Cross comparison and prognostic assessment of breast cancer multigene signatures in a large population-based contemporary clinical series

Abstract

Multigene expression signatures provide a molecular subdivision of early breast cancer associated with patient outcome. A gap remains in the validation of such signatures in clinical treatment groups of patients within population-based cohorts of unselected primary breast cancer representing contemporary disease stages and current treatments. A cohort of 3520 resectable breast cancers with RNA sequencing data included in the population-based SCAN-B initiative (ClinicalTrials.gov ID NCT02306096) were selected from a healthcare background population of 8587 patients diagnosed within the years 2010-2015. RNA profiles were classified according to 19 reported gene signatures including both gene expression subtypes (e.g. PAM50, IC10, CIT) and risk predictors (e.g. Oncotype DX, 70-gene, ROR). Classifications were analyzed in nine adjuvant clinical assessment groups: TNBC-ACT (adjuvant chemotherapy, n = 239), TNBC-untreated (n = 82), HER2+/ER- with anti-HER2+ ACT treatment (n = 110), HER2+/ER+ with anti-HER2 + ACT + endocrine treatment (n = 239), ER+/HER2-/LN- with endocrine treatment (n = 1113), ER+/HER2-/LN- with endocrine + ACT treatment (n = 243), ER+/HER2-/LN+ with endocrine treatment (n = 423), ER+/HER2-/LN+ with endocrine + ACT treatment (n = 433), and ER+/HER2-/LN- untreated (n = 200). Gene signature classification (e.g., proportion low-, high-risk) was generally well aligned with stratification based on current immunohistochemistry-based clinical practice. Most signatures did not provide any further risk stratification in TNBC and HER2+/ER- disease. Risk classifier agreement (low-, medium/intermediate-, high-risk groups) in ER+ assessment groups was on average 50-60% with occasional pair-wise comparisons having <30% agreement. Disregarding the intermediate-risk groups, the exact agreement between low- and high-risk groups was on average ~80-95%, for risk prediction signatures across all assessment groups. Outcome analyses were restricted to assessment groups of TNBC-ACT and endocrine treated ER+/HER2-/LN- and ER+/HER2-/LN+ cases. For ER+/HER2- disease, gene signatures appear to contribute additional prognostic value even at a relatively short follow-up time. Less apparent prognostic value was observed in the other groups for the tested signatures. The current study supports the usage of gene expression signatures in specific clinical treatment groups within population-based breast cancer. It also stresses the need of further development to reach higher consensus in individual patient classifications, especially for intermediate-risk patients, and the targeting of patients where current gene signatures and prognostic variables provide little support in clinical decision-making.

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1

Consort diagram of patient selection and population-based representativeness. (A) Consort diagram of patient inclusion. (B) Population representativeness for the selection process and final RNAseq cohort illustrated by proportional bar charts for important clinicopathological variables in breast cancer. For each variable, the three bars correspond to the background population (left), enrolled SCAN-B patients (center), and SCAN-B patients with RNA-seq (right) NKBC: Swedish national breast cancer quality registry.

Figure 2

Molecular subtypes and administered therapy for the study cohort. (A) Proportions of subtypes from four molecular subtype signatures in the complete RNAseq cohort. (B) Subtype proportions of the AIMS-PAM50 single sample predictor for patients diagnosed 2012, 2013, and 2014 respectively, illustrating the stability of the underlying patient demographics across inclusion years. (C) Treatment proportions according to national registry data for RNAseq patients diagnosed 2012, 2013, and 2014 respectively, illustrating the stability of the underlying patient demographics during inclusion years. ACT: adjuvant chemotherapy. Endocrine: endocrine treatment.

Figure 3

Gene signature class proportions across nine clinical assessment groups in population-based breast cancer. (A) HER2+ disease stratified into two clinical assessment groups: i) HER2+/ER− with anti-HER2 and adjuvant chemotherapy (ACT), and ii) HER2+/ER− with anti-HER2, adjuvant chemotherapy (ACT) and endocrine therapy. (B) TNBC disease stratified into two clinical assessment groups: i) TNBC with adjuvant chemotherapy, and ii) untreated TNBC. (C) ER+/HER2− disease stratified by lymph-node status (lymph-node negative: LN−, positive: LN+) and adjuvant therapy into five clinical assessment groups: (i) ER+/HER2−/LN− untreated, (ii) ER+/HER2−/LN− with endocrine therapy only, (iii) ER+/HER2−/LN+ with endocrine therapy only, (iv) ER+/HER2−/LN− with adjuvant chemotherapy (ACT) and endocrine therapy, and (v) ER+/HER2−/LN+ with adjuvant chemotherapy (ACT) and endocrine therapy.

