Brain Lower Grade Glioma: 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 4 different clustering approaches and 7 clinical features across 67 patients, 5 significant findings detected with P value < 0.05.

  • CNMF clustering analysis on array-based mRNA expression data identified 3 subtypes that correlate to 'HISTOLOGICAL.TYPE'.

  • Consensus hierarchical clustering analysis on array-based mRNA expression data identified 3 subtypes that correlate to 'HISTOLOGICAL.TYPE'.

  • 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 'AGE',  'GENDER', and 'HISTOLOGICAL.TYPE'.

Results
Overview of the results

Table 1.  Get Full Table Overview of the association between subtypes identified by 4 different clustering approaches and 7 clinical features. Shown in the table are P values from statistical tests. Thresholded by P value < 0.05, 5 significant findings detected.

Clinical
Features 		Statistical
Tests 		mRNA
CNMF
subtypes 		mRNA
cHierClus
subtypes 		MIRseq
CNMF
subtypes 		MIRseq
cHierClus
subtypes
Time to Death logrank test 0.0944 0.126 0.195 0.112
AGE ANOVA 0.326 0.467 0.285 0.0293
GENDER Fisher's exact test 0.101 0.172 0.172 0.0238
KARNOFSKY PERFORMANCE SCORE ANOVA 0.441 0.441 0.719 0.729
HISTOLOGICAL TYPE Chi-square test 0.0226 0.0136 0.469 0.00256
RADIATIONS RADIATION REGIMENINDICATION Fisher's exact test 0.384 0.757 0.79 0.339
NEOADJUVANT THERAPY Fisher's exact test 0.883 0.685 0.102 0.0789
Clustering Approach #1: 'mRNA CNMF subtypes'

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

Cluster Labels 1 2 3
Number of samples 9 10 8
'mRNA CNMF subtypes' versus 'Time to Death'

P value = 0.0944 (logrank test)

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

nPatients nDeath Duration Range (Median), Month
ALL 26 9 8.0 - 134.3 (47.3)
subtype1 9 4 10.6 - 130.8 (43.9)
subtype2 9 3 8.0 - 78.2 (41.1)
subtype3 8 2 14.4 - 134.3 (51.3)

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

'mRNA CNMF subtypes' versus 'AGE'

P value = 0.326 (ANOVA)

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

nPatients Mean (Std.Dev)
ALL 27 39.3 (9.1)
subtype1 9 39.2 (6.2)
subtype2 10 42.3 (7.6)
subtype3 8 35.8 (12.6)

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

'mRNA CNMF subtypes' versus 'GENDER'

P value = 0.101 (Fisher's exact test)

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

nPatients FEMALE MALE
ALL 9 18
subtype1 2 7
subtype2 6 4
subtype3 1 7

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

'mRNA CNMF subtypes' versus 'KARNOFSKY.PERFORMANCE.SCORE'

P value = 0.441 (ANOVA)

Table S5.  Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

nPatients Mean (Std.Dev)
ALL 17 88.8 (12.2)
subtype1 7 84.3 (16.2)
subtype2 7 92.9 (7.6)
subtype3 3 90.0 (10.0)

Figure S4.  Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

'mRNA CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.0226 (Chi-square test)

Table S6.  Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

nPatients ASTROCYTOMA OLIGOASTROCYTOMA OLIGODENDROGLIOMA
ALL 10 9 8
subtype1 7 2 0
subtype2 2 3 5
subtype3 1 4 3

Figure S5.  Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

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

P value = 0.384 (Fisher's exact test)

Table S7.  Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 19 8
subtype1 7 2
subtype2 8 2
subtype3 4 4

Figure S6.  Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'mRNA CNMF subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.883 (Fisher's exact test)

Table S8.  Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 10 17
subtype1 4 5
subtype2 3 7
subtype3 3 5

Figure S7.  Get High-res Image Clustering Approach #1: 'mRNA CNMF subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

