Classification of paclitaxel-resistant ovarian cancer cells using holographic flow cytometry through interpretable machine learning

Abstract

Progress has been made in chemotherapy drugs, but drug resistance remains a major challenge in cancer treatment. In clinical practice, although there are existing technologies to assess chemotherapy resistance, there is still a lack of rapid and convenient method to monitor different degrees of drug resistance in cancer cells. Furthermore, the cell morphological changes during the progression of drug resistance remain to be fully understood. In this study, we employed a flow imaging platform based on digital holography (DH) and machine learning to classify cancer cells with different degrees of paclitaxel sensitivity. We extracted 112 morphological features from four datasets of single-cell quantitative phase images of epithelial ovarian cancer (EOC) cells and performed classification training using five supervised machine learning algorithms. Both support vector machine (SVM) and neural network (NN) achieved over 90% classification accuracy. To understand the morphological features associated with drug resistance properties of cancer cells, we applied the SHapley Additive exPlanations (SHAP), a model interpretation framework, to quantify and rank the feature contributions for both classification models. The results reported here about this label-free and high-accuracy assessment of different degrees of drug resistance in cancer cells opens the route to exploitation for monitoring chemotherapy in clinic practices.