3 (18.9)
subtype4	117	44.8 (14.5)

Figure S70. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S76. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	283	51	110	2	47	6
subtype1	84	6	46	1	29	2
subtype2	91	29	28	1	5	2
subtype3	41	1	13	0	4	0
subtype4	67	15	23	0	9	2

Figure S71. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S77. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	142	165	169	23
subtype1	28	46	78	15
subtype2	53	57	45	2
subtype3	27	14	15	2
subtype4	34	48	31	4

Figure S72. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S78. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	226	225
subtype1	51	102
subtype2	102	29
subtype3	23	36
subtype4	50	58

Figure S73. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RNAseq CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.758 (Fisher's exact test), Q value = 0.87

Table S79. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	279	9
subtype1	95	3
subtype2	71	3
subtype3	41	0
subtype4	72	3

Figure S74. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.875 (Fisher's exact test), Q value = 0.95

Table S80. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	366	135
subtype1	121	47
subtype2	113	44
subtype3	43	16
subtype4	89	28

Figure S75. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.436 (Fisher's exact test), Q value = 0.59

Table S81. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	185	292
subtype1	56	107
subtype2	58	85
subtype3	26	31
subtype4	45	69

Figure S76. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S82. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	356	102	35
subtype1	4	135	3	26
subtype2	2	71	83	1
subtype3	1	50	3	5
subtype4	1	100	13	3

Figure S77. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RNAseq CNMF subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.984 (Fisher's exact test), Q value = 1

Table S83. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	421	17
subtype1	144	5
subtype2	129	6
subtype3	49	2
subtype4	99	4

Figure S78. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'RNAseq CNMF subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S84. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	133	18	331	1
subtype1	67	14	82	1
subtype2	24	1	123	0
subtype3	13	2	41	0
subtype4	29	1	85	0

Figure S79. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'RNAseq CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.151 (Fisher's exact test), Q value = 0.28

Table S85. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	383	52	4	30
subtype1	127	26	2	9
subtype2	125	9	0	8
subtype3	41	7	0	5
subtype4	90	10	2	8

Figure S80. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RNAseq CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 8.35e-09 (Kruskal-Wallis (anova)), Q value = 3.8e-07

Table S86. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	386	3.7 (6.2)
subtype1	139	4.4 (6.5)
subtype2	105	1.7 (4.3)
subtype3	52	5.2 (7.4)
subtype4	90	4.0 (6.3)

Figure S81. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RNAseq CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.27 (Fisher's exact test), Q value = 0.41

Table S87. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	226	265
subtype1	67	99
subtype2	74	80
subtype3	26	31
subtype4	59	55

Figure S82. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'RNAseq CNMF subtypes' versus 'TUMOR_SIZE'

P value = 0.0428 (Kruskal-Wallis (anova)), Q value = 0.094

Table S88. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	402	3.0 (1.6)
subtype1	142	3.2 (1.6)
subtype2	124	3.1 (1.6)
subtype3	43	2.5 (1.7)
subtype4	93	2.8 (1.5)

Figure S83. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 0.71 (Fisher's exact test), Q value = 0.84

Table S89. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	51	27	330
subtype1	1	17	11	120
subtype2	0	11	9	89
subtype3	0	7	1	45
subtype4	0	16	6	76

Figure S84. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #16: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.00695 (Fisher's exact test), Q value = 0.019

Table S90. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	38	359
subtype1	23	121
subtype2	9	99
subtype3	3	49
subtype4	3	90

Figure S85. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #6: 'RNAseq cHierClus subtypes'

Table S91. Description of clustering approach #6: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	105	136	86	96	78

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.411 (logrank test), Q value = 0.58

Table S92. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	499	16	0.2 - 169.3 (30.5)
subtype1	105	5	0.2 - 157.2 (34.4)
subtype2	135	6	0.2 - 166.6 (22.5)
subtype3	86	1	0.2 - 142.5 (30.8)
subtype4	95	2	0.9 - 169.3 (33.5)
subtype5	78	2	0.3 - 150.5 (25.7)

Figure S86. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.00239 (Kruskal-Wallis (anova)), Q value = 0.0071

Table S93. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	501	47.3 (15.8)
subtype1	105	51.0 (16.1)
subtype2	136	49.7 (16.0)
subtype3	86	44.5 (15.1)
subtype4	96	45.3 (14.6)
subtype5	78	43.5 (15.8)