Clustering Approach #2: 'mRNA cHierClus subtypes'

Table S9.  Get Full Table Description of clustering approach #2: 'mRNA cHierClus subtypes'

Cluster Labels 1 2 3
Number of samples 9 7 11
'mRNA cHierClus subtypes' versus 'Time to Death'

P value = 0.126 (logrank test)

Table S10.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 26 9 8.0 - 134.3 (47.3)
subtype1 9 4 10.6 - 130.8 (43.9)
subtype2 7 2 14.4 - 134.3 (52.4)
subtype3 10 3 8.0 - 78.2 (43.9)

Figure S8.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'mRNA cHierClus subtypes' versus 'AGE'

P value = 0.467 (ANOVA)

Table S11.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 27 39.3 (9.1)
subtype1 9 39.2 (6.2)
subtype2 7 36.0 (13.6)
subtype3 11 41.5 (7.6)

Figure S9.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'mRNA cHierClus subtypes' versus 'GENDER'

P value = 0.172 (Fisher's exact test)

Table S12.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 9 18
subtype1 2 7
subtype2 1 6
subtype3 6 5

Figure S10.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'mRNA cHierClus subtypes' versus 'KARNOFSKY.PERFORMANCE.SCORE'

P value = 0.441 (ANOVA)

Table S13.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

nPatients Mean (Std.Dev)
ALL 17 88.8 (12.2)
subtype1 7 84.3 (16.2)
subtype2 3 90.0 (10.0)
subtype3 7 92.9 (7.6)

Figure S11.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

'mRNA cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.0136 (Chi-square test)

Table S14.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

nPatients ASTROCYTOMA OLIGOASTROCYTOMA OLIGODENDROGLIOMA
ALL 10 9 8
subtype1 7 2 0
subtype2 0 4 3
subtype3 3 3 5

Figure S12.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

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

P value = 0.757 (Fisher's exact test)

Table S15.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 19 8
subtype1 7 2
subtype2 4 3
subtype3 8 3

Figure S13.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'mRNA cHierClus subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.685 (Fisher's exact test)

Table S16.  Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 10 17
subtype1 4 5
subtype2 3 4
subtype3 3 8

Figure S14.  Get High-res Image Clustering Approach #2: 'mRNA cHierClus subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

Clustering Approach #3: 'MIRseq CNMF subtypes'

Table S17.  Get Full Table Description of clustering approach #3: 'MIRseq CNMF subtypes'

Cluster Labels 1 2 3
Number of samples 22 25 20
'MIRseq CNMF subtypes' versus 'Time to Death'

P value = 0.195 (logrank test)

Table S18.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 66 27 1.2 - 211.2 (24.9)
subtype1 22 12 4.7 - 182.3 (19.4)
subtype2 24 9 7.7 - 134.3 (47.3)
subtype3 20 6 1.2 - 211.2 (16.3)

Figure S15.  Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq CNMF subtypes' versus 'AGE'

P value = 0.285 (ANOVA)

Table S19.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 67 43.1 (12.1)
subtype1 22 42.7 (11.2)
subtype2 25 40.8 (9.6)
subtype3 20 46.5 (15.2)

Figure S16.  Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq CNMF subtypes' versus 'GENDER'

P value = 0.172 (Fisher's exact test)

Table S20.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 33 34
subtype1 14 8
subtype2 9 16
subtype3 10 10

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

'MIRseq CNMF subtypes' versus 'KARNOFSKY.PERFORMANCE.SCORE'

P value = 0.719 (ANOVA)

Table S21.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

nPatients Mean (Std.Dev)
ALL 41 89.8 (9.6)
subtype1 13 89.2 (7.6)
subtype2 16 88.8 (12.6)
subtype3 12 91.7 (7.2)

Figure S18.  Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

'MIRseq CNMF subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.469 (Chi-square test)

