Stem cell transplant for mantle cell lymphoma in Taiwan

Mantle cell lymphoma (MCL) is a B-cell lymphoma featuring an aggressive course and a progressive relapsing pattern. International guidelines recommend early consolidative autologous stem cell transplant (auto-SCT) for eligible patients while reserving allogeneic SCT (allo-SCT) as therapy for refractory cases. Since data describing the implementation of transplants in the Asian population with MCL are limited, we aimed to analyze post-SCT outcomes of 99 MCL patients from the Taiwan Bone Marrow Transplant Registry database. The median age was 56 years, and 11% of the patients had blastoid variant MCL. Ninety-four patients received auto-SCT, while 13 patients received allo-SCT, eight of which received allo-SCT after failing auto-SCT. Before auto-SCT, 52% of the patients were in their first complete remission (CR1). Overall, 37 patients (39%) relapsed after auto-SCT. The median post-auto-SCT progression-free survival and overall survival (OS) were 43.6 months and not reached, respectively. Blastoid variant MCL, transplant not received in CR1, and disease progression within 12 months post-auto-SCT independently predicted inferior OS in multivariable analysis. The median post-allo-SCT OS was 74 months. Two patients (15%) died of MCL recurrence post-allo-SCT. Three patients with refractory diseases were salvaged with ibrutinib or venetoclax to allo-SCT. Treatment strategies incorporating novel agents warrant further optimization.

Stem cell harvest and transplantation procedures. In Taiwan, the most commonly used regimen for stem cell harvest is etoposide, methylprednisolone, cytarabine, and cisplatin (ESHAP) with or without rituximab, followed by granulocyte colony-stimulating factor (G-CSF) administration 49 . A minority of patients received high-dose cytarabine in combination with dexamethasone and cisplatin (DHAP) or other cytarabine-based regimens before harvest 50 . All except for ten patients received BCNU, etoposide, cytarabine, and melphalan (BEAM) as the conditioning chemotherapy regimen before auto-SCT 51 . Ten patients received the Benda-EAM regimen (which substitutes the BCNU in the BEAM regimen with bendamustine) as the conditioning regimen 52,53 . The mostly employed infection prophylaxis for auto-SCT is levofloxacin 750 mg or ceftibuten 400 mg daily, administered from D0 until the day absolute neutrophil counts reached 1000 cells/mm 3 . The most used conditioning regimens for allo-SCT in Taiwan are generally categorized into myeloablative and reduced-intensity regimens. The myeloablative regimens are usually backboned with busulfan IV 3.2 mg/kg/day consecutively from day − 8 to day − 5 and cyclophosphamide IV 60 mg/kg/day on day − 3 and day − 2. In contrast, the reduced-intensity regimens are based on the following scheme: fludarabine 30 mg/m 2 /day consecutively from day − 8 to day -4, busulfan IV 3.2 mg/kg/day on day − 5 and day − 4, and cyclophosphamide IV 60 mg/kg/day on day − 2. Antithymocyte globulin (ATG) 4-6 mg/kg can be added as a part of the conditioning regimen in HLA-mismatched allo-SCT scenarios. Usually, cyclosporin with methotrexate is used for graft-versus-host disease (GVHD) prevention in myeloablative allo-SCT, and cyclosporin with mycophenolate mofetil is used for reduced-intensity protocols.
Stem cell transplantation. Before auto-SCT, 49 patients (52.1%) were in their first complete remission (CR1), and 16 (17%) were in their second complete remission (CR2). Twenty-nine (30.9%) patients were in partial remission (PR), of whom 16 were in PR after 1st line therapy while 13 reached PR after two or more lines of treatment. Overall, 37 patients (39.4%) had a disease recurrence after auto-SCT at a median follow-up time of 20.9 months. Among them, 23 (62.2%) patients experienced disease recurrence within 24 months after auto-SCT whereas 18 (48.6%) has patients relapsed within 12 months post-SCT. The median time to post-auto-SCT relapse was 13.1 months (range: 0.7-84).
