EndoPredict and Mammaprint Risk Classification

EndoPredict and Mammaprint Risk Classification Comparison of risk classification between EndoPredict and Mammaprint in ER-positive/HER2-negative primary invasive breast cancer. Alberto Pelez-Garcia, Laura Yebenes, Alberto Berjon, Antonia Angulo, Pilar Zamora, Jose Ignacio Snchez-Mendez, Enrique Espinosa, Andres Redondo, Victoria Heredia, Marta Mendiola, Jaime Feliu, David Hardisson Corresponding Author: David Hardisson, MD, PhD; Department of Pathology; Hospital Universitario La Paz, IdiPAZ; Paseo de la Castellana, 261; 28046 Madrid, Spain. ABSTRACT Purpose To compare the prognostic performance of the EndoPredict assay with the MammaPrint scores obtained for the same cancer samples on 40 estrogen-receptor positive/HER2-negative breast carcinomas. Methods Formalin-fixed, paraffin-embedded invasive breast carcinoma tissues that were previously analyzed with MammaPrint as part of routine care of the patients, andwere classified as high-risk (20 patients) and low-risk (20 patients), were selected to be analyzed by the EndoPredict assay, a second generation gene expression test that combines expression of 8 genes (EP score) with two clinicopathological factors (tumor size and nodal status, EPclin score). Results The EP score classified 15 patients as low-risk and 25 patients as high-risk. EPclin re-classified 5 of the 25 EP high-risk patients into low-risk, resulting in a total of 20 high-risk and 20 low-risk tumors. EP score and MammaPrint score were significantly correlated (p=0.008). Twelve of 20 samples classified as low-risk by MammaPrint were also low-risk by EP score (60%). 17 of 20 MammaPrint high-risk tumors were also high-risk by EP score. The overall concordance between EP score and MammaPrint was 72.5%.   EPclin score also correlated with MammaPrint results (p=0.004). Discrepancies between both tests occurred in 10 cases: 5 MammaPrint low-risk patients were classified as EPclin high-risk and 5 high-risk MammaPrint were classified as low-risk by EPClin (overall concordance 75%). Conclusions This study demonstrates a moderate concordance between MammaPrint and EndoPredict. Differences in results could be explained by the inclusion of different gene sets in each platform, and the inclusion of clinical parameters, such as tumor size and nodal status, in the EndoPredict test. Keywords:  Breast cancer prognosis; gene expression signatures; EndoPredict; MammaPrint INTRODUCTION Breast cancer is the most common cancer and the second most frequent cause of cancer death among women in developed countries. Approximately 231,840 new cases of invasive breast cancer and 40,290 deaths are expected among US women in 2015 [1]. Currently, the decision on adjuvant treatment for breast cancer patients is based on risk assessment using clinicopathological criteria, such as patient age, menopausal status, axillary lymph node status, tumor size, tumor grade, estrogen receptor (ER)/progesterone receptor (PgR) expression, HER2 status, and Ki67 score. However, decision making in adjuvant treatment of women with ER-positive/HER2-negative early breast cancer remains a difficult task. Routinely, all of these patients will receive adjuvant hormonal treatment. However, a substantial proportion of these patients are also treated with adjuvant chemotherapy, although a significant part of these will not achieve a further reduction of their risk of recurrence [2].Therefore, a major challenge for clinical oncologists is to identify those patients who will not benefit for adjuvant chemotherapy, and those who are more likely to develop recurrence, so that the most appropriate therapeutic regime can be administered [2, 3]. In recent years, molecular characterization of breast cancer has contributed to broaden our understanding of breast cancer as a heterogeneous disease, and led to the development of a variety of prognostic and predictive gene signatures [4]. Morever, these assays may also be useful in recurrence prediction and treatment decision making [5]. One of the most widely used tests is the MammaPrint (MP) assay (Agendia Laboratories, Amsterdam, The Netherlands), which is a prognostic score performed by a central laboratory that was cleared by the FDA in 2007. MP was initially limited by its requirement for fresh tissue, but it is now validated for formalin-fixed, paraffin-embedded (FFPE) tissue [6]. MP measures the expression of 70 genes using a microarray platform, and reports a binary risk classification (low-risk or high-risk) for recurrence without adjuvant chemotherapy. This information is intended to spare patients at low-risk of recurrence from receiving adjuvant chemotherapy, with its attendant morbidity. It is not intended to predict the response, per se, to chemotherapy; rather, it helps to select patients who are likely to benefit from chemotherapy from a prognostic point of view [7]. More recently developed, the EndoPredict assay (EP) (Sividon Diagnostics GmbH, Cologne, Germany), is a diagnostic test based on gene expression data in combination with clinicopathological risk parameters to assess the risk of distant metastasis in patients with