Clinical Impact of the Intensity of Induction Regimens Prior to Consolidation in Treatment-Naïve Patients with Peripheral T-Cell Lymphoma (PTCL)

Introduction: Frontline treatment of peripheral T-cell lymphoma (PTCL) remains a challenge because, despite good overall response rate (ORR 60-70%) to chemotherapy, more than 50% of patients (pts) relapse quickly and die of their disease. (Mehta-Shah, ASH 2019). Two prospective clinical trials (D’Amore, JCO 2012 – non-randomized study- and Horwitz, Lancet 2019 – randomized) demonstrated improved outcomes by adding Brentuximab Vedotin (BV-CHP) or etoposide (CHOEP) to the cyclophosphamide-vincristine-adriamycine-prednisone (CHOP) backbone in selected PTCL populations. Intensive induction chemotherapy (IIC) regimens commonly used in high grade B-cell malignancies (i.e. HyperCVAD) have also been used in PTCL based on phase I-II trial data (Abramson, Ann Oncol 2014). However, IIC regimens have not been compared to CHOP or CHOP+ (CHOEP/BV-CHP) in prospective studies.

Methods: Naïve PTCL treated at Roswell Park Comprehensive Cancer Center between 1997 to 2020 were identified. Pts were divided in three groups based on front-line therapy: CHOP vs. CHOP+ (CHOEP/BV-CHP) vs. IIC (HyperCVAD, ICE). We retrospectively analyzed pre-treatment characteristics, clinical outcomes (overall and complete response rate (ORR, CR), incidence rate ratio of relapse or death within 2 years of treatment (IRR), progression free (PFS), and overall survival (OS) at 5 years) and gross toxicity indices (treatment interruptions unrelated to progression of disease (POD), and causes of death). We used the chi square test and logistic regression to compare categorical variables, the Poisson distribution to compare IRR, log rank to compare time to event curves (PFS/OS). Analysis and graphs were done with SAS and STATA software programs.

Results: We identified 151 newly diagnosed PTCL treated at RPCCC. Twenty pts who received palliative therapy or comfort care were excluded from the study. We excluded 6 NK/T-cell lymphoma nasal type (NKTL), 5 hepatosplenic gamma delta TCL (HSTCL) and 1 Adult T-cell leukemia/lymphoma (ATLL) for which CHOP-like regimens are not standard of care. In addition, we also excluded 18 ALK+ Anaplastic large cell lymphoma (ALCL) from the analysis as this good prognosis group exclusively received CHOP or CHOP+, potentially confounding the analysis. A total of 101 pts were included in the analysis representing the following subtypes: angioimmunoblastic T cell lymphoma (AITL) (N=22, 21.8%), PTCL-not otherwise specified (NOS) (N=36, 35.6%), PTCL-with T follicular helper phenotype (TFH) (N=2, 2.0%) or ALK- ALCL (N=41, 35.6%). CHOP was used in 54 pts, 27 received CHOP+ (18 CHOEP, 8 BV-CHP, 1 EPOCH), and 20 received IIC chemo (17 HyperCVAD, 3 ICE). Consolidation with HDC-ASCT was performed in 7 pts after CHOP (14%), 5 pts after CHOP+ (18%) and 8 pts after IIC treatment (40%). Baseline characteristics including sex, age, comorbidities, performance status (ECOG), international prognostic index (IPI) and histology were balanced in the CHOP, CHOP+ and IIC chemo treatment groups, after we excluded ALK+ ALCL as described above (all p values > 0.4). ORR was 69% (61% CR), 86% (81% CR) and 80% (80% CR) in the CHOP, CHOP+ and IIC groups (p=0.3). A multivariate analysis including treatment type, ECOG performance status, Charlson comorbidity score and age did not show any significant correlation with ORR (lowest p value was 0.07 for the Charlson comorbidity score). Incidence rate ratio of progression or death was 3.1, 3.4 and 4.6 per 100-person months respectively for CHOP, CHOP+ and IIC chemo (proportions: 46%, 44%, 60% respectively). PFS and OS did not show significant differences among treatment groups (Log Rank test p = 0.90 for PFS and p = 0.79 for OS). The proportion of treatment interruptions without evidence of POD was 23% in the CHOP, 28% in the CHOP+ and 40% in the IIC group (p=0.31). Of all death events per group, POD was the cause of 43%, 41% and 36% deaths respectively in the CHOP, CHOP+ and IIC groups (p=0.92), death for treatment related toxicity occurred in 21% (CHOP), 41% (CHOP+) and 45% (ICC chemo) of all deaths per group, p=0.19.

Conclusions: In our population, the intensity of front-line chemotherapy does not seem to impact on IRR PFS or OS. CHOP+ and IIC might have a tendency towards better ORR and CR rates vs CHOP. However, IIC have a tendency (non-statistically significant) of an increase in adverse events leading to treatment interruptions unrelated to POD and deaths for toxicity.