Among the 13 patients who received allo-SCT, 8 experienced relapses after prior auto-SCT (Table 2). Five patients had primary refractory diseases and were salvaged with various regimens to receive frontline allo-SCT. Only two (15.4%) patients had recurrence of MCL after allo-SCT and eventually succumbed to the disease. However, three patients died of infections after allo-SCT. Patient unlinked patient number (UPN) 13 had disease progression before the scheduled auto-SCT, and he was salvaged with ibrutinib and achieved PR prior to subsequent allo-SCT. Patient UPN 14 had a primary disease that was refractory to multiple lines of therapy and finally achieved PR after treatment with a combination of asparaginase, paclitaxel, and gemcitabine. He then received allo-SCT and was disease-free and alive until the last follow-up. Patient UPN 36 experienced relapse one month after her auto-SCT. Allo-SCT was used as salvage due to the rapid progression of the disease. She had acute GVHD and ensuing infection and died of uncontrolled infection; imaging and bone marrow examinations before death showed no evidence of disease recurrence. Patient UPN 41 had blastoid variant MCL and relapsed eight months after auto-SCT. Her lymphoma was refractory to ibrutinib; rituximab, BCNU, vincristine, methotrexate etoposide, and methylprednisolone (R-BOMES); and rituximab, bendamustine, and cytarabine (R-BAC). She finally attained PR with venetoclax and underwent allo-SCT. Patient UPN 46 had primary refractory disease, although we confirmed that he had nonblastoid MCL. He received ibrutinib with bendamustine and rituximab and was bridged to allo-SCT. Other details are shown in Table 2.
Interestingly, a trend of growing transplantation activity with time (52% of patients were transplanted after 2016) was observed despite the increasing age at transplant (before vs. after 2016: 51.8 years vs. 56.1 years, p = 0.087), which partly reflects the evolution of MCL management. Although similar proportions of patients were transplanted in CR before and after 2016 (66.7% vs. 66.6%), transplantation implemented in CR1 increased by 5% after 2016 (47.9% vs. 52.9%).  Ibrutinib. Ibrutinib was used at a dose of 560 mg/day in twelve patients as salvage therapy: in ten patients, ibrutinib was administered because of relapse after prior auto-SCT, and in two patients, ibrutinib was given as salvaging and bridging therapy to allo-SCT. The median duration of salvaging ibrutinib for post-auto-SCT relapse was 4.6 months (range: 1.9-19.7). Of the eight evaluable patients, the overall response rate (ORR) was 62.5%. Two (25%) patients achieved PR, 3 (37.5%) patients attained CR, 2 patients (25%) maintained stable disease, and one patient (12.5%) had disease progression despite treatment. The median time to response and the response duration were 2.6 months (range: 1.9-4.4) and 6.6 months (range: 1.3-18.6), respectively. Two patients (25%) lost their response after three months and 18 months, respectively. The most frequently encountered adverse events were bleeding (37.5%), pulmonary infection (25%), including Pneumocystis jiroveci pneumonia (12.5%) and pulmonary nontuberculous mycobacterial infection (12.5%), and cytopenia (25%).

Survival
The median post-auto-SCT PFS and OS of the of 86 patients receiving auto-SCT only (ASCT1) were 45.9 months and not reached (NR), respectively (Fig. 1a). The median PFS of the 8 patients who underwent auto-SCT and subsequent allo-SCT (ASCT2) due to post-auto-SCT relapse was 7.9 months (Fig. 1a). Furthermore, the median post-allo-SCT PFS and OS of the 13 patients receiving allo-SCT were 36 months and 48.9 months, respectively (Fig. 1b). Specifically, for auto-SCT, patients with blastoid variant MCL had significantly shorter PFS and OS than those without blastoid variant MCL (median PFS, 7.9 months vs. 43.6 months, p < 0.001, Fig. 2a; and median    Fig. 3a,b, respectively). Since time to progression (progression of disease within 24 months, POD24) 16,17 has been established as a prognostic factor for MCL, we subsequently explored the impact of POD24 on patients' outcomes. Conceivably, patients with POD24 had exceptionally worse survival than those without POD24 (median: 24.9 months vs. NR, p < 0.001, Fig. 4a). We next inquired whether progression of disease within 12 months (POD12) 14 also affect post-SCT survival. Strikingly, patients with POD12 also had an inferior OS compared without POD12 (median: 23.3 months vs. NR, p < 0.001, Fig. 4b). In the multivariable analysis, blastoid variant MCL, transplant not received in CR1, and POD12 independently predicted adverse post-auto-SCT survival (Table 3).
