Titulo. Classification of Acute Lymphoblastic Leukemia using Persistent Homology

Abstract: Acute Lymphoblastic Leukemia (ALL) is a pervasive childhood blood cancer. ALL are immature white blood cells that quickly replace the normal cells in bone marrow, due to their exponential growth it can be fatal if not treated. Classification of lymphoblast and normal cells is a challenging task, and domain experts also face some hurdles due to their morphological similarity. Automated computer analysis of ALL can provide vast support in this field and can save many lives. In this paper, we propose another way of classification approach that involves analyzing the shapes and extracting the topological features of ALL. In this work, we use persistent homology to capture the topological features. The proposed technique accurately and efficiently detects and classifies leukemia blast cells, with a recall of 98.2% and an F1-score of 94.6%. This approach has the potential to improve leukemia diagnosis and therapy.