BackgroundThe most common subtype of B-cell lymphomas presenting in the conjunctiva is extranodal marginal zone lymphoma (EMZL), accounting for ~80% of cases. Most patients (pts) present with localized disease, and radiation therapy (RT) is the preferred treatment strategy. We aimed to retrospectively analyze our single-institution experience to provide further insight into the characteristics and long-term outcomes of conjunctival EMZL.MethodsWe evaluated 72 pts diagnosed with conjunctival EMZL between 01/1995 and 12/2020 at the University of Miami. Pts' characteristics included age, sex, TNM-AJCC ocular lymphoma and Ann Arbor staging systems, MALT-IPI score, treatment (RT, chemotherapy, rituximab), and treatment response (complete response (CR), partial response, stable disease, progression of disease). Primary endpoints were progression-free survival (PFS) and overall survival (OS), estimated using the Kaplan-Meier method, and compared using the log rank test and the univariable Cox regression analysis. We also performed a competing risk univariable analysis (UVA) assessing predictors of cumulative incidence of relapse/progression, with death without relapse as a competing risk, using the Fine and Gray regression analysis.ResultsAmong all 72 pts, mean age was 59.9 yrs (7-93) with 38 (52.8%) being >60 yrs old, 46 (63.9%) were female, 56 (77.8%) had unilateral conjunctival disease, localized disease (T1N0M0 and Ann Arbor stage I) was present in 63 (87.5%), 6 (8.3%) had disseminated disease with more than 1 extranodal site involved, and 29 (40.3%) had a MALT-IPI of 1 or 2. After biopsy and surgical removal, 65 (90.2%) received additional treatment, while 6 (8.3%) were followed only, and in one case therapy information was not available . RT was the most common treatment (53 pts [73.6%], with 46 [86.8%] of those treated with ≥ 30 Gy). In two of these patients RT was combined with chemotherapy. Other treated pts received immunochemotherapy (12 [16.7%], including rituximab in 8). Five pts with stage II-IV disease received systemic therapy. CR was achieved in 63 (87.5%) after first line treatment, and no high-grade lymphoma transformation was seen. With a median follow up of 6.67 yrs (0.56-24.13), there were 14 relapses (19.4%). Among 53 pts treated with RT there were 8 relapses (15.1%), only one (1.9%) within the RT field. There were 23 progression events (31.9%, 14 relapses and 9 deaths without documented relapse), and a total of 14 deaths (19.4%).Mean PFS and OS were 6.69 yrs (0.49-20.37, SD 4.95) and 7.81 yrs (0.56-24.13, SD 5.40), respectively. The 10-yr PFS and OS were 68.4% (95%CI 52.8, 79.8%) and 89.4% (95%CI 77.4, 95.2%), respectively. Variables associated with shorter PFS in UVA Cox model were age > 60 yrs (HR=2.93, 95%CI 1.08, 7.95; p=0.035), high MALT-IPI (1-2) (HR=2.42, 95%CI 1.01, 5.78; p=0.048), and use of chemotherapy only (HR=2.73, 95%CI 1.13, 6.56; p=0.025). Variables associated with shorter OS included age >60 yrs (HR=9.07, 95%CI 1.17, 70.26; p=0.035) and high MALT-IPI (HR=6.19, 95%CI 1.35, 28.33; p=0.019). CR after frontline therapy was associated with longer PFS (HR=0.13, 95%CI 0.04, 0.45; p=0.001) but not OS. PFS of MALT-IPI 0 vs 1-2 was significantly longer (p=0.042), with 10-yr PFS 80.9% (95%CI 63.4%, 90.6%) vs 55.6% (95%CI 32.1%, 73.8%). Similarly, longer OS was observed in MALT-IPI 0 pts (p=0.0077; 10-yr OS 95.2% [95%CI 82.2%, 98.8%] vs 80.6% [95%CI 55.6%, 92.4%]) (Figure).A subset UVA Cox analysis of patients with Ann Arbor stage I showed longer PFS associated with CR after frontline therapy (HR 0.15, 95%CI 0.04, 0.58; p=0.006) with no significant association with age >60 yrs, high MALT-IPI or use of chemotherapy. Variables associated with shorter OS were age >60 yrs (HR 8.01, 95%CI 1.00, 63.98; p=0.05) and high MALT-IPI (HR 13.41, 95%CI 1.67, 107.99; p=0.015), similarly to the primary analysis.On univariable Fine and Gray regression models with death without relapse/progression as a competing risk, RT (SHR=0.33, 95%CI 0.12, 0.96; p=0.041) and CR-post frontline therapy (SHR=0.11, 95%CI 0.03, 0.36; p<0.001) were associated with lower risk of relapse. Conversely, chemotherapy (SHR=3.47, 95%CI 1.20, 10.0; p=0.022) was associated with higher risk of relapse.ConclusionPatients with conjunctival EMZL exhibit excellent long-term survival, and RT remains the most effective frontline therapy. MALT-IPI appropriately identifies patients at risk for treatment failure. [Display omitted] DisclosuresAlderuccio: ADC Therapeutics: Consultancy, Research Funding; Oncinfo / OncLive: Honoraria; Puma Biotechnology: Other: Family member; Inovio Pharmaceuticals: Other: Family member; Agios Pharmaceuticals: Other: Family member; Forma Therapeutics: Other: Family member. Lossos: Lymphoma Research Foundation: Membership on an entity's Board of Directors or advisory committees; Seattle Genetics: Consultancy; NCI: Research Funding; Stanford University: Patents & Royalties; Verastem: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; NIH grants: Research Funding; University of Miami: Current Employment.