Correlation between aggregated molecular cancer subtypes and selected clinical features

Overview

Introduction

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

Summary

Testing the association between subtypes identified by 8 different clustering approaches and 15 clinical features across 229 patients, 11 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 'NEOPLASM.DISEASESTAGE' and 'MULTIFOCALITY'.
3 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'EXTRATHYROIDAL.EXTENSION'.
CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that do not correlate to any clinical features.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'HISTOLOGICAL.TYPE' and 'EXTRATHYROIDAL.EXTENSION'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CNMF'. These subtypes correlate to 'AGE' and 'EXTRATHYROIDAL.EXTENSION'.
3 subtypes identified in current cancer cohort by 'MIRSEQ CHIERARCHICAL'. These subtypes do not correlate to any clinical features.
4 subtypes identified in current cancer cohort by 'MIRseq Mature CNMF subtypes'. These subtypes correlate to 'Time to Death', 'AGE', and 'EXTRATHYROIDAL.EXTENSION'.
3 subtypes identified in current cancer cohort by 'MIRseq Mature cHierClus subtypes'. These subtypes correlate to 'EXTRATHYROIDAL.EXTENSION'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 8 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, 11 significant findings detected.


Clinical Features	Statistical Tests	Copy Number Ratio CNMF subtypes	METHLYATION CNMF	RNAseq CNMF subtypes	RNAseq cHierClus subtypes	MIRSEQ CNMF	MIRSEQ CHIERARCHICAL	MIRseq Mature CNMF subtypes	MIRseq Mature cHierClus subtypes
Time to Death	logrank test	0.689 (1.00)	0.026 (1.00)	0.146 (1.00)	0.0833 (1.00)	0.00266 (0.268)	0.0269 (1.00)	0.00136 (0.14)	0.0256 (1.00)
AGE	ANOVA	0.00329 (0.329)	0.0642 (1.00)	0.132 (1.00)	0.0427 (1.00)	1.26e-05 (0.00141)	0.0207 (1.00)	0.000152 (0.0166)	0.00386 (0.382)
GENDER	Fisher's exact test	0.847 (1.00)	0.0544 (1.00)	0.167 (1.00)	0.0357 (1.00)	0.44 (1.00)	0.581 (1.00)	0.481 (1.00)	0.645 (1.00)
HISTOLOGICAL TYPE	Chi-square test	0.208 (1.00)	0.00515 (0.499)	0.0227 (1.00)	0.00013 (0.0143)	0.00407 (0.399)	0.478 (1.00)	0.0616 (1.00)	0.0166 (1.00)
RADIATIONS RADIATION REGIMENINDICATION	Fisher's exact test	0.76 (1.00)	0.427 (1.00)	0.289 (1.00)	0.105 (1.00)	0.503 (1.00)	0.574 (1.00)	0.0398 (1.00)	0.186 (1.00)
RADIATIONEXPOSURE	Fisher's exact test	0.322 (1.00)	0.644 (1.00)	0.475 (1.00)	0.812 (1.00)	1 (1.00)	0.555 (1.00)	0.279 (1.00)	0.663 (1.00)
DISTANT METASTASIS	Chi-square test	0.0565 (1.00)	0.553 (1.00)	0.912 (1.00)	0.658 (1.00)	0.00677 (0.643)	0.348 (1.00)	0.0103 (0.955)	0.129 (1.00)
EXTRATHYROIDAL EXTENSION	Chi-square test	0.0409 (1.00)	0.0011 (0.115)	0.251 (1.00)	1.8e-05 (0.00199)	0.00115 (0.12)	0.0526 (1.00)	0.00195 (0.199)	0.000714 (0.0764)
LYMPH NODE METASTASIS	Chi-square test	0.625 (1.00)	0.0145 (1.00)	0.113 (1.00)	0.056 (1.00)	0.588 (1.00)	0.657 (1.00)	0.485 (1.00)	0.72 (1.00)
COMPLETENESS OF RESECTION	Chi-square test	0.0259 (1.00)	0.0797 (1.00)	0.185 (1.00)	0.0328 (1.00)	0.00636 (0.61)	0.0968 (1.00)	0.0297 (1.00)	0.0293 (1.00)
NUMBER OF LYMPH NODES	ANOVA	0.7 (1.00)	0.0391 (1.00)	0.273 (1.00)	0.0132 (1.00)	0.216 (1.00)	0.259 (1.00)	0.668 (1.00)	0.0718 (1.00)
TUMOR STAGECODE	ANOVA
NEOPLASM DISEASESTAGE	Chi-square test	0.000219 (0.0237)	0.0317 (1.00)	0.692 (1.00)	0.162 (1.00)	0.0439 (1.00)	0.163 (1.00)	0.0426 (1.00)	0.00737 (0.693)
MULTIFOCALITY	Fisher's exact test	0.00107 (0.114)	0.612 (1.00)	0.0253 (1.00)	0.0268 (1.00)	0.265 (1.00)	0.629 (1.00)	0.0785 (1.00)	0.417 (1.00)
TUMOR SIZE	ANOVA	0.0986 (1.00)	0.0876 (1.00)	0.117 (1.00)	0.0467 (1.00)	0.986 (1.00)	0.478 (1.00)	0.708 (1.00)	0.871 (1.00)

