Correlation between aggregated molecular cancer subtypes and selected clinical features

Esophageal Carcinoma (Primary solid tumor)

28 January 2016 | analyses__2016_01_28

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 (2016): Correlation between aggregated molecular cancer subtypes and selected clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1DV1J80

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 15 clinical features across 185 patients, 76 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', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'GENDER', 'RADIATION_THERAPY', 'HISTOLOGICAL_TYPE', and 'RACE'.
6 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'PATHOLOGY_M_STAGE', 'RADIATION_THERAPY', 'HISTOLOGICAL_TYPE', 'NUMBER_PACK_YEARS_SMOKED', and 'RACE'.
CNMF clustering analysis on RPPA data identified 5 subtypes that correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'HISTOLOGICAL_TYPE', and 'RACE'.
Consensus hierarchical clustering analysis on RPPA data identified 5 subtypes that correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'HISTOLOGICAL_TYPE', 'NUMBER_OF_LYMPH_NODES', and 'RACE'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 6 subtypes that correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'RADIATION_THERAPY', 'HISTOLOGICAL_TYPE', and 'RACE'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'RADIATION_THERAPY', 'KARNOFSKY_PERFORMANCE_SCORE', 'HISTOLOGICAL_TYPE', and 'RACE'.
6 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'YEARS_TO_BIRTH', 'PATHOLOGIC_STAGE', 'PATHOLOGY_T_STAGE', 'PATHOLOGY_N_STAGE', 'RADIATION_THERAPY', 'HISTOLOGICAL_TYPE', 'NUMBER_OF_LYMPH_NODES', and 'RACE'.
4 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', 'RADIATION_THERAPY', 'HISTOLOGICAL_TYPE', and 'RACE'.
4 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', 'RADIATION_THERAPY', 'KARNOFSKY_PERFORMANCE_SCORE', 'HISTOLOGICAL_TYPE', 'NUMBER_PACK_YEARS_SMOKED', 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', 'RADIATION_THERAPY', 'HISTOLOGICAL_TYPE', 'NUMBER_PACK_YEARS_SMOKED', and 'RACE'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 10 different clustering approaches and 15 clinical features. Shown in the table are P values (Q values). Thresholded by P value < 0.05 and Q value < 0.25, 76 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	0.151 (0.226)	0.182 (0.252)	0.176 (0.247)	0.425 (0.491)	0.84 (0.869)	0.832 (0.866)	0.0806 (0.141)	0.0182 (0.0434)	0.0448 (0.0908)	0.0328 (0.0712)
YEARS TO BIRTH	Kruskal-Wallis (anova)	1.66e-05 (9.2e-05)	1.57e-05 (9.05e-05)	0.0149 (0.0366)	0.000524 (0.00218)	2.03e-05 (0.000101)	6.96e-06 (6e-05)	0.00418 (0.0123)	1.85e-06 (6e-05)	8.25e-06 (6e-05)	7.31e-07 (6e-05)
PATHOLOGIC STAGE	Fisher's exact test	0.0113 (0.0287)	0.00014 (0.000636)	0.0313 (0.069)	0.00997 (0.0258)	0.0076 (0.0207)	2e-05 (0.000101)	0.0001 (0.000469)	2e-05 (0.000101)	1e-05 (6e-05)	1e-05 (6e-05)
PATHOLOGY T STAGE	Fisher's exact test	0.00062 (0.00238)	0.00054 (0.00219)	0.00231 (0.00737)	0.00308 (0.00962)	6e-05 (0.00029)	0.00177 (0.00603)	0.0008 (0.00293)	0.00089 (0.00318)	0.00078 (0.00292)	0.00058 (0.00229)
PATHOLOGY N STAGE	Fisher's exact test	0.0142 (0.0355)	0.0498 (0.0982)	0.284 (0.36)	0.0603 (0.113)	0.0655 (0.121)	0.0048 (0.0136)	0.00754 (0.0207)	0.0434 (0.0895)	0.104 (0.174)	0.00997 (0.0258)
PATHOLOGY M STAGE	Fisher's exact test	0.275 (0.356)	0.0436 (0.0895)	0.0823 (0.142)	0.0526 (0.0999)	0.784 (0.828)	0.343 (0.425)	0.618 (0.681)	0.204 (0.273)	0.235 (0.309)	0.372 (0.457)
GENDER	Fisher's exact test	0.0239 (0.0552)	0.157 (0.23)	0.65 (0.711)	0.132 (0.206)	0.936 (0.956)	0.91 (0.935)	0.379 (0.462)	0.488 (0.554)	0.987 (1)	0.392 (0.47)
RADIATION THERAPY	Fisher's exact test	0.034 (0.073)	0.00846 (0.0227)	0.495 (0.559)	0.0897 (0.151)	0.00198 (0.00646)	0.00017 (0.00075)	0.00466 (0.0134)	0.00105 (0.00366)	0.00317 (0.0097)	0.00364 (0.0109)
KARNOFSKY PERFORMANCE SCORE	Kruskal-Wallis (anova)	0.105 (0.174)	0.514 (0.575)	0.39 (0.47)	0.0874 (0.149)	0.262 (0.342)	0.0361 (0.0762)	0.327 (0.408)	0.116 (0.189)	0.0292 (0.0663)	0.0516 (0.0993)
HISTOLOGICAL TYPE	Fisher's exact test	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)
NUMBER PACK YEARS SMOKED	Kruskal-Wallis (anova)	0.121 (0.195)	0.0463 (0.0926)	0.395 (0.47)	0.318 (0.4)	0.0686 (0.124)	0.125 (0.199)	0.194 (0.265)	0.136 (0.21)	0.0166 (0.0402)	0.0185 (0.0434)
RESIDUAL TUMOR	Fisher's exact test	0.429 (0.491)	0.527 (0.585)	0.0702 (0.124)	0.152 (0.226)	0.702 (0.752)	0.14 (0.213)	0.187 (0.257)	0.679 (0.733)	0.127 (0.201)	0.669 (0.727)
NUMBER OF LYMPH NODES	Kruskal-Wallis (anova)	0.0699 (0.124)	0.175 (0.247)	0.412 (0.483)	0.0309 (0.069)	0.283 (0.36)	0.0507 (0.0988)	0.00193 (0.00643)	0.21 (0.278)	0.158 (0.231)	0.0665 (0.122)
RACE	Fisher's exact test	1e-05 (6e-05)	1e-05 (6e-05)	0.00038 (0.00163)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)	1e-05 (6e-05)
ETHNICITY	Fisher's exact test	1 (1.00)	0.825 (0.866)	0.198 (0.267)	0.775 (0.824)	0.404 (0.478)	0.173 (0.247)	1 (1.00)	0.176 (0.247)	0.418 (0.487)	0.141 (0.213)

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	74	33	77

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

P value = 0.151 (logrank test), Q value = 0.23

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	184	76	0.1 - 122.1 (13.0)
subtype1	74	39	0.4 - 122.1 (14.1)
subtype2	33	9	1.6 - 83.2 (15.3)
subtype3	77	28	0.1 - 52.6 (12.6)

