Management of Peripheral T-cell Lymphomas and the Role of Transplant

Abstract

Peripheral T-cell lymphomas are a rare subset of non-Hodgkin lymphomas that are treated with curative intent. While therapy has been based on other aggressive lymphoid malignancies, outcomes are generally poorer than B-cell lymphomas with 5-year overall and progression free survival 30–40% and 20–30%, respectively. In effort to improve outcomes, transplant has been used in both the frontline and salvage settings. Although not studied in randomized studies, consolidation with autologous stem cell transplant in first remission has been associated with an approximately 50–60% 5-year overall survival and 40–45% 5-year progression free survival. Unfortunately, most patients relapse, and, in this setting, allogeneic transplant remains the only curative option for those who are transplant eligible. Multiple series have now shown that 3-year overall survival with allogeneic transplant is approximately 60%. However, outcomes with transplant are associated with disease control at the time of transplant.

INTRODUCTION

Peripheral T-cell lymphomas (PTCLs) are a rare group of non-Hodgkin lymphomas (NHLs) that are generally less chemoresponsive and with inferior outcomes as compared to their B-cell counterparts. The term PTCL is commonly used to refer to systemic or nodal T-cell lymphomas, comprising 19 distinct subtypes as detailed in the 2016 World Health Organization (WHO) classification of lymphoid neoplasms [1]. The most common types of PTCL constitute approximately 65% of cases, and include: PTCL, not otherwise specified (PTCL-NOS), angioimmunoblastic T-cell lymphoma (AITL), follicular T-cell lymphoma (FTCL), nodal PTCL with TFH phenotype (TFH-PTCL), anaplastic lymphoma kinase-positive (ALK+) anaplastic large cell lymphoma (ALCL), and ALK-negative (ALK-) ALCL [2].

In Western countries, PTCL accounts for 10–15% of aggressive lymphomas and 5–10% of all NHLs [3,4]. In Asia, the frequency of T- and NK-cell lymphomas is higher, accounting for 15–25% of new diagnoses of NHL [3,5]. PTCL-NOS is the most common subtype of PTCL in North America, whereas AITL is most common in Europe [6]. While there are clear geographic variations, our most comprehensive international epidemiologic study of PTCL to date found the following distribution: PTCL-NOS 26%, AITL 19%, ALK+ ALCL 7%, and ALK- ALCL 6% [4]. The median age at the time of PTCL diagnosis is 62, however several subtypes present more often in younger patients: ALK+ ALCL (33 years), hepatosplenic type (34 years), and subcutaneous panniculitis-like PTCL (33 years) [4]. These difference in distribution of patients are important to consider when evaluating studies regarding transplant in PTCL.

In general, PTCL is treated with combination chemotherapy with curative intent and often requires urgent evaluation and treatment. In most of the more common histologies, patients who are chemoresponsive are considered for an autologous transplant in first remission. While prognosis in relapsed/refractory (R/R) cases of PTCL is poor, allogeneic transplant can lead to durable remissions in patients eligible for this approach. Treatment strategies in PTCL have historically been derived from regimens designed for aggressive B-cell lymphomas. However, as our understanding of the pathophysiology of PTCL improves, we gain insight into the unique biologies of histological subtypes and how this influences their various responses to different treatments. This review will primarily focus on the management of the most common forms of PTCL, with a particular focus on the use of autologous and allogeneic stem cell transplant (SCT).

Prognosis

PTCL is a heterogenous disease and prognosis can be stratified based on several factors, notably histological subtype and prognostic indices based on clinical parameters. The International Prognostic Index (IPI) is commonly applied in PTCL [7], however was originally developed for use in patients with B-cell lymphomas [8]. Other prognostic scores specific for PTCL have been developed which include bone marrow involvement, Ki-67 expression ≥ 80%, albumin < 3.5 g/dL, platelets < 150,000, and absolute neutrophil count < 6.5 ×10⁹/L [9–12].

