Comparisons of Long-Term Survival and Safety of Haploidentical Hematopoietic Stem Cell Transplantation After CAR-T Cell Therapy or Chemotherapy in Pediatric Patients With First Relapse of B-Cell Acute Lymphoblastic Leukemia Based on MRD-Guided Treatment

Measurable residual disease (MRD) positivity before haploidentical hematopoietic stem cell transplantation (haplo-HSCT) is an independent prognostic factor in determining outcomes in patients with B-cell acute lymphoblastic leukemia (ALL). In this study, we conducted a parallel comparison of the efficacy and safety in patients with suboptimal MRD response after reinduction who underwent haplo-HSCT after chimeric antigen receptor T-cell (CAR-T) therapy or chemotherapy. Forty B-cell ALL patients who relapsed after first-line chemotherapy and with an MRD ≥0.1% after reinduction were analyzed. The median pre-HSCT MRD in the CAR-T group (n = 26) was significantly lower than that in the chemotherapy group (n = 14) (0.009% vs. 0.3%, p = 0.006). The CAR-T group exhibited a trend toward improved 3-year leukemia-free survival and a significantly improved 3-year overall survival compared to the chemotherapy group [71.8% (95% confidence interval (CI): 53.9–89.6) vs. 44.4% (95% CI: 15.4–73.4), p = 0.19 and 84.6% (95% CI: 70.6–98.5) vs. 40.0% (95% CI: 12.7–67.2), p = 0.008; respectively]. Furthermore, no increased risk of graft-versus-host disease, treatment-related mortality, or infection was observed in the CAR-T group. Our study suggests that CAR-T therapy effectively eliminates pre-HSCT MRD, resulting in better survival in the context of haplo-HSCT.


INTRODUCTION
Relapse remains the main cause of mortality in childhood B-cell acute lymphoblastic leukemia (ALL); approximately 20% of patients fail to respond to chemotherapy after relapse (1), with a 5-year survival rate of 30% (2,3). Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is an effective treatment option for patients experiencing relapse after first-line chemotherapy (4). However, tumor load has been reported to be closely related to the outcomes after transplant; patients failed to achieve complete remission (CR) before HSCT, with a reported 3-year leukemia-free survival (LFS) rate of 21%-29%. Moreover, the importance of pre-HSCT measurable residual disease (MRD) status has gained appreciation in recent years and has shown significantly inferior outcomes in ALL patients with positive pre-MRD (5). Bader et al. have reported that the cumulative incidence of relapse for pre-MRD-negative, MRD < 0.01%, MRD <0.1% and ≥0.01%, and MRD ≥0.1% was 11%, 20%, 64%, and 54%, respectively (p < 0.001) (6). Thus, more effective therapies are required for patients experiencing relapse after first-line therapy for ALL.
Previous clinical trials have shown that refractory/relapse B-cell ALL patients receiving chimeric antigen receptor T-cell (CAR-T) therapy attained higher MRD-negative CR rates than those receiving conventional chemotherapy (6). However, studies on the efficiency and safety of CAR-T therapy compared with chemotherapy on pre-HSCT MRD eradication and survival based on MRD-guided treatment in the context of haploidentical HSCT (haplo-HSCT) are lacking. Therefore, in this study, we conducted a parallel comparison of the prognosis and treatmentrelated complications among relapse patients with suboptimal MRD response after reinduction chemotherapy who underwent haplo-HSCT from either CAR-T therapy or chemotherapy.