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

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

Cluster Labels	1	2	3
Number of samples	173	27	29

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

P value = 0.689 (logrank test), Q value = 1

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	227	5	0.0 - 158.8 (15.0)
subtype1	171	4	0.2 - 158.8 (15.4)
subtype2	27	1	0.6 - 138.1 (19.0)
subtype3	29	0	0.0 - 85.1 (14.0)

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

'Copy Number Ratio CNMF subtypes' versus 'AGE'

P value = 0.00329 (ANOVA), Q value = 0.33

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

	nPatients	Mean (Std.Dev)
ALL	229	46.8 (14.9)
subtype1	173	45.4 (14.7)
subtype2	27	55.7 (15.5)
subtype3	29	46.9 (13.0)

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

'Copy Number Ratio CNMF subtypes' versus 'GENDER'

P value = 0.847 (Fisher's exact test), Q value = 1

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

nPatients	FEMALE	MALE
ALL	169	60
subtype1	129	44
subtype2	19	8
subtype3	21	8

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

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

P value = 0.208 (Chi-square test), Q value = 1

Table S5. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #4: '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	3	184	16	26
subtype1	3	140	14	16
subtype2	0	19	1	7
subtype3	0	25	1	3

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

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

P value = 0.76 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	11	218
subtype1	8	165
subtype2	2	25
subtype3	1	28

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

'Copy Number Ratio CNMF subtypes' versus 'RADIATIONEXPOSURE'

P value = 0.322 (Fisher's exact test), Q value = 1

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

nPatients	NO	YES
ALL	197	8
subtype1	147	5
subtype2	23	2
subtype3	27	1

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

'Copy Number Ratio CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.0565 (Chi-square test), Q value = 1

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

nPatients	M0	M1	MX
ALL	132	4	93
subtype1	108	4	61
subtype2	11	0	16
subtype3	13	0	16

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

'Copy Number Ratio CNMF subtypes' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.0409 (Chi-square test), Q value = 1

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

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	76	9	139
subtype1	56	6	106
subtype2	13	3	11
subtype3	7	0	22

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

'Copy Number Ratio CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.625 (Chi-square test), Q value = 1

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

nPatients	N0	N1	N1A	N1B	NX
ALL	92	27	56	33	21
subtype1	73	21	40	21	18
subtype2	8	4	8	6	1
subtype3	11	2	8	6	2

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

'Copy Number Ratio CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0259 (Chi-square test), Q value = 1

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

nPatients	R0	R1	R2	RX
ALL	182	22	1	13
subtype1	142	16	1	5
subtype2	17	5	0	4
subtype3	23	1	0	4

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

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

P value = 0.7 (ANOVA), Q value = 1

Table S12. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	188	3.5 (5.6)
subtype1	137	3.5 (5.7)
subtype2	26	4.3 (6.2)
subtype3	25	3.1 (4.4)

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

'Copy Number Ratio CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.000219 (Chi-square test), Q value = 0.024

Table S13. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	132	16	55	22	3
subtype1	107	14	36	12	3
subtype2	5	1	15	6	0
subtype3	20	1	4	4	0

Figure S12. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'Copy Number Ratio CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.00107 (Fisher's exact test), Q value = 0.11

Table S14. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	105	121
subtype1	71	100
subtype2	12	15
subtype3	22	6

Figure S13. 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 = 0.0986 (ANOVA), Q value = 1

Table S15. Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	186	2.8 (1.6)
subtype1	141	2.8 (1.6)
subtype2	22	3.4 (1.3)
subtype3	23	2.4 (1.5)

Figure S14. Get High-res Image Clustering Approach #1: 'Copy Number Ratio CNMF subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #2: 'METHLYATION CNMF'

Table S16. Get Full Table Description of clustering approach #2: 'METHLYATION CNMF'

Cluster Labels	1	2	3
Number of samples	44	88	97

'METHLYATION CNMF' versus 'Time to Death'

P value = 0.026 (logrank test), Q value = 1

Table S17. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	227	5	0.0 - 158.8 (15.0)
subtype1	43	0	1.1 - 157.2 (17.5)
subtype2	88	5	0.2 - 147.8 (17.5)
subtype3	96	0	0.0 - 158.8 (12.3)

Figure S15. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

'METHLYATION CNMF' versus 'AGE'

P value = 0.0642 (ANOVA), Q value = 1

Table S18. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	229	46.8 (14.9)
subtype1	44	46.0 (13.6)
subtype2	88	49.6 (16.6)
subtype3	97	44.6 (13.6)

Figure S16. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

'METHLYATION CNMF' versus 'GENDER'

