Development and verification of the PAM50-based Prosigna breast cancer gene signature assay

Abstract

Background: The four intrinsic subtypes of breast cancer, defined by differential expression of 50 genes (PAM50), have been shown to be predictive of risk of recurrence and benefit of hormonal therapy and chemotherapy. Here we describe the development of Prosigna™, a PAM50-based subtype classifier and risk model on the NanoString nCounter Dx Analysis System intended for decentralized testing in clinical laboratories.

Methods: 514 formalin-fixed, paraffin-embedded (FFPE) breast cancer patient samples were used to train prototypical centroids for each of the intrinsic subtypes of breast cancer on the NanoString platform. Hierarchical cluster analysis of gene expression data was used to identify the prototypical centroids defined in previous PAM50 algorithm training exercises. 304 FFPE patient samples from a well annotated clinical cohort in the absence of adjuvant systemic therapy were then used to train a subtype-based risk model (i.e. Prosigna ROR score). 232 samples from a tamoxifen-treated patient cohort were used to verify the prognostic accuracy of the algorithm prior to initiating clinical validation studies.

Results: The gene expression profiles of each of the four Prosigna subtype centroids were consistent with those previously published using the PCR-based PAM50 method. Similar to previously published classifiers, tumor samples classified as Luminal A by Prosigna had the best prognosis compared to samples classified as one of the three higher-risk tumor subtypes. The Prosigna Risk of Recurrence (ROR) score model was verified to be significantly associated with prognosis as a continuous variable and to add significant information over both commonly available IHC markers and Adjuvant! Online.

Conclusions: The results from the training and verification data sets show that the FDA-cleared and CE marked Prosigna test provides an accurate estimate of the risk of distant recurrence in hormone receptor positive breast cancer and is also capable of identifying a tumor's intrinsic subtype that is consistent with the previously published PCR-based PAM50 assay. Subsequent analytical and clinical validation studies confirm the clinical accuracy and technical precision of the Prosigna PAM50 assay in a decentralized setting.

Figures

Fig. 1

CONSORT diagram describing the breakdown for sample processing. Diagrams for (a) subtype and ROR training and (b) subtype and ROR verification

Fig. 2

Hierarchical clustering of all subtype training samples. Clustering analysis (using a Pearson’s distance metric and average linkage) was performed on the median centered normalized, Log2 transformed data. The centroid color bars below the sample dendrogram represent the significant clusters that were chosen to establish each tumor centroid. The subtype color bars represent the subtype calls using the final algorithm. Since the reduction mammoplasty normal tissue samples do not contain tumor, they were not assigned a subtype and are represented as blanks in the subtype color bars

Fig. 3

Distribution of subtypes from subtype training samples. Log2 transformed, reference sample and geomean normalized, and gene scaled nCounter data from the BC no AST, WashU, UNC cohorts assessed by the trained NanoString Prosigna algorithm

Fig. 4

DRFS Kaplan–Meier plot for subtypes for ROR training cohort. Subtype colors and numbers of patients are included in the plot along with the results from the Log Rank test

Fig. 5

Plotted pairs of the Prosigna proliferation score and the previously published [4] proliferation score. Individual points are from the algorithm training samples (n = 514). The R-squared, slope, and Y-intercept of the comparison are shown in the top left of the plot

Fig. 6

Plotted pairs of 46 gene and 50 gene ROR values. Individual points are from 514 algorithm subtype training samples. The R-squared, slope, and Y-intercept of the comparison are shown in the top left of the plot

Fig. 7

DRFS Kaplan–Meier plot for subtypes for the ROR verification cohort. Subtype colors and numbers of patients are included in the plot along with the results from the Log Rank test

Fig. 8

Boxplots showing the distribution of ROR scores for N0 BC TAM patient tumor sample. Results were grouped based on tumor classification as one of three breast cancer subtypes. The limits of the boxes represent the first and third quartile and the whiskers represent +/−1.58 IQR/sqrt(n). The horizontal dashed lines illustrate the ROR cutoffs for low/intermediate and intermediate/high risk for N0 patients. Individual data points are jittered for illustration purposes

Fig. 9

C-index of 46 and 50-gene ROR scores for distant recurrence-free survival. The limits of the boxes represent the first and third quartile and the whiskers represent +/−1.58 IQR/sqrt(n)

Fig. 10

Accuracy of the Prosigna ROR score to predict DSS and DRFS compared to other models. Different histogram colors represent whether DSS (black) or DRFS (gray) was used as the clinical endpoint to test each model