Patients
We included first relapse of B-cell ALL patients who had suboptimal MRD response after reinduction chemotherapy and underwent haplo-HSCT bridged from either CAR-T therapy or chemotherapy at Peking University People's Hospital between April 2015 and April 2020. The exclusion criteria were as follows: (1) isolated extramedullary relapse; (2) previous transplantation and/or previous CAR-T therapy; (3) significant cardiovascular, hepatic, and renal dysfunction and active infections; (4) MRD < 0.1% after reinduction chemotherapy; and (5) patients followed by matched sibling donor transplantation (MSDT). Because the uniformity of treatment protocols can help to reduce the impact of different treatment protocols and most of the donors of allo-HSCT in our institute were haplo, 3 patients followed by MSDT were excluded. The study was approved by the Ethics Committee of Peking University People's Hospital.
CAR-T cells in this study were investigational products and conducted in the setting of clinical trials (www.clinicaltrials.gov as #NCT 03050190; www.chictr.org.cn as #ChiCTR-OPN-17013507). Anti-CD19 CAR-T cells constructed with 4-1BB or CD28 costimulatory domains were generated via a lentiviral vector from fresh leukapheresis material. Lymphodepleting chemotherapy, including fludarabineand cyclophosphamidebased conditioning treatments, were administered prior to CAR-T infusion.

Statistical Analysis
Enumeration data were compared using the chi-square test or Fisher's exact test, whereas the Mann-Whitney U test was used for continuous variables. The Kaplan-Meier method was used to analyze LFS and OS. The competing risk model was used to analyze cumulative incidence, and Gray's test was used to assess the differences between cumulative incidences in univariable analyses. Multivariate analysis was performed using the Cox proportional hazards regression model. All p-values were twosided, and values <0.05 were considered statistically significant. The Statistical Program for Social Sciences (SPSS) software (version 23.0; SPSS Inc. Chicago, IL, USA) and R version 3.5.3 (R Foundation for Statistical Computing, Vienna, Austria) were used for data analysis.
In the chemotherapy group, six patients remained NR after reinduction, and the median MRD level was 7.6% (0.11%-44%) after reinduction. Before haplo-HSCT, patients in the

Long-Term Survival
Post-HSCT relapse occurred in 7 (26.9%) patients at a mean period of 12.4 months in the CAR-T group; among them, three patients died of relapse, and four patients responded to salvage therapy (two survived by HSCT, one by CAR-T, and one by chemotherapy plus donor leukocyte infusion). In the chemotherapy group, post-HSCT relapse occurred in seven (50%) patients at a mean period of 10.5 months, and seven patients died of relapse. In addition, two patients (one from each group) died of NRM.
The mean follow-up time was 37.3 years for survived patients. Among all patients, aged <10 years, isolated bone marrow relapse, pre-HSCT MRD positivity, and non-CAR-T therapy were risk factors for inferior survival ( Table 2). In the CAR-T group, the factors of isolated bone marrow relapse and pre-HSCT MRD positivity may have been abrogated by CAR-T therapy; the 3-year LFS and OS of patients with isolated bone marrow relapse were comparable with those of combined relapse

