We retrospectively reviewed the data on CD19+ R/R B-ALL patients who received hCART19s therapy at the Affiliated Hospital of Xuzhou Medical University from May 2016 to May 2020. All the enrolled patients had relapsed or refractory disease. The eligibility criteria were age less than 70 years; good organ function and evaluation of a survival longer than 3 months; and Eastern Cooperative Oncology Group (ECOG) performance status <2. All patients provided signed informed consent before the hCART19s therapy and allo-HSCT. Patients received HSCT or a second infusion of hCART19s after hCART19s infusion depending on patients’ choice, disease status, and their affordability to the treatment. Before HSCT, all patients were in MRD-negative CR after the first hCART19s infusion. In some cases of relapse after the first hCART19s infusion, the second hCART19s was infused as soon as the relapse was clinically confirmed (the detail is listed in Supplementary Table 2 ). Patients were divided into three groups: the first infusion of hCART19s without transplant or a second infusion of hCART19s (CART1 group), a second infusion of hCART19s without transplant (CART2 group), and the first infusion of hCART19s followed by HSCT (CART+HSCT group).

The study protocol was approved by the human studies review board at the Affiliated Hospital of Xuzhou Medical University (ClinicalTrials.gov # NCT02782351). The clinical investigation was conducted according to the principles of the Declaration of Helsinki.

HCART19s constructed with 4-1BB costimulatory domain were generated via a lentiviral vector as previously reported (6). All the hCART19s required quality control before discharge. After a lymphocyte-depleting chemotherapy with a fludarabine and cyclophosphamide (FC) regimen (fludarabine at 30 mg/m² per day for 3 days and cyclophosphamide at 300 mg/m² per day for 3 days), all patients provided signed informed consent before the hCART19s therapy and allo-HSCT and received a single dose of autologous hCART19s infusion at 1×10⁶ CAR-T cells/kg ( Supplementary Table 1 ). No patient experienced bridged chemotherapy from preparation of CAR-T cells to infusion, and lymphodepletion was also conducted prior to the second CAR T infusion.

All patients received marrow ablative regimens using BU/CY or modified BU/CY strategies for matched related transplantation, and anti-thymocyte globulin (ATG, rabbit) was administered for haploidentical or unrelated transplantation as a prophylactic against graft-versus-host disease (GVHD). Prophylactic regimens of GVHD were determined by the individual transplant physician based upon disease-related and transplant-related considerations composed of cyclosporine, methotrexate, and mycophenolate mofetil.

The cytokine release syndrome was graded according to the cytokine release syndrome grading system. The cytokine release syndrome was considered to be severe if it was of grade 3 or higher. Neurotoxic effects were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03. Severe neurotoxic effects were defined as a seizure of any grade or a toxic effect of grade 3 or higher.

Response to therapy was assessed using morphological analysis and multicolor flow cytometry. CR was defined as less than 5% bone marrow blasts, the absence of circulating blasts, and no extramedullary sites of disease (as assessed by means of computed tomography or positron-emission tomography), regardless of cell count recovery. MRD negativity was defined as less than 0.01% bone marrow blasts for all samples analyzed by multicolor flow cytometry. MRD detection was performed at 1, 2, 3, 6, 12, 18, 24, 36, and 48 months after hCART19s therapy or HSCT. Relapsed disease was defined as the reappearance of blasts in blood or bone marrow or in an extramedullary site after a CR. Overall survival (OS) was defined as the time from infusion to the date of death from any cause. Leukemia-free survival (LFS) was calculated from the date of CR to the date of relapse, death, or the last follow-up.

All measurement data were described using median and range and compared using Mann–Whitney U tests. Enumeration data were presented as frequency (%) and compared using chi-square tests or Fisher’s exact test. Follow-up time was estimated using the Kaplan–Meier method, whereas OS and LFS were estimated using the Kaplan–Meier method. A Cox regression model was used to obtain the hazard ratio (HR) estimates and corresponding 95% confidence intervals (CIs) for OS and LFS. The X-tile 3.6.1 software was used to determine the optimal cutoff values for MRD. All tests were two-sided, and p< 0.05 was considered statistically significant. Data were analyzed using SPSS version 26.0 and Graphpad Prism version 8.

From May 2016 to May 2020, a total of 28 patients were enrolled in this study, including 12 males and 16 females, with a median age of 22 years. The characteristics of the patients are shown in Table 1 , and detailed information is listed in Supplementary Table 1 . There was no statistical difference among the baseline data of the three groups.

Of the 28 patients receiving hCART19s infusion, 27 patients were evaluated for response at 14 or 28 days, and 1 patient died at day 14 before being evaluated. The complete remission rate was 92.6% (25/27), with 20 patients having MRD negativity. The rates of CR did not differ significantly among the aforementioned three groups after the first hCART19s infusion. Seven patients received a second hCART19s infusion, three of whom achieved second CR with MRD negativity (42.9%). Notably, as shown in Figure 1 , compared to the first infusion, the proportion of patients with CR is lower after second infusion (92.6% vs. 42.9%, respectively, p=0.02).