ER-positive/HER2-negative primary breast cancer if treated with adjuvant endocrine therapy alone [8]. This test measures the expression of eight cancer-related genes of interest (BIRC5, UBE2C, DHCR7, RBBP8, IL6ST, AZGP1, MGP and STC2) and three reference genes (CALM2, OAZ1 and RPL37A) to calculate a molecular risk score (EP score). The molecular risk score is then combined with the nodal status and tumor size resulting in a molecular-clinicopathological hybrid score (EPclin score) with improved prognostic power. Using a p redefined cutoff value, patients are stratified into low- or high-risk of distant recurrence. The test can be carried out on routinely processed and archived FFPE tissue, and is designed to be performed decentrally [9, 10]. EP was validated in three randomized endocrine phase III trials with patients with ER-positive/HER2-negative node negative and node positive breast carcinomas [5, 8]. The EP provided additional prognostic information to conventional risk factors such as grading, quantitative ER, or Ki67 and outperformed risk classification by clinical guidelines. Moreover, it could be demonstrated that EP is prognostic for early and late metastasis [5, 11].The EPclin score was also directly compared to purely clinical risk classifications (like St. Gallen, German S3, and NCCN) and found to be superior to these classifiers [11]. The objective of this study was to compare the concordance of EndoPredict results in 40 ER-positive/HER2-negative breast carcinomas which were previously tested with MammaPrint and categorized as low-risk (20 patients) or high-risk (20 patients). We further evaluate TargetPrint (Agendia Laboratories), a commercially available mRNA-based gene expression test that quantitatively determines gene expression levels of ER, PgR, and HER2. MATERIALS AND METHODS Patients and tumor samples This study involved 40 patients with ER-positive/HER2-negative early-stage breast carcinoma. All patients underwent surgery between March 2012 and December 2015 at the University Hospital La Paz, Madrid, Spain. Data on age and tumor characteristics were collected for all patients. The surgical specimens were fixed in 10% buffered formalin and embedded in paraffin. Four- µm thick sections were stained with hematoxylin-eosin for histological diagnosis. Sections (10 µm) with at least 40% of tumor cellularity were selected for the study. Immunohistochemistry for ER/PR/HER2 and Ki67 and Fluorescence in situ Hybridization (FISH) for HER2 All cases were reviewed by two breast pathologist (DH and LY) to assess tumor grade (using the Nottingham histological three-tier grading system), tumor size, nodal status, ER, PgR, HER-2, and Ki67 expression. The expression of ERÃŽ ± (clone EP1; Dako, Glostrup, Denmark, prediluted), PgR (clone PgR1294; Dako, prediluted), and Ki67 (clone MIB1; Dako, prediluted) were determined by immunohistochemistry (IHC) during routine pathologic examination. ER and PgR status was determined based on the percentage of positive nuclei in the invasive neoplastic compartment of the tissue. Tumors were classified as ER- or PgR-positive when ≠¥1% invasive tumor cells showed definite nuclear staining, regardless of staining intensity. Ki67 was evaluated as the percentage of positively stained nuclear cancer cells (regardless of staining intensity). HER2 expression was evaluated with the HercepTest kit (Dako) and scored as 0, 1+, 2+, or 3+, according to the FDA scoring system. Tumors scored as 2+ wer e re-tested with FISH using the HER2 IQFISH PharmDx kit (Dako). Mammaprint Test The MammaPrint test was performed on representative paraffin blocks at the centralized Agendia Laboratories (Amsterdam, The Netherlands) blinded for clinical and histological data as part of routine care of the patients included in this study. Additionally to MammaPrint, TargetPrint assay, an additional test that is an alternative measurement of ER, PgR, and HER2 to IHC/FISH assessment, was also performed. EndoPredict Test The same tumor tissue block used for MammaPrint testing in each case was used for EP test. RNA extraction was performed as previously described [9]. Total RNA was extracted from one 10- µm whole formalin-fixed, paraffin-embedded tissue section using a silica-coated magnetic bead-based method with Tissue Preparation Reagents (Sividon Diagnostics). Expression of eight genes-of-interest (AZGP1, BIRC5, DHCR7, IL6ST, MGP, RBBP8, STC2, UBE2C), three normalization genes (CALM2, OAZ1, RPL37A) as well as the amount of residual genomic DNA (HBB) were assessed by the EP assay (Sividon Diagnostics). Gene expression was assessed by one-step RT-qPCR using the SuperScript III PLATINUM One-Step Quantitative RT-PCR System with ROX (Invitrogen, Karlsruhe, Germany) according to manufacturers instructions in a VERSANT ® kPCR Molecular System (Siemens Healthcare Diagnostics, Erlangen, Germany). EP and EPclin scores were determined as published earlier [8, 9] using the EndoPredict Report Generator sof tware which is available online (www1.endopredict.com). The predefined cut-offs for diagnostic decisions were applied to stratify patients into low- or high-risk groups: EP low-risk (