Impact of Early Histological Transformation on Survival in Patients with Follicular Lymphoma

Abstract

Introduction: FL represents the second most common NHL with an indolent natural history and a median OS exceeding 20 y. Histologic transformation (HT) to an aggressive lymphoma is historically linked to poor outcomes with a median post-transformation survival of 2-5 y. EFS24 and POD24 are established adverse events indicating a poor overall survival. POD24 is associated with a higher incidence of HT following immunochemotherapy. However, the impact of HT within POD24 (HT24) remains unknown. Objective: Given the rarity of HT and the limited resources dedicated to this patient population, we aimed to assess and characterize the risk of HT and its outcomes, including the impact of HT24 on survival. Methods: This was a single-center retrospective observational study from the FL MSKCC database. Eligible pts were diagnosed between 01/1998-01/2022 and had a biopsy-proven grade 1-3A FL at diagnosis, with a subsequent biopsy eventually showing HT. Pts with any component of aggressive lymphoma defined as FL3B, DLBCL, or HGBCL within 3 months of diagnosis were excluded, as well as those with clinically suspected HT but not biopsy-proven. HT24 was defined as HT from 3-24 months from the initial FL diagnosis. POD24 was defined as progression within 24 months of initial diagnosis, irrespective of whether the patient underwent active surveillance or received therapeutic intervention. OS was defined as survival from the day of diagnosis. OS post-HT was defined as survival interval from HT biopsy. OS from a risk-defining event (Casulo, JCO 2015) was defined as survival from HT for HT24 or from 2 y after diagnosis for the reference group (all pts from the cohort without HT at any time). Results: 3,678 pts were identified as eligible for this analysis. At a median follow-up of 7.5 y (CI 95% 7.1,7.8), 415 pts (11%) had experienced HT. The risk of HT was associated with multiple high-burden disease features at diagnosis (i.e. advanced stage, 3 nodes above 3 cm, spleen involvement, higher FLIPI score, elevated LDH, Ki67, and higher SUVmax). In the entire cohort, the median time from initial diagnosis to HT was 4 y (CI95% 3.5; 4.7). The cumulative hazard of HT at 2, 5, and 10 y was 3%, 7%, and 10%, respectively, with an estimated rate of 1%/year over the first 10 y. During follow-up, 695 pts (19%) had died. As expected, median OS was significantly inferior for pts having experienced HT vs. non-HT pts (16 vs 22 y, p<0.001, HR 1.8). Median OS post-HT was 9.8 y (95% CI 7.2;12.5) with 2-, 5-, and 10-years OS of 73%, 61%, and 49%, respectively. Of the 415 pts with HT, 111 (27%) had experienced HT24, and those with HT24 had significantly more adverse initial characteristics ( Table) such as advanced stage, bulk, serous effusion, spleen involvement, high LDH and FLIPI score, anemia, grade 3A, and higher SUVmax. Interestingly, there were no significant risk factors specifically associated with HT>24 months. OS from a risk-defining event in HT24 was significantly inferior compared to the reference group (5-year estimate 66% vs 90%, p<0.001, HR 3.2). OS post-HT did not significantly differ between pts with HT24 vs HT>24 months (5-year estimates 66% vs. 58%, p=0.22). Similar results (51% vs. 55%, p=0.64) were observed when restricted to pts who received initial systemic therapy. POD24 from diagnosis was observed in 303 pts; of them, 192 (63%) maintained FL histology at progression (POD24 +/HT24 -). Pts with HT24 experienced a significantly lower 5-year OS compared to those POD24 +/HT24 - (66% vs. 75%, p=0.0023, HR 1.9), figure. Conclusions: Our study highlights the distinction between HT24 and HT occurring later. High-burden features at diagnosis may predict HT24, suggesting occult HT at diagnosis. We did not identify any predictors for later HT, suggesting it may represent a random event in FL history. In this sizeable single-center cohort of pts with FL, we observed a better outcome for pts with HT than previously described. Furthermore, while HT negatively impacted OS, post-HT survival outcomes were similar in pts with HT24 and those with HT after 24 mo. Strict selection criteria (excluding pts with FL 3B and those with HT documented within 3 months of diagnosis), recent progress in therapy, and tertiary center recruitment biases may account for these data. Patients with HT24 had a poorer outcome than those with POD24 without HT. Further analysis to delineate the pattern of first-line therapy and outcome of pts with HT and POD24 is in progress.