CRS was the most common nonhematological adverse event after the infusion of hCART19s, which occurred in 27 of the 28 patients, including 8 patients experiencing severe CRS. The median time of CRS occurrence was 6 days (range: 1–20). Patients with CRS were treated with nonsteroidal anti-inflammatory drugs, glucocorticoids, and tocilizumab. One patient died from grade 5 CRS at day 14 after infusion of hCART19s, and one patient died from encephalorrhagia at 24 days after infusion of hCART19s. The peak serum levels of IL-6 in patients who developed grade 3–5 CRS were higher than those with grade 0–2 CRS (n=20). There was no statistical difference in serum ferritin levels between patients with grade 0–2 CRS and grade 3–5 CRS ( Figure 2 ). Neurotoxicity occurred in one patient in the CART+HSCT group, with symptoms of convulsion. In addition, all the patients developed B cell dysfunction, manifested as CD19+ B cell deletion and hypogammaglobulinemia. III–IV hematological toxicity and other adverse events are shown in Table 2 .

The incidence of CRS after the second infusion was 7/7 (100%). No grade 3–5 CRS and neurotoxicity occurred in these patients.

Ten patients received allo-HSCT after the first hCART19s infusion, with a median time of leukocyte engraftment of 15.5 days (range: 11–25) and platelet engraftment of 20 days (range: 12–36), respectively. All patients obtained a second CR. The incidence of second CR is higher in the CART+HSCT group compared to the CART2 group (100% vs. 42.9%, p=0.015) ( Figure 1 ). One patient experienced grade 1–2 aGVHD, and two patients experienced grade 3–4 aGVHD, with a median time of 40 days (range: 26–45). No patient experienced cGVHD.

At the cutoff (January 1, 2021), the median follow-up time was 1,240 days (95% CI, 709 to 1,770). Among the 10 patients receiving allo-HSCT, the median time from the first hCART19s infusion to transplantation was 63 days. The 1-year OS rates were 70.0% (95%CI, 33.0 to 90.2), 57.1% (95% CI, 17 to 85.1), and 36.4% (95% CI, 11.2 to 62.4) in the CART+HSCT, CART2, and CART1 groups, respectively. Patients in the CART+HSCT group had a higher OS than those in the CART1 group (p=0.0063, Supplementary Figure 1A ) but did not differ from that in the CART2 group (p=0.0893, Supplementary Figure 1B ). The 1-year LFS rates were 80.0% (95%CI 40.9 to 94.6), 28.6% (95% CI, 4.1 to 61.1), and 33.3% (95% CI, 4.6 to 67.6) in the CART+HSCT, CART2, and CART1 groups, respectively. Patients in the CART+HSCT group had a higher LFS than those in the CART1 group (p=0.0018) and the CART2 group (p=0.0183) ( Supplementary Figures 1C, D ). OS and LFS in the CART1 and CART-2 groups did not differ significantly (p= 0.2327, p=0.1818, respectively, Supplementary Figures 1E, F ). When it comes to long-term survival, higher rates of long OS and LFS were achieved in the CART+HSCT group at 3 years compared to the CART2 group (58.3% vs. 0, p<0.001; 66.7% vs. 0, p<0.001) ( Figure 3 ). Similar results were obtained when the CART+HSCT group was compared to the CART1 group (58.3% vs. 0, p<0.001; 66.7% vs. 0, p<0.001) ( Figure 3 ).

In addition, three patients relapsed at 30, 44, and 399 days after allo-HSCT. For the CART2 group, the median time to first relapse was 189 days (range: 92 to 719). On the other hand, three patients gained CR after a second hCART19s infusion, of whose median time to second relapse was 92 days (range: 38–120). For the CART1 group, the median time to relapse was 149 days (range: 30–449). Between the first hCART19s infusion and the second infusion, the median relapse time was much shorter after the second infusion (92 vs. 178, p=0.039) ( Figure 1 ). Moreover, one patient relapsed with CD19 negative in the CART1 group and the CART2 group, respectively. All patients who relapsed were CD19 positive after CART+HSCT.

Considering all patients possessing positive MRD before hCART19s infusion, we used X-tile to determine the optimal cutoff value for MRD. The MRD cutoff value was with a maximum χ² log-rank value of 5.59 (p=0.017) ( Supplementary Figure 2 ). Univariate analyses revealed that age, complex chromosome set, recurrence times, treatment times, volume, and the number of infused cells had no significant effects on OS and LFS. The disease burden significantly affected OS and LFS ( Figures 4A–D ). There was also a trend toward better OS (p=0.064) for patients with MRD-negative CR versus MRD-positive CR, even when no significant difference was obtained ( Figure 4E ). Patients with MRD-negative CR achieved a longer LFS than those with MRD-positive CR (p=0.032) ( Figure 4F ). Candidate variables with a p value <0.1 on univariate analysis were included in multivariate analysis. In multivariate analyses, the proportion of BM blasts (HR 6.055; 95% CI 1.624–18.933; p=0.004) and receiving allo-HSCT (HR 0.250; 95% CI 0.068–0.915; p=0.036) were independent predictors for OS. MRD<65.6% at infusion (HR 7.905; 95% CI 1.016-61.505-0.977; p=0.048) and receiving HSCT (HR 0.139; 95% CI 0.031–0.619; p=0.010) were independent risk factors for LFS. This result showed that receiving HSCT is predictive of both OS and LFS than a second infusion of hCART19s ( Table 3 ).