Introduction

Philadelphia chromosome-positive acute lymphoblastic leukemia (Ph+ALL) is a biologically and clinically distinct variant of ALL, and the prognosis of patients with Ph+ALL is poor when treated with chemotherapy alone [1, 2]. Prior to the era of tyrosine kinase inhibitors (TKIs), the probability of disease-free survival was around 10 % [3–5] and approximately only a quarter of the patients could manage to receive allogeneic stem cell transplantation (allo-SCT) [3]. Recently, the introduction of TKIs including imatinib and dasatinib has dramatically improved the prognosis of Ph+ALL patients [6–13]. However, the vast majority of these patients treated with a combination of imatinib and chemotherapy will ultimately relapse without allo-SCT [3, 14]. In the UKALLXII/ECOG2993 study, 4-year relapse-free survival was significantly superior in the 76 patients who underwent allo-SCT with myeloablative conditioning compared with the 38 patients who did not receive allo-SCT (69 vs 18 %) [3]. Thus, allo-SCT is still recommended for patients with Ph+ALL in first remission even in the era of TKIs.

Furthermore, the majority of patients treated with chemotherapy in combination with TKIs gain hematological complete remission (CR) and proceed to allo-SCT in a favorable condition wherein molecular remission has been maintained. In patients whose leukemia burdens can be decreased as much as possible prior to allo-SCT, graft-versus-leukemia (GVL) effects function more effectively, resulting in markedly decreased leukemia relapse rates after allo-SCT. In addition, since TKIs-combined chemotherapy provides a faster and deeper response, the patients can avoid further excessive cycles of chemotherapy and then successfully receive allo-SCT with fewer comorbid conditions including infectious complications, cardiac, hepatic, and renal dysfunctions. Collectively, these advantages including pre-transplant administration of TKIs and negative results of minimal residual disease (MRD) prior to allo-SCT have dramatically improved the prognosis of Ph+ALL patients allografted [6, 7, 15, 16]. Furthermore, in such patients who have achieved minimal leukemia burdens before allo-SCT, it is possible to reduce the dose intensity of a pre-transplant conditioning regimen, which can lead to a decrease in transplant-related mortality (TRM), ultimately resulting in an improved overall survival (OS) [17].

In this study, we evaluated the effects of pre-transplant conditioning intensity, either myeloablative conditioning (MAC) or reduced intensified conditioning (RIC) on the outcomes of Ph+ALL patients who underwent allo-SCT in CR after TKIs-containing chemotherapy, and we also examined the significance of MRD status before allo-SCT.

Materials and methods

Patients

We retrospectively analyzed outcomes of 46 Japanese patients with Ph+ALL who underwent allo-SCT at CR with either MAC (n = 31) or RIC (n = 15) at four institutions of the Fukuoka Blood and Marrow Transplantation Group between 2001 and 2012 (Table 1). The median follow-up time from transplantation was 1555 days in the MAC group and 584 days in the RIC group, respectively. Patients in the MAC group were significantly younger than those in the RIC group (median 39 vs 60 years old, p < 0.01). All the patients received TKIs (imatinib or dasatinib) in combination with chemotherapy prior to allo-SCT. In Japan, imatinib had been precedently approved for the treatment of Ph+ALL compared with dasatinib; therefore, in this study, majority of the enrolled patients were treated with imatinib: 25 out of 31 patients in the MAC group and 12 out of 15 patients in the RIC group received imatinib, respectively (p = 1.00). The patient characteristics are summarized in Table 1. All the patients provided informed consent, and this study was approved by the Institutional Review Board.