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 = 1.66e-05 (Kruskal-Wallis (anova)), Q value = 9.2e-05

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

	nPatients	Mean (Std.Dev)
ALL	184	62.4 (11.9)
subtype1	74	63.9 (11.9)
subtype2	33	69.2 (12.2)
subtype3	77	58.1 (10.2)

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 = 0.0113 (Fisher's exact test), Q value = 0.029

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

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	8	5	6	1	47	31	25	14	9	7	5	4
subtype1	5	2	2	0	6	13	12	5	5	3	2	3
subtype2	3	2	1	0	9	7	3	3	0	2	0	0
subtype3	0	1	3	1	32	11	10	6	4	2	3	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 = 0.00062 (Fisher's exact test), Q value = 0.0024

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

nPatients	T0+T1	T2	T3	T4
ALL	32	43	87	5
subtype1	20	8	35	0
subtype2	6	8	14	2
subtype3	6	27	38	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 = 0.0142 (Fisher's exact test), Q value = 0.036

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

nPatients	N0	N1	N2	N3
ALL	77	68	12	8
subtype1	19	34	7	3
subtype2	19	8	0	2
subtype3	39	26	5	3

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 = 0.275 (Fisher's exact test), Q value = 0.36

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

nPatients	0	1
ALL	135	9
subtype1	43	5
subtype2	26	0
subtype3	66	4

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 = 0.0239 (Fisher's exact test), Q value = 0.055

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

nPatients	FEMALE	MALE
ALL	27	157
subtype1	7	67
subtype2	10	23
subtype3	10	67

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 = 0.034 (Fisher's exact test), Q value = 0.073

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

nPatients	NO	YES
ALL	123	43
subtype1	54	11
subtype2	23	6
subtype3	46	26

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 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.105 (Kruskal-Wallis (anova)), Q value = 0.17

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

	nPatients	Mean (Std.Dev)
ALL	68	73.8 (16.2)
subtype1	9	80.0 (15.0)
subtype2	9	81.1 (10.5)
subtype3	50	71.4 (16.8)

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

'Copy Number Ratio CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

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

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	88	96
subtype1	69	5
subtype2	17	16
subtype3	2	75

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

'Copy Number Ratio CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.121 (Kruskal-Wallis (anova)), Q value = 0.19

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

	nPatients	Mean (Std.Dev)
ALL	98	34.5 (21.5)
subtype1	38	39.7 (23.9)
subtype2	16	34.9 (21.8)
subtype3	44	29.8 (18.4)

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

'Copy Number Ratio CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.429 (Fisher's exact test), Q value = 0.49

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

nPatients	R0	R1	R2	RX
ALL	137	12	2	7
subtype1	49	7	0	3
subtype2	27	2	0	0
subtype3	61	3	2	4

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 = 0.0699 (Kruskal-Wallis (anova)), Q value = 0.12

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

	nPatients	Mean (Std.Dev)
ALL	134	2.5 (3.4)
subtype1	60	2.9 (3.4)
subtype2	25	2.4 (4.9)
subtype3	49	2.1 (2.5)

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 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

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

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	46	5	114
subtype1	3	0	55
subtype2	3	1	29
subtype3	40	4	30

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

'Copy Number Ratio CNMF subtypes' versus 'ETHNICITY'

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

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

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	88
subtype1	2	23
subtype2	0	12
subtype3	4	53

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

Clustering Approach #2: 'METHLYATION CNMF'

Table S17. Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3	4	5	6	7
Number of samples	37	42	2	41	37	17	9

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.182 (logrank test), Q value = 0.25

Table S18. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	183	77	0.1 - 122.1 (12.9)
subtype1	37	26	1.4 - 58.5 (13.5)
subtype2	42	16	0.4 - 122.1 (13.9)
subtype4	41	15	0.3 - 47.9 (13.2)
subtype5	37	12	0.1 - 68.0 (12.8)
subtype6	17	6	1.6 - 83.2 (12.6)
subtype7	9	2	2.6 - 23.4 (12.3)

Figure S16. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'YEARS_TO_BIRTH'

P value = 1.57e-05 (Kruskal-Wallis (anova)), Q value = 9e-05

Table S19. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	183	62.4 (11.9)
subtype1	37	64.3 (11.0)
subtype2	42	68.4 (10.1)
subtype4	41	55.2 (8.6)
subtype5	37	60.6 (11.4)
subtype6	17	64.3 (14.1)
subtype7	9	63.4 (18.0)

Figure S17. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'METHLYATION CNMF' versus 'PATHOLOGIC_STAGE'

P value = 0.00014 (Fisher's exact test), Q value = 0.00064

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	8	4	6	1	47	30	26	14	9	7	5	4
subtype1	2	2	1	0	0	7	6	3	1	1	2	3
subtype2	6	0	1	0	4	8	9	1	3	2	0	0
subtype4	0	1	1	0	18	5	5	2	5	1	1	1
subtype5	0	0	3	1	14	5	6	5	0	1	2	0
subtype6	0	1	0	0	8	3	0	2	0	2	0	0
subtype7	0	0	0	0	3	2	0	1	0	0	0	0

Figure S18. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 0.00054 (Fisher's exact test), Q value = 0.0022

Table S21. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	31	43	87	5
subtype1	12	3	14	0
subtype2	11	4	22	0
subtype4	4	13	22	1
subtype5	2	14	19	2
subtype6	2	7	5	2
subtype7	0	2	5	0

Figure S19. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.0498 (Fisher's exact test), Q value = 0.098

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	75	69	12	8
subtype1	7	19	2	1
subtype2	14	18	3	2
subtype4	18	15	5	1
subtype5	21	13	1	2
subtype6	11	2	0	2
subtype7	4	2	1	0

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'METHLYATION CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.0436 (Fisher's exact test), Q value = 0.09

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	135	9
subtype1	17	5
subtype2	29	0
subtype4	36	2
subtype5	32	2
subtype6	15	0
subtype7	6	0

Figure S21. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.157 (Fisher's exact test), Q value = 0.23

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	26	157
subtype1	1	36
subtype2	8	34
subtype4	5	36
subtype5	7	30
subtype6	3	14
subtype7	2	7

Figure S22. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'GENDER'

'METHLYATION CNMF' versus 'RADIATION_THERAPY'

P value = 0.00846 (Fisher's exact test), Q value = 0.023

Table S25. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	123	42
subtype1	31	3
subtype2	30	5
subtype4	24	13
subtype5	24	11
subtype6	10	5
subtype7	4	5

Figure S23. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

'METHLYATION CNMF' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.514 (Kruskal-Wallis (anova)), Q value = 0.58

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	67	73.7 (16.3)
subtype1	2	90.0 (0.0)
subtype2	4	82.5 (15.0)
subtype4	25	73.2 (17.3)
subtype5	25	72.0 (16.1)
subtype6	7	77.1 (13.8)
subtype7	4	65.0 (19.1)