One study compared the four prognostic scores published at that time (IPI, PIT, mPIT, IPTCLP) [8–11] and found the IPI to best predict complete response (CR), however identified IPTCLP to be the best independent predictor of overall survival (OS) [13]. Regardless of the prognostic score utilized, outcomes for patients with PTCL remain suboptimal even in the lowest risk-category with 5-year OS and progression free survival (PFS) rates of 50% and 33%, respectively [4]. Prognostic indices do not inform therapeutic decisions, however this may evolve as progress continues in areas of research aimed at tailoring treatment decisions based on molecular markers, interval PET-CT findings, and minimal residual disease testing [14].

The most informative data regarding the impact of histology on prognosis in PTCL is based on several registry studies [4,15,16]. These trials found 5-year OS to be 28–35% for PTCL-NOS, 31–36% for AITL, and 34–49% for ALK- ALCL. ALK+ ALCL has consistently demonstrated superior outcomes when compared with other PTCL cohorts, commonly having 5-year OS > 70% [4,17]. For this reason, patients with ALK+ ALCL are not routinely recommended to receive consolidation with autologous stem cell transplant in first remission in contrast to other forms of PTCL. With the incorporation of brentuximab vedotin into the frontline treatment for patients with ALCL and other CD30+ PTCLs, outcomes for these patients are expected to improve.

TREATMENT

Frontline treatment

For patients with the more common forms of PTCL, upfront treatment with combination chemotherapy remains the most common treatment approach and can achieve cure in a subset of patients. Anthracycline-based chemotherapy regimens, such as CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) with or without etoposide (CHOEP), have been the standard of care for decades producing an overall response rate (ORR) of approximately 80% and CR rates of approximately 40–50% [16,18]. Until the publication of ECHELON-2, however, there were no randomized trials in peripheral T-cell lymphomas that demonstrated a survival benefit.

The frontline treatment of newly diagnosed patients with PTCL significantly changed with the publication of ECHELON-2, which was a double blind, randomized study of brentuximab vedotin (BV) with CHP (cyclophosphamide, doxorubicin, prednisone) versus CHOP in 452 patients with PTCL expressing CD30 ≥ 10% [19,20]. Brentuximab vedotin is composed of a monoclonal antibody against CD30, conjugated to monomethyl auristatin E, a microtubule disrupting drug. CD30 expression is expressed universally in systemic ALCL (ALK+ and ALK-) and is variably expressed in other nodal PTCLs.

Data from 5-year follow-up demonstrated superior results of BV+CHP over CHOP, with median PFS of 62.3 months versus 23.8 months [hazard ratio (HR) 0.70, 95% CI 0.53–0.91, P = 0.008] in an intent-to-treat analysis. Patients with systemic ALCL (70% of study population) benefited the most from the addition of BV which improved the 5-year PFS from 48.4% to 60.6% (HR 0.55, 95% CI 0.39–0.79) and 5-year OS from 68.7% to 75.8% (HR 0.66, 95% CI 0.43–1.01). The PFS benefit was not seen in patients with PTCL-NOS (HR 0.79, 95% CI 0.43–1.43) or AITL (HR 1.41, 95% CI 0.64–3.11) [20]. These data lead to the approval of BV+CHP in previously untreated patients with CD30+ PTCL without a minimum threshold of CD30-positivity.

Interestingly, patients with AITL in the ECHELON-2 trial seemed to derive greater benefit from standard CHOP chemotherapy than what has been seen in prior studies. In this study, patients with AITL treated with CHOP had estimated 5-year OS and PFS of 62.5% and 48%, respectively [20]. These patients had superior outcomes compared to the PTCL-NOS subgroup treated with CHOP, who had 5-year OS and PFS of 36% and 26%, respectively.

For patients with subtypes of PTCL excluding ALCL, anthracycline-based chemotherapy regimens such as CHOP or CHOEP remains a standard option. While the addition of etoposide to CHOP has never been evaluated in a prospective randomized manner, a retrospective analysis of multiple studies of the German High-Grade Non-Hodgkin Lymphoma Study Group was done to evaluate outcomes for the subset of patients with PTCL included in prior studies of aggressive NHLs. These data demonstrated that for patients ≤ 60 years-old with normal LDH levels, the addition of etoposide to CHOP significantly improved 3-year event free survival (EFS), 75% versus 51% (P = 0.003), however did not have any effect on OS [21]. The benefit was primarily seen in patients with ALK+ ALCL and/or low-risk IPI. In this analysis, CHOEP did not provide a significant benefit to patients > 60 years-old due to increased toxicity seen with the addition of etoposide. Additional agents have been studied combination with upfront CHOP-based therapy but have had negative results. These include alemtuzumab [22], romidepsin [23], pralatrexate [24], and lenalidomide [25]. Efforts to improve on CHOP or CHOEP include a US Intergroup study to evaluate the use of hypomethylating agents or phosphoinositol-3-kinase inhibitors added to a CHOP or CHOEP backbone (A051902; NCT04803201).

