INTRODUCTION We and others have demonstrated that patients with refractory nodal mature T-Cell lymphomas (nMTCL) have worse overall survival (OS) versus relapsed disease.1–3 Among patients who relapse, responses and OS in (nMTCL) remain poor with little data to tailor decision-making based on disease kinetics.1–9 Time-to-relapse (TTR) is a significant predictor of survival and decision-making tool among many other lymphomas,10–18 however a majority of nMTCL patients will relapse prior to 24 months.14 We aimed to characterize how TTR may affect survival and explore differential second-line (2L) therapy effects based on TTR subgroups.

METHODS

This was a global retrospective cohort study using multiple international cohorts: PETAL (n=1414) and GELL (n=487).1,3,19–21 Two separate independent cohorts validated TTR12 as an OS predictor: an observational U.S. multicenter cohort (n=138) and the phase 3 randomized trial of romidepsin-CHOP versus CHOP.22 Patients with PTCL-NOS, AITL/TFHL, or ALCL with a CR to 1L were included. Patients without progression or 2L within the study period were included as not having TTR12 per landmark methods,11,23–25 and were removed for sensitivity analyses. The primary objective was OS among nMTCL who relapsed or started 2L26 within 12m from 1L (TTR12) versus without TTR12 because nearly half of patients in our cohort and previous studies of nMTCL relapsed within 12m.6,14 The secondary objective was to compare OS among 2L novel agents (NA) versus chemotherapy (CC) in those with and without TTR12.

Kaplan-Meier and Cox PH methods were used adjusting for a priori covariates.1,3 The primary analysis was a modified-landmark analysis (m-LM) with OS measured from relapse or 2L start (TTR12 group) or from 12m from 1L start (without TTR12 group) to death. Sensitivity analyses used standard 12m landmark (s-LM; excluded patients who died or lost-to-follow <12m),11,25 and time-dependent Cox (td-Cox; OS 1L start to death). For 2L analyses, OS was measured from 2L start to death.

RESULTS A total 452 were included in the final cohort and for td-Cox, 428 in m-LM, and 388 in s-LM. Of the 452 total, 165 (36.5%) had TTR12, 181 (40%) relapsed ≥12m, and 106 (23.5%) never relapsed (total without TTR12: 287 [63.5%]). The median (range) age was 58 (18-89) and 60 (15-92), PTCL-NOS included 47.3% and 43.2%, AITL/TFHL included 35.8% and 32.1%, and ALCL included 16.4% and 24.7% of patients with TTR12 and without TTR12 respectively. Patients received similar frontline regimens and HSCT utilization was comparable. Demographics in the validation cohorts largely mirrored the primary cohort.

Using m-LM, TTR12 conferred worse OS (aHR 2.14, 95%CI: 1.58-2.90; p<0.001) overall, irrespective of 1L HSCT or PIT score, and across prespecified subgroups (PTCL-NOS: aHR 2.32, 95%CI: 1.51-3.55; p<0.001, AITL/TFHL: aHR 1.92, 95%CI: 1.15-3.21; p=0.013, ALCL: aHR 3.34, 95%CI: 1.18-9.50; p=0.023, no 2L HSCT: aHR 2.27, 95%CI: 1.63-3.15; p<0.001). TTR12 patients who received any 2L HSCT had similar OS than those without TTR12 (aHR 1.85, 95%CI: 0.73-4.65; p=0.194). When excluding patients that did not relapse, TTR12 still showed poorer OS (aHR 1.60, 95%CI: 1.18-2.18; p=0.003). In the observational validation cohort, TTR12 patients had worse OS across all analyses in both univariate (HR 3.72, 95%CI 1.85-7.46; p<0.001) and multivariate models adjusting for covariates (aHR 3.60, 95%CI 1.77-7.34; p<0.001). In the phase 3 validation cohort, TTR12 patients also had worse OS (aHR 3.71, 95%CI 2.17-6.32; p<0.001). All results were consistent across s-LM and td-Cox sensitivity analyses. TTR12 predicted worse OS in parallel to, and compounding with, PIT score, with patients who had TTR12 and PIT≥2 exhibiting the worst OS. Patients with PIT score of 4 were at a significantly higher odds of developing TTR12 (OR 3.64, 95%CI 1.07-12.38; p=0.038). In TTR12 patients, 2L NA significantly improved OS versus CC (aHR 0.60, 95%CI: 0.37-0.97; p=0.038). In those without TTR12 who received 2L, there were no significant differences in OS with NA versus CC (aHR 0.82, 95%CI 0.51-1.32; p=0.407

CONCLUSION TTR12 consistently exhibited worse OS in nMTCL and highlighted differential responses to 2L, defining a unique risk group after 1L. TTR12 may be a novel endpoint for clinical trials and may better inform treatment decisions upon progression. Prospective validation and correlation with molecular alterations has been initiated and is the focus of the PETAL consortium.

This content is only available as a PDF.
2025
Sign in via your Institution