Computer vision quantitation of erythrocyte shape abnormalities provides diagnostic, prognostic, and mechanistic insight.

Examination of red blood cell (RBC) morphology in peripheral blood smears can help diagnose hematologic disease, even in resource-limited settings, but this analysis remains subjective and semi-quantitative with low throughput. Prior attempts to develop automated tools have been hampered by poor reproducibility and limited clinical validation. Here, we present a novel, open-source machine-learning approach (denoted the 'RBC-diff') to quantify abnormal RBCs in peripheral smear images and generate an RBC morphology differential. RBC-diff cell counts showed high accuracy for single-cell classification (mean AUC: 0.93) and quantitation across smears (mean R2: 0.76 compared to experts, inter-experts R2: 0.75). RBC-diff counts were concordant with clinical morphology grading for 300,000+ images and recovered expected pathophysiologic signals in diverse clinical cohorts. Criteria using RBC-diff counts distinguished thrombotic thrombocytopenic purpura and hemolytic uremic syndrome from other thrombotic microangiopathies, providing greater specificity than clinical morphology grading (72% vs. 41%, p < 0.001), while maintaining high sensitivity (94-100%). Elevated RBC-diff schistocyte counts were associated with increased 6-month all-cause mortality in a cohort of 58,950 inpatients (9.5% mortality for schist. > 1%, vs. 4.7% for schist. < 0.5%, p < 0.001) after controlling for comorbidities, demographics, clinical morphology grading, and blood count indices. The RBC-diff also enabled estimation of single-cell volume-morphology distributions, providing insight into morphology influences on routine blood count measures. Our codebase and expert-annotated images are included here to spur further advancements. These results illustrate that computer vision can enable rapid and accurate RBC morphology quantitation, which may provide value in both clinical and research contexts.