Head and Neck Squamous Cell Carcinoma: Correlation between molecular cancer subtypes and selected clinical features
Maintained by TCGA GDAC Team (Broad Institute/Dana-Farber Cancer Institute/Harvard Medical School)
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 7 different clustering approaches and 8 clinical features across 311 patients, 9 significant findings detected with P value < 0.05.

  • 8 subtypes identified in current cancer cohort by 'METHLYATION CNMF'. These subtypes correlate to 'Time to Death',  'AGE', and 'GENDER'.

  • CNMF clustering analysis on RPPA data identified 3 subtypes that correlate to 'Time to Death',  'PATHOLOGY.N', and 'NEOADJUVANT.THERAPY'.

  • Consensus hierarchical clustering analysis on RPPA data identified 3 subtypes that do not correlate to any clinical features.

  • CNMF clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'TUMOR.STAGE'.

  • Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 3 subtypes that correlate to 'GENDER'.

  • CNMF clustering analysis on sequencing-based miR expression data identified 3 subtypes that do not correlate to any clinical features.

  • Consensus hierarchical clustering analysis on sequencing-based miR expression data identified 4 subtypes that correlate to 'NEOADJUVANT.THERAPY'.

Results
Overview of the results

Table 1.  Get Full Table Overview of the association between subtypes identified by 7 different clustering approaches and 8 clinical features. Shown in the table are P values from statistical tests. Thresholded by P value < 0.05, 9 significant findings detected.

Clinical
Features 		Statistical
Tests 		METHLYATION
CNMF 		RPPA
CNMF
subtypes 		RPPA
cHierClus
subtypes 		RNAseq
CNMF
subtypes 		RNAseq
cHierClus
subtypes 		MIRseq
CNMF
subtypes 		MIRseq
cHierClus
subtypes
Time to Death logrank test 0.000207 0.0351 0.103 0.205 0.0749 0.725 0.158
AGE ANOVA 0.000106 0.434 0.303 0.925 0.383 0.73 0.688
GENDER Fisher's exact test 0.0445 0.139 0.842 0.188 0.00985 0.0643 0.637
PATHOLOGY T Chi-square test 0.0887 0.216 0.194 0.225 0.298 0.121 0.395
PATHOLOGY N Chi-square test 0.439 0.0063 0.0534 0.194 0.0772 0.109 0.191
TUMOR STAGE Chi-square test 0.272 0.0532 0.103 0.0364 0.39 0.451 0.321
RADIATIONS RADIATION REGIMENINDICATION Fisher's exact test 0.715 0.301 0.595 0.413 0.546 0.533 0.551
NEOADJUVANT THERAPY Fisher's exact test 0.753 0.023 0.0821 0.207 0.068 0.689 0.0179
Clustering Approach #1: 'METHLYATION CNMF'

Table S1.  Get Full Table Description of clustering approach #1: 'METHLYATION CNMF'

Cluster Labels 1 2 3 4 5 6 7 8
Number of samples 43 47 40 44 58 53 5 19
'METHLYATION CNMF' versus 'Time to Death'

P value = 0.000207 (logrank test)

Table S2.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 306 121 0.1 - 210.9 (14.6)
subtype1 43 15 0.1 - 142.5 (13.2)
subtype2 47 19 0.2 - 126.1 (13.6)
subtype3 39 20 0.1 - 156.5 (12.7)
subtype4 44 8 0.8 - 95.0 (22.3)
subtype5 58 30 0.5 - 135.3 (14.4)
subtype6 51 19 1.5 - 210.9 (14.3)
subtype7 5 1 5.0 - 28.6 (10.7)
subtype8 19 9 1.0 - 84.5 (18.0)

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

'METHLYATION CNMF' versus 'AGE'

P value = 0.000106 (ANOVA)

Table S3.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 309 61.1 (12.1)
subtype1 43 58.0 (10.6)
subtype2 47 55.5 (13.8)
subtype3 40 62.4 (11.1)
subtype4 44 58.1 (10.9)
subtype5 58 64.4 (10.1)
subtype6 53 63.9 (13.2)
subtype7 5 63.8 (11.7)
subtype8 19 67.4 (10.3)

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

'METHLYATION CNMF' versus 'GENDER'

P value = 0.0445 (Chi-square test)

