Safety and Efficacy of Thiotepa-Based Conditioning for Allogeneic Transplantation in AML: a Survey from the ALWP of the EBMT

Home , Thiotepa

ORIGINAL ARTICLE Safety and efﬁcacy of thiotepa-based conditioning for allogeneic transplantation in AML: a survey from the ALWP of the EBMT

S Eder1,2, M Labopin1, J Finke3, D Bunjes4, A Olivieri5, S Santarone6, A Rambaldi7, L Kanz8, G Messina9, M Mohty1,2,11 and A Nagler1,10,11 on behalf of the ALWP of the EBMT

This study evaluated the safety and efﬁcacy of thiotepa-based regimens before allogeneic stem cell transplantation in 310 adult patients with AML. Disease status at the time of transplantation was CR1 in 50%, CR2+ in 23.5% and advanced disease in 26.5%. Transplantation was performed from haploidentical (35%), matched sibling (27%), unrelated (20%) or cord blood (18%) donors. As for safety: mucositis occurred in 46.8% of the patients and the cumulative incidence (CI) of sinusoidal obstruction syndrome was 4.0%. With a median follow-up of 37 months, the CI of acute GvHD grade4II was 26.5%, whereas CI of chronic GvHD was 28.1% at 3 years. CI for non-relapse mortality at 3 years was 38.4%, 49.7% and 45.4% for patients in CR1, CR2+ and advanced disease, respectively (P = 0.10). Relapse incidence at 3 years was 20.2, 30.7 and 40.6% in these three respective groups (P = 0.002). CI for 3-year leukemia-free survival and overall survival were 41.4% and 45.6% (CR1), 19.6% and 27.7% (CR2+), and 13.9% and 13.6% (advanced disease), respectively (Po10À4 for both). Our data suggest that thiotepa-based conditioning therapy in AML is feasible, effective and safe, as investigated for sinusoidal obstruction syndrome and mucositis.

Bone Marrow Transplantation (2017) 52, 238–244; doi:10.1038/bmt.2016.239; published online 19 September 2016

INTRODUCTION lymphoma, replacing BCNU and thus reducing lung toxicity Allogeneic hematopoietic stem cell transplantation (HSCT) is a (TECAM–thiotepa, etoposide, cytoxan, ARA-C and melphalan) with 10 well-established curative therapy for hematological diseases, very good results. including AML.1 Subsequently, thiotepa was incorporated into pre-allogeneic TBI has traditionally been an integral part of pre-transplantation transplantation conditioning regimens using HLA-matched conditioning. However, it is associated with substantial toxicity donors, both for malignant and non-malignant indications. In 11 and undesirable long-term side effects.2 the 1990s, Bacigalupo et al. reported on the use of thiotepa in Thiotepa (N,N'N'-triethylenethiophosphoramide) is an alkylating patients with advanced leukemia using myeloablative condition- compound with an antineoplastic activity that has been used in ing. Thiotepa was administered at a dose of 15 mg/kg in oncology (for example, neuroblastoma, breast-, ovarian- and combination with cyclophosphamide. The transplant-related bladder cancer) for decades.3,4 Besides its antineoplastic mortality was 29% and the 2-year overall survival (OS) was 57%. activity,5,6 thiotepa has immunosuppressive properties7 and the Rosales and colleagues12 used thiotepa (5 mg/kg × 2 days), in ability to penetrate the blood-brain barrier,8 in conjunction with a addition to the standard busulfan/cyclophosphamide regimen good toxicity profile. Thiotepa is metabolized in the liver by with comparable results and moderate toxicity rates. cytochrome P450 into its metabolite, tepa. Two percent of the Later on, thiotepa was included in reduced-intensity drug is excreted unchanged in the urine. There is a direct conditioning (RIC) regimens, aiming in intensifying the anti correlation between thiotepa/tepa area under the curve and leukemic effect reducing the relapse rates that were toxicity, such as mucositis, hepatitis and brain toxicity.9 higherpostRICincomparisonwiththemyeloablative – On the basis of these favorable charactristics thiotepa has been regimens.13 15 Indeed, Bacigalupo et al.,16 were able to show subsequently incorporated into chemotherapeutic protocols used an impressive outcome of thiotepa-based RIC regimens with for hematological malignancies and HSCT. However, there are few 10-year follow-up. studies focusing on the analysis of the benefit of thiotepa in the In recent years, thiotepa has become an integral part of the TBF pre-allogeneic HSCT conditioning in a specific disease category. (thiotepa, busulfan and fludarabine) protocol, which is being used Early studies reported on the use of thiotepa in the conditioning with increasing frequency for alternative donor transplantations, regimen before autologous HSCT, mainly in patients with including haploidentical and cord blood transplants.17–20

