The outcome of first-line chemotherapy is far from optimal until today for most PTCLs subtypes. HSCT is a valuable option to achieve longer survival or cure this disease. In this study, we reported the outcomes of a multi-center retrospective study of 128 patients who underwent HSCT in 8 hospitals across China. To the best of our knowledge, this is the largest report in China. In general, the survival of patients in the auto-HSCT group was better than that in the allo-HSCT group. NRM was as expected lower in the auto-HSCT group. Relapse rates were similar between these two groups. It was difficult to draw definite conclusions because of the differences in baseline characteristics between these two groups. Compared with patients in the auto-HSCT group, patients undergoing allo-HSCT were more likely to be diagnosed with ALK-negative ALCL, advanced stage, bone marrow involvement, or relapsed/refractory disease status before transplantation. We further compared the survival of patients in different subgroups. Firstly, we found that patients with low PIT score who received auto-HSCT group in upfront setting had better outcome than patients with high PIT score. Secondly, patients with CR undergoing auto-HSCT had the best outcome (3-year OS: 88% vs 46% in the allo-HSCT group). For patients with PR, their survival was similar (3-year OS: 59% in auto-HSCT vs 48% in allo-HSCT). The survival curve of patients with PD or SD undergoing allo-HSCT was better than that of auto-HSCT recipients. Our results suggested that it was plausible to choose auto-HSCT versus allo-HSCT according PIT score and remission status before transplantation.
The efficacy of auto-HSCT had been evaluated both in retrospective and prospective studies, with reported three- to five-year OS raging from one third to more than two thirds(12, 20–23). The huge variability in survival rate was related to the heterogeneity in baseline characteristics of patients included in above studies. Among the baseline factors, remission status before HSCT was the most important factor affecting post-HSCT survival. The 3-year OS in our study was 70% (58–85%). This was higher than OS of 59% (49–68%) in the study by CIBMTR (Center for International Blood and Marrow Transplant Research), where 64 patients (56%) were in CR at transplantation(21). Also, the five-year OS was only 46% in the largest Asian study, which comprised of 104 (77%) patients in CR/PR(20). It can be partially explained by three reasons. Firstly, over 90% of patients (66/72) undergoing auto-HSCT were in CR (37) or PR (29) at transplantation in our study. Secondly, the percentage of ALK‑positive ALCL patients (19/72) was higher than that (12/135) in the Asian study. Thirdly, the median follow-up time was only 23 (4-143) months, which was shorter than that in the above studies.
Another question was which part of patients can benefit from auto-HSCT. Prognostic Index attained at diagnosis wasn’t so important for patients who received HSCT in relapsed settings. So, we only checked the role of PIT score in the upfront setting. We found that the prognosis of patients with PIT 2 or higher who received upfront auto-HSCT was very poor, which was consistent with Shigeo’s study(24). We’d rather believe that it was reasonable to avoid auto-HSCT for patient with higher PIT score in the upfront setting.
Because of the fact that the vast majority of patients will eventually relapse after upfront chemotherapies, even when auto-HSCT were followed thereafter(25, 26), or remained refractory at the beginning, allo-HSCT was mostly performed in the relapsed or refractory setting. About one half of patients could achieved long-term survival through allo-HSCT, which was confirmed in large retrospective studies from CIBMTR(21),Europe(13), Asian(20). Likewise, there were also kinds of heterogeneity in histologic subtypes, stages, remission status at transplantation, intensity of conditioning regimens, donor types and GVHD prophylaxis among previous reports. In our study, over 90% of patients (52/56) in the allo-HSCT group were diagnosed with advanced stages. And more than 80% of patients (48/56) were non-CR at transplantation. The 3-year OS of allo-HSCT recipients was 46% (34–63%), which was similar with previous studies. In particular, all allo-HSCT in our study were myeloablative conditioning regimens. There were few studies that all patients received myeloablative conditioning regimens. With regarding the impact of regimens intensities on survival, there was two studies but failed to find significant difference either in toxicity and survival between myeloablative and reduced conditioning regimens(13, 21). With progress in supportive care, GVHD prophylaxis, and donor selection system, there were 20 patients in our study that receive HLA-haploidentical HSCT.
Because of the inherent toxicity of conditioning regimens and GVHD along with allo-HSCT, NRM in the allo-HSCT group was as expected higher than that (27% vs 6%) in the auto-HSCT group. But there was no difference in relapse rates (29% vs 34% in auto-HSCT). One possible explanation was that approximately 40% patients of patients (23/56) were at progressive status before transplantation. Allo-HSCT in our study, like most previous studies, often performed as salvage treatment, compromising the effectiveness of allo-HSCT. As for causes of death, lymphoma progression or relapse was the leading factor in both groups, which indicated that prevention of relapse was also very important in the allo-HSCT.
More practical issues for physicians were how to choose the type of HSCT, HSCT timing and identification of optimal populations. With huge heterogeneity in disease itself and results of previous studies, it was no consensus about these issues until today. Hopefully with evidence from both retrospective and prospective cohort studies, auto-HSCT was recommended as consolidation in most institutes for patients who achieved a CR or PR after first-line therapies(22, 27). Two recent prospective studies also failed to demonstrate that allo-HSCT could achieve better survival than auto-HSCT for patients in first remission(28, 29). Allo-HSCT for patient in first remission still remained controversial until today. To further explore these issues, we assessed the survival of patients according to the disease status at transplantation in this study. Consistent with literature, patients with CR can benefit from auto-HSCT. For patients with PR, their survival was similar between auto-HSCT and allo-HSCT. Although no significant difference was noted, the survival curve of patients in progression underwent allo-HSCT was better than that in auto-HSCT. Based on the evidence and results above, we are proposing that it is advisable to choose auto-HSCT for patients with CR with previous treatments. And for patient with PR, either modality could be advised. For patients in progression disease, it was more plausible to proceed allo-HSCT. This was just our scenario. It was urgently to be confirmed in further prospective or randomized studies.
There were several limitations in this study. Firstly, the basal characteristics and prognosis of patients who were unable to undergo transplantation were not included in this study. It can’t reflect the whole clinical picture of PTCLs. Secondly, due to the retrospective design and small size of this study, the conclusions should be interpreted with caution. Nevertheless, this study showed patients with PTCLs can benefit from both auto-HSCT and allo-HSCT. Specifically, for patients with high PIT score in upfront setting, it was wise to avoid auto-HSCT. For patients with CR with previous treatments, it is advisable to undergo auto-HSCT. For patient with PR, either modality works well. It was wise to proceed allo-HSCT for patients in progression disease. We concluded to choose auto-HSCT versus allo-HSCT according PIT score and remission status before transplantation.