We previously published our institutional experience of 44 patients with PTCL who underwent alloBMT between January 1994 and June 2011, including 20 patients who received myeloablative conditioning (MAC) and 24 NMA/RIC. 24 Here we evaluate outcomes in 65 consecutive patients with PTCL who underwent alloBMT after January 2004 using NMA conditioning and PTCy for GVHD prophylaxis, an increasingly common approach to alloBMT as outlined below.
Multiple registry25–28 and other retrospective studies 15,29−33 have assessed alloBMT outcomes in PTCL, but its optimal application in this challenging patient population remains to be determined. At the 2020 American Society of Hematology (ASH) Meeting, Mehta-Shah et al. presented what was the largest retrospective study to date at that time with 508 patients with mature T-cell lymphoma who underwent alloBMT, 323 of which received NMA/RIC, demonstrating 2-year OS and PFS of 59.1% and 45.8%, respectively. 33 More recently, Hamadani et al. performed a collaborative study between the European Society for Blood and Marrow Transplantation (EBMT) and the Center for International Blood and Marrow Transplant Research (CIBMTR) reviewing outcomes in 1942 patients with PTCL-NOS, AITL, and ALCL who underwent alloBMT with most receiving RIC; donors were haplo (n = 237), matched sibling (n = 911), matched unrelated with in vivo T-cell depletion (n = 468), and matched unrelated without in vivo T-cell depletion (n = 326). Their results showed a 3-year OS of 59–64% and 3-year PFS of 48–52%, with similar survival outcomes among the 4 cohorts. 28
PTCy-based GVHD prophylaxis is a well-established method for expanding the donor pool without compromising safety in patients with lymphoma undergoing alloBMT,24,34−37 though published accounts in PTCL are often limited to haplo donors. Castagna et al. reported outcomes in 68 patients with PTCL who underwent alloBMT with NMA/RIC, including 29 patients from haplo donors who received PTCy for GVHD prophylaxis and 39 patients from matched sibling (n = 20) or unrelated donors (n = 19) who received cyclosporine A; graft source was PB in 85% of cases. Their results showed a 4-year OS and 4-year PFS of 75% and 70%, respectively, with 2-year cumulative incidence of relapse of 21% and 2-year NRM of 9%. They also noted lower aGVHD and 4-year NRM in alloBMT from related donors;32 this finding was not seen our study, where 10 patients underwent alloBMT from unrelated donors (7 mMUD, 3 full) without increased GVHD and no instances of NRM. However, it is worth noting all patients in our study received PTCy, rather than the haploidentical donors only. PTCy-based GVHD prophylaxis was also used in patients who underwent alloBMT from a haplo donor (n = 237) in the previously mentioned study by Hamadani et al. 28
The most striking finding in our study is the difference in outcomes based on graft source; compared to BM, PB allografts were associated with significantly improved PFS and OS, lower relapse, and no significant difference in NRM or GVHD. In a meta-analysis involving 9 randomized controlled trials (RCTs) assessing alloBMT graft source for hematologic malignancies in adults, PB was associated with faster engraftment but increased risk of GVHD compared to BM. 38 A Lymphoma Working Party (LWP)-EBMT study of 474 lymphoma patients (including 88 with PTCL) who underwent alloBMT using haplo donors, PTCy, and either RIC (77%) or MAC (23%) showed similar results with improved engraftment but increased GVHD. 39 PB allografts have also been associated with reduced risk of relapse and improved survival in patients with Hodgkin lymphoma who underwent alloBMT with RIC. 40 While some studies have found no association between graft source and survival outcomes in PTCL, 28 source may be influenced by conditioning regimen, 41 donor selection, 28 and histology/disease characteristics. 26 In our study, all patients received NMA conditioning with PTCy, regardless of histology or other clinical features. Based on our results, PB allografts in this setting are safe and may offer superior outcomes in PTCL. Since the use of PB allografts was accompanied by increased TBI dosing to 400 cGy in our study (15 of 19 patients), we cannot assume the benefits seen in this cohort were due to graft source alone; increased TBI dosing may have contributed as well. Our study also demonstrated improved PFS and OS in those patients in CR vs. PR prior to alloBMT, a finding consistent with prior studies of PTCL patients. 15,26,28,33
One of the limitations of our study is the smaller sample size, which limited more extensive statistical analysis. The median follow up among survivors was also significantly shorter in those receiving PB (2.2 years) vs. BM (4.7 years) after alloBMT. While there were no differences in histology, disease status prior to alloBMT, or prior therapies between these 2 cohorts, longer follow up and a larger cohort size are needed to evaluate outcome differences with adjustment of potential confounders.
In conclusion, alloBMT using NMA conditioning with PTCy for GVHD prophylaxis is well-tolerated in patients with PTCL. Our preliminary data suggest increasing TBI dose to 400 cGy and using PB allografts may offer improved disease control and better survival outcomes, though longer follow up and additional studies are needed.