Moreover, as the recruitment of patients spanned an extended period, we further analyzed patients receiving transplantation in different eras to examine the possible chronologic effect. Patients were separated by the median calendar year of transplantation, 2016. The post-SCT outcomes of patients who received transplantation more contemporarily were better than their counterparts regarding PFS and OS (NR vs. 40.8 months, p = 0.039, and NR vs. 73.9 months, p = 0.087, respectively, Supplementary Fig. S1).

Discussion
International guidelines recommend various induction regimens, mainly cytarabine-containing regimens, for transplant-eligible MCL patients and underscore the importance of frontline auto-SCT 28,29 . The survival benefit of early consolidation followed by auto-SCT was first reported in a prospective randomized trial of the European MCL Network, notwithstanding that there was only a PFS, but not an OS, benefit 10 . In Asia, Miura et al. analyzed the outcomes of 64 newly diagnosed Japanese patients with MCL 54 . Sixteen patients in the study received various intensive chemotherapies and ensuing auto-SCT. Before auto-SCT, nine patients were in CR1, one was in CR2, and six were refractory. The survival was rather promising, with a 5-year OS of 93%, compared to the  www.nature.com/scientificreports/ 5-year OS of less than 50% in their counterparts. In another retrospective study, 16% of 501 Japanese patients received auto-SCT following induction chemotherapy with a rituximab-high-dose cytarabine combination 55 . This approach yielded HRs of 0.24 and 0.43 for PFS and OS, respectively, compared to R-CHOP treatment. A survival benefit of frontline auto-SCT was also observed in 97 transplant-eligible patients in a recently published Taiwanese study 32 . This multi-institutional study identified auto-SCT, gastric MCL involvement, blastoid variant MCL, and POD12 as independent prognostic factors. Patients who underwent auto-SCT in CR1 also had a better OS than those who underwent auto-SCT in CR2 or PR (median OS: NR vs. 71 months, p = 0.027). In accordance with previously mentioned studies, approximately half of the patients in the present study received auto-SCT in CR1 and had a superior post-SCT survival than those who received auto-SCT in CR2 or PR. While early auto-SCT improved the long-term outcome of MCL patients, survival plateaus have not been observed. Disease recurrence frequently intercedes in the post-transplantation course, which imposes a challenge against a durable remission for MCL patients 28 . Recently, Visco et al. modelled the relationship between early POD (POD24) 16 and risk of death in 188 relapsed or refractory MCL patients from the Fondazione Italiana Linfomi series. The early POD patients had significantly shorter OS than the late POD group (median, 12 months vs. NR, p < 0.0001). Early POD was also confirmed prognostically detrimental in multivariable analysis with a hazard ratio (HR) of 3.9 (95% confidence interval, 2.16-7.06) and then validated in 93 patients from the MCL Younger study. Interestingly, in 41 patients that received allo-SCT at any time during the study period, patients with early POD had comparable OS to patients with a late POD (p = 0.72), implying that detrimental effects of early POD might be ameliorated by allo-SCT. In a retrospective analysis of the European Society for Blood & Marrow Transplantation (EBMT) registry 56 , there was no significant difference in post-allo-SCT PFS (p = 0.173) and OS (p = 0.336) between patients with and without POD24, further demonstrating the potential of allo-SCT overcoming the poor prognosis of POD24. In the present cohort, patients with POD24 also had a reduced survival than those without, corresponding to that in Visco et al. 's study. On the other hand, although POD12 was associated with worse survival in our study, further validation in larger cohorts may be warranted.
The role of maintenance therapy with rituximab after auto-SCT has been established in the randomized, phase III LyMA trial, in which the rituximab group had superior event-free survival, PFS, and OS at 4 years post randomization 57 . A systematic review and meta-analysis also indicated that rituximab maintenance therapy improved post-auto-SCT PFS and OS in MCL patients 58 . However, post-SCT maintenance therapy in our analysis did not seem to affect post-SCT survival as much as it did in the previous studies. Admittedly, since the Taiwan National Health Insurance does not reimburse maintenance rituximab for MCL, the rate of patients receiving post-auto-SCT maintenance rituximab in this study was limited, thereby precluding a sound comparison.