Table S4.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 85 224
subtype1 7 36
subtype2 17 30
subtype3 15 25
subtype4 5 39
subtype5 15 43
subtype6 18 35
subtype7 1 4
subtype8 7 12

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

'METHLYATION CNMF' versus 'PATHOLOGY.T'

P value = 0.0887 (Chi-square test)

Table S5.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 24 78 63 102
subtype1 2 10 11 17
subtype2 7 14 12 10
subtype3 1 10 5 19
subtype4 4 9 5 8
subtype5 1 13 13 23
subtype6 8 18 8 17
subtype7 1 0 2 1
subtype8 0 4 7 7

Figure S4.  Get High-res Image Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #4: 'PATHOLOGY.T'

'METHLYATION CNMF' versus 'PATHOLOGY.N'

P value = 0.439 (Chi-square test)

Table S6.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 101 32 100 5
subtype1 21 4 11 1
subtype2 19 4 14 0
subtype3 15 4 9 2
subtype4 9 5 9 0
subtype5 12 5 24 2
subtype6 18 8 22 0
subtype7 1 0 3 0
subtype8 6 2 8 0

Figure S5.  Get High-res Image Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #5: 'PATHOLOGY.N'

'METHLYATION CNMF' versus 'TUMOR.STAGE'

P value = 0.272 (Chi-square test)

Table S7.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 17 47 42 156
subtype1 1 7 8 24
subtype2 6 9 10 17
subtype3 1 8 2 24
subtype4 2 3 6 13
subtype5 1 8 6 34
subtype6 5 9 7 29
subtype7 1 0 0 3
subtype8 0 3 3 12

Figure S6.  Get High-res Image Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.715 (Chi-square test)

Table S8.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 78 231
subtype1 7 36
subtype2 10 37
subtype3 11 29
subtype4 10 34
subtype5 16 42
subtype6 17 36
subtype7 2 3
subtype8 5 14

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

'METHLYATION CNMF' versus 'NEOADJUVANT.THERAPY'

P value = 0.753 (Chi-square test)

Table S9.  Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 48 261
subtype1 5 38
subtype2 5 42
subtype3 6 34
subtype4 7 37
subtype5 10 48
subtype6 10 43
subtype7 2 3
subtype8 3 16

Figure S8.  Get High-res Image Clustering Approach #1: 'METHLYATION CNMF' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Clustering Approach #2: 'RPPA CNMF subtypes'

Table S10.  Get Full Table Description of clustering approach #2: 'RPPA CNMF subtypes'

Cluster Labels 1 2 3
Number of samples 82 85 45
'RPPA CNMF subtypes' versus 'Time to Death'

P value = 0.0351 (logrank test)

Table S11.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 212 107 0.1 - 210.9 (13.2)
subtype1 82 42 1.5 - 129.2 (12.1)
subtype2 85 47 0.1 - 210.9 (13.1)
subtype3 45 18 2.1 - 156.5 (20.5)

Figure S9.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA CNMF subtypes' versus 'AGE'

P value = 0.434 (ANOVA)

Table S12.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 212 62.1 (12.2)
subtype1 82 61.1 (12.1)
subtype2 85 63.4 (12.1)
subtype3 45 61.4 (12.7)

Figure S10.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RPPA CNMF subtypes' versus 'GENDER'

P value = 0.139 (Fisher's exact test)

Table S13.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 62 150
subtype1 19 63
subtype2 25 60
subtype3 18 27

Figure S11.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'RPPA CNMF subtypes' versus 'PATHOLOGY.T'

P value = 0.216 (Chi-square test)

Table S14.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 13 59 53 79
subtype1 3 24 16 34
subtype2 6 18 27 32
subtype3 4 17 10 13

Figure S12.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

'RPPA CNMF subtypes' versus 'PATHOLOGY.N'

P value = 0.0063 (Chi-square test)

Table S15.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 73 20 79 4
subtype1 27 9 30 1
subtype2 19 8 40 2
subtype3 27 3 9 1

Figure S13.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

'RPPA CNMF subtypes' versus 'TUMOR.STAGE'

P value = 0.0532 (Chi-square test)

Table S16.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 9 39 31 121
subtype1 1 17 12 47
subtype2 4 9 12 55
subtype3 4 13 7 19

Figure S14.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.301 (Fisher's exact test)