1EBMT Office Paris, Hôpital Saint-Antoine, Paris, France; 2Service d’Hématologie Clinique et de Thérapie Cellulaire, Hôpital Saint-Antoine, Paris, France; 3Department of Medicine— Hematology, Oncology, University of Freiburg, Freiburg, Germany; 4Klinik fuer Innere Medizin III, Universitätsklinikum Ulm, Ulm, Germany; 5Department of Hematology, Azienda Ospedali Riuniti di Ancona, Ancona-Torrete, Italy; 6Department of Hematology, Ospedale Civile, Pescara, Italy; 7USC Ematologia, Azienda Ospedaliera Papa Giovanni XXIII, Bergamo, Italy; 8Abteilung II, Universität Tübingen, Tübingen, Germany; 9Centro Trapianti Midollo Osseo, Azienda Ospedaliera ‘BMM’, Reggio Calabria, Italy and 10Chaim Sheba Medical Center, Tel-Hashomer, Israel. Correspondence: Dr S Eder, Service d’Hématologie Clinique et de Thérapie Cellulaire, Hôpital Saint-Antoine, 184, rue du Faubourg Saint-Antoine, Paris 75012, France. E-mail: [email protected] 11The two last authors share senior authorship. Received 20 April 2016; revised 22 July 2016; accepted 27 July 2016; published online 19 September 2016 Safety and efficacy of thiotepa for allo-HSCT in AML S Eder et al 239 The aim of the current analysis was to investigate the potential Disease status at HSCT was CR1 in 50% and CR2+ in 23.5%, impact of thiotepa as part of the conditioning therapy for although 26.5% of the patients had an advanced disease at the AML patients, undergoing allogeneic transplantation from time of transplantation. HLA-matched or alternative donors, with respect to efficacy and Thiotepa was mostly combined with fludarabine (43%), safety. fludarabine+busulfan (32%) or cyclophosphamide (25%; Table 1). The median dose of thiotepa was 10 mg/kg (range, 4–20). Anti-GvHD prophylaxis included cyclosporine A alone or in PATIENTS AND METHODS combination with methotrexate and/or mycophenolate mofetil, Study design and data collection in vitro T-cell depletion or post-transplantation cyclophosphamide This is a retrospective multicenter analysis based on the registry data of the (Table 2). Acute Leukemia Working Party (ALWP) of the European Society for Blood and Marrow Transplantation (EBMT) registry. The study was designed and Transplantation outcome 4 approved by the ALWP. The EBMT is a voluntary working group of 600 Engraftment. The median follow-up time was 37 months (range, transplant centers that are required to report all consecutive HSCT 6–169). CI of neutrophil engraftment (defined as ⩾ 0.5 × 109/L) procedures and their follow-up once a year. Audits are routinely performed was 88.7% (95% confidence interval (CI): 84.5–81.8) at day +30 to determine the accuracy of the data. Since 1990, patients provide post HSCT. Median time to engraftment was 16 days (range, 8–38). informed consent authorizing the use of their personal information for fi ⩾ 9 research purposes. CI of platelet engraftment (de ned as 20 × 10 /L) was 86.2% Eligible patients for this analysis were adults (age 418 years) with (95% CI: 81.6-89.7) at day +60 post HSCT. de novo or secondary AML who underwent a first allogeneic (optionally after autologous) HSCT with a thiotepa-based regimen between 1992 Organ toxicity and GvHD. Mucositis (all grades) occurred in 46.8% and 2012 (median year of HSCT 2008). Any donors were included of patients with a median day of onset at day 3 after HSCT. (HLA-matched siblings, unrelated or haploidentical donors and cord blood The CI of SOS was low: 4.0% (95% CI: 2.2–6.6). The median day transplantation). of onset was day 8 post HSCT. It was severe in only in one patient, Variables collected included recipient and donor characteristics in whom it resulted in multiorgan failure and death. (age, gender, CMV and serostatus), disease status at transplant, The incidence of acute GvHD (gradeXII) was 26.5% (95% CI: transplant-related factors, including conditioning regimen, immunosup- 21.6–31.6) at day 100 post HSCT, whereas CI of chronic GvHD was pression (in vivo T-cell depletion vs none), stem cell source, prophylaxis of 28.1% (95% CI: 23.1–33.4) at 3 years (56% of the patients had a GvHD and outcome variables (acute and chronic GvHD, relapse, limited and 44% an extensive disease). non-relapse mortality (NRM), leukemia-free survival (LFS), OS, toxicity (sinusoidal obstruction syndrome (SOS) and mucositis) and causes NRM and cause of death. Overall, the NRM was 19.0% (95% CI: of death. 14.9–23.6) and 43.0% (95% CI: 37.1–48.7) at day 100 and 3 years post HSCT, respectively. Statistical analysis The 3-year NRM was 38.4% (95% CI: 27.4–49.2), 49.7% (95%CI: The primary end points of the study were LFS and OS. Secondary end 38.9–59.6) and 45.4% (95% CI: 34.4–55.7) for patients in CR1, CR2+ points included: disease relapse incidence (RI), NRM, engraftment, and advanced disease, respectively (P = 0.10; Figure 1a). incidence, and severity of acute and chronic GvHD. The starting point Three-year NRM was 32.3% (95% CI: 23.7–41.2), 29.2% for time-to-event analysis was ‘date of transplantation’. OS was defined as (95% CI: 20.9–38.1), 55.7% (95% CI: 47.4–63.1) and 49.3% the time to death from any cause. Surviving patients were censored at the (95% CI: 40.6–57.4) for matched sibling, unrelated, haploidentical fi time of last follow-up. LFS was de ned as survival without relapse or and cord blood transplantation, respectively (P = 0.01; Table 3). fi progression. RI was de ned as time to onset of leukemia recurrence. NRM A total of 204 of 310 patients (66%) died. The main cause of was defined as death without relapse/progression. Conditioning intensity, – death was infection (35%), original disease (31%), GvHD (10%), engraftment, GvHD and SOS were defined, and classified as described.21 25 The probabilities for OS and LFS were estimated from the time of HSCT interstitial pneumonia (6%), SOS (2%) and other transplant-related using the Kaplan–Meier product-limit estimate and were compared by the causes (15%). Two patients died due to a secondary malignancy log-rank test in univariate analysis. Estimates of engraftment, SOS, acute (1%; Table 4). and chronic GvHD, NRM and RI were calculated using cumulative incidence (CI) rates to accommodate competing risks, and were compared by Gray’s Relapse. The RI at 3 years was 28.1% (95% CI: 23.1–33.3). It was test. Death was the competing event for engraftment, SOS, GvHD and 20.2% (95% CI: 14.1–27.1) in patients transplanted in CR1, 30.7% relapse, relapse was the competing event for NRM. (95% CI: 20.3–41.7) in those transplanted in CR2+ (NS) and 40.6% Univariate comparisons were done using the log-rank test for OS, LFS (95% CI: 29.8–51.2) in patients transplanted in advanced disease and the Gray’s test for RI, NRM and GvHD CIs. Multivariate analyses were (P = 0.002; Figure 1b). performed using Cox proportional hazards model. The type I error rate was fixed at 0.05 for the determination of factors associated with time-to-event Leukemia-free and overall survival. Three-year LFS and OS outcomes. The P-value represents the difference between groups, using correlated with disease status at HSCT: 41.4% (95% CI: 33.1–49.8) either the log-rank or the Gray test. Statistical analyses were performed and 45.6% (95% CI: 37.1–54.1) for patients transplanted in CR1, with SPSS 22.0 (IBM Corp., Armonk, NY, USA) and R 3.1.1 software packages 19.6% (95% CI: 10.2–28.9) and 27.7% (95% CI: 17.2–38.2) for (R Development Core Team, Vienna, Austria). patients transplanted in CR2+, and 13.9% (95% CI: 6–21.9) and 13.6% (95% CI: 5.5–21.8) for patients transplanted in advanced RESULTS disease, respectively (Po10–4; Figure 1c and d) Patients and disease characteristics Outcome according to time distribution. The median year of A total of 310 patients with AML (13% secondary) were identified. transplant was 2008, whereas 127 patients were transplanted from Patient and transplant characteristics are shown in Table 1. 1992 to 2007 and 183 patients from 2008 to 2012. We could see The median age was 46.5 (range, 18–72) years; 54% were males an improvement for NRM, LFS and OS for HSCT performed in the and 46% females. Fourteen percent of the patients had a previous more recent years (Table 3). We then did an analysis according to autologous HSCT. The median interval from diagnosis to HSCT was the donor type for transplants between 2008 and 2012: no 222 days. Male donor gender comprised 56% of patients, and 24% differences could be observed for NRM and LFS, whereas there male recipients received grafts from a female donor. was a significantly difference when looking to whole period. On