Table 1 Patient characteristics
Full size table

Conditioning regimen, stem cell source, and graft-versus-host disease prophylaxis

In this study, the MAC regimen included total body irradiation (TBI, 12 Gy) plus intravenous cyclophosphamide (CY, 120 mg/kg) or either TBI/CY plus intravenous etoposide or cytarabine (AraC) in 27 patients; intravenous busulfan (BU, 12.8 mg/kg) and cyclophosphamide (CY, 120 mg/kg) in 3 patients; and TBI and cytarabine (AraC; 120 mg/kg) in 1 patient. The RIC regimen included fludarabine (FLU 125–180 mg/m2) and BU (6.4 mg/kg) in 6 patients and fludarabine (FLU 125 mg/m2) and melphalan (MEL, 80 mg/m2) with or without TBI 2-4 Gy in 9 patients.

In the MAC group, 5 patients received G-CSF-mobilized peripheral stem cells, 22 patients received bone marrow cells, and 4 patients received cord blood stem cells. In the RIC group, 1 patient received G-CSF-mobilized peripheral stem cells, 9 patients received bone marrow cells, and 5 patients received cord blood stem cells. In the MAC group, all the 5 peripheral blood stem cell transplantation (PBSCT) and 2 bone marrow transplantation (BMT) recipients were transplanted from related donors, and the other 20 BMT recipients were from unrelated donors. In the RIC group, 1 PBSCT recipient was transplanted from related donor and all the 9 BMT recipients were from unrelated donors. A combination of a calcineurin inhibitor (cyclosporine or tacrolimus) and methotrexate was used for graft-versus-host disease (GVHD) prophylaxis except in five cord blood stem cell transplantation (CBT) recipients with RIC who received tacrolimus and mycophenolate mofetil for GVHD prophylaxis.

Minimal residual disease analysis

The MRD status was evaluated by polymerase chain reaction (PCR) analysis of the BCR-ABL fusion transcript. MRD was considered positive with any value above the threshold of the test sensitivity. MRD was evaluated within a 30-day period before transplantation.

Statistical analysis

We evaluated the probabilities of OS, leukemia-free survival (LFS), non-relapse mortality (NRM), and relapse incidence (RI) in this study. We analyzed OS and LFS using the log-rank test and RI and NRM using the Gray test. Chi-square testing was used for univariate comparison to examine categorical variables, and the Mann–Whitney U test was used to compare numerical variables. The p values <0.05 were considered to be statistically significant. All the statistical analyses were performed using EZR, a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [18].

Results

Engraftment and graft-versus-host disease

Engraftment was achieved in 26 of the 31 patients from the MAC group and 12 of 15 patients from the RIC group. The median time for neutrophil count recovery was 15.5 (range 11–23 days) and 16.5 (range 9–22 days) in the MAC and RIC group, respectively (p = 0.86). Engraftment was not evaluated in 4 patients (1 peripheral stem cell recipient, 2 bone marrow cell recipients, and 1 CBT recipient) in the MAC group because of severe infection (n = 3) and regimen-related toxicity (n = 1). Engraftment was also not evaluated in 1 CBT recipient in the RIC group because of regimen-related toxicity. Primary engraftment failure occurred in 1 CBT recipient in the MAC group and 2 CBT recipients in the RIC group. Grade ΙΙ to IV acute GVHD was documented in 14 of the 31 patients (45 %) in the MAC group and 6 of the 15 patients (30 %) of the RIC group (p = 0.82, Table 2). Eight out of the 21 patients (38 %) in the MAC group and 3 out of 14 (21 %) in the RIC group developed chronic GVHD, which was comparable (p = 0.44, Table 2).

Table 2 GVHD and cause of death
Full size table

OS, LFS, NRM and RI

Two-year OS rates were 56.5 ± 9.2 and 62.6 ± 13.5 % in the MAC and RIC groups, respectively (p = 0.53, Fig. 1a). Two-year LFS rates were similar between the MAC and RIC groups (50.0 ± 9.2 vs 64.6 ± 12.9 %, p = 0.29, Fig. 1b). Two-year NRM was also similar between the MAC and RIC groups (38.6 ± 6.5 vs 35.4 ± 7.0 %, p = 0.62, Fig. 1c). Five of the 31 patients relapsed in the MAC group, while 1 of the 15 patients in the RIC group relapsed at 2 years, giving an RI of 18.5 ± 1.7 % in the MAC group and 7.7 ± 0.6 % in the RIC group (p = 0.28, Fig. 1d). Two patients received TKIs (dasatinib) for the treatment of molecular relapse after allo-SCT [19].

