Predicting Cancer Chemosensitivity Based on Intensity/Distribution Profiles of Cells Loaded with a Fluorescent Sphingolipid Analogue

Cancer is a group of heterogeneous and complex diseases, with limited therapeutic options due to the recurrent emergence of drug resistance. Sphingolipids are bioactive molecules that participate in signaling of cell death or proliferation. Because there is no laboratory test to rapidly predict a tumor chemosensitivity, we propose the use of a sphingomyelin fluorescent analogue as a chemotherapy response sensor. Through kinetic live cell imaging experiments, we extracted 1611 fluorescence features with single cell time resolved resolution. After comparing the variations in this reporter, induced by different chemotherapies, it was possible to reduce the system complexity to elaborate a decision tree algorithm based in only 3 fluorescence features capable of predicting chemosensitivity with a 73% of accuracy. This approach serves as a proof of principle for the possible future implementation of a chemosensitivity test that could be used with patient primary tumors, and thus contribute to personalized therapy against cancer.