Optical Diffraction Tomography and Interpretable Machine Learning Reveal Biophysical Signatures of Gametocyte-Stage Malaria in Red Blood Cells

Gametocyte stage-infected RBCs exhibit distinct morphology, with significantly reduced sphericity and increased eccentricity compared to those of ring stage-infected or normal RBCs. Supervised models trained on the combined feature space achieve 88.3% accuracy in multiclass classification (normal, ring, gametocyte) and 98.1% accuracy in binary classification (normal vs infected). SHAP analysis reveals deep-learning-derived features as the strongest predictors for normal and ring-stage RBCs, with handcrafted features critically influencing gametocyte-stage classification, reflecting pronounced morphological changes in gametocyte-stage RBCs.

Our results establish ODT as a powerful modality for quantitative, label-free malaria staging and demonstrate a generalizable and interpretable framework for stage-resolved phenotyping with potential applications in high-throughput analysis, culture monitoring, and future diagnostic workflows.