Bladder Urothelial Carinoma: 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 4 clinical features across 56 patients, 2 significant findings detected with P value < 0.05.

CNMF clustering analysis on sequencing-based mRNA expression data identified 4 subtypes that correlate to 'Time to Death'.
Consensus hierarchical clustering analysis on sequencing-based mRNA expression data identified 2 subtypes that do not correlate to any clinical features.
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 'Time to Death'.

Results

Overview of the results

Table 1. Get Full Table Overview of the association between subtypes identified by 4 different clustering approaches and 4 clinical features. Shown in the table are P values from statistical tests. Thresholded by P value < 0.05, 2 significant findings detected.


Clinical Features	Time to Death	AGE	GENDER	KARNOFSKY PERFORMANCE SCORE
Statistical Tests	logrank test	ANOVA	Fisher's exact test	ANOVA
RNAseq CNMF subtypes	0.0369	0.284	0.187	0.746
RNAseq cHierClus subtypes	0.799	0.733	0.398	1
MIRseq CNMF subtypes	0.387	0.423	0.405	0.557
MIRseq cHierClus subtypes	0.0258	0.367	0.734	0.376

Clustering Approach #1: 'RNAseq CNMF subtypes'

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

Cluster Labels	1	2	3	4
Number of samples	22	9	16	6

'RNAseq CNMF subtypes' versus 'Time to Death'

P value = 0.0369 (logrank test)

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

	nPatients	nDeath	Duration Range (Median), Month
ALL	53	19	0.4 - 118.9 (8.3)
subtype1	22	6	0.5 - 100.5 (7.4)
subtype2	9	6	5.1 - 118.9 (12.2)
subtype3	16	5	0.4 - 75.3 (8.9)
subtype4	6	2	1.8 - 10.6 (3.8)

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

'RNAseq CNMF subtypes' versus 'AGE'

P value = 0.284 (ANOVA)

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

	nPatients	Mean (Std.Dev)
ALL	53	68.4 (10.2)
subtype1	22	66.8 (10.6)
subtype2	9	67.1 (9.9)
subtype3	16	68.5 (10.0)
subtype4	6	75.8 (8.5)

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

'RNAseq CNMF subtypes' versus 'GENDER'

P value = 0.187 (Fisher's exact test)

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

nPatients	FEMALE	MALE
ALL	31	22
subtype1	16	6
subtype2	6	3
subtype3	7	9
subtype4	2	4

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

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

P value = 0.746 (ANOVA)

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

	nPatients	Mean (Std.Dev)
ALL	11	80.0 (16.7)
subtype1	4	82.5 (15.0)
subtype2	2	85.0 (7.1)
subtype3	4	77.5 (25.0)
subtype4	1	70.0 (NA)

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

Clustering Approach #2: 'RNAseq cHierClus subtypes'

Table S6. Get Full Table Description of clustering approach #2: 'RNAseq cHierClus subtypes'

Cluster Labels	1	2
Number of samples	22	31

'RNAseq cHierClus subtypes' versus 'Time to Death'

P value = 0.799 (logrank test)

Table S7. Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	53	19	0.4 - 118.9 (8.3)
subtype1	22	10	0.4 - 75.3 (8.9)
subtype2	31	9	0.5 - 118.9 (7.0)

Figure S5. Get High-res Image Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'RNAseq cHierClus subtypes' versus 'AGE'

P value = 0.733 (t-test)

Table S8. Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	53	68.4 (10.2)
subtype1	22	67.8 (10.4)
subtype2	31	68.8 (10.3)

Figure S6. Get High-res Image Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'RNAseq cHierClus subtypes' versus 'GENDER'

P value = 0.398 (Fisher's exact test)

Table S9. Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	31	22
subtype1	11	11
subtype2	20	11

Figure S7. Get High-res Image Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

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

P value = 1 (t-test)

Table S10. Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

	nPatients	Mean (Std.Dev)
ALL	11	80.0 (16.7)
subtype1	6	80.0 (20.0)
subtype2	5	80.0 (14.1)

Figure S8. Get High-res Image Clustering Approach #2: 'RNAseq cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