Table 1. Patents' and transplant characteristics

CR1 CR2+ Advanced disease Global

155 pts, 50% 73 pts, 23.5% 82 pts, 26.5% 310 pts, 100%

Median age of pts (range) 47 (18–69) 42 (18–72) 48 (19–70) 46.5 Median year of HSCT (range) 2009 (1996–2012) 2008 (1992–2012) 2007 (1996–2012) 2008 1992–1999 9% 12% 15% 11% 2000–2004 13% 12% 32% 18% 2005–2009 32% 45% 22% 32% 2010–2012 47% 30% 32% 39% Interval diagnosis—HSCT, in days 171 (82–822) 699 (168–2678) 280 (58–1501) 222 Patient sex Male 51.0% 58.9% 53.7% 53.5% Female 49.0% 41.1% 46.3% 46.5% Donor sex Male 53.5% 60.9% 55.6% 55.7% Female 46.5% 39.1% 44.4% 44.3% Female donor to male recipient 25.2% 21.7% 23.5% 23.9% Donor MSD 31.6% 21.9% 20.7% 26.5% UD 16.1% 20.5% 26.8% 20.0% haplo 31.0% 37.0% 41.5% 35.2% CB 21.3% 20.5% 11.0% 18.4% Cytogenetics good 5.8% 19.2% 11.0% 10.3% intermediate 51.0% 58.9% 50.0% 52.6% poor 14.8% 6.8% 18.3% 13.9% NA/failed 6.5% 9.6% 6.1% 7.1% sec. AML 15.5% 6.8% 12.2% 12.6% FAB classiﬁcation M0 9.4% 11.8% 8.7% 9% M1 17.3% 19.1% 20.3% 16% M2 26.8% 23.5% 30.4% 24% M3 0.0% 2.9% 1.4% 20% M4 28.3% 23.5% 17.4% 15% M5 12.6% 19.1% 15.9% 4% M6 4.7% 0.0% 5.8% 1% M7 0.8% 0.0% 0.0% 13% Patient CMV positive 78.1% 74.6% 78.9% 77.5% Donor CMV positiv 53.3% 57.8% 55.1% 54.9% secondary AML 15.5% 6.8% 12.2% 12.6% previous autologous HSCT 1.9% 30.1% 22.0% 13.9% MAC conditioning 71.6% 72.6% 65.9% 70.3% TBI 33.5% 19.2% 29.3% 29.0% in-vivo T-cell depletion 70.8% 66.7% 69.2% 69.4% Conditioning, Cy + Thiotepa MAC: Cy + Thiotepa 20 pts most frequent combination 25.4% MAC: Cy + Thiotepa + TBI 29 pts RIC: Cy + Thiotepa 22 pts Flu + Thiotepa MAC: Flu + Thiotepa 32 pts 43.0% MAC: Flu + Thiotepa + TBI 50 pts RIC: Flu + Thiotepa 38 pts Flu + Bu + Thiotepa MAC: Flu + Bu + Thiotepa 69 pts 31.5% RIC: Flu + Bu + Thiotepa 19 pts Abbreviations: Bu = busulfan; CB = cord blood; Cy = cyclophosphamide; Flu = ﬂudarabine; haplo = haploidentical; HSCT = hematopoietic stem cell transplantation; MAC = myeloablative conditioning; MSD = matched sibling donor; NA = not applicable; pts = patients; RIC = reduced-intensity conditioning; sec. AML = secondary AML; UD = unrelated donor.