P value = 0.0544 (Fisher's exact test), Q value = 1

Table S19. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	169	60
subtype1	32	12
subtype2	58	30
subtype3	79	18

Figure S17. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

'METHLYATION CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 0.00515 (Chi-square test), Q value = 0.5

Table S20. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: '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	3	184	16	26
subtype1	0	34	2	8
subtype2	1	70	2	15
subtype3	2	80	12	3

Figure S18. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

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

P value = 0.427 (Fisher's exact test), Q value = 1

Table S21. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	11	218
subtype1	1	43
subtype2	3	85
subtype3	7	90

Figure S19. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'METHLYATION CNMF' versus 'RADIATIONEXPOSURE'

P value = 0.644 (Fisher's exact test), Q value = 1

Table S22. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	197	8
subtype1	39	2
subtype2	77	4
subtype3	81	2

Figure S20. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'METHLYATION CNMF' versus 'DISTANT.METASTASIS'

P value = 0.553 (Chi-square test), Q value = 1

Table S23. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	132	4	93
subtype1	25	0	19
subtype2	48	3	37
subtype3	59	1	37

Figure S21. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'METHLYATION CNMF' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.0011 (Chi-square test), Q value = 0.12

Table S24. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	76	9	139
subtype1	16	0	28
subtype2	32	9	44
subtype3	28	0	67

Figure S22. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'METHLYATION CNMF' versus 'LYMPH.NODE.METASTASIS'

P value = 0.0145 (Chi-square test), Q value = 1

Table S25. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	92	27	56	33	21
subtype1	18	3	13	7	3
subtype2	25	15	26	17	5
subtype3	49	9	17	9	13

Figure S23. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'METHLYATION CNMF' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0797 (Chi-square test), Q value = 1

Table S26. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	182	22	1	13
subtype1	33	5	1	3
subtype2	66	13	0	6
subtype3	83	4	0	4

Figure S24. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0391 (ANOVA), Q value = 1

Table S27. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	188	3.5 (5.6)
subtype1	40	2.8 (4.1)
subtype2	72	4.9 (5.8)
subtype3	76	2.7 (6.0)

Figure S25. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'METHLYATION CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0317 (Chi-square test), Q value = 1

Table S28. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	132	16	55	22	3
subtype1	28	2	10	4	0
subtype2	42	6	22	16	2
subtype3	62	8	23	2	1

Figure S26. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'METHLYATION CNMF' versus 'MULTIFOCALITY'

P value = 0.612 (Fisher's exact test), Q value = 1

Table S29. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	105	121
subtype1	18	26
subtype2	43	43
subtype3	44	52

Figure S27. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

'METHLYATION CNMF' versus 'TUMOR.SIZE'

P value = 0.0876 (ANOVA), Q value = 1

Table S30. Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	186	2.8 (1.6)
subtype1	35	2.6 (1.6)
subtype2	75	3.1 (1.5)
subtype3	76	2.6 (1.7)

Figure S28. Get High-res Image Clustering Approach #2: 'METHLYATION CNMF' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #3: 'RNAseq CNMF subtypes'

Table S31. Get Full Table Description of clustering approach #3: 'RNAseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	57	77	89

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.146 (logrank test), Q value = 1

Table S32. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	221	5	0.0 - 158.8 (14.7)
subtype1	57	3	0.2 - 157.2 (11.2)
subtype2	76	1	0.2 - 147.8 (17.4)
subtype3	88	1	0.0 - 158.8 (12.6)

Figure S29. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.132 (ANOVA), Q value = 1

Table S33. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	223	47.0 (15.0)
subtype1	57	50.4 (14.9)
subtype2	77	45.7 (16.2)
subtype3	89	45.8 (13.8)

Figure S30. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.167 (Fisher's exact test), Q value = 1

Table S34. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	165	58
subtype1	45	12
subtype2	51	26
subtype3	69	20

Figure S31. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.0227 (Chi-square test), Q value = 1

Table S35. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #4: '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	3	178	16	26
subtype1	0	44	4	9
subtype2	1	60	2	14
subtype3	2	74	10	3

Figure S32. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

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

P value = 0.289 (Fisher's exact test), Q value = 1

Table S36. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	10	213
subtype1	1	56
subtype2	6	71
subtype3	3	86

Figure S33. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq CNMF subtypes' versus 'RADIATIONEXPOSURE'

P value = 0.475 (Fisher's exact test), Q value = 1

Table S37. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	191	8
subtype1	49	1
subtype2	66	2
subtype3	76	5

Figure S34. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'RNAseq CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.912 (Chi-square test), Q value = 1

Table S38. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	130	4	89
subtype1	35	1	21
subtype2	45	2	30
subtype3	50	1	38

Figure S35. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'RNAseq CNMF subtypes' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.251 (Chi-square test), Q value = 1

Table S39. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	75	9	134
subtype1	23	2	32
subtype2	29	4	40
subtype3	23	3	62

