Abstract PO-080: Predicting relapse in patients with triple negative breast cancer (TNBC) using a deep-learning approach

The abundance and/or precise location of tumor infiltrating lymphocytes (TILs), especially CD8+ T cells, can serve as a prognostic indicator in various types of solid tumors. However, it is often difficult to select an appropriate algorithm in order to stratify patients into well-defined risk groups. More importantly, patient stratification results often depends on the selection of tumor regions, where subjective judgement could affect the final results. On the other hand, machine-learning approaches can help to stratify patients in an objective and automatic fashion. Based on immunofluorescence (IF) images of CD8+ T lymphocytes and cancer cells, we develop a machine-learning approach which can predict the risk of relapse for patients with Triple Negative Breast Cancer (TNBC). Tumor-section images from 9 patients with poor outcome and 15 patients with good outcome were used as a training set. Tumor-section images of 29 patients in an independent cohort were used to test the predictive power of our algorithm. One of the key innovations is dissecting the section images into patches in a size of 640 µm x 640 µm for training and test, which allows one to make use of the information in the section images despite the small number of patients. In the test cohort, 6 (out of 29) patients who belong to the poor-outcome group were all correctly identified by our algorithm; for the 23 (out of 29) patients who belong to the good-outcome group, 17 were correctly predicted with some evidence that improvement is possible if other measures, such as the grade of tumors, are factored in. Our approach does not involve arbitrarily defined metrics and can be applied to other types of cancer in which the abundance/location of CD8+ T lymphocytes/other types of cells is an indicator of prognosis. Furthermore, we showed that using limited parts of the tumor section image for predictions would give rise to inaccurate results, which suggests that tumor heterogeneity should be carefully taken into account for a rigorous evaluation of the outcome. In summary, despite the limited number of patients, we demonstrated that the deep-learning approach can make good use information in the infiltration pattern of CD8+ T lymphocytes and thereby enable prognosis. Additional data collection efforts should be made to eventually enable a more rigorous analysis. Citation Format: Guangyuan Yu, Xuefei Li, Ting-Fang He, Tina Gruosso, Dongmei Zuo, Margarita Souleimanova, Valentina Munoz Ramos, Atilla Omeroglu, Sarkis Meterissian, Marie-Christine Guiot, Li Yang, Yuan Yuan, Morag Park, Peter P. Lee, Herbert Levine. Predicting relapse in patients with triple negative breast cancer (TNBC) using a deep-learning approach [abstract]. In: Proceedings of the AACR Virtual Special Conference on Tumor Heterogeneity: From Single Cells to Clinical Impact; 2020 Sep 17-18. Philadelphia (PA): AACR; Cancer Res 2020;80(21 Suppl):Abstract nr PO-080.