Redefining the Treatment Continuum: A Toxicity- and Response-Adapted Framework for Sequencing Autologous CAR-T, Allogeneic Stem Cell Transplantation, and Bispecific T-Cell Engagers in Relapsed/Refractory Hematologic Malignancies

Objective: The aim of this study is to propose a biomarker- and toxicity-informed framework reframing "which therapy is superior" as "which sequence, at which disease timepoint, optimizes durable remission while preserving future options."

Methods: We conducted a conceptual synthesis of patient-level published data from three CAR-T registration trials (ZUMA-1, TRANSFORM, BELINDA), late-cycle allo-HSCT registries, and phase I/II BiTE studies, deriving four quantifiable, cross-experimental decision variables: pre-therapy disease burden and its trajectory (MRD by next-generation flow, cutoff ≥10⁻⁴); cumulative toxicity burden (grade ≥3 CRS/ICANS event count and duration); immune reconstitution kinetics (absolute lymphocyte and NK-cell counts at days +30 and +90, flow cytometry); and graft/product availability (donor identification time, CAR-T manufacturing lead time).

Statistical Plan: Multivariate logistic regression and Fine–Gray competing-risk regression will identify checkpoint-associated predictors of PFS and non-relapse mortality independent of prior line count. Validation will proceed as a multi-center observational cohort (n=200) with pre-specified subgroup analyses, bootstrap resampling (n=500) for internal optimism-correction, and external temporal validation on a 24-month hold-out cohort. Discrimination will use time-dependent AUC for PFS at 12 and 24 months; calibration will be performed via the Greenwood–Nam–D'Agostino test.

Proposed Framework: (1) MRD-negative relapse favors cellular therapy; MRD-positive relapse favors transplant-first or BiTE debulking. (2) Episodes with ≥2 and ≥3 grade CRS/ICANS contraindicate repeat cellular therapy, prompting BiTE salvage; <2 permits sequential CAR-T or consolidative transplant. (3) Day +90 lymphocyte <0.5×10⁹/L or NK-cell <100/µL delays transplant pending immune-stimulatory bridging. (4) BiTEs bridge to transplant/salvage only when CD3+ count exceeds 200/µL, preserving CAR-T expansion capacity.

Conclusion: This is, to our knowledge, the first toxicity- and immune-fitness-adapted continuum unifying these modalities, with prospective validation of reconstitution kinetics as a timing biomarker poised to reshape tumor board decisions and trial design.