Abstract:Background: Machine learning (ML) in translational medicine has led to prediction of clinical outcomes and identification of new biomarkers. We employ ML in prediction of pathologic complete response (pCR) in high-risk breast cancer patients in the neoadjuvant I-SPY2 TRIAL where not all novel agents have strong predictive biomarkers. Leveraging a ML approach using progressively expanded candidate genes, we explore the limitations of using only known mechanisms of action in predicting pCR, and the extent to which biology outside known drug action improves response prediction in the first 10 arms of the trial. Methods: ML random forest models were developed in I-SPY2 patients (n=982) with pre-treatment gene expression and pCR data across 10 treatment arms (PMID: 35623341), including inhibitors of HER2: neratinib (N), pertuzumab (P), TDM1/P; AKT (MK-2206); IGF1R (ganitumab); HSP90 (ganetespib); PARP/DNA repair (veliparib/carboplatin, VC); ANG1/2 (trebananib, T); immune checkpoints (PD1-inh); and Control (Ctr). Each HR/HER2 receptor/treatment arm subset (m=27) was evaluated independently. We employed a three-pronged feature-selection approach using (1) genes restricted to known mechanism of action of individual I-SPY2 agents (k=10 to 88 genes); (2) genes expanded to include targeted pathways for all 10 agents/combinations (k=282); and (3) an unbiased whole genome approach (k=17,990). Samples were partitioned with 75% used for training and cross-validation, and 25% held out as test sets. Predictive ML models were defined as those with performance ≥ 0.90 based on different performance metrics (e.g., AUC, sensitivity, specificity). Results: For each of the 27 subtype-treatment subsets, at least one high performing model was identified. In 6 subtype-treatment subsets, mechanism of action genes were sufficient to predict pCR: AKT/PI3K/HER genes in HR+HER2- N and HR-HER2+ P; DNA repair genes in HR+HER2- VC; angiogenesis-associated genes in HR+HER2+ T; and immune-associated genes in both HR+HER2- and HR-HER2- PD1-inh subsets. Expanded targeted pathway models were required to identify predictive models in 8 additional subtype-treatment pairs from the N, T-DM1/P, MK-2206, VC, T, and HER2+ Ctr arms, with significant contribution of DNA repair, immune, and HSP90 genes for multiple arms. A genome-wide approach was required for the remaining 13 subtype-treatment pairs with no previous models from the N, P, MK-2206, ganitumab, ganetespib, T, and HER2- Ctr arms. Even for subtype-treatment pairs where mechanism of action gene sets was sufficient for reasonable models, expanded gene sets resulted in improved performance. For instance, metabolism genes improved model performance for HR-HER2+ in N and Ctr, and for HR+HER2- in the PD1-inh arm; and mitochondrial and protein folding dysfunction genes improved response prediction in HR-HER2- in the ganetespib arm. Conclusion: Our study identifies mechanism of action biomarkers associated with response to each drug and elucidates possible off-target effects contributing to observed drug sensitivity and resistance. Citation Format: Rosalyn W. Sayaman, Denise M. Wolf, Christina Yau, Julia Wulfkhule, Emanuel F. Petricoin, Lamorna Brown-Swigart, Tam Binh Bui, Gillian L. Hirst, Diane Heditsian, W. Fraser Symmans, Angela DeMichele, Mark LaBarge, Laura J. Esserman, Laura van 't Veer. Machine learning elucidates biology of response within and outside the mechanisms of action of therapeutic agents in the I-SPY2 breast cancer TRIAL [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Breast Cancer Research; 2023 Oct 19-22; San Diego, California. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_1) nr A066.

Abstract 6446: ResponderIDTM KRAS: Biology-driven machine learning to personalize KRAS inhibitor therapeutics

Abstract 3321: KRASmulti inhibitor BI 3706674, an orally bioavailable, direct inhibitor of diverse oncogenic KRAS variants drives tumor regression in KRASG12V-driven preclinical models

Abstract 2146: A quantitative in vivo pharmacogenomics platform uncovers biomarkers of therapy response

Abstract 3317: KRASmulti inhibitor BI 3706674 shows efficacy in KRAS-driven preclinical models of cancer that supports clinical testing in patients with tumors harbouring KRASG12V mutations and KRAS wild-type amplifications

Abstract 1214: Investigating therapeutic strategies to promote immune rejection of KRASG12C inhibitor-resistant subpopulations in lung cancer

Abstract 4684: A comprehensive platform for identification of KRAS-specific synthetic lethal targets using patient-derived cells

Abstract B080: Non-genetic determinants driving sub-clonal resistance to KRAS G12C combination therapies in KRAS mutant non-small cell lung cancer

Abstract 3867: Chromatin modification driving sub-clonal resistance to KRAS G12C combination therapies in KRAS mutant non-small cell lung cancer

Abstract 6909: KRASG12C inhibitor resistance in patient-derived non-small cell lung cancer models

Abstract 1138: A single-cell spatially resolved atlas of immune-mediated control of lung adenocarcinoma using 1,000-plex single-cell spatial molecular imaging in a transgenic mouse model

Abstract LB_C17: Spatial tumor immune microenvironment analysis of KRAS-mutated Non-Small Cell Lung Cancer and immunotherapy outcome

Abstract 2494: RAS-Bio a unique pan-cancer biobank for RAS-driven tumors

Abstract 1940: Identifying resistant mechanisms to direct KRAS inhibitors in NSCLC

Abstract 3255: Probing and overcoming KRASG12C inhibitor resistance by combination with a pan-KRAS SOS1 inhibitor

Abstract 2667: Trial in progress: Phase 1 study of BI 1823911, an irreversible KRASG12C inhibitor targeting KRAS in its GDP-loaded state, as monotherapy and in combination with the pan-KRAS SOS1 inhibitor BI 1701963 in solid tumors expressing KRASG12C mutation

Abstract 1943: Discovery and characterization of potential drivers of resistance to the KRAS G12D inhibitor, MRTX1133, in cancer cell line models

An Integrative Pharmacogenomics Analysis Identifies Therapeutic Targets in KRAS-Mutant Lung Cancer

Identifying Genes Associated with Resistance to KRAS G12C Inhibitors Via Machine Learning Methods.

Abstract 6841: Mutation specific mechanisms of resistance to oncogenic KRAS inhibition

Abstract 4265: Humanized 3D tumor models that are mutationally aligned with AACR GENIE patients predict IMM-1-104 activity in RAS-addicted tumors

Abstract A066: Machine learning elucidates biology of response within and outside the mechanisms of action of therapeutic agents in the I-SPY2 breast cancer TRIAL