Figure S36. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'RNAseq CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.113 (Chi-square test), Q value = 1

Table S40. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	89	27	54	33	20
subtype1	19	4	16	13	5
subtype2	29	16	17	9	6
subtype3	41	7	21	11	9

Figure S37. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'RNAseq CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.185 (Chi-square test), Q value = 1

Table S41. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	177	21	1	13
subtype1	40	9	1	5
subtype2	64	7	0	3
subtype3	73	5	0	5

Figure S38. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.273 (ANOVA), Q value = 1

Table S42. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	182	3.6 (5.7)
subtype1	51	4.0 (5.4)
subtype2	61	4.3 (7.3)
subtype3	70	2.8 (4.1)

Figure S39. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'RNAseq CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.692 (Chi-square test), Q value = 1

Table S43. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	127	16	54	22	3
subtype1	30	2	16	8	1
subtype2	44	5	21	6	1
subtype3	53	9	17	8	1

Figure S40. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'RNAseq CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.0253 (Fisher's exact test), Q value = 1

Table S44. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	103	117
subtype1	25	32
subtype2	27	47
subtype3	51	38

Figure S41. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'RNAseq CNMF subtypes' versus 'TUMOR.SIZE'

P value = 0.117 (ANOVA), Q value = 1

Table S45. Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	182	2.8 (1.6)
subtype1	45	2.5 (1.6)
subtype2	67	3.1 (1.6)
subtype3	70	2.7 (1.6)

Figure S42. Get High-res Image Clustering Approach #3: 'RNAseq CNMF subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #4: 'RNAseq cHierClus subtypes'

Table S46. Get Full Table Description of clustering approach #4: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	90	67	66

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0833 (logrank test), Q value = 1

Table S47. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	221	5	0.0 - 158.8 (14.7)
subtype1	89	0	0.0 - 158.8 (12.5)
subtype2	66	4	0.4 - 157.2 (14.6)
subtype3	66	1	0.2 - 147.8 (17.3)

Figure S43. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.0427 (ANOVA), Q value = 1

Table S48. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	223	47.0 (15.0)
subtype1	90	46.2 (13.0)
subtype2	67	50.6 (16.1)
subtype3	66	44.3 (15.9)

Figure S44. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.0357 (Fisher's exact test), Q value = 1

Table S49. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	165	58
subtype1	72	18
subtype2	52	15
subtype3	41	25

Figure S45. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.00013 (Chi-square test), Q value = 0.014

Table S50. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: '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	3	178	16	26
subtype1	1	72	13	4
subtype2	0	49	2	16
subtype3	2	57	1	6

Figure S46. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

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

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

Table S51. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	10	213
subtype1	1	89
subtype2	5	62
subtype3	4	62

Figure S47. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq cHierClus subtypes' versus 'RADIATIONEXPOSURE'

P value = 0.812 (Fisher's exact test), Q value = 1

Table S52. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	191	8
subtype1	80	3
subtype2	60	2
subtype3	51	3

Figure S48. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'RNAseq cHierClus subtypes' versus 'DISTANT.METASTASIS'

P value = 0.658 (Chi-square test), Q value = 1

Table S53. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	130	4	89
subtype1	55	1	34
subtype2	35	1	31
subtype3	40	2	24

Figure S49. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'RNAseq cHierClus subtypes' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 1.8e-05 (Chi-square test), Q value = 0.002

Table S54. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	75	9	134
subtype1	19	0	70
subtype2	35	4	26
subtype3	21	5	38

Figure S50. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'RNAseq cHierClus subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.056 (Chi-square test), Q value = 1

Table S55. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	89	27	54	33	20
subtype1	45	5	23	9	8
subtype2	17	12	17	14	7
subtype3	27	10	14	10	5

Figure S51. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'RNAseq cHierClus subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0328 (Chi-square test), Q value = 1

Table S56. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	177	21	1	13
subtype1	78	3	0	6
subtype2	46	12	1	5
subtype3	53	6	0	2

Figure S52. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.0132 (ANOVA), Q value = 1

Table S57. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	182	3.6 (5.7)
subtype1	71	2.6 (4.5)
subtype2	60	5.4 (7.4)
subtype3	51	3.0 (4.4)

Figure S53. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'RNAseq cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.162 (Chi-square test), Q value = 1

Table S58. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	127	16	54	22	3
subtype1	55	8	19	6	1
subtype2	31	2	24	9	1
subtype3	41	6	11	7	1

Figure S54. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'RNAseq cHierClus subtypes' versus 'MULTIFOCALITY'

P value = 0.0268 (Fisher's exact test), Q value = 1

Table S59. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	103	117
subtype1	52	38
subtype2	26	41
subtype3	25	38

Figure S55. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'RNAseq cHierClus subtypes' versus 'TUMOR.SIZE'

P value = 0.0467 (ANOVA), Q value = 1

Table S60. Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	182	2.8 (1.6)
subtype1	69	2.4 (1.5)
subtype2	59	3.1 (1.7)
subtype3	54	2.9 (1.6)