Table S22.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

nPatients ASTROCYTOMA OLIGOASTROCYTOMA OLIGODENDROGLIOMA
ALL 24 20 23
subtype1 8 7 7
subtype2 10 9 6
subtype3 6 4 10

Figure S19.  Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

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

P value = 0.79 (Fisher's exact test)

Table S23.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 19 48
subtype1 5 17
subtype2 8 17
subtype3 6 14

Figure S20.  Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq CNMF subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.102 (Fisher's exact test)

Table S24.  Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 28 39
subtype1 13 9
subtype2 7 18
subtype3 8 12

Figure S21.  Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

Clustering Approach #4: 'MIRseq cHierClus subtypes'

Table S25.  Get Full Table Description of clustering approach #4: 'MIRseq cHierClus subtypes'

Cluster Labels 1 2 3 4
Number of samples 25 15 13 14
'MIRseq cHierClus subtypes' versus 'Time to Death'

P value = 0.112 (logrank test)

Table S26.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

nPatients nDeath Duration Range (Median), Month
ALL 66 27 1.2 - 211.2 (24.9)
subtype1 24 9 7.7 - 134.3 (47.3)
subtype2 15 8 6.4 - 211.2 (13.4)
subtype3 13 6 4.7 - 114.0 (18.9)
subtype4 14 4 1.2 - 182.3 (24.1)

Figure S22.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq cHierClus subtypes' versus 'AGE'

P value = 0.0293 (ANOVA)

Table S27.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

nPatients Mean (Std.Dev)
ALL 67 43.1 (12.1)
subtype1 25 40.1 (9.1)
subtype2 15 47.8 (12.4)
subtype3 13 38.0 (9.3)
subtype4 14 48.3 (15.6)

Figure S23.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq cHierClus subtypes' versus 'GENDER'

P value = 0.0238 (Fisher's exact test)

Table S28.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients FEMALE MALE
ALL 33 34
subtype1 8 17
subtype2 11 4
subtype3 9 4
subtype4 5 9

Figure S24.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

'MIRseq cHierClus subtypes' versus 'KARNOFSKY.PERFORMANCE.SCORE'

P value = 0.729 (ANOVA)

Table S29.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

nPatients Mean (Std.Dev)
ALL 41 89.8 (9.6)
subtype1 16 88.8 (12.6)
subtype2 12 88.3 (5.8)
subtype3 8 92.5 (4.6)
subtype4 5 92.0 (13.0)

Figure S25.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

'MIRseq cHierClus subtypes' versus 'HISTOLOGICAL.TYPE'

P value = 0.00256 (Chi-square test)

Table S30.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

nPatients ASTROCYTOMA OLIGOASTROCYTOMA OLIGODENDROGLIOMA
ALL 24 20 23
subtype1 11 9 5
subtype2 9 4 2
subtype3 4 4 5
subtype4 0 3 11

Figure S26.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #5: 'HISTOLOGICAL.TYPE'

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

P value = 0.339 (Fisher's exact test)

Table S31.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

nPatients NO YES
ALL 19 48
subtype1 8 17
subtype2 2 13
subtype3 3 10
subtype4 6 8

Figure S27.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #6: 'RADIATIONS.RADIATION.REGIMENINDICATION'

'MIRseq cHierClus subtypes' versus 'NEOADJUVANT.THERAPY'

P value = 0.0789 (Fisher's exact test)

Table S32.  Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

nPatients NO YES
ALL 28 39
subtype1 6 19
subtype2 6 9
subtype3 8 5
subtype4 8 6

Figure S28.  Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #7: 'NEOADJUVANT.THERAPY'

Methods & Data
Input

  • Cluster data file = LGG.mergedcluster.txt

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

  • Number of patients = 67

  • Number of clustering approaches = 4

  • Number of selected clinical features = 7

  • 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

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] 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)
Copyright © 2011 Broad Institute TCGA GDAC as part of the TCGA Research Network. All rights reserved.
Made with Nozzle