Abstract
Immunotherapy has revolutionized hematologic cancer treatment, yet responses remain unpredictable due to primary resistance, relapse, and life-threatening toxicities. Conventional biomarkers fail to capture the complexity of tumor-immune interactions, necessitating integrative approaches. This review explores how multi-omics technologies, genomics, transcriptomics, proteomics, metabolomics, spatial omics, and microbiome profiling, decode the molecular drivers of immunotherapy efficacy and adverse events in hematologic malignancies. We highlight key advances: genomics reveals neoantigen landscapes and HLA diversity shaping checkpoint inhibitor responses; transcriptomics identifies T-cell exhaustion signatures predictive of CAR-T failure; metabolomics uncovers lactate-driven immunosuppression in AML; and spatial omics maps immune architectures linked to Hodgkin lymphoma outcomes. Supervised machine learning algorithms (e.g., random forest, support vector machines) integrate these layers to build predictive models for cytokine release syndrome (CRS) and resistance, while longitudinal ctDNA monitoring enables dynamic therapy adaptation. Emerging frontiers like CRISPR-based epitope editing, digital twins for in silico clinical trials, and non-coding RNA biomarkers further refine precision strategies. Despite challenges in data integration, tumor plasticity, and ethical frameworks, multi-omics is accelerating biomarker-driven trial designs (e.g., basket trials with omics stratification) and patient-centric tools (wearable sensors for real-time metabolite tracking). This review distinguishes itself by synthesizing these rapid technological advances not only to predict outcomes but also to chart a forward-looking roadmap for their clinical translation, offering a unique perspective on overcoming the current barriers to precision immuno-oncology. Together, these advances promise to transform immunotherapy from empirical to precision medicine, optimizing outcomes for leukemia, lymphoma, and myeloma patients.