Figure S56. Get High-res Image Clustering Approach #4: 'RNAseq cHierClus subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #5: 'MIRSEQ CNMF'

Table S61. Get Full Table Description of clustering approach #5: 'MIRSEQ CNMF'

Cluster Labels	1	2	3
Number of samples	78	69	81

'MIRSEQ CNMF' versus 'Time to Death'

P value = 0.00266 (logrank test), Q value = 0.27

Table S62. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	226	5	0.0 - 158.8 (15.1)
subtype1	78	0	0.0 - 158.8 (21.1)
subtype2	67	0	0.6 - 147.8 (14.5)
subtype3	81	5	0.2 - 157.2 (11.2)

Figure S57. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CNMF' versus 'AGE'

P value = 1.26e-05 (ANOVA), Q value = 0.0014

Table S63. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	228	46.9 (14.9)
subtype1	78	40.7 (12.1)
subtype2	69	48.7 (15.0)
subtype3	81	51.3 (15.3)

Figure S58. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #2: 'AGE'

'MIRSEQ CNMF' versus 'GENDER'

P value = 0.44 (Fisher's exact test), Q value = 1

Table S64. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	168	60
subtype1	60	18
subtype2	47	22
subtype3	61	20

Figure S59. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #3: 'GENDER'

'MIRSEQ CNMF' versus 'HISTOLOGICAL.TYPE'

P value = 0.00407 (Chi-square test), Q value = 0.4

Table S65. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #4: '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	3	183	16	26
subtype1	1	66	10	1
subtype2	0	56	3	10
subtype3	2	61	3	15

Figure S60. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

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

P value = 0.503 (Fisher's exact test), Q value = 1

Table S66. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	11	217
subtype1	5	73
subtype2	4	65
subtype3	2	79

Figure S61. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CNMF' versus 'RADIATIONEXPOSURE'

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

Table S67. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	196	8
subtype1	67	3
subtype2	59	2
subtype3	70	3

Figure S62. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'MIRSEQ CNMF' versus 'DISTANT.METASTASIS'

P value = 0.00677 (Chi-square test), Q value = 0.64

Table S68. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	131	4	93
subtype1	35	1	42
subtype2	37	2	30
subtype3	59	1	21

Figure S63. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'MIRSEQ CNMF' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.00115 (Chi-square test), Q value = 0.12

Table S69. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	75	9	139
subtype1	15	1	61
subtype2	24	2	40
subtype3	36	6	38

Figure S64. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'MIRSEQ CNMF' versus 'LYMPH.NODE.METASTASIS'

P value = 0.588 (Chi-square test), Q value = 1

Table S70. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	92	27	55	33	21
subtype1	33	10	17	8	10
subtype2	28	9	18	8	6
subtype3	31	8	20	17	5

Figure S65. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'MIRSEQ CNMF' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.00636 (Chi-square test), Q value = 0.61

Table S71. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	181	22	1	13
subtype1	68	1	0	5
subtype2	56	5	0	4
subtype3	57	16	1	4

Figure S66. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.216 (ANOVA), Q value = 1

Table S72. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	187	3.5 (5.6)
subtype1	61	3.1 (4.6)
subtype2	58	2.8 (4.5)
subtype3	68	4.5 (7.1)

Figure S67. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CNMF' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0439 (Chi-square test), Q value = 1

Table S73. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	131	16	55	22	3
subtype1	56	5	12	4	0
subtype2	38	6	16	7	2
subtype3	37	5	27	11	1

Figure S68. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'MIRSEQ CNMF' versus 'MULTIFOCALITY'

P value = 0.265 (Fisher's exact test), Q value = 1

Table S74. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	104	121
subtype1	41	36
subtype2	27	40
subtype3	36	45

Figure S69. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRSEQ CNMF' versus 'TUMOR.SIZE'

P value = 0.986 (ANOVA), Q value = 1

Table S75. Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	185	2.8 (1.6)
subtype1	60	2.8 (1.6)
subtype2	61	2.8 (1.5)
subtype3	64	2.8 (1.7)

Figure S70. Get High-res Image Clustering Approach #5: 'MIRSEQ CNMF' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #6: 'MIRSEQ CHIERARCHICAL'

Table S76. Get Full Table Description of clustering approach #6: 'MIRSEQ CHIERARCHICAL'

Cluster Labels	1	2	3
Number of samples	30	105	93

'MIRSEQ CHIERARCHICAL' versus 'Time to Death'

P value = 0.0269 (logrank test), Q value = 1

Table S77. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	226	5	0.0 - 158.8 (15.1)
subtype1	29	0	1.0 - 147.4 (14.5)
subtype2	104	0	0.2 - 147.8 (15.7)
subtype3	93	5	0.0 - 158.8 (14.6)

Figure S71. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #1: 'Time to Death'

'MIRSEQ CHIERARCHICAL' versus 'AGE'