Table S17.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 56 156
subtype1 25 57
subtype2 23 62
subtype3 8 37

Figure S15.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RPPA CNMF subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.023 (Fisher's exact test)

Table S18.  Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 36 176
subtype1 18 64
subtype2 16 69
subtype3 2 43

Figure S16.  Get High-res Image Clustering Approach #2: 'RPPA CNMF subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Clustering Approach #3: 'RPPA cHierClus subtypes'

Table S19.  Get Full Table Description of clustering approach #3: 'RPPA cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 115 81 16
'RPPA cHierClus subtypes' versus 'Time to Death'

P value = 0.103 (logrank test)

Table S20.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 212 107 0.1 - 210.9 (13.2)
subtype1 115 56 0.1 - 210.9 (13.2)
subtype2 81 42 1.5 - 156.5 (13.6)
subtype3 16 9 3.3 - 52.3 (11.6)

Figure S17.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RPPA cHierClus subtypes' versus 'AGE'

P value = 0.303 (ANOVA)

Table S21.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 212 62.1 (12.2)
subtype1 115 63.3 (11.2)
subtype2 81 61.1 (13.1)
subtype3 16 59.4 (14.3)

Figure S18.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RPPA cHierClus subtypes' versus 'GENDER'

P value = 0.842 (Fisher's exact test)

Table S22.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 62 150
subtype1 35 80
subtype2 22 59
subtype3 5 11

Figure S19.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RPPA cHierClus subtypes' versus 'PATHOLOGY.T'

P value = 0.194 (Chi-square test)

Table S23.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 13 59 53 79
subtype1 7 25 36 42
subtype2 5 30 13 30
subtype3 1 4 4 7

Figure S20.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

'RPPA cHierClus subtypes' versus 'PATHOLOGY.N'

P value = 0.0534 (Chi-square test)

Table S24.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 73 20 79 4
subtype1 39 8 46 3
subtype2 32 11 22 1
subtype3 2 1 11 0

Figure S21.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

'RPPA cHierClus subtypes' versus 'TUMOR.STAGE'

P value = 0.103 (Chi-square test)

Table S25.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 9 39 31 121
subtype1 6 16 18 69
subtype2 3 21 13 39
subtype3 0 2 0 13

Figure S22.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.595 (Fisher's exact test)

Table S26.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 56 156
subtype1 30 85
subtype2 20 61
subtype3 6 10

Figure S23.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RPPA cHierClus subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.0821 (Fisher's exact test)

Table S27.  Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 36 176
subtype1 19 96
subtype2 11 70
subtype3 6 10

Figure S24.  Get High-res Image Clustering Approach #3: 'RPPA cHierClus subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Clustering Approach #4: 'RNAseq CNMF subtypes'

Table S28.  Get Full Table Description of clustering approach #4: 'RNAseq CNMF subtypes'

Cluster Labels 1 2 3
Number of samples 118 93 91
'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.205 (logrank test)

Table S29.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 299 120 0.1 - 210.9 (14.8)
subtype1 118 42 0.1 - 135.3 (14.2)
subtype2 91 40 0.2 - 142.5 (15.7)
subtype3 90 38 0.1 - 210.9 (14.0)

Figure S25.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.925 (ANOVA)

Table S30.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 302 61.0 (12.2)
subtype1 118 61.3 (11.1)
subtype2 93 60.6 (12.8)
subtype3 91 61.2 (12.9)

Figure S26.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.188 (Fisher's exact test)

Table S31.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 81 221
subtype1 25 93
subtype2 27 66
subtype3 29 62

Figure S27.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.T'

P value = 0.225 (Chi-square test)

Table S32.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 23 78 60 101
subtype1 6 25 24 36
subtype2 13 25 19 29
subtype3 4 28 17 36

Figure S28.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

'RNAseq CNMF subtypes' versus 'PATHOLOGY.N'

P value = 0.194 (Chi-square test)

Table S33.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 100 32 97 5
subtype1 34 10 32 3
subtype2 28 10 42 1
subtype3 38 12 23 1

Figure S29.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

'RNAseq CNMF subtypes' versus 'TUMOR.STAGE'

P value = 0.0364 (Chi-square test)

Table S34.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 16 47 41 153
subtype1 2 15 17 54
subtype2 11 12 10 52
subtype3 3 20 14 47

