The Number of Extranodal Sites Matters! a Proposal for a Feasible Prognostic Model in Nodal PTCL. a Collaborative Study grupo De Estudio Latino-Americano De Linfoproliferativos (GELL) & T-Cell Brazil Project (TCBP)

Introduction

Peripheral T cell lymphoma (PTCL) accounts for 10-15% of all non-Hodgkin lymphomas. As previously described by our group (Fischer et al, 2023 ASH Meeting), Latin America has a distinct epidemiological PTCL distribution, with a high frequency of adult T cell leukemia/lymphoma and extranodal NKT cell lymphoma cases, likely influenced by distinct genetic profiles and viral epidemiology. Survival rates have not improved over the past 20 years, with an estimated OS of 40% at 3 years. Treatment advances have also been limited, except for BV-CHP which access is limited in low- and middle-income countries. Given the above, it is crucial to understand the risk factors related to poor outcomes in PTCL. The IPI, which includes extranodal (EN) site as a variable, has been validated for PTCL. However, the specific impact of EN involvement in nodal PTCL (such as PTCL-NOS, AITL, ALCL ALK+/ALK-) and its biological implications remain unclear. Simplification could improve the reproducibility and applicability of these models, especially in low- and middle-income countries.

Objective

We aim to evaluate the number of EN sites in nodal PTCL lymphomas as a risk factor or surrogate for outcomes such as OS and PFS in a Latin America cohort.

Methodology

We compiled data from patients (pts) aged ≥18 years with newly diagnosed nodal PTCL-NOS, AITL and ALCL ALK+/ALK-, from the retrospective registry of the Grupo de Estudio Latinoamericano de Linfoproliferativos (n=339, 2000-2023), the T-cell Brazil Project (n=427, 2015-2022, ambispective). Clinical information included coded patient and site identifiers, gender, age, date and site of diagnostic biopsy, sites of disease, EN (number) and Ann Arbor stage. Initial therapy and response; details on response and relapse rates; subsequent lines of therapy and vital status were also recorded. Our primary endpoints were OS and PFS. REDcap Platform (by Vanderbilt) has been used to collect and store data and for descriptive analysis the IBM-SPSS version 24 was applied. Kaplan-Meier method estimated the OS, whereas Log-Rank tests to compare its curves. This trial is registered at Clinical trials (NCT03207789).

Results

We enrolled 766 pts (427 TCBP; 339 GELL). We grouped pts according to the number of EN involvement: No EN (n=383); one EN (EN1 n=168) and ≥2 EN involvement (EN2 n=215). In the entire cohort, there was a slight male predominance (61%); median age 56 years (range 15-92 y/o); 74% stage III/IV; 69% IPI ≥2. Most frequently histology was PTCL-NOS (60%), followed by ALCL ALK- (19%) and AITL (12%). The majority had B symptoms (61%) and elevated LDH (55%). CHOEP was the most frequent chemotherapy used (47%), followed by CHOP (34%). Transplant was performed as consolidation in only 16% of the cases. Almost half of patients achieved complete response (CR) after first line (47%), although 33% relapsed. Clinical characteristics were similar between No EN, EN1 and EN2 groups except for stage III/IV (58% vs 79% vs 96%; p<.0001); IPI ≥2 (58% vs 59% vs 99%; p<.0001); ECOG>1 (58% vs 92% vs 99%; p<.0001); bone marrow involvement (16% vs 24% vs 63%%; p <.0001), elevated LDH (54% vs 47% vs 61%; p=.03) and albumin <3.5 g/dL (33% vs 38% vs 58%; p <.0001), respectively. As previously reported, better OS and PFS were seen in ALCL ALK+ compared to ALK- and PTCL-NOS. Among No EN, EN1 and EN2 groups, the CR rates after 1^st line (54% vs 46% vs 37%, p=0.002), 60-month OS (45% vs. 44% vs. 27%, p <.0001) and PFS (31% vs. 34% vs. 18%; p <.0001) were inferior in the EN2 group, suggesting the number of EN sites involved assessed by CT and PET-CT matters and may be a feasible clinical surrogate for outcomes in this patient population.

Conclusion

PTCL remains an unmet medical need due to suboptimal treatment responses and outcomes. Thus, new biomarkers are essential to better stratify patients’ risk for PTCL mortality. In this large cohort of PTCL patients, we found the number of EN involvement as a useful and practical prognostic tool to predict response to therapy and outcomes. EN2 was associated to inferior response rates and shorter OS and PFS. Our findings indicate distinct biological characteristics in EN PTCL presentation that merit further molecular investigation, potentially enhancing tailored therapies. Registries are essential given the rarity and poor prognosis associated with these diseases as well as for hypotheses generation.