P value = 0.0207 (ANOVA), Q value = 1

Table S78. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	228	46.9 (14.9)
subtype1	30	48.3 (14.0)
subtype2	105	44.0 (14.4)
subtype3	93	49.7 (15.2)

Figure S72. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #2: 'AGE'

'MIRSEQ CHIERARCHICAL' versus 'GENDER'

P value = 0.581 (Fisher's exact test), Q value = 1

Table S79. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	168	60
subtype1	20	10
subtype2	77	28
subtype3	71	22

Figure S73. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #3: 'GENDER'

'MIRSEQ CHIERARCHICAL' versus 'HISTOLOGICAL.TYPE'

P value = 0.478 (Chi-square test), Q value = 1

Table S80. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: '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	3	183	16	26
subtype1	0	24	1	5
subtype2	1	86	10	8
subtype3	2	73	5	13

Figure S74. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRSEQ CHIERARCHICAL' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.574 (Fisher's exact test), Q value = 1

Table S81. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	11	217
subtype1	1	29
subtype2	7	98
subtype3	3	90

Figure S75. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRSEQ CHIERARCHICAL' versus 'RADIATIONEXPOSURE'

P value = 0.555 (Fisher's exact test), Q value = 1

Table S82. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	196	8
subtype1	29	0
subtype2	87	5
subtype3	80	3

Figure S76. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'MIRSEQ CHIERARCHICAL' versus 'DISTANT.METASTASIS'

P value = 0.348 (Chi-square test), Q value = 1

Table S83. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	131	4	93
subtype1	16	0	14
subtype2	55	3	47
subtype3	60	1	32

Figure S77. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'MIRSEQ CHIERARCHICAL' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.0526 (Chi-square test), Q value = 1

Table S84. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	75	9	139
subtype1	15	0	14
subtype2	27	3	71
subtype3	33	6	54

Figure S78. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'MIRSEQ CHIERARCHICAL' versus 'LYMPH.NODE.METASTASIS'

P value = 0.657 (Chi-square test), Q value = 1

Table S85. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	92	27	55	33	21
subtype1	13	4	6	3	4
subtype2	45	14	26	11	9
subtype3	34	9	23	19	8

Figure S79. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'MIRSEQ CHIERARCHICAL' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0968 (Chi-square test), Q value = 1

Table S86. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	181	22	1	13
subtype1	24	3	1	1
subtype2	87	6	0	6
subtype3	70	13	0	6

Figure S80. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

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

P value = 0.259 (ANOVA), Q value = 1

Table S87. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	187	3.5 (5.6)
subtype1	25	4.6 (9.1)
subtype2	86	2.8 (4.4)
subtype3	76	3.9 (5.4)

Figure S81. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'MIRSEQ CHIERARCHICAL' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.163 (Chi-square test), Q value = 1

Table S88. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	131	16	55	22	3
subtype1	17	2	10	1	0
subtype2	67	10	18	8	2
subtype3	47	4	27	13	1

Figure S82. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'MIRSEQ CHIERARCHICAL' versus 'MULTIFOCALITY'

P value = 0.629 (Fisher's exact test), Q value = 1

Table S89. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	104	121
subtype1	12	18
subtype2	46	56
subtype3	46	47

Figure S83. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRSEQ CHIERARCHICAL' versus 'TUMOR.SIZE'

P value = 0.478 (ANOVA), Q value = 1

Table S90. Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	185	2.8 (1.6)
subtype1	25	3.1 (1.6)
subtype2	85	2.9 (1.6)
subtype3	75	2.7 (1.6)

Figure S84. Get High-res Image Clustering Approach #6: 'MIRSEQ CHIERARCHICAL' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #7: 'MIRseq Mature CNMF subtypes'

Table S91. Get Full Table Description of clustering approach #7: 'MIRseq Mature CNMF subtypes'

Cluster Labels	1	2	3	4
Number of samples	65	40	62	61

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

P value = 0.00136 (logrank test), Q value = 0.14

Table S92. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	226	5	0.0 - 158.8 (15.1)
subtype1	65	0	0.0 - 158.8 (20.4)
subtype2	39	0	1.0 - 147.8 (18.2)
subtype3	62	5	0.2 - 157.2 (10.7)
subtype4	60	0	0.6 - 131.2 (11.5)

Figure S85. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature CNMF subtypes' versus 'AGE'

P value = 0.000152 (ANOVA), Q value = 0.017

Table S93. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	228	46.9 (14.9)
subtype1	65	41.0 (12.2)
subtype2	40	45.5 (16.2)
subtype3	62	52.4 (16.2)
subtype4	61	48.5 (13.0)

Figure S86. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature CNMF subtypes' versus 'GENDER'

P value = 0.481 (Fisher's exact test), Q value = 1

Table S94. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	168	60
subtype1	48	17
subtype2	26	14
subtype3	49	13
subtype4	45	16

