Topological Data Analysis Applied to Radiomics (topiomics) Data in Recurrent Prostate Cancer

Non-invasive imaging techniques, particularly Positron Emission Tomography (PET), have revolutionized medical diagnostics. Radiomics, an emerging field, aims to extract quantitative data from tomographic images to reveal subtle tissue alterations. Despite challenges in standardization, radiomics holds promise in linking imaging features to clinical outcomes, crucial in metastatic cancer treatment. This study explores Topological Data Analysis (TDA) to dissect intra-tumor heterogeneity in metastatic prostate cancer. Patient representation methods, including hierarchical dendrograms and persistence diagrams, are compared. Significant differences in clinical variables and treatment responses are observed across patient clusters. This research contributes to advancing precision oncology in metastatic cancer subtyping.