Paper Details

ABSTRACT

Background Allogeneic hematopoietic cell transplantation (allo-HCT) can provide durable disease control in mature T- and NK-cell lymphomas, but long-term competing-risk outcomes, conditional survival, and the determinants of relapse versus nonrelapse mortality (NRM) remain incompletely defined. We evaluated a three-center U.S. cohort, focusing on pre-HCT disease status

Methods We retrospectively analyzed 140 patients with systemic mature T- or NK-cell lymphoma who underwent allo-HCT at Moffitt Cancer Center, Mayo Clinic, or the University of Miami between 2002 and 2025; primary cutaneous lymphomas were excluded. Overall survival (OS) and progression-free survival (PFS) were estimated by Kaplan-Meier methods, and relapse/progression and NRM as competing risks by Aalen-Johansen methods. Associations of pre-HCT complete response (CR), partial response (PR), or stable/progressive disease (SD/PD) with outcomes were evaluated using center-stratified Cox models and Fine-Gray models adjusted for center. A clinically specified fixed-covariate model was designated as the primary adjusted model; stepwise models were used as sensitivity analyses. A 2-year landmark analysis evaluated subsequent outcomes among patients alive and progression-free at 2 years. Analyses were exploratory, without multiplicity adjustment, and used complete cases.

Results Median age was 53 years (range 22-73), 59% were male, and patients had received a median of 3 prior lines. Histologies were peripheral T-cell lymphoma, not otherwise specified (PTCL-NOS) (43%), angioimmunoblastic T-cell lymphoma (22%), anaplastic large-cell lymphoma (14%), NK/T-cell lymphoma (9%), adult T-cell leukemia/lymphoma (ATLL) (5%), and other subtypes (hepatosplenic and enteropathy-associated T-cell lymphoma; 8%). Primary refractory disease was present in 51%, 29% had undergone prior autologous HCT, and reduced-intensity conditioning was used in 65%. Among 136 evaluable patients, pre-HCT status was CR in 60%, PR in 33%, and SD/PD in 7%.

After median follow-up of 6.3 years, 60 deaths, 27 relapses/progressions, and 43 nonrelapse deaths occurred. Five-year OS, PFS, relapse, and NRM were 62% (95% CI, 54-71), 54% (95% CI, 46-63), 20% (95% CI, 13-27), and 26% (95% CI, 19-34), respectively. Relapse plateaued at approximately 20% by year 3, whereas NRM continued to 34% (95% CI, 25-43) at 10 years.

Among 77 patients alive and progression-free at 2 years, conditional OS from HCT was 93% at 5 years and 79% at 10 years; subsequent relapse plateaued at 3%, whereas NRM increased to 19%. The pre-HCT status gradient was no longer evident: only 2 subsequent relapses occurred, both after PR (CR, 0 of 48; SD/PD, 0 of 4).

Pre-HCT disease status was specifically associated with relapse. Relative to CR, relapse risk was higher with PR (subdistribution hazard ratio [sHR], 4.79; 95% CI, 2.02-11.35; P<.001) and SD/PD (sHR, 6.46; 1.72-24.33; P=.006), with 5-year relapse incidences of 9%, 37%, and 40%, respectively. Disease status was not significantly associated with OS or NRM. In the fixed multivariable relapse model, PR (sHR, 4.17; 1.83-9.48; P<.001) and SD/PD (sHR, 9.02; 2.33-34.97; P=.002) remained adverse. In contrast, later transplant era was independently associated with improved OS (2016-2025 vs 2015 or earlier: HR, 0.36; 95% CI, 0.18-0.71; P=.003), whereas 3 or more prior lines (HR, 2.02; 1.12-3.63; P=.019) and rare/aggressive histology (ATLL, NK/T-cell, or other; HR, 2.10; 1.08-4.09; P=.030) were associated with inferior OS. Conditioning intensity was not associated with OS (reduced-intensity vs myeloablative: HR, 1.07; 0.57-1.98; P=.84).

Conclusions Allo-HCT produced durable disease control across a heterogeneous population of mature T- and NK-cell lymphomas. CR before transplantation was associated with substantially lower relapse, while relapse occurred predominantly early, and late failure among 2-year progression-free survivors was predominantly NRM. These findings support efforts to achieve optimal disease control before allo-HCT when feasible and emphasize reduction of late NRM as a complementary priority.