Abstract 99: Evolution of OncoKB, a precision oncology knowledgebase

Abstract Genomic sequencing of tumors is a routine part of cancer patient care. To address the need for a comprehensive resource that annotates the oncogenic effects and clinical actionability of somatic alterations in cancer, we developed OncoKB, a precision oncology knowledgebase. Since its introduction in 2016, OncoKB has grown to include annotations for >500 alterations in 682 genes. This includes 42 Level 1 genes (included in the FDA drug label), 12 Level 2 genes (included in the NCCN guidelines), and 29 Level 3A genes (predictive of drug response in well-powered clinical studies). We evaluated changes in the clinical actionability landscape and evolution of the OncoKB annotation rules and processes by comparing the AACR Project GENIE cohort (v8.1) annotated with the OncoKB version from February 2018 to that from November 2020. Even within a short window of time, this comparison reveals a significant shift of the proportion of samples that harbor a standard care alteration (Level 1 and 2), increasing from ~9% in 2018 to ~24% in 2020, and a Level 3A alteration, decreasing from ~11% to ~5%. This shift is partially attributable to the FDA approval of a PI3K inhibitor in PIK3CA-mutant ER+/HER2- breast cancer, approval of RAF inhibitors in BRAF V600E mutant anaplastic thyroid cancer and colorectal cancer, approval of NTRK-inhibitors in NTRK fusion-positive solid tumors, FGFR-inhibitor approval in FGFR2 fusion-positive bladder cancer and cholangiocarcinoma, RET-inhibitor approval in RET fusion-positive thyroid cancer and non-small cell lung cancer and expansion of indications for PARP inhibitors to include prostate cancer. The tumor agnostic approval of the checkpoint blockade inhibitor, pembrolizumab, in TMB-H solid tumors additionally gave rise to a ~4% increase in Level 1 samples. Analysis of the AACR Project GENIE cohort also revealed significant changes to the OncoKB process, including those reflected in the updated OncoKB Levels of Evidence v2.0. The refined levels system deprioritized the significance of standard care biomarkers when present in indications outside of the FDA-approved/NCCN listed indication. This change was based on clinical data demonstrating that patients with investigational predictive biomarkers for a specific tumor type based on compelling clinical evidence from phase 3 trials (currently Level 3A) are more likely to experience clinical benefit compared to patients with predictive biomarkers that are considered standard care in a different tumor type (previously Level 2B, currently Level 3B), and is consistent with guidelines published by ASCO/AMP/CAP and ESMO. Knowledgebases such as OncoKB have emerged as key informational resources that the clinical oncology and scientific communities have used to rapidly connect sequencing results to clinical actionability. Their utility depends on their ability to stay abreast of clinical data and seamlessly adapt their rules and processes to the evolving field of precision medicine. Citation Format: Sarah P. Suehnholz, Ritika Kundra, Hongxin Zhang, Shaleigh Smith, Moriah Nissan, Yifu Yao, Lindsay LaFave, Kinisha Gala, Linde Miles, Maria E. Arcila, Marc Ladanyi, Michael F. Berger, Ahmet Zehir, Aijaz Syed, Julia Rudolph, Paul Sabbatini, Ross Levine, Ahmet Dogan, Jianjiong Gao, David Solit, Nikolaus Schultz, Debyani Chakravarty. Evolution of OncoKB, a precision oncology knowledgebase [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 99.