Figure S87. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'RNAseq cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S94. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	283	51	110	2	47	6
subtype1	46	2	39	1	16	1
subtype2	77	29	22	1	4	2
subtype3	57	3	18	0	8	0
subtype4	54	12	19	0	8	2
subtype5	49	5	12	0	11	1

Figure S88. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S95. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	142	165	169	23
subtype1	17	21	57	10
subtype2	43	55	36	2
subtype3	38	22	23	1
subtype4	26	41	25	4
subtype5	18	26	28	6

Figure S89. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S96. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	226	225
subtype1	29	68
subtype2	91	19
subtype3	38	45
subtype4	43	45
subtype5	25	48

Figure S90. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.403 (Fisher's exact test), Q value = 0.58

Table S97. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	279	9
subtype1	56	1
subtype2	57	3
subtype3	60	0
subtype4	58	3
subtype5	48	2

Figure S91. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.174 (Fisher's exact test), Q value = 0.31

Table S98. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	366	135
subtype1	84	21
subtype2	96	40
subtype3	61	25
subtype4	74	22
subtype5	51	27

Figure S92. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.212 (Fisher's exact test), Q value = 0.36

Table S99. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	185	292
subtype1	32	72
subtype2	48	78
subtype3	38	41
subtype4	38	56
subtype5	29	45

Figure S93. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S100. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	356	102	35
subtype1	1	79	2	23
subtype2	2	57	77	0
subtype3	2	68	10	6
subtype4	0	83	11	2
subtype5	3	69	2	4

Figure S94. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'RNAseq cHierClus subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.253 (Fisher's exact test), Q value = 0.4

Table S101. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	421	17
subtype1	95	3
subtype2	111	5
subtype3	66	6
subtype4	85	1
subtype5	64	2

Figure S95. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'RNAseq cHierClus subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S102. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	133	18	331	1
subtype1	53	10	38	1
subtype2	16	1	110	0
subtype3	20	1	62	0
subtype4	22	1	71	0
subtype5	22	5	50	0

Figure S96. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'RNAseq cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.0198 (Fisher's exact test), Q value = 0.05

Table S103. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	383	52	4	30
subtype1	70	23	2	6
subtype2	104	10	0	8
subtype3	70	4	0	6
subtype4	76	8	2	6
subtype5	63	7	0	4

Figure S97. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'RNAseq cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 8.87e-09 (Kruskal-Wallis (anova)), Q value = 3.8e-07

Table S104. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	386	3.7 (6.2)
subtype1	92	4.7 (6.6)
subtype2	87	1.4 (3.5)
subtype3	72	4.8 (8.5)
subtype4	72	3.8 (6.3)
subtype5	63	3.8 (4.6)

Figure S98. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RNAseq cHierClus subtypes' versus 'MULTIFOCALITY'

P value = 0.244 (Fisher's exact test), Q value = 0.39

Table S105. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	226	265
subtype1	39	66
subtype2	64	69
subtype3	43	41
subtype4	47	46
subtype5	33	43

Figure S99. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'RNAseq cHierClus subtypes' versus 'TUMOR_SIZE'

P value = 0.00195 (Kruskal-Wallis (anova)), Q value = 0.0059

Table S106. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	402	3.0 (1.6)
subtype1	93	3.1 (1.5)
subtype2	112	3.2 (1.5)
subtype3	62	2.3 (1.5)
subtype4	76	2.9 (1.5)
subtype5	59	3.2 (1.8)

Figure S100. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 0.00106 (Fisher's exact test), Q value = 0.0034

Table S107. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	51	27	330
subtype1	1	4	11	77
subtype2	0	7	8	75
subtype3	0	17	1	60
subtype4	0	12	6	60
subtype5	0	11	1	58

Figure S101. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #16: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.0314 (Fisher's exact test), Q value = 0.073

Table S108. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	38	359
subtype1	15	77
subtype2	9	82
subtype3	3	71
subtype4	3	71
subtype5	8	58

Figure S102. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #7: 'MIRSEQ CNMF'

Table S109. Description of clustering approach #7: 'MIRSEQ CNMF'

Cluster Labels	1	2	3
Number of samples	182	153	167

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.0336 (logrank test), Q value = 0.076