Clustering Approach #3: 'MIRseq CNMF subtypes'

Table S11. Get Full Table Description of clustering approach #3: 'MIRseq CNMF subtypes'

Cluster Labels	1	2	3
Number of samples	10	27	15

'MIRseq CNMF subtypes' versus 'Time to Death'

P value = 0.387 (logrank test)

Table S12. Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	52	21	0.4 - 118.9 (8.0)
subtype1	10	3	1.5 - 118.9 (7.8)
subtype2	27	14	0.4 - 49.2 (8.6)
subtype3	15	4	0.5 - 100.5 (7.2)

Figure S9. Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq CNMF subtypes' versus 'AGE'

P value = 0.423 (ANOVA)

Table S13. Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	52	68.7 (10.0)
subtype1	10	71.1 (9.0)
subtype2	27	66.9 (10.0)
subtype3	15	70.2 (10.8)

Figure S10. Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq CNMF subtypes' versus 'GENDER'

P value = 0.405 (Fisher's exact test)

Table S14. Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	33	19
subtype1	8	2
subtype2	17	10
subtype3	8	7

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

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

P value = 0.557 (ANOVA)

Table S15. Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

	nPatients	Mean (Std.Dev)
ALL	10	79.0 (17.3)
subtype2	7	77.1 (19.8)
subtype3	3	83.3 (11.5)

Figure S12. Get High-res Image Clustering Approach #3: 'MIRseq CNMF subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

Clustering Approach #4: 'MIRseq cHierClus subtypes'

Table S16. Get Full Table Description of clustering approach #4: 'MIRseq cHierClus subtypes'

Cluster Labels	1	2	3	4
Number of samples	9	12	12	19

'MIRseq cHierClus subtypes' versus 'Time to Death'

P value = 0.0258 (logrank test)

Table S17. Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

	nPatients	nDeath	Duration Range (Median), Month
ALL	52	21	0.4 - 118.9 (8.0)
subtype1	9	7	2.0 - 26.9 (9.0)
subtype2	12	3	0.4 - 49.2 (10.0)
subtype3	12	4	1.5 - 118.9 (9.5)
subtype4	19	7	0.5 - 100.5 (7.2)

Figure S13. Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #1: 'Time to Death'

'MIRseq cHierClus subtypes' versus 'AGE'

P value = 0.367 (ANOVA)

Table S18. Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

	nPatients	Mean (Std.Dev)
ALL	52	68.7 (10.0)
subtype1	9	69.6 (7.5)
subtype2	12	64.2 (11.8)
subtype3	12	70.6 (8.5)
subtype4	19	69.9 (10.6)

Figure S14. Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #2: 'AGE'

'MIRseq cHierClus subtypes' versus 'GENDER'

P value = 0.734 (Fisher's exact test)

Table S19. Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

nPatients	FEMALE	MALE
ALL	33	19
subtype1	5	4
subtype2	8	4
subtype3	9	3
subtype4	11	8

Figure S15. Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #3: 'GENDER'

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

P value = 0.376 (ANOVA)

Table S20. Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

	nPatients	Mean (Std.Dev)
ALL	10	79.0 (17.3)
subtype1	1	80.0 (NA)
subtype2	5	74.0 (23.0)
subtype4	4	85.0 (10.0)

Figure S16. Get High-res Image Clustering Approach #4: 'MIRseq cHierClus subtypes' versus Clinical Feature #4: 'KARNOFSKY.PERFORMANCE.SCORE'

Methods & Data

Input

Cluster data file = BLCA.mergedcluster.txt
Clinical data file = BLCA.clin.merged.picked.txt
Number of patients = 56
Number of clustering approaches = 4
Number of selected clinical features = 4
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

Student's t-test analysis

For continuous numerical clinical features, two-tailed Student's t test with unequal variance (Lehmann and Romano 2005) was applied to compare the clinical values between two tumor subtypes using 't.test' function in R

Download Results

This is an experimental feature. Location of data archives could not be determined.

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] Lehmann and Romano, Testing Statistical Hypotheses (3E ed.), New York: Springer. ISBN 0387988645 (2005)