Cross-Species Morphology Learning Enables Nucleic Acid-Independent Detection of Live Mutant Blood Cells

In hematology/oncology clinics, molecular diagnostics based on nucleic acid sequencing or hybridization are routinely employed to detect malignancy-associated genetic mutations and are instrumental in therapeutic stratification and prognostication. However, their limited cost-efficiency constrains their use in pre-malignant screening, specifically, the detection of rare circulating mutant blood cells in asymptomatic individuals. In both neonates and adults, the presence of malignancy-associated mutations in peripheral blood correlates with an elevated risk of future neoplastic transformation, with certain mutations, such as KMT2A rearrangements, exhibiting near-complete penetrance. If feasible, pre-malignant screening could enable early intervention and even disease prevention. Here, we introduce a high-throughput, single-cell computer vision platform capable of identifying mutant peripheral blood cells by recognizing mutation-specific morphological features. The morphology recognition module was developed through cross-species learning from murine to human datasets, enabling a generalizable and cost-effective approach for detecting mutations in live blood cells. The platform holds promise for translation into pre-malignant screening applications in asymptomatic neonates and adults. Furthermore, it provides a novel single-cell morphological data modality that complements existing molecular layers, including genomic, epigenomic, transcriptomic, and proteomic information.