Table S110. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	500	16	0.2 - 169.3 (30.5)
subtype1	182	10	0.2 - 158.8 (31.4)
subtype2	152	5	0.5 - 166.6 (24.8)
subtype3	166	1	0.2 - 169.3 (30.2)

Figure S103. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.056 (Kruskal-Wallis (anova)), Q value = 0.12

Table S111. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	502	47.3 (15.8)
subtype1	182	49.4 (16.0)
subtype2	153	47.1 (16.2)
subtype3	167	45.2 (14.8)

Figure S104. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CNMF' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S112. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	283	51	111	2	47	6
subtype1	90	6	56	1	26	2
subtype2	94	28	22	1	5	2
subtype3	99	17	33	0	16	2

Figure S105. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 5e-05 (Fisher's exact test), Q value = 0.00019

Table S113. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	143	166	168	23
subtype1	56	38	71	15
subtype2	43	68	40	2
subtype3	44	60	57	6

Figure S106. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S114. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	227	225
subtype1	65	108
subtype2	97	28
subtype3	65	89

Figure S107. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.466 (Fisher's exact test), Q value = 0.62

Table S115. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	279	9
subtype1	122	2
subtype2	66	3
subtype3	91	4

Figure S108. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.841 (Fisher's exact test), Q value = 0.93

Table S116. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	367	135
subtype1	135	47
subtype2	109	44
subtype3	123	44

Figure S109. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY'

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

Table S117. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	186	292
subtype1	67	105
subtype2	56	88
subtype3	63	99

Figure S110. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRSEQ CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S118. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	356	102	36
subtype1	4	144	8	26
subtype2	2	67	84	0
subtype3	2	145	10	10

Figure S111. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRSEQ CNMF' versus 'RADIATION_EXPOSURE'

P value = 0.951 (Fisher's exact test), Q value = 0.98

Table S119. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	422	17
subtype1	154	7
subtype2	127	5
subtype3	141	5

Figure S112. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'MIRSEQ CNMF' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S120. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	132	18	333	1
subtype1	64	14	97	1
subtype2	17	0	128	0
subtype3	51	4	108	0

Figure S113. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'MIRSEQ CNMF' versus 'RESIDUAL_TUMOR'

P value = 0.178 (Fisher's exact test), Q value = 0.31

Table S121. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	384	52	4	30
subtype1	132	26	3	12
subtype2	120	10	0	10
subtype3	132	16	1	8

Figure S114. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRSEQ CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 8.43e-09 (Kruskal-Wallis (anova)), Q value = 3.8e-07

Table S122. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	387	3.6 (6.2)
subtype1	150	5.0 (7.3)
subtype2	101	1.6 (3.5)
subtype3	136	3.6 (6.0)

Figure S115. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CNMF' versus 'MULTIFOCALITY'

P value = 0.817 (Fisher's exact test), Q value = 0.92

Table S123. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	226	266
subtype1	82	99
subtype2	66	82
subtype3	78	85

Figure S116. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRSEQ CNMF' versus 'TUMOR_SIZE'

P value = 0.032 (Kruskal-Wallis (anova)), Q value = 0.074

Table S124. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	402	3.0 (1.6)
subtype1	142	2.8 (1.7)
subtype2	127	3.2 (1.5)
subtype3	133	2.9 (1.5)

Figure S117. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #15: 'TUMOR_SIZE'

'MIRSEQ CNMF' versus 'RACE'

P value = 0.361 (Fisher's exact test), Q value = 0.53

Table S125. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	52	27	330
subtype1	0	25	8	132
subtype2	0	9	10	87
subtype3	1	18	9	111

Figure S118. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #16: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

P value = 0.715 (Fisher's exact test), Q value = 0.84

Table S126. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	38	360
subtype1	13	144
subtype2	10	96
subtype3	15	120

Figure S119. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #8: 'MIRSEQ CHIERARCHICAL'

Table S127. Description of clustering approach #8: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3
Number of samples	179	130	193

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.00883 (logrank test), Q value = 0.023

Table S128. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	500	16	0.2 - 169.3 (30.5)
subtype1	179	10	0.2 - 158.8 (31.0)
subtype2	129	5	0.5 - 166.6 (22.2)
subtype3	192	1	0.2 - 169.3 (32.5)

Figure S120. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 0.00138 (Kruskal-Wallis (anova)), Q value = 0.0043