Figure S30.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.413 (Fisher's exact test)

Table S35.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 75 227
subtype1 25 93
subtype2 27 66
subtype3 23 68

Figure S31.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq CNMF subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.207 (Fisher's exact test)

Table S36.  Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 46 256
subtype1 17 101
subtype2 19 74
subtype3 10 81

Figure S32.  Get High-res Image Clustering Approach #4: 'RNAseq CNMF subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Clustering Approach #5: 'RNAseq cHierClus subtypes'

Table S37.  Get Full Table Description of clustering approach #5: 'RNAseq cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 97 85 120
'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0749 (logrank test)

Table S38.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 299 120 0.1 - 210.9 (14.8)
subtype1 95 43 1.5 - 142.5 (15.1)
subtype2 84 36 0.1 - 210.9 (14.0)
subtype3 120 41 0.1 - 135.3 (15.1)

Figure S33.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.383 (ANOVA)

Table S39.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 302 61.0 (12.2)
subtype1 97 60.0 (13.3)
subtype2 85 62.5 (12.9)
subtype3 120 60.9 (10.5)

Figure S34.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.00985 (Fisher's exact test)

Table S40.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 81 221
subtype1 31 66
subtype2 29 56
subtype3 21 99

Figure S35.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.T'

P value = 0.298 (Chi-square test)

Table S41.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 23 78 60 101
subtype1 13 27 22 30
subtype2 4 26 15 32
subtype3 6 25 23 39

Figure S36.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

'RNAseq cHierClus subtypes' versus 'PATHOLOGY.N'

P value = 0.0772 (Chi-square test)

Table S42.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 100 32 97 5
subtype1 27 12 45 1
subtype2 37 10 20 1
subtype3 36 10 32 3

Figure S37.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

'RNAseq cHierClus subtypes' versus 'TUMOR.STAGE'

P value = 0.39 (Chi-square test)

Table S43.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 16 47 41 153
subtype1 9 14 12 56
subtype2 4 18 13 41
subtype3 3 15 16 56

Figure S38.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.546 (Fisher's exact test)

Table S44.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 75 227
subtype1 28 69
subtype2 20 65
subtype3 27 93

Figure S39.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'RNAseq cHierClus subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.068 (Fisher's exact test)

Table S45.  Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 46 256
subtype1 21 76
subtype2 8 77
subtype3 17 103

Figure S40.  Get High-res Image Clustering Approach #5: 'RNAseq cHierClus subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Clustering Approach #6: 'MIRseq CNMF subtypes'

Table S46.  Get Full Table Description of clustering approach #6: 'MIRseq CNMF subtypes'

Cluster Labels 1 2 3
Number of samples 83 131 94
'MIRseq CNMF subtypes' versus 'Time to Death'

P value = 0.725 (logrank test)

Table S47.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 305 122 0.1 - 210.9 (15.0)
subtype1 82 30 0.1 - 210.9 (12.1)
subtype2 131 52 0.1 - 142.5 (14.3)
subtype3 92 40 1.8 - 126.1 (18.1)

Figure S41.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq CNMF subtypes' versus 'AGE'

P value = 0.73 (ANOVA)

Table S48.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 308 61.1 (12.1)
subtype1 83 60.4 (12.5)
subtype2 131 61.7 (11.4)
subtype3 94 60.9 (12.7)

Figure S42.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq CNMF subtypes' versus 'GENDER'

P value = 0.0643 (Fisher's exact test)

Table S49.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 86 222
subtype1 16 67
subtype2 37 94
subtype3 33 61

Figure S43.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

'MIRseq CNMF subtypes' versus 'PATHOLOGY.T'

P value = 0.121 (Chi-square test)

Table S50.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 24 78 62 103
subtype1 5 23 11 30
subtype2 7 26 29 47
subtype3 12 29 22 26

Figure S44.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

'MIRseq CNMF subtypes' versus 'PATHOLOGY.N'

P value = 0.109 (Chi-square test)

Table S51.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 99 32 101 5
subtype1 30 8 21 2
subtype2 44 9 41 3
subtype3 25 15 39 0

Figure S45.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

'MIRseq CNMF subtypes' versus 'TUMOR.STAGE'

P value = 0.451 (Chi-square test)

Table S52.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 17 46 41 158
subtype1 4 15 8 41
subtype2 4 16 18 68
subtype3 9 15 15 49