While real-world data have potentially demonstrated the benefit of upfront auto-SCT in some patients, a high recurrence rate, even after consolidation treatment, has been observed in patients harboring high-risk factors, including blastoid variant MCL or TP53 mutation 15,59,60 . Hence, an approach with intensified chemotherapy followed by frontline auto-SCT for young and fit patients might be justified, particularly in the era of novel agents and improved safety and accessibility of allo-SCT.
Although current data supporting allo-SCT as upfront therapy are lacking, earlier studies suggested at least some benefits for patients with chemorefractory MCL. Hamadani et al. investigated the outcomes of 202 patients from the Center for International Blood and Marrow Transplant Research (CIBMTR) database and found that approximately 25% of patients with refractory MCL could attain durable remission after allo-HCT 18 . In a study by Fenske et al., 519 chemotherapy-sensitive MCL patients received auto-SCT or allo-SCT with reduced-intensity conditioning at different time points during the disease course 61 . Auto-SCT and allo-SCT resulted in comparable 5-year OS rates (61% vs. 62%, p = 0.951). Both auto-SCT and allo-SCT in frontline settings were demonstrated to be beneficial for survival in multivariable analysis. In a recent prospective and multicenter study, 24 of 25 patients received upfront allo-SCT and were engrafted without therapy-related mortality by day 100 62 . With a www.nature.com/scientificreports/ median follow-up of 60.5 months, there were three (12.5%) deaths from MCL. The PFS and OS were 56% and 76% at 5 years, implying that frontline allo-SCT is feasible and should be considered for selected patients, such as those with refractory disease or a high risk of progression (i.e., POD24). In this study, two patients had a relapse after allo-SCT, and 80% of the patients receiving upfront allo-SCT were still alive at a median follow-up of 21.5 months. Nonetheless, too few patients in this cohort received allo-SCT, hindering the results regarding the impact of allo-SCT. Therefore, larger-scale, prospective, and randomized controlled trials are warranted to support this approach.
In our cohort, the ORR of patients receiving ibrutinib for post-auto-SCT relapsed MCL was 62.5%, in agreement with that in trial settings [63][64][65] as well as in real-world analysis 66 . There were no unexpected adverse events, and all events were controlled by tapering of doses. Nevertheless, the limited number of patients using BTK inhibitor could make the data much less convincing than it could be. Moreover, in the scenario where patients who remain in remission with ibrutinib and hence hesitate about SCT, it may also render bias in this transplantation registry-based analysis.
Ibrutinib has been shown efficacious in relapsed MCL after first-line treatment and auto-SCT, with ORRs spawning from 69 to 74% and CR rates ranging from 27 to 32% 56,66 . Nonetheless, its effectiveness in early-POD patients has not been conclusive. In a real-world analysis, 211 patients received ibrutinib as second-line therapy for relapsed or refractory MCL 66 . POD24 was a significant adverse predictor of PFS and OS in a univariable regression model, yet this association was lost in multivariable analysis. In Burney et al. 's study, there was no difference in the duration of response to ibrutinib between patients with and without POD24 (median: 10.4 months vs. 9.8, p = 0.9). After 2 years of initiating ibrutinib, the POD24 group had markedly shorter PFS and OS than patients without POD24 (p = 0.017 and p = 0.019, respectively). In the MANTLE-FIRST study 17 , curiously, ibrutinib was associated with significantly better OS in patients with POD24 compared with other second-line therapies (HR 2.41 for RB, 2.78 for R-BAC, and 2.17 for others). By contrast, in patients without POD24, such improvement in survival was not observed. As novel agents have become more broadly used in real-world practice, a treatment strategy that incorporates conventional chemotherapies, stem cell transplants, and targeted therapies needs to be optimized for this heterogeneous disease.