Figure S24. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'METHLYATION CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S27. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	89	94
subtype1	37	0
subtype2	42	0
subtype4	0	41
subtype5	0	37
subtype6	7	10
subtype7	3	6

Figure S25. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'METHLYATION CNMF' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.0463 (Kruskal-Wallis (anova)), Q value = 0.093

Table S28. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	96	34.7 (21.7)
subtype1	14	46.9 (16.8)
subtype2	23	37.6 (26.1)
subtype4	26	30.2 (19.5)
subtype5	22	29.4 (17.5)
subtype6	7	41.3 (29.6)
subtype7	4	20.9 (13.0)

Figure S26. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'METHLYATION CNMF' versus 'RESIDUAL_TUMOR'

P value = 0.527 (Fisher's exact test), Q value = 0.59

Table S29. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	135	13	2	7
subtype1	21	5	0	0
subtype2	31	3	0	2
subtype4	34	3	1	1
subtype5	29	2	1	3
subtype6	15	0	0	0
subtype7	5	0	0	1

Figure S27. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'METHLYATION CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.175 (Kruskal-Wallis (anova)), Q value = 0.25

Table S30. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	134	2.5 (3.4)
subtype1	28	3.1 (3.0)
subtype2	37	2.2 (3.1)
subtype4	26	2.6 (2.4)
subtype5	27	1.7 (2.6)
subtype6	11	4.5 (7.5)
subtype7	5	1.4 (1.7)

Figure S28. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'METHLYATION CNMF' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S31. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	46	4	113
subtype1	0	0	29
subtype2	1	0	33
subtype4	24	3	12
subtype5	16	0	20
subtype6	3	0	14
subtype7	2	1	5

Figure S29. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'RACE'

'METHLYATION CNMF' versus 'ETHNICITY'

P value = 0.825 (Fisher's exact test), Q value = 0.87

Table S32. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	86
subtype1	2	12
subtype2	1	9
subtype4	2	29
subtype5	1	20
subtype6	0	9
subtype7	0	7

Figure S30. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #3: 'RPPA CNMF subtypes'

Table S33. Description of clustering approach #3: 'RPPA CNMF subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	26	28	30	34	8

'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.176 (logrank test), Q value = 0.25

Table S34. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	126	44	0.1 - 122.1 (13.2)
subtype1	26	13	2.1 - 46.1 (12.8)
subtype2	28	7	0.5 - 122.1 (13.7)
subtype3	30	11	0.3 - 52.6 (12.8)
subtype4	34	9	0.1 - 68.0 (12.6)
subtype5	8	4	4.5 - 46.2 (25.1)

Figure S31. 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 = 0.0149 (Kruskal-Wallis (anova)), Q value = 0.037

Table S35. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	126	62.7 (11.5)
subtype1	26	68.0 (10.6)
subtype2	28	63.2 (11.5)
subtype3	30	63.1 (10.1)
subtype4	34	58.1 (12.0)
subtype5	8	62.2 (11.2)

Figure S32. 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 = 0.0313 (Fisher's exact test), Q value = 0.069

Table S36. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	6	3	4	1	42	23	17	12	7	2	3	1
subtype1	3	2	1	0	3	5	4	3	2	1	0	1
subtype2	3	0	1	1	9	5	2	4	0	1	0	0
subtype3	0	1	1	0	6	7	6	2	3	0	2	0
subtype4	0	0	1	0	21	4	5	2	1	0	0	0
subtype5	0	0	0	0	3	2	0	1	1	0	1	0

Figure S33. 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 = 0.00231 (Fisher's exact test), Q value = 0.0074

Table S37. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	18	35	69	2
subtype1	9	4	13	0
subtype2	4	10	13	0
subtype3	4	4	20	1
subtype4	0	16	17	1
subtype5	1	1	6	0

Figure S34. 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 = 0.284 (Fisher's exact test), Q value = 0.36

Table S38. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	62	49	9	3
subtype1	9	13	3	1
subtype2	14	10	1	1
subtype3	12	13	3	1
subtype4	24	9	1	0
subtype5	3	4	1	0

Figure S35. 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 = 0.0823 (Fisher's exact test), Q value = 0.14

Table S39. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	105	4
subtype1	18	1
subtype2	25	0
subtype3	25	2
subtype4	32	0
subtype5	5	1

Figure S36. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.65 (Fisher's exact test), Q value = 0.71

Table S40. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	18	108
subtype1	4	22
subtype2	6	22
subtype3	3	27
subtype4	5	29
subtype5	0	8

Figure S37. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.495 (Fisher's exact test), Q value = 0.56

Table S41. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	84	32
subtype1	19	4
subtype2	19	6
subtype3	21	7
subtype4	20	12
subtype5	5	3

Figure S38. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RPPA CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.39 (Kruskal-Wallis (anova)), Q value = 0.47

Table S42. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	48	69.2 (16.2)
subtype1	4	60.0 (28.3)
subtype2	7	72.9 (19.8)
subtype3	11	69.1 (14.5)
subtype4	21	67.1 (13.5)
subtype5	5	80.0 (14.1)

Figure S39. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'RPPA CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S43. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	48	78
subtype1	22	4
subtype2	12	16
subtype3	10	20
subtype4	1	33
subtype5	3	5

Figure S40. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'RPPA CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.395 (Kruskal-Wallis (anova)), Q value = 0.47

Table S44. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	66	36.4 (20.6)
subtype1	11	34.7 (17.0)
subtype2	13	32.5 (19.2)
subtype3	17	48.0 (27.9)
subtype4	21	30.2 (14.1)
subtype5	4	37.2 (12.8)

Figure S41. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'RPPA CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.0702 (Fisher's exact test), Q value = 0.12

Table S45. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	101	10	1	7
subtype1	20	3	0	3
subtype2	27	0	0	0
subtype3	20	4	0	2
subtype4	29	2	0	2
subtype5	5	1	1	0

Figure S42. 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 = 0.412 (Kruskal-Wallis (anova)), Q value = 0.48

Table S46. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	97	2.2 (2.8)
subtype1	24	2.8 (3.3)
subtype2	20	1.8 (2.5)
subtype3	27	2.1 (3.0)
subtype4	19	1.6 (2.2)
subtype5	7	2.9 (2.3)

Figure S43. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RPPA CNMF subtypes' versus 'RACE'

P value = 0.00038 (Fisher's exact test), Q value = 0.0016

Table S47. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	43	2	73
subtype1	2	0	20
subtype2	7	1	18
subtype3	11	0	18
subtype4	21	0	12
subtype5	2	1	5

Figure S44. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #14: 'RACE'

'RPPA CNMF subtypes' versus 'ETHNICITY'

P value = 0.198 (Fisher's exact test), Q value = 0.27

Table S48. Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	2	68
subtype1	0	7
subtype2	0	14
subtype3	0	19
subtype4	1	25
subtype5	1	3

Figure S45. Get High-res Image Clustering Approach #3: 'RPPA CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #4: 'RPPA cHierClus subtypes'

