Abstract / Summary
Diffuse large B-cell lymphoma (DLBCL) is curable in many patients, yet pretreatment risk arises from distinct biological and clinical domains, including disease burden, molecular pathogenesis, extranodal anatomy, immune context, host reserve and treatment context. Baseline evaluation should therefore establish diagnostic certainty, quantify anatomical and metabolic burden, define clinically relevant cytogenetic and genomic features, determine whether circulating tumour DNA (ctDNA) can be monitored longitudinally, assess the risk of central nervous system relapse, and establish whether curative-intent therapy can be delivered safely and completely. This narrative review integrates evidence on the International Prognostic Index family, cell-of-origin and genetic-subtype assignment, MYC/BCL2 expression and relevant rearrangements, bone marrow involvement, positron emission tomography/computed tomography-derived tumour burden, liquid biopsy, tumour microenvironment signatures and geriatric assessment. It also examines how contemporary frontline and relapsed/refractory trials reshape the interpretation of baseline risk. Across these domains, broad clinical enrichment has often failed to identify treatment-responsive subgroups, whereas strategies anchored to a defined mechanism, measurable burden or early molecular response appear more promising. We propose a practical, domain-based reporting framework that distinguishes relapse biology from treatment-delivery risk and links each adverse feature to an explicit management action. Further progress will require prospective validation, calibrated absolute-risk estimates, harmonised imaging and molecular workflows, and trials in which a prespecified baseline marker triggers a prespecified intervention.