Low-intensity cytoreduction with low-dose cytarabine prior to blinatumomab for pediatric and AYA R/R ALL integrated with multi-omics analysis to discover predictive biomarkers: Initial results from JPLSG-ALL-R19-blin

Abstract

Abstract Purpose: The prognosis for relapsed acute lymphoblastic leukemia (ALL) remains poor, highlighting the need for novel therapeutic strategies. Over the past five years, blinatumomab has demonstrated safety and efficacy in both frontline and relapsed settings, including in heavily pretreated or refractory cases, with overall response rates (ORR) of approximately 30–50%. However, in many relapsed cases, especially those with intensive prior treatments or comorbidities, strong cytotoxic chemotherapy is often not feasible. Thus, gentler regimens that can reduce leukemia burden while maintaining blinatumomab's efficacy and safety are critically needed. Simultaneously, identifying predictive factors through comprehensive analyses of the tumor immune microenvironment and robust, disease-specific biomarkers is essential to optimize treatment strategies. To address these challenges, we conducted a phase I/II study (JPLSG-ALL-R19-BLIN) combining low-dose cytarabine and blinatumomab in pediatric and AYA patients with relapsed or refractory B-ALL, together with exploratory multi-omics analyses aimed at discovering predictive biomarkers. Methods: Pediatric and AYA patients (≤25 years) with second or later relapse, refractory relapse, or relapse after stem cell transplantation (SCT) were included. Patients with first relapse of ALL were also eligible if they were intolerant to conventional chemotherapy and required less toxic reinduction therapy because of comorbidities. Patients received continuous infusion of cytarabine at 20 mg/m²/day for 7 days, followed by two cycles of blinatumomab. In total, 20 patients were enrolled. Comprehensive analyses of treatment response, immune microenvironment in bone marrow (BM) and peripheral blood (PB) using mass cytometry, and gut microbiota and metabolites were performed to identify potential predictive biomarkers. The gut microbiota was analyzed by 16S rRNA gene sequencing, and gut metabolites were profiled using gas chromatography–mass spectrometry and capillary electrophoresis time-of-flight mass spectrometry. Results: Among the initial 20 evaluable patients, the male-to-female ratio was balanced (10:10), with a median age of 10 years (range, 4–17 years). The cohort included one patient with induction failure at first diagnosis, 10 patients with first relapse, 8 with second relapse, and 1 with third or later relapse. Notably, seven patients had Down syndrome.　The overall response rate (ORR) was 50%, consistent with prior blinatumomab studies. At baseline (TP0), median bone marrow blast percentages were 64.9% (range, 5.6–91.4%) in the CR group and 88.6% (range, 6.2–99.4%) in the non-CR group, with no significant difference observed (P = 0.22). In contrast, TP2 bone marrow assessment (day 15 of cycle 1) showed a strong association with subsequent achievement of CR at the end of 2 cycles of blinatumomab, with blast percentages of 0% (range, 0–0%) in the CR group and 58.0% (range, 43.8–84.6%) in the non-CR group (P < 0.01), suggesting its utility as an early predictive marker of response. Mass cytometry analysis of immune cells revealed that a higher proportion of total T cells in pre-treatment BM or PB was observed in complete responders than in non-responders, and that there was no obvious impaired T-cell activation following cytarabine preconditioning. In the gut microbiota analysis at TP0, specific bacteria, including Clostridium paraputrificum, Eggerthella, and Flavonifractor were enriched in responders, suggesting a potential link between gut microbiota and blinatumomab efficacy. In addition, gut metabolome analysis at TP0 revealed distinct differences in the metabolite composition between the CR group and non-CR group. Conclusions: Low-dose cytarabine cytoreduction followed by blinatumomab represents a promising low-intensity induction strategy for relapsed/refractory pediatric B-ALL, particularly in patients who cannot tolerate intensive chemotherapy due to prior treatments or comorbidities. Integrated immunological and microbiome analyses may provide novel predictive biomarkers for blinatumomab response and guide future personalized immunotherapy approaches.