Table S49. Description of clustering approach #4: 'RPPA cHierClus subtypes'

Cluster Labels	1	2	3	4	5
Number of samples	23	21	22	26	34

'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.425 (logrank test), Q value = 0.49

Table S50. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	126	44	0.1 - 122.1 (13.2)
subtype1	23	11	1.4 - 46.2 (18.4)
subtype2	21	6	0.5 - 122.1 (15.5)
subtype3	22	10	2.7 - 52.6 (13.0)
subtype4	26	7	0.3 - 52.3 (12.5)
subtype5	34	10	0.1 - 68.0 (12.9)

Figure S46. 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 = 0.000524 (Kruskal-Wallis (anova)), Q value = 0.0022

Table S51. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	126	62.7 (11.5)
subtype1	23	71.4 (9.4)
subtype2	21	64.0 (10.6)
subtype3	22	62.8 (10.0)
subtype4	26	58.1 (11.6)
subtype5	34	59.5 (11.2)

Figure S47. 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 = 0.00997 (Fisher's exact test), Q value = 0.026

Table S52. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	6	3	4	1	42	23	17	12	7	2	3	1
subtype1	3	2	1	0	1	5	3	2	2	2	0	0
subtype2	3	0	1	0	6	5	2	2	0	0	0	0
subtype3	0	0	0	0	4	5	4	3	2	0	2	1
subtype4	0	0	1	1	12	3	6	2	1	0	0	0
subtype5	0	1	1	0	19	5	2	3	2	0	1	0

Figure S48. 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 = 0.00308 (Fisher's exact test), Q value = 0.0096

Table S53. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	18	35	69	2
subtype1	9	3	10	0
subtype2	4	7	9	0
subtype3	3	3	15	1
subtype4	0	8	18	0
subtype5	2	14	17	1

Figure S49. 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 = 0.0603 (Fisher's exact test), Q value = 0.11

Table S54. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	62	49	9	3
subtype1	6	10	4	2
subtype2	11	9	0	0
subtype3	8	11	2	1
subtype4	14	10	1	0
subtype5	23	9	2	0

Figure S50. 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 = 0.0526 (Fisher's exact test), Q value = 0.1

Table S55. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	105	4
subtype1	13	0
subtype2	20	0
subtype3	16	3
subtype4	26	0
subtype5	30	1

Figure S51. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.132 (Fisher's exact test), Q value = 0.21

Table S56. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	18	108
subtype1	4	19
subtype2	5	16
subtype3	1	21
subtype4	1	25
subtype5	7	27

Figure S52. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RPPA cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.0897 (Fisher's exact test), Q value = 0.15

Table S57. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	84	32
subtype1	19	2
subtype2	15	4
subtype3	13	8
subtype4	13	10
subtype5	24	8

Figure S53. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RPPA cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.0874 (Kruskal-Wallis (anova)), Q value = 0.15

Table S58. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	48	69.2 (16.2)
subtype1	1	90.0 (NA)
subtype2	5	82.0 (13.0)
subtype3	6	66.7 (17.5)
subtype4	15	63.3 (15.4)
subtype5	21	70.0 (15.8)

Figure S54. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'RPPA cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S59. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	48	78
subtype1	23	0
subtype2	10	11
subtype3	12	10
subtype4	2	24
subtype5	1	33

Figure S55. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'RPPA cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.318 (Kruskal-Wallis (anova)), Q value = 0.4

Table S60. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	66	36.4 (20.6)
subtype1	11	37.2 (20.2)
subtype2	9	34.3 (20.9)
subtype3	14	45.5 (21.2)
subtype4	16	32.8 (24.8)
subtype5	16	32.6 (14.9)

Figure S56. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'RPPA cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.152 (Fisher's exact test), Q value = 0.23

Table S61. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	101	10	1	7
subtype1	18	2	0	2
subtype2	20	0	0	0
subtype3	14	4	0	1
subtype4	24	1	0	0
subtype5	25	3	1	4

Figure S57. 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 = 0.0309 (Kruskal-Wallis (anova)), Q value = 0.069

Table S62. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	97	2.2 (2.8)
subtype1	23	3.9 (4.0)
subtype2	18	1.2 (1.8)
subtype3	21	1.9 (2.3)
subtype4	12	2.5 (2.4)
subtype5	23	1.3 (1.8)

Figure S58. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RPPA cHierClus subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S63. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	43	2	73
subtype1	0	0	19
subtype2	3	1	15
subtype3	6	0	15
subtype4	18	0	7
subtype5	16	1	17

Figure S59. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #14: 'RACE'

'RPPA cHierClus subtypes' versus 'ETHNICITY'

P value = 0.775 (Fisher's exact test), Q value = 0.82

Table S64. Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	2	68
subtype1	0	4
subtype2	0	9
subtype3	0	10
subtype4	0	21
subtype5	2	24

Figure S60. Get High-res Image Clustering Approach #4: 'RPPA cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #5: 'RNAseq CNMF subtypes'

Table S65. Description of clustering approach #5: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3	4	5	6
Number of samples	74	14	42	29	9	16

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.84 (logrank test), Q value = 0.87

Table S66. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	184	77	0.1 - 122.1 (13.0)
subtype1	74	37	0.4 - 122.1 (13.7)
subtype2	14	4	0.1 - 58.6 (12.5)
subtype3	42	13	0.5 - 68.0 (12.6)
subtype4	29	11	0.3 - 44.8 (12.4)
subtype5	9	4	2.6 - 83.2 (20.8)
subtype6	16	8	6.3 - 54.0 (19.4)

Figure S61. 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 = 2.03e-05 (Kruskal-Wallis (anova)), Q value = 1e-04

Table S67. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	184	62.5 (11.9)
subtype1	74	67.6 (10.5)
subtype2	14	60.1 (15.7)
subtype3	42	59.9 (10.6)
subtype4	29	55.8 (8.5)
subtype5	9	59.2 (11.3)
subtype6	16	61.3 (14.8)

Figure S62. 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 = 0.0076 (Fisher's exact test), Q value = 0.021

Table S68. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	8	5	6	1	46	31	26	14	9	7	5	4
subtype1	8	2	1	0	4	15	13	4	3	4	2	3
subtype2	0	0	0	0	6	2	3	1	0	0	0	0
subtype3	0	1	3	1	18	7	5	3	1	1	2	0
subtype4	0	0	1	0	11	4	5	2	3	1	1	0
subtype5	0	0	0	0	4	1	0	3	0	1	0	0
subtype6	0	2	1	0	3	2	0	1	2	0	0	1

Figure S63. 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 = 6e-05 (Fisher's exact test), Q value = 0.00029

Table S69. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	32	42	88	5
subtype1	24	6	34	0
subtype2	0	4	8	0
subtype3	3	14	23	2
subtype4	2	10	15	1
subtype5	0	4	3	2
subtype6	3	4	5	0

Figure S64. 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 = 0.0655 (Fisher's exact test), Q value = 0.12