DISCUSSION
Although the prognosis of pediatric ALL has been proven, >85% of patients survived without relapse (11). However, survival following relapse is dismal; the Children's Oncology Group has reported that the 5-year OS rate for relapse patients is 36% (12), and the 10-year OS for children treated in the ALL-REZ BFM 90 trial was 36% (13). Multiple studies have demonstrated that the interval between remission and relapse and the site of relapse are important predictors of outcome (14)(15)(16). Moreover, MRD response to reinduction has been defined as the strongest prognostic factor for relapsed ALL; the ALL-REZ BFM P95/96 trial showed that MRD <0.1% after reinduction predicted a 10year event-free survival (EFS) of >70% and MRD ≥0.1% after reinduction predicted a 10-year EFS of <20% with conventional chemotherapy (17). Therefore, an MRD-based strategy to intensify treatment with allo-HSCT in patients with MRD ≥0.1% after reinduction was implemented in the ALL-REZ BFM 2002 trial and an improved outcome (8-year EFS, 64% vs. 18% in the historical control) was observed (18). However, some studies have suggested that pre-HSCT MRD positivity is associated with subsequent relapse and poor survival (8,19). The International BFM Study Group reported the first prospective study to evaluate the effect of pre-HSCT MRD status and found that the 4-year EFS was the highest at 64% for undetectable MRD and 48% for MRD <0.01%; a multivariate Cox regression model confirmed that pre-HSCT MRD positivity (>0.01%) was the only independent prognostic factor for relapsed ALL (6). We found the same results in this study; patients with pre-HSCT MRD positivity had significantly lower 3-year LFS than those with pre-HSCT MRD negativity (33% vs. 71%, p = 0.041).
Therefore, we aimed to explore a powerful strategy to eliminate pre-HSCT MRD under the guidance of MRD in the context of haplo-HSCT. Blinatumomab has been proven to be active in relapsed and refractory adult and pediatric ALL (20,21). COG AALL1331 reported that blinatumomab has a deeper MRD clearance, improved LFS and OS, and lower toxicity compared with chemotherapy. In patients with MRD ≥0.1% after reinduction, blinatumomab was used followed by HSCT to eliminate pre-HSCT MRD in some institutes (4). However, studies on intensifying treatment with CAR-T pre-haplo-HSCT in patients with MRD ≥0.1% after reinduction are lacking. In the present study, 35.7% of patients in the chemotherapy group failed to achieve pre-HSCT MRD-negative CR, and the median level of pre-HSCT MRD was 0.3 (0%-3.8%). In the CAR-T group, 19.2% of patients failed to achieve pre-HSCT MRDnegative CR, and the median level of pre-HSCT MRD was 0.009 (0%-0.17%). Simultaneously, the CAR-T group exhibited a trend toward improved 3-year LFS and a significantly improved 3-year OS compared with the chemotherapy group (71.8% vs. 44.4%, p = 0.19) (84.6% vs. 40.0%, p = 0.008). We believe that a deeper MRD clearance effectively led to improved survival. Our data showed that MRD-based strategies, including CAR-T, improved the prognosis of relapse patients. Moreover, our study observed that patients aged <10 years had an inferior survival compared to those aged ≥10 years. Isolated bone marrow relapse was found to have inferior survival compared to bone marrow combined extramedullary relapse, which is consistent with previous studies, and may in part be due to the added TBI in patients with extramedullary relapse. Safety was another concern regarding the application of CAR-T pre-HSCT; consistent with previous studies, no increased NRM was observed in the CAR-T group in the present study (22,23). The median duration of engraftment after haplo-HSCT, the cumulative 100-day incidence of aGVHD grades II-IV, and the 3-year incidence of total and severe cGVHD in the CAR-T group were also similar to those of the chemotherapy group. Moreover, several studies have reported that CRS occurred in nearly 30%-94% of patients from 1 to 22 days after CAR-T cell infusion (24,25). Maude et al. observed that 37%-78% of patients show cytopenia 1 month after CAR-T cell infusion (24). Therefore, we designed a regimen that bridged haplo-HSCT after at least 1 month of CAR-T infusion to reduce treatment-related complications. Park et al. have reported that 3 of 53 patients experienced infection-related mortality after CAR-T infusion. In the present study, no increased risk of viremia was detected in the CAR-T treatment protocol combined with haplo-HSCT.
The strengths of this study include homogeneous patient groups using consistent conditioning regimens, stem cell sources, and supportive care algorithms. However, our study has the following limitations: (1) it was not a prospective randomized study, (2) the sample size was limited, and (3) the choice of covariates for the multivariate analysis was constrained by the small number of observed events.
In conclusion, our study indicates that CAR-T therapy effectively eliminates pre-HSCT MRD, resulting in better survival in the context of haplo-HSCT. Moreover, no additional treatment-related complications were observed.

AUTHOR CONTRIBUTORS
LPZ and XH designed the research and revised the paper. GH and YFC analyzed the data and wrote the paper. YZ, YJC, PS, YJ, AL, YW, SJ, LJZ, YS, CY, LX, XZ, KL, and YW collected and analyzed data. All authors contributed to the article and approved the submitted version.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.