In vivo CAR T-cell therapy: a systematic review of clinical trials

Abstract Background In vivo chimeric antigen receptor (CAR) T-cell therapy directly programs endogenous T cells within the patient rather than generating the cells through conventional ex vivo manufacturing. While promising, the clinical landscape of this emerging technology is fragmented. Therefore, this systematic review evaluated its clinical use in hematologic malignancies and autoimmune diseases. Methods PubMed, Scopus, Web of Science, and abstracts from major conferences (e.g., ASH, ASCO, EHA) were searched from inception through 31 July 2026 for clinical trials reporting efficacy and safety outcomes of in vivo CAR T-cell therapy. Study characteristics, delivery platforms, kinetics, efficacy, and adverse events were extracted and assessed. Results Seven clinical trials involving 41 patients were included. Thirty-three patients had hematologic malignancies, including B-cell non-Hodgkin lymphoma and relapsed or refractory multiple myeloma, and eight had autoimmune diseases. Four studies used lipid nanoparticle platforms delivering CAR messenger RNA, whereas three used integrating lentiviral vectors. In vivo CAR T-cell generation was demonstrated in all studies. Clinical activity was evaluable in 37 patients and included responses and minimal residual disease negativity in multiple myeloma, objective responses in B-cell lymphoma, and improvement in disease activity across autoimmune cohorts. Cytokine release syndrome occurred in 24 of 31 patients in studies reporting numerical incidence and was usually low grade, although severe events occurred. Neurotoxicity was uncommon, with one grade 1 event reported. Conclusions In vivo CAR T-cell therapy has demonstrated biological and clinical feasibility across malignant and autoimmune diseases. However, small heterogeneous cohorts, short follow-up, and platform-specific toxicities preclude conclusions regarding comparative effectiveness or durability. Future trials should prioritize standardized reporting of cellular kinetics and toxicity grading to enable cross-platform comparisons and truly assess the long-term durability of this promising approach.