the contrary, no differences in LFS and OS could be observed over allogeneic HSCT for AML is feasible and effective, with the main the time (Table 3). outcomes being comparable to those achieved with other regimens. Multivariate analysis. A multivariate analysis adjusted for age, Notably, the incidence of side effects was relatively low: SOS disease status, gender matching, donor type, secondary AML, was observed in only 4.0%, leading to death in only one patient. intensity of conditioning and cytogenetics was subsequently This incidence was lower than previously reported with for other performed. Age was found as a signiﬁcant risk factor for NRM, LFS conditioning regimens (mean incidence of 13.7%).26,27 Moreover, and OS. Disease status had a statistically signiﬁcant impact on the 4.0% incidence is somewhat similar to the 7.8% incidence NRM, RI, LFS and OS. NRM was elevated for haploidentical and recently reported by Nagler et al.28 in an EBMT study analyzing cord blood transplantations. Finally, haploidentical transplantation HSCT following IV-busulfan-based conditioning. had a higher RI (Table 5). Likewise, we observed a rather low incidence of mucositis (46.8%) in comparison with the previous studies, which reported an incidence of up to 99% (67.4% experienced grade III or IV).29 DISCUSSION Being a retrospective study, we have no information on the This large registry survey, conducted by the ALWP of the EBMT, relationship between the area under the curve and thiotepa suggests that thiotepa-based conditioning therapy before and/or tepa, which is known to induce elevated transaminases

Table 2. GvHD prevention by donor type and by status

MSD UD Haplo CB CR1 CR2+ Active disease Total

CsA 5 17 0 17 16 8 15 39 6.4% 28.3% 0.00% 30.9% 10.6% 11.1% 19.2% 13.00% CsA+MTX 41 23 2 2 30 20 18 68 52.6% 38.3% 1.9% 3.6% 19.9% 27.8% 23.1% 22.6% CsA+MMF 4 6 0 26 23 9 4 36 5.1% 10.00% 0.00% 47.3% 15.2% 12.5% 5.1% 12.00% CsA+MTX+MMF 0 0 15 1 12 2 2 16 0.00% 0.00% 13.9% 1.8% 7.9% 2.8% 2.6% 5.3% PT-Cy 1 0 17 1 12 5 2 19 1.3% 0.00% 15.7% 1.8% 7.9% 6.9% 2.6% 6.3% In vitro TCD 19 13 67 3 49 24 29 102 24.4% 21.7% 62.00% 5.5% 32.5% 33.3% 37.2% 33.9% Other 8 1 7 5 9 4 8 21 10.3% 1.7% 6.5% 9.1% 6.00% 5.6% 10.3% 7.00% Total 78 60 108 55 151 72 78 301 Abbreviations: CB = cord blood; CsA = cyclosporine A; haplo = haploidentical; MMF = mycophenolate mofetil; MSD = matched sibling donor; MTX = methotrexate; PT-Cy = post-transplantation cyclophosphamide; TCD = T-cell depletion; UD = unrelated donor.

a NRM b RI

1.0 CR1 1.0 CR1 CR2+ CR2+ 0.8 Active disease 0.8 Active disease

0.6 0.6

0.4 0.4

0.2 0.2

Cumulative incidence of NRM 0.0 0.0 Cumulative incidence of relapse 012345 012345 Time from transplant (years) Time from transplant (years)

c LFS d OS

1.0 CR1 1.0 CR1 CR2+ CR2+ 0.8 Active disease 0.8 Active disease

0.6 0.6

0.4 0.4 Overall survival 0.2 0.2 Leukemia free survival

0.0 0.0 0 1 2 3 4 5 0 1 2 3 4 5 Time from transplant (years) Time from transplant (years) Figure 1. Outcome according to remission status. (a) Cumulative incidence for NRM, non-relapse mortality. (b) Cumulative incidence for RI, relapse incidence. (c) LFS, leukemia-free survival. (d) OS, overall survival. and mucositis. In our study, cyclophosphamide was combined leading cause of death in haploidentical transplantations with a with thiotepa in 25% of our patients; this agent shows auto- thiotepa-based conditioning.31,32 Recurrence of leukemia was the induction and is inhibited by thiotepa which complicates a second most frequent cause of death, which is in line with the fact pharmacokinetic guided dosing schedule.30 that 50% of our patients were not in CR1 and ~ 25% had an Of note, infections were the leading cause of death in our study, advanced disease status at time of transplantation. being accounted for 34.7% of all deaths. The incidence was as Our survey indicated that thiotepa is used with high prevalence high as 51.4% for patients transplanted with cord blood graft and in alternative donor transplantations. In this regard, the TBF 37.7% for those receiving a haploidentical graft. Severe immuno- protocol originated for cord blood transplantation (thiotepa suppression and slow immune reconstitution were most probably 10 mg/kg, busulfan 9.6 mg/kg, ﬂudarabine 150 mg/m2 and the main contributing factors (in haploidentical transplants the rabbit antithymocyte globulin 8 mg/kg),19 is being increasingly percentage of in vitro T-cell depletion was 63%). Our data are in used in Europe, both for cord blood and for haploidentical line with previous publications that reported infections as the transplantation.