Transplant in First Remission

In patients who are chemotherapy responsive, we consider consolidation with high-dose therapy (HDT) and autologous stem cell transplantation (auto-SCT) in first remission for most fit patients with PTCL [26,27]. For patients with ALK+ ALCL, consolidation with auto-SCT is not recommended as these patients have a more favorable outcome compared with patients with other nodal PTCL subtypes [17].

Several groups have worked to identify which patients will benefit from consolidation with auto-SCT through prospective, non-randomized trials (Table 1). One of the largest studies from Germany enrolled 111 patients with newly diagnosed PTCL, excluding patients with ALK+ ALCL, administered induction chemotherapy with CHOP and then if the patient achieved CR or partial response (PR), they proceeded to HDT and auto-SCT. Seventy-five (68%) of patients received auto-SCT with the main reason for not proceeding to transplant being progressive/persistent disease. For the intent-to-treat population (ITTP) of 111 patients, estimated 5-year OS and PFS was 44% and 39%, respectively. The estimated 5-year OS for patients who underwent auto-SCT was 57%; treatment-related mortality (TRM) was 3.6% as 4 patients died of treatment-related infectious complications [28,29].

Table 1.

Studies evaluating autologous hematopoietic transplantation in patients with PTCL. Studies included in this table were restricted to those with more than 30 patients.

Study	Year	Cases (n)	PTCL Subtypes	Median Follow-Up (months)	Survival
Registry Studies
Rodriguez et al. [64] (GEL-TAMO)	2003	37	All subtypes including ALCL	37	DFS: 79% at 5 years OS: 80% at 5 years
Ellin et al. [16] (Swedish)	2014	128	All subtypes including ALCL (ALK− n = 24)	97.3	PFS: 41% at 5 years OS: 48% at 5 years
Park et al. [32] (COMPLETE)	2019	119 (auto-SCT n = 36)	All subtypes including ALCL (ALK− n = 42)	33.6	OS: 87.6% at 2 years Improved PFS and OS for AITL PFS: 68.8% at 2 years OS: 93.3% at 2 years
Garcia-Sancho et al. [30] GELTAMO/FIL	2022	174 (auto-SCT n =103)	All subtypes excluding ALK+ ALCL	66	PFS 63.8% at 5 years OS: 74%% at 5 years
Retrospective Series
Rodriguez et al. [64]	2003	35	All subtypes	37.5	37% at 5 years
Feyler et al. [65]	2007	64	All subtypes	37	53% at 3 years
Rodriguez et al.[66]	2007	74	All subtypes	67	68% at 5 years
Kyriakou et al. [67]	2008	146	AITL	31	59% at 4 years
Yang et al.[68]	2009	64	PTCL-NOS	29.7	53% at 3 years
Numata et al. [69]	2010	39	All subtypes	78	62.2% at 5 years
Beitinjaneh et al. [59]	2015	47	All subtypes	35	76% at 4 years
Han et al. [70]	2017	52	All subtypes	34	71.1% at 5 years
Wu et al. [71]	2018	47	All subtypes	23.6	89.8% at 2 years
Fossard et al. [33]	2018	269	All subtypes	54	59.2% at 5 years
Mehta-Shah et al. [38]	2019	99	ALCL, ALK+ excluded	48	48% at 4 years

The Nordic Lymphoma Group published a large prospective study to date on the use of auto-SCT for consolidation in PTCL, which included 160 patients in the ITTP. Of the 131 patients who achieved CR/PR, 115 (72% of the ITTP) underwent HDT and auto-SCT. Overall TRM was 4% as 7 patients died from infection or hemorrhage. The 5-year OS and PFS of the ITTP was 51% and 44%, respectively [18]. A retrospective study of the Swedish Lymphoma Registry further confirmed these findings in an ITT analysis of 252 cases of nodal PTCL and enteropathy-associated TCL, excluding ALK+ ALCL [16]. They demonstrated that, compared to patients treated without intent to pursue auto-SCT in CR1, upfront auto-SCT was associated with a superior 5-year OS of 48% versus 26% and PFS of 41% versus 20%. In this cohort, 68% of patients planned for auto-SCT were able to receive the procedure.