Fig. 1
figure 1

Transplantation outcomes (2 years) according to the conditioning intensity: myloablative conditioning (MAC) vs reduced-intensity conditioning (RIC). a OS, 57 vs 63 % (p = 0.53), b LFS, 50 vs 65 % (p = 0.29), c NRM, 39 vs 35 % (p = 0.62), d RI, 19 vs 8 % (p = 0.28)

Full size image

In total, 16 of the 31 patients (52 %) died in the MAC group: 2 of disease progression, 7 of infectious complications, 1 of GVHD, 2 of idiopathic intestinal pneumonia, 1 of acute respiratory distress syndrome, 1 of liver dysfunction, 1 of pleural hemorrhage and 1 of EB virus-associated lympho-proliferative disease. Pathogen of infections included Enterococcus faecium, Klebsiella oxytoca, Pseudomonas aeruginosa, Staphylococcus haemolyticus, Aspergillus and RS virus. There is one case of sepsis lacking the information of microorganism in the MAC group. In contrast, 6 of the 15 patients (20 %) died in the RIC group: 1 of disease progression, 2 of infection, 1 of idiopathic intestinal pneumonia, 1 of heart failure and 1 of multiple organ failure (Table 2). Infection included sepsis of Enterobactor cloaca and multiple pathogen-induced bacterial pneumonia. There was no significant difference in rate and cause of NRM among the RIC and MAC groups (p = 0.27, p = 0.53).

MRD

We also evaluated the influence of pretransplant qualitative MRD status on OS and LFS rates in 44 patients (29 in the MAC group and 15 in the RIC group) (Fig. 2). 19 of the 29 patients (66 %) and 12 of the 15 patients (80 %) achieved molecular remission in the MAC and RIC groups, respectively (p = 0.45, Table 1). In total, 31 of the 44 patients were MRD negative before allo-SCT. There is no significant difference in relapse rates according to pretransplant MRD status (p = 0.50). 2-year OS and 2-year LFS rates were similar between the MRD-negative and MRD-positive patients (p = 0.28 and 0.48, respectively, Fig. 2a, b). The same trend was observed in the MAC group (p = 0.16 and 0.29, respectively, Fig. 2c, d) as well as in the RIC group (p = 0.94 and 0.99, respectively, Fig. 2e, f).

Fig. 2
figure 2

Transplantation outcomes (2 years) according to the MRD status prior to allo-SCT: MRD positive vs MRD negative. a OS in all patients, 54 vs 77 % (p = 0.28), b LFS in all patients, 51 vs 69 % (p = 0.48), c OS in MAC group, 47 vs 80 % (p = 0.16), d LFS in MAC group, 42 vs 70 % (p = 0.29), e OS in RIC group, 63 vs 67 % (p = 0.94), f LFS in RIC group: 65 vs 67 % (p = 0.99)

