Abstract PD5-02: Effective and globally reproducible digital pathologist training program on PD-L1 immunohistochemistry scoring on immune cells as a predictive biomarker for cancer immunotherapy in triple negative breast cancer

Abstract

Abstract BACKGROUND: The VENTANA PD-L1 (SP142) Assay (SP142 assay) is the FDA-approved companion diagnostic for the combination of TECENTRIQ (atezolizumab) and nab-paclitaxel for the treatment of patients with unresectable, locally advanced, or metastatic triple-negative breast cancer (TNBC) whose tumors express PD-L1 - defined as PD-L1 stained tumor-infiltrating immune cells (IC) of any intensity covering ≥ 1 percent of the tumor area. Accurate and reproducible assessment of IC PD-L1 staining with the SP142 assay by pathologists is important to ensure the right treatment decisions for patients are made. Analytical studies have demonstrated variable agreement rates for PD-L1 IC scoring with the SP142 assay, ranging from low, moderate, strong and near-perfect agreement depending on tumor type, cut-off and prior training. The Roche International Pathologist Training Program has trained over 1000 pathologists globally for the SP142 assay in the indications of non-small cell lung cancer, urothelial carcinoma and TNBC. Here we present results from the TNBC training program. METHODS: The glass-slide based training program for pathologists participating in clinical trials leading to the approval of the SP142 assay was adapted to enable live training to be conducted using a novel digital training platform (Pathomation) which was customized for Roche with a rapid, user-friendly interface, facilitating monitoring of scores and reporting. Feedback and endorsement from a Pathologist Training Expert Committee led to a global Train-the Trainer program which enabled regional training sessions to be conducted using the digital platform. Training content was developed around the FDA-approved 1% IC cut-off and was conducted over a 1-day program. Didactic instruction was followed by case reviews and a final proficiency test consisting of 28 cases, with a passing score of ≥85%. RESULTS: Between 17th November, 2018 and 19th June 2019, 432 pathologists from 58 countries participated in the TNBC training program for the SP142 assay. This includes 1 PTEC, 10 Train-the Trainer Sessions and 24 Regional Training sessions. The passing rate for trainees was 99.1%. Overall percent agreement (OPA) was 98.2%, with positive percent agreement (PPA) 99.4% and negative percent agreement (NPA) 96.6%. CONCLUSION: This is the first formal PD-L1 training program for practicing pathologists to be developed for the assessment of PD-L1 on ICs in TNBC. The SP142 assay training utilizing a novel digital platform demonstrates robust, reproducible and excellent pathologist concordance scores in this real-world program. Citation Format: Eslie Dennis, Mark Kockx, Greg Harlow, Zhuangyu Cai, Ken Bloom, Ehab ElGabry. Effective and globally reproducible digital pathologist training program on PD-L1 immunohistochemistry scoring on immune cells as a predictive biomarker for cancer immunotherapy in triple negative breast cancer [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr PD5-02.