More recently, the GELTAMO/FIL groups published a combined analysis from their registries of 174 patients with PTCL excluding ALK+ ALCL who were in CR at the completion of frontline therapy. In this group, 103 patients underwent consolidative auto-SCT and 71 did not. At a median follow up of 65.5 months (range 4.1–176.7 months), those who underwent consolidative transplant had superior 5-year PFS (63% versus 49%) and 5-year OS (74% versus 62%). These results were then validated in a multivariate analysis, which demonstrated first line auto-SCT was associated with significantly prolonged PFS (HR 0.57, 95% CI 0.35–0.93) and OS (HR 0.57, 95% CI 0.33–0.99) [30].

Post-hoc analysis of the ECHELON-2 data examining outcomes in patients with ALK- ALCL and non-ALCL histologies who achieved CR with BV-CHP and proceeded to consolidative auto-SCT suggested that auto-SCT continues to provide PFS benefit in this population [31]. Although only 19% of these patients underwent a consolidative transplant, the cohort of patients in CR after BV+CHP that went on to auto-SCT had estimated 3-year PFS of 76% versus 53% that did not receive auto-SCT consolidation. When adjusting for age and regional practices, the HR for auto-SCT was 0.39 (95% CI: 0.18–0.82). Interestingly, patients with ALK- ALCL benefited less from auto-SCT consolidation than patients with other forms of CD30+ PTCL included in the trial (multivariate HR 0.52 versus HR 0.32). While the study was underpowered for this analysis and the decision to pursue transplant was at the discretion of the treating physician, the analysis supports to use of consolidative transplant in the era of brentuximab-based therapy.

While it is our approach to consider autologous transplant in first remission, there are analyses that suggest that the benefit is less robust. In an analysis of 119 patients in the COMPLETE registry, there was no statistically significant difference in outcomes of the 36 patients who underwent auto-SCT versus the 83 patients who did not [32]. In a similar analysis from the LYSA group who were analyzed by an intent to treat with transplant analysis, the 5-year PFS was 40.5% versus 46.3% amongst patients with ITT versus no ITT. Similarly, 5-year OS was 60.9% versus 59.2% amongst patients with ITT versus no ITT [33]. These studies are often limited by low proportions of patients who underwent autologous transplant.

The role for allogeneic SCT in first-line therapy remains highly controversial and is generally not recommended. One of the few randomized, phase III trials in PTCL was recently published and compared allo-SCT to auto-SCT for consolidation in first remission [34]. The trial enrolled 104 patients ≤ 60 years-old with poor prognosis (stage II-IV or age-adjusted IPI > 0) with T-cell lymphoma, excluding ALK+ ALCL, at centers in France and Germany. Of the patients randomized to each arm, 41 (63%) underwent auto-SCT and 26 (53%) underwent allo-SCT. With median follow-up of 42 months, there was no significant difference in EFS, PFS, or OS between the two treatment groups. In the ITTP, the 3-year OS and PFS for auto-SCT versus allo-SCT was 70% versus 57% and 39% versus 43%, respectively. While none of the 21 responding patients proceeding to allo-SCT relapsed, 8 of 26 patients (31%) who received allo-SCT died of transplant-related toxicities. In contrast, 13 of 36 (36%) responding patients proceeding to auto-SCT relapsed, however no patients receiving auto-SCT died of transplant-related causes. Overall, the strong graft-versus lymphoma effect was counterbalanced by transplant-related mortality.