Table 3. Three-year outcome according to donor (a), time distribution (b) and donor for transplantation after 2008 (c)

(a) ﬁ ayo hoeafral-STi AML in allo-HSCT for thiotepa of cacy

– MSD 82 pts, 26.5% UD 62 pts, 20% Haplo 109 pts, 35.2% CB 57 pts, 18.4% Global 310 pts, 100% P-value 4 07McilnPbihr iie,pr fSrne Nature. Springer of part Limited, Publishers Macmillan 2017 © 244 NRM 32.3% (95% CI: 23.7–41.2) 29.2% (95% CI: 20.9–38.1) 55.7% (95% CI: 47.4–63.1) 49.3% (95% CI: 40.6–57.4) 43.0% (95% CI: 37.1–48.7) 0.006 RI 34.3% (95% CI: 24.1–44.7) 33.7% (95% CI: 21.8–46.1) 25.4% (95% CI: 17.4–34.1) 18.2% (95% CI: 9.2–29.5) 28.1% (95% CI: 23.1–33.3) 0.010 LFS 33.4% (95% CI: 23.1–43.8) 37.1% (95% CI: 24.6–49.5) 19.0% (95% CI: 10.6–27.3) 32.6% (95% CI: 19.2–45.9) 28.9% (95% CI: 23.5–34.2) 0.301 OS 38.0% (95% CI: 27.3–48.6) 45.6% (95% CI: 33–58.3) 22.4% (95% CI: 13.6–31.1) 34.3% (95% CI: 20.9–47.7) 32.9% (95% CI: 27.3–38.5) 0.115 Chronic GvHD 42.7% (95%CI: 31.5–53.4) 22.8% (95% CI: 13.2–34.1) 18.9% (95% CI: 11.8–27.3] 29.5 % (95% CI: 17.8–42.1) 28.1% (95% CI: 23.1–33.4) 0.006 Eder S (b) tal et 1992–2007 2008–2012 P-value

NRM 52.8% (95% CI: 43.7–61) 34.8% (95% CI: 23–46.9) 0.003 RI 27.6% (95% CI: 20.1–35.5) 28.8% (95% CI: 22.1–35.7) 0.839 LFS 19.7% (95% CI: 12.8–26.6) 36.4% (95% CI: 28.8–44) 0.004 OS 23.6% (95% CI: 16.2–31) 40.7% (95% CI: 32.8–48.5) 0.005 Chronic GvHD 26.2% (95%CI: 18.8–34.2) 28.6% (95% CI: 22.1–35.6) 0.561

(c)

MSD UD Haplo CB P-value

NRM 23.6% (95% CI: 10–40.5) 24.7% (95% CI: 12.7–38.9) 45% (95% CI: 36.5–53) 38.8% (95% CI: 27.1–50.3) 0.090 RI 37.5% (95% CI: 21.5–53.5) 33.6% (95% CI: 19.3–48.6) 26.4% (95% CI: 15.2–39) 19.9% (95% CI: 9.1–33.8) 0.284 LFS 38.9% (95% CI: 22.4–55.4) 41.6% (95% CI: 26.3–57) 28.6% (95% CI: 15.6–41.6) 41.3% (95% CI: 24.8–57.7) 0.541 OS 37.9% (95% CI: 21.3–54.5) 54% (95% CI: 39–69) 33.3% (95% CI: 19.7–46.9) 43.9% (95% CI: 27.5–60.3) 0.458 Chronic GvHD 51.3% (95%CI: 32.2–67.3) 29.3% (95% CI: 16.6–43.1) 16.5% (95% CI: 7.7–28.1) 28 % (95% CI: 14.8–42.8) 0.005 Abbreviations: CB = cord blood; CI = conﬁdence interval; haplo = haploidentical; LFS = leukemia-free survival; MSD = matched sibling donor; NRM = non-relapse mortality; OS = overall survival; pts = patients; RI = relapse incidence; UD = unrelated donor. Safety and efﬁcacy of thiotepa for allo-HSCT in AML S Eder et al 243

Table 4. Cause of death (percentage of cause of death for patients who died)

CR1 (%) CR2+ (%) Advanced disease (%) MSD (%) UD (%) Haplo (%) CB (%) Global (%)

Infection 32.1 39.2 34.3 17.6 36.4 37.7 51.4 34.7 Original disease 26.9 27.5 38.8 47.1 36.4 23.4 20.0 31.1 Other HSCT related 12.8 15.7 6.0 15.7 0.0 14.3 8.6 11.2 GvHD 12.8 5.9 9.0 11.8 12.1 5.2 14.3 9.7 IP 7.7 7.8 3.0 2.0 3.0 13.0 0.0 6.1 SOS 2.6 2.0 3.0 0.0 3.0 2.6 5.7 2.6 Haemorhage 1.3 2.0 3.0 3.9 6.1 0.0 0.0 2.0 Failure/rejection 1.3 0.0 1.5 0.0 0.0 2.6 0.0 1.0 Second malignancy 2.6 0.0 0.0 0.0 3.0 1.3 0.0 1.0 Cardiac toxicity 0.0 0.0 1.5 2.0 0.0 0.0 0.0 0.5 Abbreviations: CB = cord blood; haplo = haploidentical; HSCT = hematopoietic stem cell transplantation; MSD = matched sibling donor; SOS = sinusoidal obstruction syndrome; UD = unrelated donor.

