Abstract PS4-21: Comparison of risk assessment in primary ER+, HER2- Breast Cancer in a real-world data set: Classical pathological parameters vs. 12-gene molecular assay (EndoPredict)

Abstract

Abstract Background: Risk assessment on molecular level is becoming more common in modern pathology to determine the recurrence risk for patients diagnosed with estrogen receptor positive (ER+), HER2 negative (HER2-) breast cancer (BC). The gene expression test EndoPredict (EP) was trained and validated to predict a 10-year risk of distant recurrence to support therapy decision regarding endocrine therapy alone or in combination with chemotherapy. The EP test provides the 12-gene molecular score (12-gene MS) and the EPclin-Score (EPclin) combining the molecular score with tumor size and nodal status. In this project we investigated the correlation of 12-gene MS and EPclin scores with classical pathological markers like tumor grading and proliferation. Methods: Retrospectively, we investigated EP test results in a total of 1652 patients tested in clinical routine from 2017 to 2020 at the Institute of Pathology, Charité University Hospital, Berlin. Consecutive cases with valid EP test result and available tumor grading and proliferation (Ki67) status were included in the dataset. 12-gene MS and EPclin were classified as low or high risk based on validated cutoff values at 5 and 3.32867, respectively. Ki67 cut-offs were set according to St. Gallen guidelines at 20% for binary classification (Bustreo et al., 2016) and Federal Joint Committee (G-BA) guidelines, using 10% and 30% for three classes (low, intermediate, high). Results: In our dataset with 1652 cases, 1242 (75.2%) cases were detected as 12-gene MS high risk and 1026 (62.1%) as EPclin high risk score. Mean Ki67 expression was 17.5% (95%CI 16.9 - 17.9). As expected we found a strong association between risk scores and clinical parameter with p-values ≤ 0.001: In the Ki67 binary low group (N=1203, Ki67≤20%) 695 (57.8%) patients had a EPclin high risk score while in the Ki67 binary high expression group (>20%) still 118 (26.3%) patients had a EPclin low risk score. Similar results were found using three Ki67 classes: In the Ki67 low (N=557, <10%) group 299 (53.7%) patients had EPclin high results and in the Ki67 high (N=101, >30%) group, 28 (27.7%) patients were classified as EPclin low (p<0.001). Regarding tumor grading we observed a correlation between poorly differentiated breast cancer (G3) and a higher EPclin risk score. Nevertheless, in Grade 1 tumors (N=140) 57.9% of patients had a EPclin high risk score (p=0.001). In comparison, in 25% (N=46) of G3 cases EPclin risk score was low. Similar results were seen using 12-gene-MS, not shown here. Conclusions: In this study we could show that EP risk scores are distributed differently among Ki67 expression groups, especially in Ki67 low (<10%) and high (>30%) tumors with a substantial proportion of patients with EPclin high risk results in Ki67 low tumors and vice versa. Our research group is currently collecting a long time follow up of all patients. Citation Format: Paul Jank, Carsten Denkert, Judith L Lindner, Annika Lehmann, Berit M Pfitzner, Jens-Uwe Blohmer, David Horst, Ralf Kronenwett, Wolfgang D Schmitt. Comparison of risk assessment in primary ER+, HER2- Breast Cancer in a real-world data set: Classical pathological parameters vs. 12-gene molecular assay (EndoPredict) [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS4-21.