Synergistic integration of CAR-T cell pharmacokinetics and NGS-MRD evolution enables precision risk stratification in relapsed/refractory B-cell acute lymphoblastic leukemia

Abstract Background: Predicting long-term treatment response to CAR-T therapy for R/R B-ALL remains a major challenge, making it difficult to effectively stratify patients and make personalized treatment decisions after therapy. Since CAR-T expansion kinetics and MRD dynamics play a crucial role in treatment effectiveness, this study aims to explore how each of these factors individually affects outcomes for R/R B-ALL patients. It also seeks to develop a comprehensive predictive model by combining NGS-MRD trajectories with CAR-T expansion parameters. Methods: This study included 27 patients with relapsed/refractory B-cell acute lymphoblastic leukemia (r/r B-ALL) enrolled in a phase 2, multicenter study of inaticabtagene autoleucel (Inati-cel) for adult CD19+ r/r B-ALL (ClincialTrials.gov Identifier: NCT04684147). All participants were treated with this CD19 CAR-T therapy and had available continuous next-generation sequencing minimal residual disease (NGS-MRD) data and CAR-T expansion parameters. Patients were categorized based on their NGS-MRD trajectory over 1-12 months, and RMST Comparison methods were used to assess the impact of different NGS-MRD trajectories on relapse-free survival (RFS). A Cox proportional hazards model was employed to examine the relationship between CAR-T expansion parameters and RFS for each NGS-MRD trajectory group. Results: The dynamic changes in NGS-MRD over 1 to 12 months were classified into three patterns: (1) consistently negative trajectory, (2) low-level positivity, and (3) a rapid increase to high levels. The median RFS for the consistently negative group was 23 months, while the low-level positive and rapidly increasing groups had 14.63 months and 8.37 months shorter RFS, respectively, compared to the consistently negative group (RMST Comparison; pattern 2 vs. pattern 1, P=0.0012; pattern 3 vs. pattern 1, P <0.0001). This highlights the clinical significance of NGS-MRD trajectories in disease prognosis. Furthermore, multivariate Cox regression revealed that the area under the curve (AUC) of CAR-T expansion over 28 days (AUC0_28d) was an independent factor affecting RFS. Higher CAR-T exposure in the first 28 days was associated with a protective effect on RFS (HR, 0.1003; 95% CI,0.0167-0.6006; P=0.0118). In subgroups based on CAR-T exposure, the impact of NGS-MRD trajectories on RFS was more pronounced in the low expansion group (low AUC) compared to the high expansion group (high AUC). In the high expansion group, there was no significant difference in prognosis between the low-level positive and consistently negative NGS-MRD groups, suggesting that when CAR-T expansion is sufficient, minor changes in NGS-MRD have limited impact on prognosis (P=0.4756). Conversely, in the low expansion group, the difference in prognosis between the low-level positive and consistently negative groups was significant. Conclusions: Both the 28-day cumulative CAR-T expansion area under the curve (AUC) and NGS-MRD trajectory act as independent predictors of treatment outcomes in patients with relapsed/refractory B-cell acute lymphoblastic leukemia (R/R B-ALL). When these two factors are integrated into a comprehensive predictive model, we can more accurately pinpoint high-risk patients and develop tailored treatment strategies, which could ultimately improve long-term survival for this group of patients.