Table 5. Multivariate analysis Notably, an EBMT/Eurocord study, comparing single versus double cord blood transplantations administering different HR 95% CI P-value conditioning regimens, showed that, at least with single umbilical NRM cord blood transplantation, a conditioning regimen consisting of Age (by deacade) 1.18 1.02–1.38 0.03 TBF might replace the classical myeloablative regimens with CR1 (ref) 1 adequate cell dose (42.5 × 107/kg).20 CR2/3 1.74 1.08–2.79 0.02 Active disease 1.88 1.19–2.95 0.006 Our 3-year NRM of 43% needs to be seen in the context of the Female donor to male recipient 0.94 0.60–1.48 0.8 transplant characteristics of our cohort, with 450% being MSD (ref) 1 transplanted from alternative donors. Indeed, the 3-year NRM in UD 0.95 0.50–1.80 0.88 Haplo 1.98 1.20–3.25 0.007 the HLA-matched group was substantially lower. This figure is CB 1.86 1.03–3.36 0.04 higher than the one reported, for a better group of patients, Sec. AML 0.78 0.42–1.47 0.45 RIC conditioning 0.73 0.47–1.12 0.15 in our recently published study, where we compared thiotepa- Poor cytogenetics 1.37 0.85–2.18 0.19 based conditioning to TBI-based conditioning for HSCT in adults with AML. In that study, we observed a 2-year RI Age (by deacade) 1.07 0.90–1.29 0.44 NRM of 23.9% for adults in CR1 with a sibling or unrelated CR1 (ref) 1 matched donor, receiving thiotepa mediated myeloablative CR2/3 1.84 1.02–3.30 0.04 conditioning.33 Active disease 3.8 2.26–6.40 o0.0001 Female donor to male recipient 0.95 0.57–1.60 0.86 Finally, we showed a rather low incidence of chronic MSD (ref) 1 GvHD, notably in CR2+ 22.6% (95% CI: 13.6–33.1) and advanced UD 0.69 0.38–1.25 0.22 – Haplo 0.51 0.28–0.91 0.02 disease status 22.4% (95% CI: 13.9 32.1) with a global 3-year CB 0.61 0.29–1.29 0.19 CI of 28.1% (95% CI: 23.1–33.4). This was rather low compared Sec. AML 1.01 0.53–1.93 0.97 with the previous studies: Nagler et al.34 could show an RIC conditioning 1.50 0.93–2.42 0.09 Poor cytogenetics 1.07 0.58–1.97 0.82 occurrence of chronic GvHD after 2 years of 38% after a busulfan/cyclophosphamide-based regimen and 50% after a LFS cyclophosphamide/TBI-based regimen for patients in CR1 Age (by deacade) 1.14 1.01–1.28 0.03 CR1 (ref) 1 and CR2. CR2/3 1.78 1.24–2.57 0.002 Our study suffers from several limitations: being a registry and Active disease 2.51 1.79–3.52 o0.0001 retrospective study, our cohort is rather heterogeneous. We have Female donor to male recipient 0.96 0.69–1.34 0.81 MSD (ref) 1 no detailed information on the grade of mucositis, the criteria UD 0.84 0.54–1.29 0.42 used for SOS diagnosis and its grading, as well as laboratory tests Haplo 1.13 0.79–1.64 0.5 CB 1.18 0.75–1.85 0.48 such as pharmacokinetics studies of thiotepa, including area Sec. AML 0.89 0.57–1.40 0.62 under the curve or immune reconstitution. RIC conditioning 0.99 0.72–1.36 0.95 – In conclusion, in this registry-based study of AML patients who Poor cytogenetics 1.25 0.86 1.80 0.24 received thiotepa before allogeneic HSCT, with more that 50% of OS patients transplanted from alternative donors, we could demon- Age (by deacade) 1.14 1.01–1.29 0.03 strate that thiotepa is feasible and safe with low incidences of CR1 (ref) 1 CR2/3 1.73 1.19–2.52 0.004 mucositis and SOS combined with good transplantation outcome. Active disease 2.45 1.73–3.45 o0.0001 These data should be confirmed in a well-designed multicenter Female donor to male recipient 1.02 0.73–1.44 0.89 MSD (ref) 1 prospective study. UD 0.91 0;58–1.42 0.66 Haplo 1.40 0.97–2.04 0.08 CB 1.42 0.89–2.24 0.14 CONFLICT OF INTEREST Sec. AML 0.80 0.50–1.28 0.35 RIC conditioning 0.92 0.67–1.28 0.63 MM received lectures honoraria and research support from Riemser whose product is Poor cytogenetics 1.24 0.86–1.80 0.25 discussed in this work. The remaining authors declare no conflict of interest. Abbreviations: CB = cord blood; CI = confidence interval; haplo = haploidentical; HR = hazard ratio; LFS = leukemia-free survival; MSD = matched ACKNOWLEDGEMENTS sibling donor; NRM = non-relapse mortality; OS = overall survival; ref = reference; RI = relapse incidence; RIC = reduced-intensity condition- We thank all national registries, all EBMT centers and all data managers for their great ing; sec. AML = secondary acute myeloid leukemia; UD = unrelated donor. work and help. See Supplementary Appendix for list of institutions reporting data included in this study.