Table S70. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	76	69	12	8
subtype1	20	35	5	4
subtype2	7	5	0	0
subtype3	27	10	2	2
subtype4	11	13	3	1
subtype5	6	1	0	1
subtype6	5	5	2	0

Figure S65. 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 = 0.784 (Fisher's exact test), Q value = 0.83

Table S71. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	135	9
subtype1	43	5
subtype2	12	0
subtype3	36	2
subtype4	26	1
subtype5	8	0
subtype6	10	1

Figure S66. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.936 (Fisher's exact test), Q value = 0.96

Table S72. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	26	158
subtype1	10	64
subtype2	2	12
subtype3	7	35
subtype4	3	26
subtype5	2	7
subtype6	2	14

Figure S67. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq CNMF subtypes' versus 'RADIATION_THERAPY'

P value = 0.00198 (Fisher's exact test), Q value = 0.0065

Table S73. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	123	43
subtype1	56	7
subtype2	11	3
subtype3	23	16
subtype4	19	8
subtype5	3	5
subtype6	11	4

Figure S68. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RNAseq CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.262 (Kruskal-Wallis (anova)), Q value = 0.34

Table S74. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	67	73.6 (16.2)
subtype1	6	85.0 (12.2)
subtype2	8	66.2 (22.6)
subtype3	26	70.4 (15.9)
subtype4	18	75.6 (15.0)
subtype5	4	77.5 (12.6)
subtype6	5	78.0 (13.0)

Figure S69. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S75. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	89	95
subtype1	74	0
subtype2	5	9
subtype3	0	42
subtype4	0	29
subtype5	4	5
subtype6	6	10

Figure S70. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'RNAseq CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.0686 (Kruskal-Wallis (anova)), Q value = 0.12

Table S76. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	97	34.7 (21.5)
subtype1	34	41.7 (24.2)
subtype2	3	14.5 (0.9)
subtype3	27	29.3 (17.2)
subtype4	16	30.1 (21.9)
subtype5	3	41.2 (17.3)
subtype6	14	36.5 (20.5)

Figure S71. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'RNAseq CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.702 (Fisher's exact test), Q value = 0.75

Table S77. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	136	13	2	7
subtype1	52	7	0	2
subtype2	11	0	0	0
subtype3	35	1	1	3
subtype4	22	3	1	1
subtype5	7	1	0	0
subtype6	9	1	0	1

Figure S72. 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 = 0.283 (Kruskal-Wallis (anova)), Q value = 0.36

Table S78. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	134	2.5 (3.4)
subtype1	64	3.0 (4.0)
subtype2	7	1.6 (1.5)
subtype3	28	1.6 (2.6)
subtype4	20	2.5 (2.1)
subtype5	5	3.0 (6.2)
subtype6	10	2.7 (2.9)

Figure S73. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RNAseq CNMF subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S79. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	46	5	113
subtype1	1	0	57
subtype2	6	0	7
subtype3	19	1	21
subtype4	17	3	8
subtype5	2	0	7
subtype6	1	1	13

Figure S74. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'RACE'

'RNAseq CNMF subtypes' versus 'ETHNICITY'

P value = 0.404 (Fisher's exact test), Q value = 0.48

Table S80. Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	88
subtype1	2	17
subtype2	1	8
subtype3	0	27
subtype4	2	21
subtype5	0	6
subtype6	0	9

Figure S75. Get High-res Image Clustering Approach #5: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #6: 'RNAseq cHierClus subtypes'

Table S81. Description of clustering approach #6: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	68	20	96

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.832 (logrank test), Q value = 0.87

Table S82. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	184	77	0.1 - 122.1 (13.0)
subtype1	68	34	1.4 - 122.1 (15.3)
subtype2	20	10	0.4 - 54.0 (10.9)
subtype3	96	33	0.1 - 68.0 (12.6)

Figure S76. 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 = 6.96e-06 (Kruskal-Wallis (anova)), Q value = 6e-05

Table S83. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	184	62.5 (11.9)
subtype1	68	67.1 (11.7)
subtype2	20	65.5 (13.3)
subtype3	96	58.6 (10.5)

Figure S77. 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 = 2e-05 (Fisher's exact test), Q value = 1e-04

Table S84. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	8	5	6	1	46	31	26	14	9	7	5	4
subtype1	6	2	1	0	3	16	13	4	1	5	2	3
subtype2	2	0	1	0	2	2	1	1	3	0	0	0
subtype3	0	3	4	1	41	13	12	9	5	2	3	1

Figure S78. 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 = 0.00177 (Fisher's exact test), Q value = 0.006

Table S85. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	32	42	88	5
subtype1	18	9	30	1
subtype2	6	2	7	0
subtype3	8	31	51	4

Figure S79. 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 = 0.0048 (Fisher's exact test), Q value = 0.014

Table S86. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	76	69	12	8
subtype1	17	32	3	5
subtype2	5	7	3	0
subtype3	54	30	6	3

Figure S80. 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 = 0.343 (Fisher's exact test), Q value = 0.43

Table S87. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	135	9
subtype1	42	5
subtype2	9	0
subtype3	84	4

Figure S81. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.91 (Fisher's exact test), Q value = 0.93

Table S88. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	26	158
subtype1	10	58
subtype2	2	18
subtype3	14	82

Figure S82. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

'RNAseq cHierClus subtypes' versus 'RADIATION_THERAPY'

P value = 0.00017 (Fisher's exact test), Q value = 0.00075

Table S89. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	123	43
subtype1	48	9
subtype2	19	0
subtype3	56	34

Figure S83. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'RNAseq cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.0361 (Kruskal-Wallis (anova)), Q value = 0.076

Table S90. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	67	73.6 (16.2)
subtype1	6	85.0 (12.2)
subtype2	1	90.0 (NA)
subtype3	60	72.2 (16.2)

Figure S84. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S91. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	89	95
subtype1	68	0
subtype2	20	0
subtype3	1	95

Figure S85. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'RNAseq cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.125 (Kruskal-Wallis (anova)), Q value = 0.2

Table S92. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	97	34.7 (21.5)
subtype1	27	37.6 (25.6)
subtype2	14	43.8 (21.9)
subtype3	56	31.0 (18.6)

Figure S86. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'RNAseq cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.14 (Fisher's exact test), Q value = 0.21

Table S93. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	136	13	2	7
subtype1	46	8	0	1
subtype2	12	0	0	2
subtype3	78	5	2	4

Figure S87. 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 = 0.0507 (Kruskal-Wallis (anova)), Q value = 0.099

Table S94. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	134	2.5 (3.4)
subtype1	56	3.5 (4.4)
subtype2	15	1.7 (2.1)
subtype3	63	1.9 (2.4)

Figure S88. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'RNAseq cHierClus subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S95. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	46	5	113
subtype1	1	0	54
subtype2	0	0	17
subtype3	45	5	42

Figure S89. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'RACE'

'RNAseq cHierClus subtypes' versus 'ETHNICITY'

P value = 0.173 (Fisher's exact test), Q value = 0.25

Table S96. Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	5	88
subtype1	2	14
subtype2	1	9
subtype3	2	65