Figure S87. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'MIRseq Mature CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.0616 (Chi-square test), Q value = 1

Table S95. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: '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	3	183	16	26
subtype1	1	57	6	1
subtype2	0	31	1	8
subtype3	1	46	3	12
subtype4	1	49	6	5

Figure S88. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRseq Mature CNMF subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.0398 (Fisher's exact test), Q value = 1

Table S96. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	11	217
subtype1	5	60
subtype2	4	36
subtype3	2	60
subtype4	0	61

Figure S89. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq Mature CNMF subtypes' versus 'RADIATIONEXPOSURE'

P value = 0.279 (Fisher's exact test), Q value = 1

Table S97. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	196	8
subtype1	52	3
subtype2	35	2
subtype3	53	3
subtype4	56	0

Figure S90. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'MIRseq Mature CNMF subtypes' versus 'DISTANT.METASTASIS'

P value = 0.0103 (Chi-square test), Q value = 0.95

Table S98. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	131	4	93
subtype1	27	2	36
subtype2	19	0	21
subtype3	44	1	17
subtype4	41	1	19

Figure S91. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'MIRseq Mature CNMF subtypes' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.00195 (Chi-square test), Q value = 0.2

Table S99. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	75	9	139
subtype1	14	1	48
subtype2	15	1	24
subtype3	30	6	26
subtype4	16	1	41

Figure S92. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'MIRseq Mature CNMF subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.485 (Chi-square test), Q value = 1

Table S100. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	92	27	55	33	21
subtype1	28	8	12	7	10
subtype2	16	5	11	6	2
subtype3	23	8	13	14	4
subtype4	25	6	19	6	5

Figure S93. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'MIRseq Mature CNMF subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0297 (Chi-square test), Q value = 1

Table S101. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	181	22	1	13
subtype1	53	2	0	5
subtype2	32	2	1	3
subtype3	45	13	0	3
subtype4	51	5	0	2

Figure S94. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

'MIRseq Mature CNMF subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.668 (ANOVA), Q value = 1

Table S102. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	187	3.5 (5.6)
subtype1	50	3.1 (4.7)
subtype2	35	2.8 (4.6)
subtype3	53	4.1 (5.6)
subtype4	49	3.8 (7.1)

Figure S95. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature CNMF subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.0426 (Chi-square test), Q value = 1

Table S103. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	131	16	55	22	3
subtype1	45	6	9	3	1
subtype2	23	3	9	5	0
subtype3	29	0	22	10	1
subtype4	34	7	15	4	1

Figure S96. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'MIRseq Mature CNMF subtypes' versus 'MULTIFOCALITY'

P value = 0.0785 (Fisher's exact test), Q value = 1

Table S104. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	104	121
subtype1	35	29
subtype2	12	27
subtype3	26	36
subtype4	31	29

Figure S97. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRseq Mature CNMF subtypes' versus 'TUMOR.SIZE'

P value = 0.708 (ANOVA), Q value = 1

Table S105. Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	185	2.8 (1.6)
subtype1	50	3.0 (1.6)
subtype2	35	2.9 (1.4)
subtype3	49	2.7 (1.6)
subtype4	51	2.7 (1.7)

Figure S98. Get High-res Image Clustering Approach #7: 'MIRseq Mature CNMF subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

Clustering Approach #8: 'MIRseq Mature cHierClus subtypes'

Table S106. Get Full Table Description of clustering approach #8: 'MIRseq Mature cHierClus subtypes'

Cluster Labels	1	2	3
Number of samples	108	21	99

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

P value = 0.0256 (logrank test), Q value = 1

Table S107. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	226	5	0.0 - 158.8 (15.1)
subtype1	107	0	0.2 - 158.8 (15.4)
subtype2	21	0	0.0 - 116.9 (20.9)
subtype3	98	5	0.2 - 157.2 (12.9)

Figure S99. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq Mature cHierClus subtypes' versus 'AGE'

P value = 0.00386 (ANOVA), Q value = 0.38

Table S108. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	228	46.9 (14.9)
subtype1	108	44.6 (14.2)
subtype2	21	41.6 (13.8)
subtype3	99	50.5 (15.1)

Figure S100. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq Mature cHierClus subtypes' versus 'GENDER'

P value = 0.645 (Fisher's exact test), Q value = 1

Table S109. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	168	60
subtype1	79	29
subtype2	14	7
subtype3	75	24

Figure S101. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'MIRseq Mature cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.0166 (Chi-square test), Q value = 1

Table S110. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: '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	3	183	16	26
subtype1	2	88	11	7
subtype2	0	21	0	0
subtype3	1	74	5	19

Figure S102. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #4: 'HISTOLOGICAL.TYPE'

'MIRseq Mature cHierClus subtypes' versus 'RADIATIONS.RADIATION.REGIMENINDICATION'

P value = 0.186 (Fisher's exact test), Q value = 1

Table S111. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients	NO	YES
ALL	11	217
subtype1	8	100
subtype2	1	20
subtype3	2	97