Full size image

Discussion

Several groups retrospectively analyzed for factors affecting outcomes after allo-SCT in the patients with Ph+ALL who received RIC or MAC pre-transplant regimen and the previous reports are summarized in Table 3 [7, 16, 20, 21]. Bachanova et al. have reported that in the RIC group (n = 67), pre-transplant TKIs were administered to 76 % of the patients, with 39 % MRD-negative prior to sllo-SCT. In contrast, 78 % received pre-transplant TKIs in the MAC group (n = 130) and 49 % were MRD negative, which were statistically equal to those of the RIC group (p = 0.71 and p = 0.79). In this study, the 1-year TRM was significantly superior in the RIC group compared with the MAC group (13 vs 36 %, p < 0.001), whereas the 3-year relapse rate was higher in the RIC group than in the MAC group (49 vs 28 %, p = 0.058), resulting in the counterbalanced similar outcomes in terms of 3-year OS among the two groups (39 vs 35 %, p = 0.62). Multivariate analyses also demonstrated that pre-transplant TKI had a significant impact on the relapse rate [HR = 1.88 (1.11–3.17), p = 0.018], while the MRD status was not statistically associated with the relapse rate [HR = 1.60 (0.96–2.67), p = 0.13] [20]. Brissot et al. have retrospectively analyzed the results from 473 Ph+ALL patients allografted. The multivariate analyses revealed that pre-transplant TKIs had a significant impact on relapse-free [HR = 0.56 (0.36–0.87), p = 0.01], and the RIC conditioning was closely associated with RI [HR = 1.69 (1.08–2.65), p = 0.02]. This study also found that there was no significant difference in OS, LFS and RI between MRD-positive and MRD-negative allo-SCT recipients [7].

Table 3 Comparison of pre-transplant conditioning intensity of allogeneic stem cell transplantation for Ph+ALL
Full size table

In the present study, we retrospectively analyzed transplant outcomes of 46 patients with Ph+ALL to clarify the impact of conditioning intensity on the outcome after allo-SCT. All patients had been treated with a TKIs-combined chemotherapy and then allografted during hematological CR with either RIC or MAC. In our study, there was no significant difference in 2-year OS (62 vs 56 %, p = 0.53), 2-year LFS (65 vs 50 %, p = 0.29), and NRM (35 vs 39 %, p = 0.62) between the RIC and MAC groups. We also found no significant impact of the MRD status prior to allo-SCT on 2-year OS (77 vs 54 %, p = 0.28), and 2-year LFS (69 vs 51 %, p = 0.38) among MRD-negative and MRD-positive groups. NRM in our study was relatively higher than that from the previous reports; however, there was no significant difference in rate and cause of NRM among the RIC and MAC groups. Despite the small number of cases retrospectively studied, our results suggest that if the patients have been treated by TKIs-combined chemotherapy and maintained hematological CR prior to allo-SCT, even though their MRD was detected by sensitive PCR, the RIC regimen could still provide a favorable outcome for such patients, equivalent to the MAC group.

In general, the MRD status prior to allo-SCT is regarded as one of the most important indicators to determine whether MAC or RIC conditioning would be recommended for each patient. A combination of TKIs and chemotherapy can induce both a marked and fast response, and currently, a 0.01 % threshold has been widely accepted to define an MRD-positive status [22, 23]. However, our results and previous reports revealed that the MRD status prior to allo-SCT was not predictive for the transplantation outcome in the era of TKIs, although other reports in the pre-TKIs era suggested that MRD was closely correlated with the treatment outcome [24, 25]. It is possible that this MRD threshold prior to allo-SCT might not be useful for completely discriminating between a warning sign of impending relapse and the process of MRD disappearance. Ph+ALL cells possess aggressive proliferative potential, thus if few Ph+ leukemia cells survive prior to allo-SCT, these cells can rapidly grow at a short interval before conditioning regimen. Therefore, it would be important to quantitatively evaluate MRD level not only at one point prior to allo-SCT but consecutive several points from induction therapy to allo-SCT. In addition, recent interests have focused on the post-transplant administration of TKIs for the prevention of relapse or eradicating persistent MRD [19, 26–31]. Therefore, a serial quantitative monitoring of MRD would be useful for predicting relapse or cure in the patients with Ph+ALL.

We have shown that the intensity of pre-transplant conditioning did not affect the outcomes of Ph+ALL patients who received allo-SCT during hematological CR after TKIs-combined chemotherapy. Prospective randomized trials comparing an MAC versus RIC regimen for allo-SCT might be useful for examining the effects of intensity of conditioning on the clinical outcomes of patients with Ph+ALL in the era of TKIs.