Click the image below to view the full size version of this cover.
Created by: Salvador Chulián
Issue 271: Does blood have a shape?”. We consider the Shape of Cancer Relapse: using bone marrow samples from acute lymphoblastic leukemia (ALL) patients, we use methods from topological data analysis (TDA), which quantifies shapes in data, to predict known outcomes at the time of diagnosis. Given that ALL therapy fails in approximately 20% of these patients, we were able to distinguish isolated data islands and empty spaces in the immunophenotypic space of CD10, CD20, CD38 and CD45. In the light of TDA analyses and machine learning, we reduced the four-dimensional shape characteristics, and were able to perform high-accuracy predictions of relapse in children and adolescents with ALL. This work also covers the issue August 2023 in PLoS Computational Biology.