Figure S90. Get High-res Image Clustering Approach #6: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #7: 'MIRSEQ CNMF'

Table S97. Description of clustering approach #7: 'MIRSEQ CNMF'

Cluster Labels	1	2	3	4	5	6
Number of samples	71	15	10	18	28	42

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.0806 (logrank test), Q value = 0.14

Table S98. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	184	77	0.1 - 122.1 (13.2)
subtype1	71	35	0.4 - 122.1 (14.3)
subtype2	15	9	0.1 - 52.3 (10.8)
subtype3	10	6	1.4 - 44.8 (5.3)
subtype4	18	6	5.3 - 58.6 (12.8)
subtype5	28	8	0.3 - 83.2 (14.2)
subtype6	42	13	0.8 - 68.0 (13.0)

Figure S91. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'YEARS_TO_BIRTH'

P value = 0.00418 (Kruskal-Wallis (anova)), Q value = 0.012

Table S99. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	184	62.5 (11.9)
subtype1	71	66.0 (11.6)
subtype2	15	58.7 (12.1)
subtype3	10	63.5 (12.5)
subtype4	18	60.8 (10.9)
subtype5	28	63.9 (13.4)
subtype6	42	57.6 (9.8)

Figure S92. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CNMF' versus 'PATHOLOGIC_STAGE'

P value = 1e-04 (Fisher's exact test), Q value = 0.00047

Table S100. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	8	5	6	1	46	31	26	14	9	7	5	4
subtype1	8	2	2	0	3	13	12	2	4	4	1	3
subtype2	0	0	0	0	4	1	3	1	0	2	1	0
subtype3	0	0	0	0	1	2	2	2	1	0	1	0
subtype4	0	2	0	1	10	2	1	0	1	0	0	0
subtype5	0	1	1	0	8	8	3	5	0	1	0	1
subtype6	0	0	3	0	20	5	5	4	3	0	2	0

Figure S93. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_T_STAGE'

P value = 8e-04 (Fisher's exact test), Q value = 0.0029

Table S101. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	32	43	87	5
subtype1	21	6	31	0
subtype2	0	2	9	1
subtype3	1	2	6	0
subtype4	4	7	7	0
subtype5	3	10	13	2
subtype6	3	16	21	2

Figure S94. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_N_STAGE'

P value = 0.00754 (Fisher's exact test), Q value = 0.021

Table S102. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	76	69	12	8
subtype1	19	30	5	4
subtype2	5	5	0	2
subtype3	1	5	2	1
subtype4	13	4	1	0
subtype5	13	13	0	1
subtype6	25	12	4	0

Figure S95. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CNMF' versus 'PATHOLOGY_M_STAGE'

P value = 0.618 (Fisher's exact test), Q value = 0.68

Table S103. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	135	9
subtype1	41	4
subtype2	10	1
subtype3	6	1
subtype4	16	0
subtype5	25	1
subtype6	37	2

Figure S96. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.379 (Fisher's exact test), Q value = 0.46

Table S104. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	27	157
subtype1	10	61
subtype2	0	15
subtype3	1	9
subtype4	5	13
subtype5	4	24
subtype6	7	35

Figure S97. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CNMF' versus 'RADIATION_THERAPY'

P value = 0.00466 (Fisher's exact test), Q value = 0.013

Table S105. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	124	42
subtype1	54	7
subtype2	8	6
subtype3	8	1
subtype4	14	3
subtype5	15	10
subtype6	25	15

Figure S98. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRSEQ CNMF' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.327 (Kruskal-Wallis (anova)), Q value = 0.41

Table S106. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	67	74.0 (16.2)
subtype1	5	84.0 (13.4)
subtype2	8	80.0 (10.7)
subtype3	3	76.7 (15.3)
subtype4	10	64.0 (22.2)
subtype5	14	72.9 (17.3)
subtype6	27	74.4 (14.2)

Figure S99. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'MIRSEQ CNMF' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S107. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	89	95
subtype1	71	0
subtype2	4	11
subtype3	5	5
subtype4	2	16
subtype5	7	21
subtype6	0	42

Figure S100. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'MIRSEQ CNMF' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.194 (Kruskal-Wallis (anova)), Q value = 0.26

Table S108. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	97	34.7 (21.5)
subtype1	33	41.3 (24.8)
subtype2	12	25.7 (12.0)
subtype3	5	28.0 (12.0)
subtype4	8	33.4 (13.1)
subtype5	14	40.6 (25.4)
subtype6	25	28.6 (19.5)

Figure S101. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'MIRSEQ CNMF' versus 'RESIDUAL_TUMOR'

P value = 0.187 (Fisher's exact test), Q value = 0.26

Table S109. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	136	13	2	7
subtype1	47	6	0	3
subtype2	8	3	0	0
subtype3	7	0	0	1
subtype4	17	0	0	0
subtype5	22	3	0	0
subtype6	35	1	2	3

Figure S102. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRSEQ CNMF' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.00193 (Kruskal-Wallis (anova)), Q value = 0.0064

Table S110. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	135	2.5 (3.4)
subtype1	58	3.2 (4.1)
subtype2	8	4.5 (2.9)
subtype3	8	4.1 (2.8)
subtype4	10	1.1 (1.6)
subtype5	21	2.0 (3.4)
subtype6	30	1.2 (1.6)

Figure S103. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CNMF' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S111. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	45	5	114
subtype1	1	0	56
subtype2	3	3	6
subtype3	3	0	6
subtype4	10	0	7
subtype5	9	1	17
subtype6	19	1	22

Figure S104. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #14: 'RACE'

'MIRSEQ CNMF' versus 'ETHNICITY'

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

Table S112. Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	87
subtype1	2	20
subtype2	0	7
subtype3	0	4
subtype4	1	14
subtype5	1	15
subtype6	2	27

Figure S105. Get High-res Image Clustering Approach #7: 'MIRSEQ CNMF' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #8: 'MIRSEQ CHIERARCHICAL'

Table S113. Description of clustering approach #8: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3	4
Number of samples	63	26	62	33

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.0182 (logrank test), Q value = 0.043

Table S114. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	184	77	0.1 - 122.1 (13.2)
subtype1	63	38	0.4 - 122.1 (13.5)
subtype2	26	7	3.0 - 83.2 (19.1)
subtype3	62	22	0.1 - 35.2 (12.5)
subtype4	33	10	0.8 - 68.0 (15.5)

Figure S106. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'YEARS_TO_BIRTH'

P value = 1.85e-06 (Kruskal-Wallis (anova)), Q value = 6e-05

Table S115. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	184	62.5 (11.9)
subtype1	63	65.1 (11.9)
subtype2	26	71.2 (11.5)
subtype3	62	58.1 (11.1)
subtype4	33	59.3 (8.6)

Figure S107. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGIC_STAGE'

