Integrating genomic and ex vivo drug sensitivity profiling to predict treatment outcomes in Korean patients with non-Hodgkin lymphoma
摘要
Non-Hodgkin lymphoma (NHL) exhibits marked interpatient heterogeneity in treatment response, yet links between tumor genotype and pharmacologic phenotype in primary samples remain incompletely defined. We integrate whole-genome sequencing of patient-derived tumor tissue with ex vivo drug sensitivity profiling across standard chemotherapeutics in a cohort of 32 patients encompassing aggressive and indolent NHL, and relate these data to clinical outcomes. Ex vivo sensitivities span more than 1,000-fold IC₅₀ ranges across drugs. HLA-DQB1 is the most frequently mutated with recurrent co-mutations in immune-regulatory genes. Mutation-wise association analyses reveal gene-specific drug effects, including increased actinomycin sensitivity with EGFR mutations, resistance to dexamethasone, prednisone and mechlorethamine with ITK mutations, and multi-drug resistance patterns linked to epigenetic regulators (KMT2D, SETD2). EP300 mutations are associated with enhanced prednisone sensitivity across subtypes. Clinically, FAT4 and CD22 mutations correlate with inferior progression-free survival. FAT4-mutated samples show reduced ex vivo sensitivity to cyclophosphamide while CD22-mutated samples demonstrate enhanced sensitivity to vinblastine, suggesting opportunities for regimen optimization. These data define an integrative functional-genomic framework that uncovers mutation–drug interactions with outcome correlates and motivate prospective studies testing genotype-informed therapy selection in NHL.