There have been efforts to try to select patients for transplant who are most likely to benefit. Data from a retrospective series suggested that patients with DUSP22 rearranged ALK- ALCL have a more favorable prognosis compared to the remaining subset of patients with ALK- ALCL and may not benefit additionally from HDT and auto-SCT to achieve durable remissions [35,36]. However, this finding was refuted by a subsequent smaller case series that examined the impact of DUSP22 on prognosis in ALK- ALCL and concluded that these patients actually have worse outcomes than patients with ALK+ ALCL [37]. Therefore, the utility of DUSP22 to inform decisions about consolidative transplant remain controversial. Interim PET/CT has also been studied as a method of selecting patients most likely to benefit from transplant. In a 99 patient single institution series of patients treated with the intent to transplant in first remission, those who had a negative interim PET after 4 cycles by the Lugano Criteria (Deuville 1–3) had a 4-year OS and PFS of 83% and 58%, respectively, whereas no patients were alive or disease free at 4 years if they had a positive interim PET [38]. This suggests that interim PET predicts outcomes in an intent to transplant population.

There are rarer subtypes of PTCL, such as hepatosplenic T-cell lymphoma (HSTCL), adult T-cell leukemia/lymphoma, and primary cutaneous gamma-delta T-cell lymphoma, that do not often have durable responses to therapy without consolidation with an allogeneic transplant. Therefore, in fit patients who achieve disease control with these subtypes, we consider a consolidative allogenic transplant in first remission. A retrospective analysis of patients with HSTCL found that 12 of the 13 patients who achieved long-term survival were treated with allo-SCT [39]. Similarly a retrospective analysis of 80 patients with primary cutaneous gamma delta T-cell lymphomas showed improved long term survival in those consolidated with allogeneic transplant in first remission [40]. An analysis of the largest registry for ATLL from Japan has also suggested that early transplant (<100 days from diagnosis) is associated with better outcomes [41]. For these patients, the risk of TRM is offset by the even greater risk of relapsed or persistent disease which is universally fatal.

Taken together, these findings indicate that auto-SCT can induce durable remissions, and is possibly curative, for a substantial portion of patients with PTCL able to proceed to auto-SCT. A major hurdle for patients with PTCL, however, is that only about two-thirds of patients can proceed to HDT and auto-SCT, either due to failure of induction therapy or early relapse. IPI continues to define risk groups in PTCL, however treatment approaches should not be tailored to IPI score as even patients with low-risk disease have poor outcomes with intensive treatment. As approximately 20% of patients are likely to benefit from a consolidative autologous transplant, selection of these patients remains an area of research. Strategies to evaluate the use of mutational profiling, cell free DNA and minimal residual disease monitoring to help select patients for transplant are under evaluation [42,43]. Finally, more recent data regarding allo-SCT in the frontline setting does not support its use as the benefit of the graft-versus-lymphoma effect was offset by high TRM.

Relapsed/Refractory PTCL

Despite most patients initially having chemo-sensitive disease, relapse in nodal PTCL is common, with median time from end of treatment to relapse of 8 to 12 months [44,45]. Patients with relapsed or refractory PTCL have had a median PFS of 3.3 to 9.6 months and median OS of 5.5 months [46]. Patients who were originally chemosensitive and have relapsed disease have longer survival compared to patients with refractory disease, with median overall survival of 29.1 versus 12.3 months from time of diagnosis [45]. While outcomes in relapsed/refractory PTCL remain poor, durable remissions can be achieved for those who are eligible for allogeneic transplant. However, patients are often limited by comorbidities, age, lack of donor, or inability to achieve sufficient disease control.

While the focus of this review is to discuss the role of transplant, bridging therapy leading to remission is critical. There are four agents that have been FDA approved for PTCL in the relapsed/refractory setting: pralatrexate, belinostat, brentuximab vedotin and romidepsin which is now NCCN compendium listed. With the exception of brentuximab vedotin for those with CD30 expressing disease, these agents have modest activity in patients with R/R PTCL and demonstrate response rates from 20 to 35% as single agents [47–49].

