Rituximab‐containing combined modality therapy in limited stage follicular lymphoma: mature follow up and derivation of a novel prognostic score from the trog99.03 trial

Introduction: The TROG99.03 represents the only randomised phase III trial of combined modality therapy (CMT) in limited-stage follicular lymphoma ‘LSFL’, reporting a prolonged progression-free survival (PFS) in the CMT arm (MacManus, JCO, 2018). Here, we report extended follow up of this study providing mature data of patients treated with a rituximab-containing CMT regimen and the development of a new gene-expression based prognostic score. Methods: Patients with LSFL, grade 1–3a were randomised (1:1) to either involved-field radiotherapy alone (IFRT) (30–36Gy) or to CMT consisting of identical IFRT followed by 6 cycles of CVP. Reflecting evolving clinical practice, from 2006 onwards (i.e., ‘modern-era’), PET staging was increasingly utilised, and rituximab added to the CMT arm. Digital multiplex gene expression by Nanostring was performed on diagnostic biopsies based on genes previously identified to differentiate LSFL from advanced stage FL ‘ASFL’ (AM Staiger, Blood, 2020). Results: 150 patients were recruited between 2000 and 2012 with 31/75 patients in each arm recruited in the ‘modern-era’. At median follow-up 11.3 years, PFS remained superior for CMT compared to RT (HR 0.6; p = 0.043). Although no significant difference in OS was observed (HR 0.45, p = 0.11), compared with IFRT, patients in the CMT arm experienced fewer composite (deaths and histological transformation ‘HT’) events (HR 0.25; p = 0.045). With additional follow up no new non-malignant late toxicities were observed and incidence of secondary malignancies were similar between both arms (11 IFRT vs. 10 CMT, p = 0.99). Patients treated with a rituximab regimen (i.e., IFRT+R-CVP) had a markedly superior PFS compared to those treated without rituximab (i.e., IFRT, or IFRT+CVP), 8 year PFS rates 81% versus 52%, HR 0.42 p = 0.013 (Figure 1). Amongst PET staged patients the difference between R-CVP/IFRT versus IFRT increased (HR 0.35 p =0.027) suggesting this effect was not due to stage migration. No clinical factors were significantly associated with PFS on multivariate analysis. Nor were prognostic associations found for expression level of any individual genes. However, by penalised Cox regression an 8-gene Lasso-weighted prognosticator was identified, termed the ‘Bio-LSFL-score’. Genes (CACNA2D2, CD69, GZMB, IL7R, MYCT1, SLP1, TNFRSF14, TNFRSF25) reflected both B-cells and the microenvironment. The Bio-LSFL-score was highly significant for PFS (HR 0.25, p < 0.0001) with 100% patients in the high-risk group relapsing before 8 years. Conclusions: Particularly when incorporating rituximab, LSFL patients demonstrated significant increases in PFS and reduction in the rates of death and/or HT when treated with CMT compared with IFRT alone. A novel gene expression prognosticator was identified which in this trial cohort showed ‘ASFL-like’ behaviour by identifying LSFL patients unlikely to experience durable remissions. Keywords: diagnostic and prognostic biomarkers, indolent non-Hodgkin lymphoma, radiation therapy Conflicts of interests pertinent to the abstract C. Cheah Consultant or advisory role: Roche, Janssen, Gilead, Astra Zeneca, Lilly, TG Therapeutics, Beigene, Novartis, Menarini, Daizai, Abbvie, Genmab, BMS Research funding: BMS, Roche, Abbvie, MSD, Lilly J. F. Seymour Consultant or advisory role: Abbvie, Astra Zeneca, Celgene/BMS, Genentech, Genor Bio, Gilead, Janssen, Morphosys, Roche, Sunesis, TG Therapeutics Research funding: Celgene/BMS M. K. Gandhi Research funding: Beigene, Janssen