Asian Journal of Research in Medical and Pharmaceutical Sciences

Chimeric antigen receptor (CAR) T-cell therapy has altered the treatment of several B-cell malignancies and multiple myeloma, yet a substantial proportion of treated patients relapse, and activity in solid tumours remains limited. The field has responded with an expanding repertoire of engineering strategies, including multi-antigen targeting, affinity and signalling redesign, logic-gated receptors, cytokine armouring, transcriptional and metabolic reprogramming, precise gene editing, allogeneic and in vivo generated products, and alternative effector cells. This review examines that repertoire critically, taking resistance rather than platform novelty as its organising principle. Literature was identified through structured searching of open scholarly databases and indexes, supplemented by targeted retrieval of foundational reports and by checking for corrections and retraction notices, with each included source verified for bibliographic identity and for its relationship to the claim it supports. Evidence was appraised against a four-level scheme distinguishing clinical demonstration of durability, clinical demonstration of activity, mechanistic demonstration in human material or immunocompetent models, and mechanistic plausibility alone. Three findings emerge. Resistance is not a single phenomenon but a set of mechanistically distinct failure modes, several of which were conflated in earlier accounts, and antigen density behaves as a continuous determinant rather than a binary presence or absence. Most next-generation strategies currently rest on level three or level four evidence, so the gap between mechanistic demonstration and demonstrated durability is wider than the volume of publication suggests. Strategies that intensify effector function tend to approach a shared toxicity and manufacturing ceiling, which constrains how far intensification can be pursued independently of delivery and access considerations. Unresolved questions concern the predictive value of pre-infusion product attributes, the durability of logic-gated and armoured designs in human disease, the long-term safety of multiplexed editing, and whether decentralised manufacturing can reconcile biological sophistication with equitable availability. Confidence in current conclusions about next-generation designs remains limited by short follow-up, small single-arm cohorts and uneven geographical representation of the evidence.