Table S129. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	502	47.3 (15.8)
subtype1	179	49.7 (16.2)
subtype2	130	49.1 (15.5)
subtype3	193	43.9 (15.0)

Figure S121. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S130. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	283	51	111	2	47	6
subtype1	88	5	57	1	25	2
subtype2	75	27	20	1	4	2
subtype3	120	19	34	0	18	2

Figure S122. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 5e-05 (Fisher's exact test), Q value = 0.00019

Table S131. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	143	166	168	23
subtype1	55	36	71	15
subtype2	41	53	35	1
subtype3	47	77	62	7

Figure S123. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S132. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	227	225
subtype1	66	103
subtype2	86	19
subtype3	75	103

Figure S124. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.336 (Fisher's exact test), Q value = 0.49

Table S133. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	279	9
subtype1	124	2
subtype2	55	3
subtype3	100	4

Figure S125. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.636 (Fisher's exact test), Q value = 0.79

Table S134. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	367	135
subtype1	135	44
subtype2	92	38
subtype3	140	53

Figure S126. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY'

P value = 0.711 (Fisher's exact test), Q value = 0.84

Table S135. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	186	292
subtype1	70	99
subtype2	45	76
subtype3	71	117

Figure S127. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S136. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	356	102	36
subtype1	3	141	9	26
subtype2	2	51	77	0
subtype3	3	164	16	10

Figure S128. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_EXPOSURE'

P value = 0.905 (Fisher's exact test), Q value = 0.97

Table S137. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	422	17
subtype1	153	6
subtype2	107	5
subtype3	162	6

Figure S129. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'MIRSEQ CHIERARCHICAL' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S138. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	132	18	333	1
subtype1	67	14	91	1
subtype2	15	0	107	0
subtype3	50	4	135	0

Figure S130. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'MIRSEQ CHIERARCHICAL' versus 'RESIDUAL_TUMOR'

P value = 0.193 (Fisher's exact test), Q value = 0.33

Table S139. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	384	52	4	30
subtype1	129	26	3	13
subtype2	101	9	0	7
subtype3	154	17	1	10

Figure S131. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_OF_LYMPH_NODES'

P value = 4.47e-09 (Kruskal-Wallis (anova)), Q value = 3.8e-07

Table S140. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	387	3.6 (6.2)
subtype1	144	4.7 (6.4)
subtype2	85	1.7 (4.6)
subtype3	158	3.8 (6.5)

Figure S132. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CHIERARCHICAL' versus 'MULTIFOCALITY'

P value = 0.603 (Fisher's exact test), Q value = 0.78

Table S141. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	226	266
subtype1	76	101
subtype2	61	67
subtype3	89	98

Figure S133. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRSEQ CHIERARCHICAL' versus 'TUMOR_SIZE'

P value = 0.0378 (Kruskal-Wallis (anova)), Q value = 0.085

Table S142. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	402	3.0 (1.6)
subtype1	139	2.8 (1.6)
subtype2	109	3.2 (1.6)
subtype3	154	3.0 (1.5)

Figure S134. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'TUMOR_SIZE'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 0.202 (Fisher's exact test), Q value = 0.34

Table S143. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	52	27	330
subtype1	0	26	7	130
subtype2	0	6	8	71
subtype3	1	20	12	129

Figure S135. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #16: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.797 (Fisher's exact test), Q value = 0.9

Table S144. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	38	360
subtype1	13	142
subtype2	8	76
subtype3	17	142

Figure S136. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #9: 'MIRseq Mature CNMF subtypes'

Table S145. Description of clustering approach #9: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	161	165	146

'MIRseq Mature CNMF subtypes' versus 'Time to Death'

P value = 0.0299 (logrank test), Q value = 0.07

Table S146. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	470	16	0.2 - 169.3 (30.2)
subtype1	161	10	0.2 - 157.2 (30.6)
subtype2	165	2	0.3 - 169.3 (31.6)
subtype3	144	4	0.4 - 150.5 (29.1)

Figure S137. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.0242 (Kruskal-Wallis (anova)), Q value = 0.06

Table S147. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	472	47.3 (15.8)
subtype1	161	49.7 (16.5)
subtype2	165	44.9 (15.8)
subtype3	146	47.4 (14.6)