P value = 2e-05 (Fisher's exact test), Q value = 1e-04

Table S116. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	8	5	6	1	46	31	26	14	9	7	5	4
subtype1	7	2	2	0	1	11	12	2	3	3	1	3
subtype2	1	0	0	0	4	7	3	3	1	2	1	0
subtype3	0	2	2	1	24	9	9	6	4	2	1	0
subtype4	0	1	2	0	17	4	2	3	1	0	2	1

Figure S108. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_T_STAGE'

P value = 0.00089 (Fisher's exact test), Q value = 0.0032

Table S117. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	32	43	87	5
subtype1	19	8	23	0
subtype2	5	3	15	1
subtype3	4	21	31	4
subtype4	4	11	18	0

Figure S109. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_N_STAGE'

P value = 0.0434 (Fisher's exact test), Q value = 0.09

Table S118. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	76	69	12	8
subtype1	14	29	4	3
subtype2	8	11	2	2
subtype3	33	19	5	2
subtype4	21	10	1	1

Figure S110. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'PATHOLOGY_M_STAGE'

P value = 0.204 (Fisher's exact test), Q value = 0.27

Table S119. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	135	9
subtype1	33	4
subtype2	19	1
subtype3	55	1
subtype4	28	3

Figure S111. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.488 (Fisher's exact test), Q value = 0.55

Table S120. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	27	157
subtype1	7	56
subtype2	5	21
subtype3	8	54
subtype4	7	26

Figure S112. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'RADIATION_THERAPY'

P value = 0.00105 (Fisher's exact test), Q value = 0.0037

Table S121. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	124	42
subtype1	49	6
subtype2	19	3
subtype3	34	24
subtype4	22	9

Figure S113. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRSEQ CHIERARCHICAL' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.116 (Kruskal-Wallis (anova)), Q value = 0.19

Table S122. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	67	74.0 (16.2)
subtype1	6	85.0 (12.2)
subtype2	1	90.0 (NA)
subtype3	37	72.4 (15.5)
subtype4	23	73.0 (17.7)

Figure S114. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S123. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	89	95
subtype1	63	0
subtype2	26	0
subtype3	0	62
subtype4	0	33

Figure S115. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.136 (Kruskal-Wallis (anova)), Q value = 0.21

Table S124. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	97	34.7 (21.5)
subtype1	28	35.7 (20.2)
subtype2	13	48.5 (30.4)
subtype3	35	33.3 (19.0)
subtype4	21	27.1 (17.7)

Figure S116. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'MIRSEQ CHIERARCHICAL' versus 'RESIDUAL_TUMOR'

P value = 0.679 (Fisher's exact test), Q value = 0.73

Table S125. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	136	13	2	7
subtype1	41	5	0	1
subtype2	18	3	0	2
subtype3	48	3	2	2
subtype4	29	2	0	2

Figure S117. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

'MIRSEQ CHIERARCHICAL' versus 'NUMBER_OF_LYMPH_NODES'

P value = 0.21 (Kruskal-Wallis (anova)), Q value = 0.28

Table S126. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	135	2.5 (3.4)
subtype1	49	2.9 (3.3)
subtype2	23	3.5 (5.4)
subtype3	40	1.8 (2.2)
subtype4	23	2.0 (2.8)

Figure S118. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRSEQ CHIERARCHICAL' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S127. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	45	5	114
subtype1	1	0	49
subtype2	0	0	22
subtype3	30	4	26
subtype4	14	1	17

Figure S119. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'RACE'

'MIRSEQ CHIERARCHICAL' versus 'ETHNICITY'

P value = 0.176 (Fisher's exact test), Q value = 0.25

Table S128. Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	87
subtype1	3	17
subtype2	0	6
subtype3	1	46
subtype4	2	18

Figure S120. Get High-res Image Clustering Approach #8: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #9: 'MIRseq Mature CNMF subtypes'

Table S129. Description of clustering approach #9: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3	4
Number of samples	65	67	35	12

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

P value = 0.0448 (logrank test), Q value = 0.091

Table S130. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	179	77	0.1 - 122.1 (13.2)
subtype1	65	35	0.4 - 122.1 (13.9)
subtype2	67	25	0.1 - 68.0 (12.7)
subtype3	35	11	0.3 - 83.2 (15.5)
subtype4	12	6	0.8 - 19.7 (8.2)

Figure S121. 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 = 8.25e-06 (Kruskal-Wallis (anova)), Q value = 6e-05

Table S131. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	179	62.5 (12.0)
subtype1	65	66.4 (12.2)
subtype2	67	56.8 (9.7)
subtype3	35	65.5 (12.7)
subtype4	12	64.8 (9.0)

Figure S122. 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 = 6e-05

Table S132. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	7	5	5	1	46	30	25	14	9	7	5	4
subtype1	7	2	2	0	2	12	12	3	3	2	2	3
subtype2	0	2	2	1	31	7	8	3	5	2	3	1
subtype3	0	1	1	0	12	9	3	5	1	2	0	0
subtype4	0	0	0	0	1	2	2	3	0	1	0	0

Figure S123. 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 = 0.00078 (Fisher's exact test), Q value = 0.0029

Table S133. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	31	42	85	5
subtype1	20	8	27	0
subtype2	7	22	35	1
subtype3	4	9	19	2
subtype4	0	3	4	2

Figure S124. 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 = 0.104 (Fisher's exact test), Q value = 0.17

Table S134. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	73	68	12	8
subtype1	16	32	5	2
subtype2	35	21	5	3
subtype3	18	12	1	2
subtype4	4	3	1	1

Figure S125. 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 = 0.235 (Fisher's exact test), Q value = 0.31

Table S135. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	132	9
subtype1	36	5
subtype2	57	4
subtype3	31	0
subtype4	8	0

Figure S126. 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 = 0.987 (Fisher's exact test), Q value = 1

Table S136. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	25	154
subtype1	10	55
subtype2	9	58
subtype3	5	30
subtype4	1	11

Figure S127. 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 = 0.00317 (Fisher's exact test), Q value = 0.0097

Table S137. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	122	40
subtype1	51	6
subtype2	40	22
subtype3	21	11
subtype4	10	1

Figure S128. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRseq Mature CNMF subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.0292 (Kruskal-Wallis (anova)), Q value = 0.066

Table S138. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	65	74.2 (16.4)
subtype1	6	85.0 (12.2)
subtype2	40	71.5 (16.7)
subtype3	13	71.5 (16.8)
subtype4	6	86.7 (5.2)

Figure S129. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S139. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	87	92
subtype1	65	0
subtype2	1	66
subtype3	16	19
subtype4	5	7

Figure S130. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'MIRseq Mature CNMF subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.0166 (Kruskal-Wallis (anova)), Q value = 0.04

Table S140. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	94	35.1 (21.6)
subtype1	31	36.7 (19.2)
subtype2	40	27.0 (16.1)
subtype3	18	47.2 (27.5)
subtype4	5	45.6 (29.9)

Figure S131. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'MIRseq Mature CNMF subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.127 (Fisher's exact test), Q value = 0.2