© 2017 Macmillan Publishers Limited, part of Springer Nature. Bone Marrow Transplantation (2017) 238 – 244 Safety and efficacy of thiotepa for allo-HSCT in AML S Eder et al 244 AUTHOR CONTRIBUTIONS 18 Sanz J, Sanz MA, Saavedra S, Lorenzo I, Montesinos P, Senent L et al. Cord blood SE, MM, AN and ML designed the study. ML collected the data from the transplantation from unrelated donors in adults with high-risk acute myeloid 16 – database of the ALWP of the EBMT (chaired by MM and AN). JF, DB, AO, SS, AR, leukemia. Biol Blood Marrow Transplant 2010; :86 94. 19 Sanz J, Boluda JC, Martín C, González M, Ferrá C, Serrano D et al. Single-unit LK and GM were the principal investigators of the seven highest recruiting umbilical cord blood transplantation from unrelated donors in patients with centers of the study. ML, SE, MM and AN analyzed and interpreted the data. hematological malignancy using busulfan, thiotepa, fludarabine and ATG as SE, ML, MM and AN wrote or revised the manuscript, and all authors reviewed myeloablative conditioning regimen. Bone Marrow Transplant 2012; 47: fi its nal version. 1287–1293. 20 Ruggeri A, Sanz G, Bittencourt H, Sanz J, Rambaldi A, Volt F et al. Comparison of outcomes after single or double cord blood transplantation in adults with acute REFERENCES leukemia using different types of myeloablative conditioning regimen, a retro- 1 Appelbaum FR. Hematopoietic-cell transplantation at 50. N Engl J Med 2007; 357: spective study on behalf of Eurocord and the Acute Leukemia Working Party 1472–1475. of EBMT. Leukemia 2014; 28: 779–786. 2 Mohty B, Mohty M. Long-term complications and side effects after allogeneic 21 Giralt S, Ballen K, Rizzo D, Bacigalupo A, Horowitz M, Pasquini M et al. Reduced- hematopoietic stem cell transplantation: an update. Blood Cancer J 2011; 1:e16. intensity conditioning regimen workshop: defining the dose spectrum. Report of 3 Kamani N, August CS, Bunin N, Leahey A, Bayever E, Goldwein J et al. A study of a workshop convened by the center for international blood and marrow trans- thiotepa, etoposide and fractionated total body irradiation as a preparative plant research. Biol Blood Marrow Transplant 2009; 15:367–369. regimen prior to bone marrow transplantation for poor prognosis patients with 22 Laughlin MJ, Barker J, Bambach B, Koc ON, Rizzieri DA, Wagner JE et al. Hema- neuroblastoma. Bone Marrow Transplant 1996; 17:911–916. topoietic engraftment and survival in adult recipients of umbilical-cord blood 4 Bitran JD, Samuels B, Klein L, Hanauer S, Johnson L, Martinec J et al. Tandem from unrelated donors. N Engl J Med 2001; 344:1815–1822. high-dose chemotherapy supported by hematopoietic progenitor cells yields 23 Przepiorka D, Weisdorf D, Martin P, Klingemann HG, Beatty P, Hows J et al. 1994 prolonged survival in stage IV breast cancer. Bone Marrow Transplant 1996; 17: Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant 1995; 157–162. 15:825–828. 0 00 5 Ng SF, Waxman DJ. N,N ,N -triethylenethiophosphoramide (thio-TEPA) oxygena- 24 Filipovich AH, Weisdorf D, Pavletic S, Socie G, Wingard JR, Lee SJ et al. National tion by constitutive hepatic P450 enzymes and modulation of drug metabolism Institutes of Health consensus development project on criteria for clinical trials in and clearance in vivo by P450-inducing agents. Cancer Res 1991; 51: 2340–2345. chronic graft-versus-host disease: I. Diagnosis and staging working group report. 6 Emanuel NM, Konovalova NP, Djachkovskaya RF. Toxicity, antitumor activity, and Biol Blood Marrow Transplant 2005; 11:945–956. 60 pharmacokinetics of spin-labeled thiotepa analogs. Cancer Treat Rep 1976; : 25 McDonald GB, Sharma P, Matthews DE, Shulman HM, Thomas ED. Venocclusive – 1605 1609. disease of the liver after bone marrow transplantation: diagnosis, incidence, and 7 Down JD, Westerhof GR, Boudewijn A, Setroikromo R, Ploemacher RE. Thiotepa predisposing factors. Hepatology 1984; 4: 116–122. improves allogeneic bone marrow engraftment without enhancing stem cell 26 Coppell JA, Richardson PG, Soiffer R, Martin PL, Kernan NA, Chen A et al. 21 – depletion in irradiated mice. Bone Marrow Transplant 1998; :327 330. Hepatic veno-occlusive disease following stem cell transplantation: incidence, 8 Heideman RL, Cole DE, Balis F, Sato J, Reaman GH, Packer RJ et al. Phase I and clinical course, and outcome. Biol Blood Marrow Transplant 2010; 16: fl pharmacokinetic evaluation of thiotepa in the cerebrospinal uid and plasma of 157–168. pediatric patients: evidence for dose-dependent plasma clearance of thiotepa. 