Impact of extranodal involvement in stage I aggressive B-cell lymphoma in the rituximab era: An individual patient data meta-analysis of two Phase III dshnhl/GLA trials

Abstract Background: The prognostic impact of extranodal stage I large B-cell lymphoma (LBCL) in the rituximab era remains unclear, and the role of consolidative radiotherapy (RT) for all patients without primary refractory disease is debated. We performed an individual patient data meta-analysis of two prospective phase III DSHNHL/GLA trials for patients <61 years old (FLYER, UNFOLDER) to characterize clinical outcomes of extranodal stage I LBCL and evaluate the role of RT. Methods: Patients with stage I LBCL treated within the FLYER (18-60 y, IPI:0, no bulk (≥7.5cm); NCT00278421) or UNFOLDER (18-60y, IPI: 0 w bulk, or IPI: 1 w/wo bulk; NCT00278408) trials were included. All patients received R-CHOP-based chemoimmunotherapy (4/6 cycles); in UNFOLDER administration of RT was randomized. In both trials patients with testicular lymphoma were assigned to RT per protocol. Primary endpoints for this analysis were overall survival (OS) and progression-free survival (PFS), compared between nodal and extranodal presentations as well as for RT both planned and unplanned using Kaplan-Meier and Cox regression analyses. Subgroup analyses evaluated the impact of bulky disease (≥7.5cm), age-adjusted (aa)IPI, LDH, extranodal sites with n≥10 in the whole cohort as well as in the extranodal only cohorts. Multivariate analyses were performed. Results: Of 1283 patients, 505 (39%) had stage I disease and were selected for further analysis; of those, 137 (27%) had extranodal involvement. Median follow-up was 66 months. Common extranodal sites included bone (n=18, 13.1%), thyroid gland (n=11, 8%), mouth region (n=10, 7.3%), soft tissue (n=10, 7.3%), stomach (n=10, 7.3%) and testis (n=8, 5.8%). Extranodal disease was not associated with inferior OS (HR 0.8, 95% CI: 0.3–1.9; p = 0.6) or PFS (HR 0.9, 95% CI: 0.5–1.7; p = 0.86) compared to nodal disease in a multivariate model containing aaIPI and bulk. Elevated LDH was identified as an independent predictor of worse OS in stage I patients (HR 2.4, p = 0.02) with LDH, bulk, age and treatment arm in the multivariate model. Further we evaluated the effect of planned RT applied or not according to protocol or due to other reasons besides treatment failure or excessive toxicity. In a multivariate model including planned RT per the above definition, aaIPI, bulk, extranodal sites with n≥10 and testis, OS was not statistically different with RT (HR 2.2, p= 0.09) whereas PFS was worse (HR 2.6, CI 1.1–5.9; P = 0.02). Results were similar when examining RT due to any cause vs no RT. Sensitivity analyses were performed which excluded a) all PMBCL and b) testicular LBCL. Upon exclusion of testicular involvement this PFS difference is no longer statistically significant (HR 2.3 p= 0.05). Exclusion of PBCML did not affect the results. Cumulatively RT both planned and unplanned did not significantly improve OS or PFS in extranodal stage I disease. In the subgroup of extranodal disease, soft tissue involvement was significantly associated with inferior PFS (HR 6.6; 95% CI 1.3–31.9; P = 0.018) and OS (HR 13.5; 95% CI 1.1–158.6; P = 0.03) compared to other in a multivariate analysis including aaIPI and bulk. Testicular involvement was similarly associated with a trend toward inferior PFS/OS compared to nodal disease; however, these differences did not reach statistical significance (PFS: P = 0.15; OS: P = 0.06 Conclusion: In this meta-analysis of two randomized trials, extranodal involvement in stage I LBCL was not independently associated with inferior survival in the rituximab era. Routine RT may not be required for most extranodal sites except for testicular involvement. Prospective studies using modern PET-guided strategies are warranted to refine site-specific treatment approaches for extranodal stage I LBCL.