Figure S138. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S148. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	267	49	104	2	42	6
subtype1	78	6	52	1	22	2
subtype2	105	27	21	0	9	2
subtype3	84	16	31	1	11	2

Figure S139. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 4e-05 (Fisher's exact test), Q value = 0.00016

Table S149. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	136	156	156	22
subtype1	42	34	69	15
subtype2	47	67	46	5
subtype3	47	55	41	2

Figure S140. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 0.00064 (Fisher's exact test), Q value = 0.0022

Table S150. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	213	212
subtype1	59	95
subtype2	88	58
subtype3	66	59

Figure S141. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRseq Mature CNMF subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.436 (Fisher's exact test), Q value = 0.59

Table S151. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	263	9
subtype1	108	2
subtype2	72	4
subtype3	83	3

Figure S142. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.948 (Fisher's exact test), Q value = 0.98

Table S152. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	345	127
subtype1	119	42
subtype2	119	46
subtype3	107	39

Figure S143. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.286 (Fisher's exact test), Q value = 0.43

Table S153. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	175	273
subtype1	61	92
subtype2	68	91
subtype3	46	90

Figure S144. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S154. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	340	92	32
subtype1	3	126	7	25
subtype2	2	113	47	3
subtype3	3	101	38	4

Figure S145. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRseq Mature CNMF subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.411 (Fisher's exact test), Q value = 0.58

Table S155. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	397	16
subtype1	135	5
subtype2	136	8
subtype3	126	3

Figure S146. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'MIRseq Mature CNMF subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S156. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	125	17	312	1
subtype1	60	14	79	1
subtype2	31	2	127	0
subtype3	34	1	106	0

Figure S147. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'MIRseq Mature CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.029 (Fisher's exact test), Q value = 0.07

Table S157. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	360	49	4	29
subtype1	112	25	3	9
subtype2	138	9	1	11
subtype3	110	15	0	9

Figure S148. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRseq Mature CNMF subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.0113 (Kruskal-Wallis (anova)), Q value = 0.029

Table S158. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	363	3.7 (6.2)
subtype1	136	4.6 (6.9)
subtype2	117	3.0 (4.7)
subtype3	110	3.4 (6.4)

Figure S149. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.841 (Fisher's exact test), Q value = 0.93

Table S159. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	215	249
subtype1	75	86
subtype2	71	88
subtype3	69	75

Figure S150. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRseq Mature CNMF subtypes' versus 'TUMOR_SIZE'

P value = 0.161 (Kruskal-Wallis (anova)), Q value = 0.29

Table S160. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	378	3.0 (1.6)
subtype1	126	2.9 (1.7)
subtype2	133	3.1 (1.5)
subtype3	119	2.9 (1.6)

Figure S151. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 0.0461 (Fisher's exact test), Q value = 0.1

Table S161. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	49	26	312
subtype1	0	18	6	122
subtype2	0	12	7	109
subtype3	1	19	13	81

Figure S152. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #16: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.975 (Fisher's exact test), Q value = 0.99

Table S162. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	35	342
subtype1	14	130
subtype2	12	118
subtype3	9	94

Figure S153. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #17: 'ETHNICITY'

Clustering Approach #10: 'MIRseq Mature cHierClus subtypes'

Table S163. Description of clustering approach #10: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	165	128	179

'MIRseq Mature cHierClus subtypes' versus 'Time to Death'

P value = 0.00422 (logrank test), Q value = 0.012

Table S164. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	470	16	0.2 - 169.3 (30.2)
subtype1	165	10	0.2 - 157.2 (31.0)
subtype2	127	5	0.5 - 166.6 (21.7)
subtype3	178	1	0.3 - 169.3 (32.8)

Figure S154. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'YEARS_TO_BIRTH'

P value = 0.000229 (Kruskal-Wallis (anova)), Q value = 0.00081

Table S165. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	472	47.3 (15.8)
subtype1	165	50.1 (16.3)
subtype2	128	49.2 (15.6)
subtype3	179	43.4 (14.6)

Figure S155. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGIC_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S166. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IV	STAGE IVA	STAGE IVC
ALL	267	49	104	2	42	6
subtype1	77	4	58	1	23	2
subtype2	71	26	24	1	3	2
subtype3	119	19	22	0	16	2

Figure S156. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_T_STAGE'