Figure S46.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.533 (Fisher's exact test)

Table S53.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 77 231
subtype1 21 62
subtype2 29 102
subtype3 27 67

Figure S47.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq CNMF subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.689 (Fisher's exact test)

Table S54.  Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 48 260
subtype1 11 72
subtype2 20 111
subtype3 17 77

Figure S48.  Get High-res Image Clustering Approach #6: 'MIRseq CNMF subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Clustering Approach #7: 'MIRseq cHierClus subtypes'

Table S55.  Get Full Table Description of clustering approach #7: 'MIRseq cHierClus subtypes'

Cluster Labels 1 2 3 4
Number of samples 13 12 167 116
'MIRseq cHierClus subtypes' versus 'Time to Death'

P value = 0.158 (logrank test)

Table S56.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 305 122 0.1 - 210.9 (15.0)
subtype1 13 5 0.2 - 56.7 (12.0)
subtype2 12 5 9.0 - 84.4 (12.8)
subtype3 166 61 0.1 - 210.9 (14.0)
subtype4 114 51 0.5 - 156.5 (17.1)

Figure S49.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq cHierClus subtypes' versus 'AGE'

P value = 0.688 (ANOVA)

Table S57.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 308 61.1 (12.1)
subtype1 13 58.4 (13.2)
subtype2 12 60.8 (12.0)
subtype3 167 61.8 (11.5)
subtype4 116 60.4 (12.9)

Figure S50.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq cHierClus subtypes' versus 'GENDER'

P value = 0.637 (Fisher's exact test)

Table S58.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 86 222
subtype1 4 9
subtype2 4 8
subtype3 42 125
subtype4 36 80

Figure S51.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'MIRseq cHierClus subtypes' versus 'PATHOLOGY.T'

P value = 0.395 (Chi-square test)

Table S59.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

nPatients T1 T2 T3 T4
ALL 24 78 62 103
subtype1 2 5 2 3
subtype2 0 3 2 5
subtype3 8 38 30 60
subtype4 14 32 28 35

Figure S52.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #4: 'PATHOLOGY.T'

'MIRseq cHierClus subtypes' versus 'PATHOLOGY.N'

P value = 0.191 (Chi-square test)

Table S60.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

nPatients N0 N1 N2 N3
ALL 99 32 101 5
subtype1 8 1 3 0
subtype2 5 0 5 0
subtype3 54 17 42 4
subtype4 32 14 51 1

Figure S53.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #5: 'PATHOLOGY.N'

'MIRseq cHierClus subtypes' versus 'TUMOR.STAGE'

P value = 0.321 (Chi-square test)

Table S61.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

nPatients I II III IV
ALL 17 46 41 158
subtype1 2 4 2 4
subtype2 0 1 1 8
subtype3 5 24 23 80
subtype4 10 17 15 66

Figure S54.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #6: 'TUMOR.STAGE'

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

P value = 0.551 (Fisher's exact test)

Table S62.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 77 231
subtype1 3 10
subtype2 4 8
subtype3 37 130
subtype4 33 83

Figure S55.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #7: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq cHierClus subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.0179 (Fisher's exact test)

Table S63.  Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 48 260
subtype1 0 13
subtype2 3 9
subtype3 19 148
subtype4 26 90

Figure S56.  Get High-res Image Clustering Approach #7: 'MIRseq cHierClus subtypes' versus Clinical Feature #8: 'NEOADJUVANT.THERAPY'

Methods & Data
Input

  • Cluster data file = HNSC.mergedcluster.txt

  • Clinical data file = HNSC.clin.merged.picked.txt

  • Number of patients = 311

  • Number of clustering approaches = 7

  • Number of selected clinical features = 8

  • 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

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

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

Download Results

This is an experimental feature. The full results of the analysis summarized in this report can be downloaded from the TCGA Data Coordination Center.

References
[1] Brunet et al., Metagenes and molecular pattern discovery using matrix factorization, PNAS 101(12):4164-9 (2004)
[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] Greenwood and Nikulin, A guide to chi-squared testing, Wiley, New York. ISBN 047155779X (1996)
[6] 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)
Copyright © 2012 Broad Institute TCGA GDAC as part of the TCGA Research Network. All rights reserved.
Made with Nozzle