P1122: a prospective, non-interventional study of real-world treatment and outcome in secondary cns lymphoma

Topic: 19. Aggressive Non-Hodgkin lymphoma - Clinical Background: Secondary central nervous system lymphoma (SCNSL) confers a dismal prognosis and treatment advances are constrained by the lack of prospective studies and real-world treatment evidence. Aims: To determine real-world patient characteristics, treatments and outcomes of SCNSL. Methods: Patients with SCNSL of all entities were included at first diagnosis from June 2011 to June 2022 and patient characteristics, treatment data, and outcomes were prospectively collected in the Secondary CNS Lymphoma Registry (SCNSL-R) (NCT05114330). Kaplan Meier estimator was applied to estimate OS for baseline covariables at the time of SCNSL diagnosis, and log-rank test was used for comparison of survival curves. To assess, if HDT-ASCT had a positive effect on OS in patients achieving a CR or PR upon induction therapy, we applied a multivariable Cox proportional hazards model controlling for known prognostic factors in SCNSL by modelling HDT-ASCT treatment as a time-dependent covariable. Results: 279 patients from 47 institutions in Germany, Switzerland, Israel and Austria were enrolled and 243 patients (median age: 66 years; range: 23-86) were available for analysis. Of those, 49 (20%) patients presented with synchronous (cohort I) and 194 (80%) with metachronous SCNSL (cohort II). The predominant histology was diffuse large B-cell lymphoma (DLBCL, 68%), followed by follicular lymphoma grade I-IIIA (n=16, 7%) and mantle cell lymphoma (n=13, 5%). At initial diagnosis before CNS involvement, cohort II patients most frequently presented with Ann Arbor stage IV (n=110; 57%) and extranodal involvement (n=156, 80%). Bone marrow was the most commonly affected extranodal site at initial lymphoma diagnosis in cohort II (n=56, 29%), followed by involvement of kidney/adrenal, gastrointestinal (each n=25, 13%), lung/pleura, and testes (each n=23, 12%) Median overall survival (OS) from diagnosis of CNS involvement was 17·2 months (95% CI 12-27·5), with longer OS in cohort I (60·6 months, 95% CI 45·5-not estimable (NE)) than cohort II (11·4 months, 95% CI 7·8-17·7). Predominant induction regimens included R-CHOP/high-dose MTX (cohort I) and high-dose MTX/cytarabine (cohort II). Rituximab was used in 166 (68%) of patients with B-cell lymphoma. The overall response rate to induction therapy was 54%, with 84 (35%) and 46 (19%) patients achieving a complete response (CR) and partial response (PR), respectively. When CNS relapse of DLBCL was accompanied by a systemic relapse, patients had inferior outcomes compared to those with isolated CNS involvement (17·2 months (95% CI 9·7-30·6) vs 6·6 months (95% CI 3·8-9), p=0·0026). A total of 103 patients (42%) underwent HDT-ASCT, with BCNU/Etoposide/Thiotepa being the most prevalent regimen in 25 (24%) patients. Patients who achieved partial response (PR) or better benefitted from consolidating high-dose therapy and autologous hematopoietic stem cell transplantation (HDT-ASCT) (HR adjusted 0·49, 95% CI 0·24-0·97, p = 0·0418). Summary/Conclusion: This study is the largest prospective cohort of SCNSL patients providing a comprehensive overview of an international real-world treatment landscape and outcomes. Prognosis was better in patients with SCNSL involvement at initial diagnosis (cohort I) and consolidating HDT-ASCT prolonged survival in patients with PR or better. For rational clinical trial design, a precise understanding of the real-world treatment landscape of SCNSL is crucial and provided by our prospective registry study. Keywords: Real world data, CNS lymphoma, Clinical trial, DLBCL