Geometric Patterns of Cells After breast cancers, gynecological malignancies are the most common cancer in women, with 114,000 cases occurring annually in the United States. Considering the costs and side-effects of chemotherapy, there is an unmet clinical need in gynecological cancer treatment. However, artificial intelligence and deep-learning models with high accuracy characterize the geographical patterns of spatial arrangements between tumor-infiltrating lymphocytes (TILs) and cancer cells based of debulking surgery tumor specimen. This figure reveals that computationally-derived features from the architecture of TILs and tumor cells are indeed associated with survival time in gynecological cancers and suggest that dispersion of TILs throughout the tumor is associated with better treatment response. These findings could aid in identifying therapy-refractory patients and further enable personalized treatment decision making.