Romidepsin achieved an ORR of 25% with a median duration of response of 13 to 17 months in the studies leading to its approval in R/R PTCL [48–50]. Belinostat similarly showed an ORR of 26%, CR rate of 10.8% and median duration of response of 13.6 months. Notably, patients who achieved a CR had a median duration of response of > 29 months [48]. Pralatrexate, an antifolate, demonstrated an ORR of 29%, median duration of response of 10 months, median PFS of 4 months and median OS of 15 months in a heavily pretreated group of patients with R/R PTCL [47]. Brentuximab vedotin demonstrated a response rate of 83% in ALCL with a median duration of response of 20.8 months, but when evaluated in patients with AITL and PTCL-NOS carried a response rate of 41% with median duration of response of 2.6 months [51,52]. Combination chemotherapies can also lead to remission as well, but a recent analysis of the COMPLETE registry showed that the overall response rate to second line therapy as deemed by the treating physician was 60% with single agents compared to 44% in multiagent therapy [45]. Multiple novel therapies are under investigation in peripheral T-cell lymphomas including phosphoinositol-3-kinase inhibitors (NCT03770000, NCT03372057), EZH2 inhibitors (NCT04703192), and JAK/STAT inhibitors (NCT04105010) as single agents or in combination with other therapies.

Transplant in Relapsed/Refractory Disease

Evidence for the use of autologous or allogeneic SCT for patients with R/R PTCL has historically been limited to retrospective studies and registry data. (Table 2)

Table 2.

Studies of allogeneic hematopoietic stem cell transplantation in patients with relapsed or refractory PTCL. Studies included were restricted to those with more than 20 patients.

Study	N / median age (range), y	Follow-up (range), mo.	Histological subtype	Status before allo-SCT	Donor type	Conditioning regimen	TRM/NRM	DFS	OS
Hamadani, 2022 [53] Retrospective / International (CIBMTR, EBMT)	1942 / 52 (18–77)	40 (1–134)	PTCL-NOS 50% AITL 30% ALCL 20%	CR 50% PR 28% PD 19%	Haplo 12.2% MSD 47% MUD TCD+ 24% MUD TCD- 17%	MAC 33.5% RIC 65%	NRM at 3 years Haplo 22% MSD 21% MUD TCD+ 24% MUD TCD- 23%	PFS at 3 years Haplo 50% MSD 50% MUD TCD+ 48% MUD TCD- 52%	At 3 years Haplo 60% MSD 63% MUD TCD+ 59% MUD TCD- 64%
Castagna, 2021 [54] Retrospective / Italy	68 / 56 (20–70)	55 (12–116)	PTCL-NOS 37% AITL 26% ALCL 19% Other 17%	CR 74% PR 21% SD/PD 5%	Haplo 43% MSD 29% MUD 25% mMUD 3%	NMAC 40% RIC 60%	NRM at 4 years Haplo 7% MSD 0% (m)MUD 21%	PFS at 4 years Haplo 72% MSD 69% (m)MUD 68%	At 4 years Haplo 76% MSD 80% (m)MUD 68%