27 Carreras E, Díaz-Beyá M, Rosiñol L, Martínez C, Fernández-Avilés F, Rovira M. The 49 – Cancer Res 1989; : 736 741. incidence of veno-occlusive disease following allogeneic hematopoietic stem cell 9 Nieto Y, Vaughan WP. Pharmacokinetics of high-dose chemotherapy. Bone transplantation has diminished and the outcome improved over the last decade. Marrow Transplant 2004; 33:259–269. Biol Blood Marrow Transplant 2011; 17: 1713–1720. 10 Nagler A, Ackerstein A, Or R, Naparstek E, Slavin S. Immunotherapy with recom- 28 Nagler A, Labopin M, Berger R, Bunjes D, Campos A, Socié G et al. Allogeneic binant human interleukin-2 and recombinant interferon-alpha in lymphoma hematopoietic SCT for adults AML using i.v. BU in the conditioning regimen: patients postautologous marrow or stem cell transplantation. Blood 1997; 89: outcomes and risk factors for the occurrence of hepatic sinusoidal obstructive 3951–3959. syndrome. Bone Marrow Transplant 2014; 49: 628–633. 11 Bacigalupo A, Van Lint MT, Valbonesi M, Lercari G, Carlier P, Lamparelli T et al. 29 Wardley AM, Jayson GC, Swindell R, Morgenstern GR, Chang J, Bloor R et al. Thiotepa cyclophosphamide followed by granulocyte colony-stimulating factor Prospective evaluation of oral mucositis in patients receiving myeloablative mobilized allogeneic peripheral blood cells in adults with advanced leukemia. conditioning regimens and haemopoietic progenitor rescue. Br J Haematol 2000; Blood 1996; 88: 353–357. 110:292–299. 12 Rosales F, Naparstek E, Varadi G, Or R, Slavin S, Nagler A. The role of thiotepa in 30 Huitema AD, Spaander M, Mathĵt RA, Tibben MM, Holtkamp MJ, Beijnen JH et al. allogeneic stem cell transplantation in patients with leukemia. Leuk Res 1999; 23: Relationship between exposure and toxicity in high-dose chemotherapy with 947–952. cyclophosphamide, thiotepa and carboplatin. Ann Oncol 2002; 13: 13 Raiola AM, Van Lint MT, Lamparelli T, Gualandi F, Mordini N, Berisso G et al. 374–384. Reduced intensity thiotepa-cyclophosphamide conditioning for allogeneic 31 Aversa F, Terenzi A, Tabilio A, Falzetti F, Carotti A, Ballanti S et al. Full haplotype- haemopoietic stem cell transplants (HSCT) in patients up to 60 years of age. Br J Haematol 2000; 109:716–721. mismatched hematopoietic stem-cell transplantation: a phase II study in patients 23 – 14 Corradini P, Tarella C, Olivieri A, Gianni AM, Voena C, Zallio F et al. Reduced- with acute leukemia at high risk of relapse. J Clin Oncol 2005; : 3447 3454. intensity conditioning followed by allografting of hematopoietic cells can pro- 32 Walker I, Shehata N, Cantin G, Couture F, Dhédin N, Barty R et al. Canadian duce clinical and molecular remissions in patients with poor-risk hematologic multicenter pilot trial of haploidentical donor transplantation. Blood Cells Mol Dis 33 – malignancies. Blood 2002; 99:75–82. 2004; : 222 226. 15 Alessandrino EP, Bernasconi P, Colombo AA, Caldera D, Malcovati L, Troletti D 33 Eder S, Labopin M, Arcese W, Or R, Majolino I, Bacigalupo A et al. Thiotepa-based et al. Reduced-intensity conditioning regimen with thiotepa and fludarabine versus total body irradiation-based myeloablative conditioning prior to allogeneic fi followed by allogeneic blood stem cell transplantation in haematological malig- stem cell transplantation for acute myeloid leukaemia in rst complete remission: nancies. Bone Marrow Transplant 2004; 34: 1039–1045. a retrospective analysis from the Acute Leukemia Working Party of the European 16 Bacigalupo A, Raiola AM, Lamparelli T, Gualandi F, Occhini D, Bregante S et al. Group for Blood and Marrow Transplantation. Eur J Haematol 2015; 96:90–97. Thiotepa-based reduced intensity conditioning regimen: a 10 year follow up. Bone 34 Nagler A, Rocha V, Labopin M, Unal A, Ben Othman T, Campos A et al. Allogeneic Marrow Transplant 2007; 40: 1091–1093. hematopoietic stem-cell transplantation for acute myeloid leukemia in remission: 17 Lamparelli T, van Lint MT, Gualandi F, Raiola AM, Barbanti M, Sacchi N et al. comparison of intravenous busulfan plus cyclophosphamide (Cy) versus total- Alternative donor transplants for patients with advanced hematologic body irradiation plus Cy as conditioning regimen--a report from the acute malignancies, conditioned with thiotepa, cyclophosphamide and antithymocyte leukemia working party of the European group for blood and marrow globulin. Bone Marrow Transplant 2000; 26: 1305–1311. transplantation. J Clin Oncol 2013; 31: 3549–3556.

Supplementary Information accompanies this paper on Bone Marrow Transplantation website (http://www.nature.com/bmt)