Linking the Companion Diagnostic Function of Cancer Genome Profiling to Therapy: Investigation of TMB-high Predictors.

e13033 Background: Based on the companion diagnostic function of the Comprehensive Cancer Genome Profiling (CGP) test in Japan, pembrolizumab is the standard treatment for patients with Tumor Mutational Burden (TMB)-high (H) solid tumors. In this report, we attempted to extract predictive factors of TMB-H from CGP test results. Methods: We carried out a retrospective cohort study of 42 patients with malignant breast tumors (excluding circulating tumor DNA in the blood) who underwent CGP testing between June 1, 2019, and November 30, 2023, at our hospital and affiliated institutions. We analyzed the association between clinicopathological factors, the number of analyzed cancer-related gene mutations (SNV, InDel), and TMB-H using univariate and multivariate statistical analysis. Results: Of the 42 cases (27 positive and 15 negative for estrogen receptor (ER), respectively), seven (16.7%) were TMB-H. Samples significantly associated with TMB-H according to univariate analysis were submitted as pathology specimens (non-primary (16.7%) vs. primary (0%), p = 0.011), ER (positive (16.7%) vs. negative (0%), p = 0.044), letrozole (LET) (yes (11.9%) vs. no (4.8%), p = 0.008), and previous treatment (after 6th line (14.3%) vs. before (2.4%), p = 0.041). Multivariate analysis showed that LET had a significant effect on TMB-H in Aromatase inhibitor agents (LET + Anastrozole + Exemestane) (LET: OR, 19.9; 95% CI, 2.2–183.5; p = 0.008) and the number of cancer-related gene mutations (PIK3CA+TP53) were significantly affected (PIK3CA: OR, 3.8; 95% CI, 1.1–13.2; p = 0.032. TP53: OR, 0.08; 95% CI, 0.01–0.66; p = 0.015). Single regression analysis showed a significant association between the number of PIK3CA mutations and LET. (β, 0.31; SE, 0.11; p = 0.008). Conclusions: ER positivity, history of LET use, prior therapy after the 6th line, and submission of pathology specimens from non-primary tumors to CGP testing are predictive factors for TMB-H. In multivariate analysis, the history of LET use and the number of PIK3CA mutations were significantly associated with TMB-H. Research shows that PIK3CA mutations are most frequent in patients who have undergone preoperative endocrine therapy with LET (Breast Cancer Res Treat. 2020 Nov;184(1):123–133.), suggesting that an increased number of PIK3CA mutations due to LET contributed to TMB-H. Although immune checkpoint inhibitors (ICIs) are not indicated for ER-positive breast cancer, TMB-H may be a predictor of ICI efficacy in such cases. The limitations of our research were the small sample size, and the study population only included Japanese. For the next steps needed to continue in this research, clinical trials should take place to test whether TMB-H is a predictor of ICIs in a wider population of ER-positive breast cancer.