Mamez, 2020 [72] Retrospective / France, Belgium, Switzerland	147 / 49.5 (16–69)	72 (69–80)	PTCL-NOS 39% AITL 29% ALCL 15% Other 17%	CR 58% PR 21% PD 21%	MSD 45% MUD 36% mMUD 5% Cord 12% Haplo 2%	MAC 38% RIC 62%	TRM at 4 years 30% for CR/PR2+ 40% for PD	GRFS at 2-years 45% for CR/PR2+ 30% for PD	At 4 years 61% for CR/PR2+ 37% for PD
McIlroy, 2020 [73] Retrospective / United Kingdom	21 / 45 (24–72)	95	PTCL-NOS 35% AITL 20% ALCL 25%	CR/PR 90% PD 10%	MSD 52% MUD 48%	MAC 24% RIC 76%	NRM at 1-year 10%	NR	5-year OS 65%
Wulf, 2019 [74] Retrospective / Germany	84 / 50 (17–74)	14.5 (1.5–114)	PTCL-NOS 36% AITL 20% ALCL 18% Other 26%	CR 15% PR 42% SD 17% PD 26%	MRD 26% MUD 63% mMUD 11%	MAC 100%	NRM at 1-year 13.1% 3-years 32.3% 5-years 46%	PFS at 3 years 37% GRFS at 1 year 51%	3-year OS 38%
Beitinjaneh, 2015 [59] Retrospective / United States	35 / 43 (22–73)	45 (9–90)	PTCL-NOS 43% AITL 3% ALCL 11% Other 43%	CR/PR 51% Refractory 49%	MRD 54% MUD 26% mMUD 20%	NMAC 28% MAC 72%	NRM at 4 years 40% for entire allo-SCT cohort	PFS at 4 years 28%	4-year OS 36% for all R/R 53% for CR2/3
Smith, 2013 [58] Retrospective / United States (CIBMTR)	126 / 38 (5–60)	NR	PTCL-NOS 50% AITL 10% ALCL 40%	CR1 14% CR2+ 16% PIF 36% PD 31%	MRD 60% MUD 24% mMUD 16%	MAC 59% RIC 36% Unknown 5%	NRM at 3 years MAC 34% RIC 27%	PFS at 3 years MAC 29% RIC 32%	At 3 years MAC 31% RIC 50%
Dodero, 2012 [57] Retrospective / Italy	52 / 47 (15–64)	67	PTCL-NOS 45% AITL 17% ALCL 21% Other 17%	CR 35% PR 40% PIF 25%	MRD 64% MUD 25% Haplo 11%	RIC 100%	NRM at 5 years 12%	PFS at 5 years 40%	At 3 years 50%
Zain, 2011 [75] Retrospective / United States	24 / 40 (7–72)	49 (16–100)	PTCL-NOS 33% AITL 17% ALCL 25% Other 25%	CR 25% PR 21% PD 21% PIF 33%	MRD 79% MUD 21%	MAC 42% RIC/NMAC 58%	NRM at 1-year 19% 5-year 29%	PFS at 5 years 50%	At 5 years 60%
Jacobsen, 2011 [76] Retrospective / United States	52 / 46 (24–72)	49 (20–157)	PTCL-NOS 38% AITL 10% ALCL 12% Other 40%	CR1 19% CR2+ 25% PR 31% PD 10% PIF 15%	MRD 46% MUD 37% mMRD 4% mMUD 13%	MAC 60% RIC 40%	NRM at 3 years RIC 14% MAC 36%	PFS at 3 years 45%	At 3 years 52%
Le Gouill, 2008 [56] Retrospective / France	77 / 36 (12–61)	43 (3.5–195)	PTCL-NOS 35% AITL 14% ALCL 35% Other 16%	CR 40% PR 30% SD/PD/PIF 30%	MRD 78% MUD 13% mMUD 9%	MAC 74% RIC 26%	TRM at 5-years 34%	EFS at 5 years 53%	At 5 years 57%