Table S141. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	132	13	2	7
subtype1	44	6	0	2
subtype2	57	2	1	3
subtype3	23	5	0	1
subtype4	8	0	1	1

Figure S132. 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 = 0.158 (Kruskal-Wallis (anova)), Q value = 0.23

Table S142. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	132	2.6 (3.4)
subtype1	54	2.8 (3.0)
subtype2	41	2.5 (2.7)
subtype3	27	2.1 (4.6)
subtype4	10	3.1 (4.8)

Figure S133. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature CNMF subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S143. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	44	4	111
subtype1	1	0	49
subtype2	36	3	24
subtype3	6	1	27
subtype4	1	0	11

Figure S134. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'RACE'

'MIRseq Mature CNMF subtypes' versus 'ETHNICITY'

P value = 0.418 (Fisher's exact test), Q value = 0.49

Table S144. Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	84
subtype1	2	16
subtype2	2	45
subtype3	1	17
subtype4	1	6

Figure S135. Get High-res Image Clustering Approach #9: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'ETHNICITY'

Clustering Approach #10: 'MIRseq Mature cHierClus subtypes'

Table S145. Description of clustering approach #10: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	56	29	94

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

P value = 0.0328 (logrank test), Q value = 0.071

Table S146. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	179	77	0.1 - 122.1 (13.2)
subtype1	56	36	1.4 - 122.1 (13.4)
subtype2	29	9	0.4 - 83.2 (18.2)
subtype3	94	32	0.1 - 68.0 (12.7)

Figure S136. 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 = 7.31e-07 (Kruskal-Wallis (anova)), Q value = 6e-05

Table S147. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'YEARS_TO_BIRTH'

	nPatients	Mean (Std.Dev)
ALL	179	62.5 (12.0)
subtype1	56	65.0 (12.1)
subtype2	29	70.7 (10.8)
subtype3	94	58.6 (10.7)

Figure S137. 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 = 6e-05

Table S148. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'PATHOLOGIC_STAGE'

nPatients	STAGE I	STAGE IA	STAGE IB	STAGE II	STAGE IIA	STAGE IIB	STAGE III	STAGE IIIA	STAGE IIIB	STAGE IIIC	STAGE IV	STAGE IVA
ALL	7	5	5	1	46	30	25	14	9	7	5	4
subtype1	5	2	2	0	1	10	11	2	3	3	1	3
subtype2	2	0	0	0	4	7	4	2	1	2	1	0
subtype3	0	3	3	1	41	13	10	10	5	2	3	1

Figure S138. 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 = 0.00058 (Fisher's exact test), Q value = 0.0023

Table S149. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY_T_STAGE'

nPatients	T0+T1	T2	T3	T4
ALL	31	42	85	5
subtype1	16	7	22	0
subtype2	7	3	16	0
subtype3	8	32	47	5

Figure S139. 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 = 0.00997 (Fisher's exact test), Q value = 0.026

Table S150. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY_N_STAGE'

nPatients	N0	N1	N2	N3
ALL	73	68	12	8
subtype1	11	27	4	3
subtype2	10	12	2	2
subtype3	52	29	6	3

Figure S140. 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 = 0.46

Table S151. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'PATHOLOGY_M_STAGE'

nPatients	0	1
ALL	132	9
subtype1	29	4
subtype2	21	1
subtype3	82	4

Figure S141. 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 = 0.392 (Fisher's exact test), Q value = 0.47

Table S152. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'GENDER'

nPatients	FEMALE	MALE
ALL	25	154
subtype1	5	51
subtype2	5	24
subtype3	15	79

Figure S142. 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 = 0.00364 (Fisher's exact test), Q value = 0.011

Table S153. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

nPatients	NO	YES
ALL	122	40
subtype1	44	6
subtype2	21	3
subtype3	57	31

Figure S143. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'RADIATION_THERAPY'

'MIRseq Mature cHierClus subtypes' versus 'KARNOFSKY_PERFORMANCE_SCORE'

P value = 0.0516 (Kruskal-Wallis (anova)), Q value = 0.099

Table S154. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

	nPatients	Mean (Std.Dev)
ALL	65	74.2 (16.4)
subtype1	6	85.0 (12.2)
subtype3	59	73.1 (16.4)

Figure S144. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'KARNOFSKY_PERFORMANCE_SCORE'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL_TYPE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S155. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

nPatients	ESOPHAGUS ADENOCARCINOMA, NOS	ESOPHAGUS SQUAMOUS CELL CARCINOMA
ALL	87	92
subtype1	56	0
subtype2	29	0
subtype3	2	92

Figure S145. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'HISTOLOGICAL_TYPE'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER_PACK_YEARS_SMOKED'

P value = 0.0185 (Kruskal-Wallis (anova)), Q value = 0.043

Table S156. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

	nPatients	Mean (Std.Dev)
ALL	94	35.1 (21.6)
subtype1	24	35.2 (19.9)
subtype2	15	50.9 (27.1)
subtype3	55	30.7 (18.9)

Figure S146. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'NUMBER_PACK_YEARS_SMOKED'

'MIRseq Mature cHierClus subtypes' versus 'RESIDUAL_TUMOR'

P value = 0.669 (Fisher's exact test), Q value = 0.73

Table S157. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #12: 'RESIDUAL_TUMOR'

nPatients	R0	R1	R2	RX
ALL	132	13	2	7
subtype1	36	5	0	1
subtype2	21	3	0	2
subtype3	75	5	2	4

Figure S147. 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 = 0.0665 (Kruskal-Wallis (anova)), Q value = 0.12

Table S158. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

	nPatients	Mean (Std.Dev)
ALL	132	2.6 (3.4)
subtype1	44	3.1 (3.4)
subtype2	26	3.3 (5.1)
subtype3	62	1.9 (2.4)

Figure S148. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NUMBER_OF_LYMPH_NODES'

'MIRseq Mature cHierClus subtypes' versus 'RACE'

P value = 1e-05 (Fisher's exact test), Q value = 6e-05

Table S159. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'RACE'

nPatients	ASIAN	BLACK OR AFRICAN AMERICAN	WHITE
ALL	44	4	111
subtype1	1	0	43
subtype2	0	0	24
subtype3	43	4	44

Figure S149. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'RACE'

'MIRseq Mature cHierClus subtypes' versus 'ETHNICITY'

P value = 0.141 (Fisher's exact test), Q value = 0.21

Table S160. Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

nPatients	HISPANIC OR LATINO	NOT HISPANIC OR LATINO
ALL	6	84
subtype1	3	15
subtype2	0	6
subtype3	3	63

Figure S150. Get High-res Image Clustering Approach #10: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'ETHNICITY'

Methods & Data

Input

Cluster data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_mergedClustering/ESCA-TP/22541001/ESCA-TP.mergedcluster.txt
Clinical data file = /xchip/cga/gdac-prod/tcga-gdac/jobResults/Append_Data/ESCA-TP/22506419/ESCA-TP.merged_data.txt
Number of patients = 185
Number of clustering approaches = 10
Number of selected clinical features = 15
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