Figure S103. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #5: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq Mature cHierClus subtypes' versus 'RADIATIONEXPOSURE'

P value = 0.663 (Fisher's exact test), Q value = 1

Table S112. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

nPatients	NO	YES
ALL	196	8
subtype1	89	5
subtype2	20	0
subtype3	87	3

Figure S104. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONEXPOSURE'

'MIRseq Mature cHierClus subtypes' versus 'DISTANT.METASTASIS'

P value = 0.129 (Chi-square test), Q value = 1

Table S113. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

nPatients	M0	M1	MX
ALL	131	4	93
subtype1	59	3	46
subtype2	8	0	13
subtype3	64	1	34

Figure S105. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #7: 'DISTANT.METASTASIS'

'MIRseq Mature cHierClus subtypes' versus 'EXTRATHYROIDAL.EXTENSION'

P value = 0.000714 (Chi-square test), Q value = 0.076

Table S114. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

nPatients	MINIMAL (T3)	MODERATE/ADVANCED (T4A)	NONE
ALL	75	9	139
subtype1	26	3	75
subtype2	3	0	18
subtype3	46	6	46

Figure S106. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #8: 'EXTRATHYROIDAL.EXTENSION'

'MIRseq Mature cHierClus subtypes' versus 'LYMPH.NODE.METASTASIS'

P value = 0.72 (Chi-square test), Q value = 1

Table S115. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

nPatients	N0	N1	N1A	N1B	NX
ALL	92	27	55	33	21
subtype1	46	15	27	11	9
subtype2	9	1	5	3	3
subtype3	37	11	23	19	9

Figure S107. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #9: 'LYMPH.NODE.METASTASIS'

'MIRseq Mature cHierClus subtypes' versus 'COMPLETENESS.OF.RESECTION'

P value = 0.0293 (Chi-square test), Q value = 1

Table S116. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

nPatients	R0	R1	R2	RX
ALL	181	22	1	13
subtype1	89	5	0	7
subtype2	19	0	0	0
subtype3	73	17	1	6

Figure S108. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #10: 'COMPLETENESS.OF.RESECTION'

'MIRseq Mature cHierClus subtypes' versus 'NUMBER.OF.LYMPH.NODES'

P value = 0.0718 (ANOVA), Q value = 1

Table S117. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

	nPatients	Mean (Std.Dev)
ALL	187	3.5 (5.6)
subtype1	90	2.8 (4.3)
subtype2	14	2.1 (2.9)
subtype3	83	4.6 (6.9)

Figure S109. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #11: 'NUMBER.OF.LYMPH.NODES'

'MIRseq Mature cHierClus subtypes' versus 'NEOPLASM.DISEASESTAGE'

P value = 0.00737 (Chi-square test), Q value = 0.69

Table S118. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

nPatients	STAGE I	STAGE II	STAGE III	STAGE IVA	STAGE IVC
ALL	131	16	55	22	3
subtype1	67	11	19	8	2
subtype2	15	3	1	2	0
subtype3	49	2	35	12	1

Figure S110. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #13: 'NEOPLASM.DISEASESTAGE'

'MIRseq Mature cHierClus subtypes' versus 'MULTIFOCALITY'

P value = 0.417 (Fisher's exact test), Q value = 1

Table S119. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

nPatients	MULTIFOCAL	UNIFOCAL
ALL	104	121
subtype1	52	53
subtype2	11	10
subtype3	41	58

Figure S111. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #14: 'MULTIFOCALITY'

'MIRseq Mature cHierClus subtypes' versus 'TUMOR.SIZE'

P value = 0.871 (ANOVA), Q value = 1

Table S120. Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

	nPatients	Mean (Std.Dev)
ALL	185	2.8 (1.6)
subtype1	87	2.9 (1.6)
subtype2	17	2.9 (1.4)
subtype3	81	2.8 (1.7)

Figure S112. Get High-res Image Clustering Approach #8: 'MIRseq Mature cHierClus subtypes' versus Clinical Feature #15: 'TUMOR.SIZE'

Methods & Data

Input

Cluster data file = THCA-Mut_BRAF.mergedcluster.txt
Clinical data file = THCA-Mut_BRAF.clin.merged.picked.txt
Number of patients = 229
Number of clustering approaches = 8
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

ANOVA analysis

For continuous numerical clinical features, one-way analysis of variance (Howell 2002) was applied to compare the clinical values between tumor subtypes using 'anova' 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

Chi-square test

For multi-class clinical features (nominal or ordinal), Chi-square tests (Greenwood and Nikulin 1996) were used to estimate the P values using the 'chisq.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.

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] Howell, D, Statistical Methods for Psychology. (5th ed.), Duxbury Press:324-5 (2002)

[5] 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)

[6] Greenwood and Nikulin, A guide to chi-squared testing, Wiley, New York. ISBN 047155779X (1996)

[7] 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