This study's limitations lie in its registry-based nature and, crucially, its lack of TP53 genotyping data. The study could also be improved with more data, including data regarding the regimens and responses of each line therapy. Furthermore, this study enrolled patients over 20 years, thereby introducing various inherent confounding factors into our analysis. While early transplantation has been evoked since the early 2010s 8,14 , the lack of consensus on treatment for MCL in Taiwan makes the decision to an upfront transplantation more heterogeneous in real-world practice. Although there isn't a comparison arm for patients who did not receive transplantation in this registry-based study, a previous Taiwanese multi-institutional study indicated that 14  Nowadays, allo-SCT remains the curable potential in patients with relapsed/refractory MCL 61,67 . Aside from transplantation, the chimeric antigen receptor T-cells (CAR -T) may represent an alternative for long-term survival, notably after BTK inhibitors have failed 68 . In the ZUMA-2 study 68 , the ORR of 60 patients evaluated for response was 93%, with a CR rate of 67%. Remissions were durable in a majority of patients while only 15% of patients experienced grade 3 or higher but non-fatal cytokine release syndrome. In the TRANSCEND-NHL-001 Trial 69 , the ORR was 84% in 32 evaluable patients, and 59% achieved a CR. Notably, patients with a history of auto-SCT or allo-SCT were eligible in the TRANSCEND-NHL-001 study while those with prior allo-HCT was not included in the Zuma-2 study. While many trials are currently enrolling patients, the above data suggest the high efficacy of CD19-targeted CAR-T therapy for relapsed or refractory MCL, further underpinning the importance of integrating novel agents, cellular therapies, and transplantation. Consideration of disease refractoriness following relapse after transplantation or novel agents, and importantly, cost-effectiveness, is also required. Although tisagenlecleucel has been recently approved in Taiwan for the treatment of B-cell acute lymphoblastic leukemia in children and young adults 70 and relapsed or refractory diffuse large B-cell lymphoma 71 , CAR-T therapies for MCL patients are still restricted within a clinical trial setting. However, with the impressive results from previous studies, a profound change in the treatment landscape for MCL is anticipated following the introduction of CAR-T therapy.
In summary, this study depicted the implementation of transplantation for MCL patients in an Asian population, emphasizing the survival benefit conferred by early consolidative auto-SCT compared to later auto-SCT and the prognostic impact of blastoid variant MCL and POD24. In addition, novel agents such as ibrutinib or venetoclax may play a role in bridging high-risk patients to subsequent allo-SCT, which was also demonstrated to be feasible in the frontline setting for selected patients. While the results presented are mainly confirmational, these data simultaneously convey the need for additional Asian population-based registry and clinical trials to understand this currently incurable disease and improve patient outcomes.

Patients and methods
Data source. This retrospective observational study reviewed and analyzed data from the Taiwan Blood and Marrow Transplantation Registry (TBMTR). The TBMTR is maintained by the Taiwan Society of Blood and Marrow Transplantation (TSBMT), which has been tasked with registering clinical information of blood and bone marrow transplant recipients in Taiwan since 2009. Currently, 17 hospitals contribute to the registry, and the collection and analysis of data from the TBMTR is approved by the institutional review board of each participating hospital. All contributing centers that were registered with the TBMTR Data Center are listed in Supplementary data S1. All experimental protocols were approved by Institutional review board or Ethical com- Statistical analysis. We utilized the Mann-Whitney U test to compare the medians and distributions of continuous variables. Fisher's exact test or the χ 2 test was performed to examine the differences between discrete variables. Progression-free survival (PFS) was the duration from the date of reinfusion of autologous stem cells to the date of first documented disease progression, last follow-up, allo-SCT, or death from any cause, whichever occurred first. Overall survival (OS) was the duration from the date of reinfusion of autologous stem cells to the date of last follow-up, allo-SCT, or death from any cause, whichever occurred first. We plotted the survival curves with Kaplan-Meier analysis and calculated the statistical significance with the log-rank test. The Cox proportional hazards model was used in the univariate and multivariable regression analyses. P values < 0.05 were considered statistically significant. We considered biologically relevant factors (MIPI classification) and parameters with p < 0.1 in the univariate Cox regression analysis as covariates in the multivariable analysis. All statistical analyses and imaging were performed with IBM SPSS Statistics 23 for Windows and R software.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.