Abstract 5960: Machine learning assisted prognostication model based on genomic expression in the tumor microenvironment of stage II and III breast cancer

Background: Immune cells and stroma in the tumor microenvironment (TME) has an important role in patient prognosis and responses to therapy. Only few mathematical models exist to prognosticate patients, based on mRNA expressivity in the TME. Methods: Clinical outcomes data and mRNA-seq of 246 patients with stage 2 and 98 patients with stage 3 estrogen receptor (ER) positive (+) and HER2 negative (-) breast cancer were obtained from TCGA. 26 gene groups composed of 191 genes* enriched in cellular and non-cellular elements of TME, mutational burden (MB), and clinical data were analyzed by Kaplan-Meier (KM) analysis and multivariate nonlinear regression assisted by machine learning to achieve confined optimization with model-data minimization among multiple distribution functions. *Due to character limit, more details about these genes will be at actual presentation. Results: Prognostication was modeled with higher risk score (RS) representing worse prognosis in stage 2 and 3 ER+HER2- breast cancer. In stage 2 patients, six genes (C15orf53, PDGFB, IL10, HS3ST2, GPNMB, PADI4) and seven genes (FCRL3, IFNGR2, ICAM2, CXCR4, HLA-DMB, LGMN, ICOSLG) out of 191 genes associated with poor prognosis were identified (p Similarly, in stage 3 patients, 15 genes (CD8A, CD8B, FCRL3, GZMK, CD3E, CCL5, TP53, ICAM3, CD247, IFNG, IFNGR1, ICAM4, SHH, HLA-DOB, CXCR3) and 5 genes (LOXL2, PHEX, ACTA2, MEGF9, TNFSF4) out of 191 genes associated with good and poor prognosis were identified. Genomic expression of the 15 and 5 gene groups were labeled as G and P, respectively. RS = 9.3185 - 0.3250 × (Age at diagnosis0.0001) - 8.2979 × (P/G0.0051). Based on RS, patients were clustered into 2 groups; high and low RS groups, showing two KM curves with P = 0.05, HR = 2.878 (95% CI 1.903 - 3.471), confirming the validity of RS modeling. MB was not associated with survival in both groups. Conclusions: Machine learning-assisted mathematical modeling of RS and gene analysis identified TME-related genes and gene groups that are strongly associated with worse prognosis in stage 2 and 3 ER+HER2- breast cancer. RS could potentially prognosticate patients in the clinic with available genomic profiles. Citation Format: Yara Abdou, Jessica Jerez, Sunyoung Lee. Machine learning assisted prognostication model based on genomic expression in the tumor microenvironment of stage II and III breast cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5960.