TRM: treatment related mortality; NRM: non-relapse mortality; DFS: disease free survival; EFS: event free survival; PFS: progression free survival; GRFS: graft versus host disease-free relapse-free survival; OS: overall survival; PTCL-NOS: peripheral T-cell lymphoma not otherwise specified; AITL: angioimmunoblastic T-cell lymphoma; ALCL: anaplastic large cell lymphoma; CR: complete response; PR: partial response; SD: stable disease; PD: progressive disease; PIF: primary induction failure; MSD: matched sibling donor; MUD: matched unrelated donor; mMUD: mismatched unrelated donor; Haplo: haploidentical donor; mMRD: mismatched related donor; MUD TCD-: matched unrelated donor SCT without in vivo T-cell depletion; MUD TCD+: matched unrelated donor SCT with in vivo T-cell depletion; MAC: myeloablative conditioning; NMAC: non-myeloablative conditioning; RIC: reduced intensity conditioning; NR: not reported.

A recent collaboration by the European Society for Blood and Marrow Transplantation (EBMT) and the Center for International Blood and Marrow Transplant Research (CIBMTR) evaluated the impact of donor source on outcomes for nearly 2,000 patients with PTCL who received allo-SCT. At a median follow up of 38 months, the 3-year overall survival was approximately 60% and 3-year PFS was approximately 50%. Patient with AITL had better outcomes with allogeneic transplant compared to non-AITL histologies. Patient were also categorized by graft source: post-transplant cyclophosphamide (ptCY) based haploidentical (haplo-SCT), matched sibling donor (MSD), matched unrelated donor with in vivo T-cell depletion (MUD TCD+), and MUD without T-cell depletion (MUD TCD-). Among the four donor cohorts, there was no significant difference in 3-year OS or PFS, incidence of relapse, or non-relapse mortality [53]. These findings are in line with single-center experiences [54], and support the use of haplo-SCT with ptCY for patients with R/R PTCL. Similar to prior studies [55–59], they found active disease at the time of transplant and decreased performance status to predict for worse outcomes. While T-cell lymphoma was the most common cause of death in this series, infections accounted for 19% of deaths.

A separate multicenter analysis of academic centers in the United States including 508 patients showed that overall survival following allogeneic transplant was 59.1% at 2 years and 50.8% at 5 years [60]. Similarly, PFS was 45.8% at 2 years and 39.5% at 5 years. Outcomes for those with AITL had a trend towards superior 5-year PFS compared to PTCL-NOS and ALCL (47.3% versus 52.3%). One-year treatment related mortality (TRM) was 11.2%. Three-year TRM was 22% and did not differ based on graft type. Disease status at the time of transplant was associated with outcome, emphasizing the need for adequate disease control at the time of transplant. The 3-year PFS for patients in CR, PR or with refractory/untreated relapse was 57%, 47% and 36% respectively in this series and the respective 3-year OS estimates were 68%, 59% and 49%.

Recently, a meta-analysis was performed which aggregated data from 30 trials published between 2001 and 2020 to analyze the benefit of transplant for patients with R/R PTCL. The study included 880 patients who underwent allo-SCT and 885 patients who underwent auto-SCT. In the allo-SCT group, nine trials with a total of 388 patients were assessed for 3-year OS and demonstrated pooled 3-year OS of 50% for all patients with R/R PTCL treated with allo-SCT. The auto-SCT group included nine studies with 614 patients, resulting in pooled 3-year OS of 55%. While this would suggest that auto-SCT was superior to allo-SCT, the analysis included patients undergoing auto-SCT in first remission and known chemosensitive disease as well. Therefore, when considering these confounding factors, allo-SCT prolonged the 3-year OS overall, especially for patients who did not achieve CR prior to transplant [61]. Patients that received allo-SCT were more often chemo-refractory and allo-SCT served as salvage therapy to provide an additional survival advantage in this subgroup.

Retrospective studies directly comparing outcomes for patients that received auto-SCT or allo-SCT found little to no relative benefit of allo-SCT over auto-SCT in the R/R setting as the value from the graft-versus-lymphoma effect was offset by the increased TRM [58,59,62]. The 3 to 5-year TRM quoted in these studies ranges from 33 to 40% for allo-SCT and 6 to 17% for auto-SCT. These relatively high rates of TRM reflect transplant practices from over 20 years ago, and therefore do not accurately represent modern prophylaxis and supportive care algorithms associated with improved outcomes [63].

CONCLUSION

Overall, randomized controlled trials in PTCL regarding the role of transplant are lacking and much of our current treatment approach is based on phase II trials, retrospective series, and expert opinion. While the backbone of induction remains CHOP-like chemotherapy, the incorporation of brentuximab vedotin to frontline treatment algorithms will continue to improve outcomes for patients with ALCL. Multiple studies have now shown similar results regarding the use of auto-SCT in first remission which improves PFS and OS by about 20% compared to those not treated with an intent to transplant approach. Given that we have limited tools to help select patients for transplant, organizations have suggested universally offering consolidative transplant for eligible patients in first remission [27]. Unfortunately, most patients with PTCL will relapse after induction or be refractory to initial therapy. For these patients, the best hope for long-term survival is achieving disease control and then proceeding to allo-SCT. Recent retrospective data has demonstrated that haplo-SCT with ptCY produces equivalent results as traditional donor sources and should be utilized in this setting. While progress is being made, the paucity of quality evidence limits a better understanding of this area of lymphoma and more research is needed. Given the increasing data suggesting the separate biologic underpinnings of different histologies of PTCL, in the future, we hope that we will be able to personalize therapy based on subtype as well as other biomarkers such as minimal residual disease, mutational profiling and radiographic parameters.