Figure 4

Outcome analyses for clinical assessment groups and gene signatures in ER+/HER2− disease. (A) Kaplan-Meier plot of OS for the nine clinical assessment groups using all available samples. Percentages in parentheses represent proportion of entire cohort. (B) Table of events per clinical assessment group, outlining the number of cases with events, and a note on whether a group is kept for subsequent outcome analysis. (C) Forest plot of hazard ratios (HR) with 95% confidence interval for each signature class from multivariable Cox regression analysis using tumor size, patient age, lymph node status (where applicable), and tumor grade as covariates in the 1113 ER+/HER2−/LN− tumors with endocrine treatment only. Signature classes smaller than 8% of the total population are excluded from multivariable analysis. If not otherwise stated, the reference group is the low-risk group for a signature. Significant classes marked (sig). Bottom: selected Kaplan-Meier plots for the ROR-S and HDPP signatures in these cases. (D) Similar forest plot as in C but for the 423 ER+/HER2−/LN+ tumors with endocrine treatment only. Significant classes marked (sig). Bottom: selected Kaplan-Meier plots for the ROR-Tot and Oncotype DX signatures in these cases. * indicates significance level of a likelihood ratio test. ACT: adjuvant chemotherapy. Endo: endocrine treatment. mAB: anti-HER2 blockade. P-values in Kaplan-Meier plots were calculated using the log-rank test.

Figure 5

Outcome analyses in TNBC and HER2+ disease. Signature classes A) Forest plot displaying hazard ratios (HR) with 95% confidence interval for respective signature class from multivariable Cox regression analysis using tumor size, patient age, lymph node status, and tumor grade as covariates in the 239 TNBC tumors with adjuvant chemotherapy. Significant classes marked (sig). For several signatures only one class existed, thus no values were calculated. Right: selected Kaplan-Meier plots for the TNBCtype and Genius signatures in these cases. (B) Kaplan-Meier plots of PAM50 subtypes defined through the single sample predictor AIMS or a centroid-based approach in HER2+/ER− disease treated with combined HER2-blocade and adjuvant chemotherapy. (C) Kaplan-Meier plots of PAM50 subtypes defined through the single sample predictor AIMS or a centroid-based approach in HER2+/ER+ disease treated with combined HER2-blocade, endocrine therapy, and adjuvant chemotherapy. ACT: adjuvant chemotherapy. Endo: endocrine treatment. P-values in Kaplan-Meier plots were calculated using the log-rank test.

Figure 6

Signature class consensus for risk prediction signatures. (A) Distribution of exact agreements between risk prediction signature pairs summarized by clinical assessment group. Analyzed risk prediction signatures include ROR variants, Oncotype DX, Gene70, GGI, Endopredict, Genius, Gene76 and SDPP. For each compared signature pair the exact classification agreement using all available classes (low-, medium/intermediate-, and high-risk) was calculated. Next, all agreement values from all signature combinations were summarized into a box plot for each assessment group. (B) Same analysis as in A, but now only for comparisons after omitting medium/intermediate-risk classified samples. I.e., all patients with a medium/intermediate-risk prediction in a signature pair comparison were omitted before calculating the exact agreement. (C) Percentage of exact risk class agreement for risk prediction signature pairs in ER+/HER2−/LN− endocrine treated samples using all available signature classes in the individual comparisons. The heatmap corresponds to all values included in the corresponding assessment group box plot in A. (D) Similar display as in C, both now for low-risk and high-risk classified samples only in ER+/HER2−/LN− endocrine treated patients. In this analysis, all patients with a medium/intermediate-risk prediction in a signature pair comparison were omitted before calculating the exact agreement. (E) Specific agreement charts for Oncotype DX versus ROR-Tot (left), Gene70 versus Oncotype DX (center), and Gene76 versus ROR-Tot (right) in ER+/HER2−/LN− endocrine treated samples similar as described. Briefly, rectangles are drawn for each level of the test outcomes, i.e., low-, medium/intermediate-, and high-risk, based on the row and column cumulative totals. The boundaries of the rectangles along both axes represent the number of tumors categorized as that outcome group for each test. Black squares within the rectangles represent exact agreement between the levels of the two tests and are of size based on the cell frequencies and located according to the cumulative totals of the previous levels. Gray rectangles represent partial agreement, where the scores from one test are within one level of those from the other test, i.e., a low-risk prediction in one test but medium/intermediate-risk in the other test. White areas within the rectangle reflect disagreement by more than one level, i.e., low-risk in one test and high-risk in the other test. (F) Percentage of exact risk class agreement for risk prediction signature pairs in the HER2+ER− assessment group using all available signature classes in the individual comparisons. (G) Percentage of exact risk class agreement for risk prediction signature pairs in the HER2+ER+ assessment group using all available signature classes in the individual comparisons. ACT: adjuvant chemotherapy. Endo: endocrine treatment.

Figure 7

Signature classification consensus and transcriptional programs in clinical assessment groups. AIPS was used to derive activation status of gene signatures related to gene ontology terms. (A) Signature classifications and AIPS heatmap for 1563 ER+/HER2−/LN− tumors. (B) Signature classifications and AIPS heatmap for 321 TNBC tumors.