P value = 3e-05 (Fisher's exact test), Q value = 0.00012

Table S167. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T1	T2	T3	T4
ALL	136	156	156	22
subtype1	47	33	68	15
subtype2	39	54	34	1
subtype3	50	69	54	6

Figure S157. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_N_STAGE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S168. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	0	1
ALL	213	212
subtype1	59	98
subtype2	80	23
subtype3	74	91

Figure S158. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'PATHOLOGY_M_STAGE'

P value = 0.372 (Fisher's exact test), Q value = 0.54

Table S169. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	263	9
subtype1	114	2
subtype2	52	3
subtype3	97	4

Figure S159. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.213 (Fisher's exact test), Q value = 0.36

Table S170. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	345	127
subtype1	125	40
subtype2	86	42
subtype3	134	45

Figure S160. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.934 (Fisher's exact test), Q value = 0.98

Table S171. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	175	273
subtype1	60	94
subtype2	45	74
subtype3	70	105

Figure S161. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S172. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

nPatients	OTHER SPECIFY	THYROID PAPILLARY CARCINOMA - CLASSICAL/USUAL	THYROID PAPILLARY CARCINOMA - FOLLICULAR (>= 99% FOLLICULAR PATTERNED)	THYROID PAPILLARY CARCINOMA - TALL CELL (>= 50% TALL CELL FEATURES)
ALL	8	340	92	32
subtype1	3	130	7	25
subtype2	2	57	69	0
subtype3	3	153	16	7

Figure S162. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'HISTOLOGICAL_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATION_EXPOSURE'

P value = 0.852 (Fisher's exact test), Q value = 0.94

Table S173. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

nPatients	NO	YES
ALL	397	16
subtype1	139	6
subtype2	107	5
subtype3	151	5

Figure S163. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'RADIATION_EXPOSURE'

'MIRseq Mature cHierClus subtypes' versus 'EXTRATHYROIDAL_EXTENSION'

P value = 1e-05 (Fisher's exact test), Q value = 4.2e-05

Table S174. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE	VERY ADVANCED (T4B)
ALL	125	17	312	1
subtype1	63	14	81	1
subtype2	17	0	104	0
subtype3	45	3	127	0

Figure S164. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'EXTRATHYROIDAL_EXTENSION'

'MIRseq Mature cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.0474 (Fisher's exact test), Q value = 0.1

Table S175. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	360	49	4	29
subtype1	115	27	2	11
subtype2	98	9	0	9
subtype3	147	13	2	9

Figure S165. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_OF_LYMPH_NODES'

P value = 1.43e-07 (Kruskal-Wallis (anova)), Q value = 3.5e-06

Table S176. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	363	3.7 (6.2)
subtype1	136	4.6 (6.2)
subtype2	82	1.5 (3.2)
subtype3	145	4.1 (7.0)

Figure S166. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature cHierClus subtypes' versus 'MULTIFOCALITY'

P value = 0.639 (Fisher's exact test), Q value = 0.79

Table S177. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	215	249
subtype1	72	92
subtype2	62	63
subtype3	81	94

Figure S167. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRseq Mature cHierClus subtypes' versus 'TUMOR_SIZE'

P value = 0.028 (Kruskal-Wallis (anova)), Q value = 0.068

Table S178. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

	nPatients	Mean (Std.Dev)
ALL	378	3.0 (1.6)
subtype1	126	2.7 (1.6)
subtype2	107	3.2 (1.6)
subtype3	145	3.0 (1.6)

Figure S168. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'TUMOR_SIZE'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 0.087 (Fisher's exact test), Q value = 0.18

Table S179. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #16: 'RACE'

nPatients	AMERICAN INDIAN OR ALASKA NATIVE	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	1	49	26	312
subtype1	0	25	5	122
subtype2	0	7	9	70
subtype3	1	17	12	120

Figure S169. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #16: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.251 (Fisher's exact test), Q value = 0.4

Table S180. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #17: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	35	342
subtype1	10	136
subtype2	7	80
subtype3	18	126

Figure S170. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #17: 'ETHNICITY'

Methods & Data

Input

Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/THCA-TP/20140899/THCA-TP.mergedcluster.txt
Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/THCA-TP/19775590/THCA-TP.merged_data.txt
Number of patients = 503
Number of clustering approaches = 10
Number of selected clinical features = 17
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