Cumulative Incidence, Predictors, and Outcomes of Transformation to Diffuse Large B-Cell Lymphoma in a Prospective Cohort of Patients with Marginal Zone Lymphoma
Côme Bommier,Brian K. Link,Brianna Gysbers,Matthew J. Maurer,Melissa C. Larson,Arushi Khurana,Yucai Wang,Carrie Thompson,Dai Chihara,Juan Pablo Alderuccio,Jean L. Koff,Izidore S. Lossos,Grzegorz S. Nowakowski,Andrew L. Feldman,Thomas M. Habermann,Rebecca L. King,James R. Cerhan
DOI: https://doi.org/10.1182/blood-2023-186626
IF: 20.3
2023-11-28
Blood
Abstract:Background: Although marginal zone lymphoma (MZL) typically exhibits an indolent course, transformation to diffuse large B-cell lymphomas (DLBCL) is associated with a poorer outcome and remains a clinical challenge in managing patients with MZL. The limited existing literature reports various cumulative incidence rates, ranging from 2.5% and 4.7% (Finnish Cancer Registry, 1995-2018) to 6.6% and 8.4% (Florida Cancer Registry, 1995-2016) at 5 and 10-years, respectively, likely in part driven by varying study designs, ascertainment, and study population. Similarly, reported outcomes after transformed MZL (tMZL) have been variable. We evaluated the cumulative incidence of histological transformation (HT) to LBCL in a well-characterized prospective cohort of MZL patients from the upper Midwest, USA, and overall survival (OS) after transformation. Methods: Patients ≥ 18 years old with newly diagnosed MZL were prospectively enrolled in the University of Iowa/Mayo Clinic Molecular Epidemiology Resource (MER) cohort from 2002-2015. Clinical and treatment data were abstracted from medical records using a standard protocol. Patients were actively followed for retreatment, transformation, and death through 2022. Pathology was classified according to WHO 4/4R, including all transformations. The cumulative incidence of HT to LBCL was determined using death as a competing risk. Time to HT (TTHT) was defined as time from the date of MZL diagnosis to the date of transformation. We evaluated potential predictors of HT using a stepwise approach, and derived hazard ratios (HR) and 95% confidence intervals (CI) from Cox proportional hazards models. To assess predictors of OS after HT, patients ≥ 18 years old with newly diagnosed tMZL (into DLBCL) at Mayo Clinic between 1995 and 2002 (N=7) were combined with the MER cases. Results: We identified 524 patients with MZL in the MER cohort: 362 extranodal (EMZL; lung 18%, stomach 16%, orbit 13%, parotid 10%, skin 10%), 84 splenic (SMZL), and 78 nodal (NMZL). The median age at diagnosis was 63 years (range 18-92); 56% were female, and 83% were non-Hispanic white. During a median follow-up of 18.8 years (95% CI 16.7, NA), there were 25 HT (including 8 patients under observation) and 158 deaths. Eight Ten-year OS was 67% and 77% in the groups with and without HT, respectively. The overall cumulative incidence of HT was 1.5% at 2 years after diagnosis, 2.9% at 5 years, 4.5% at 10 years, and 5.7% at 15 years (Panel 1). The cumulative incidence of HT varied by subtype (e.g., at 5 years: 2%, 4% and 6%, in EMZL, NMZL and SMZL, respectively; Panel 2). The median time to HT also varied by subtype: 76 months (interquartile range (IQR) 43, 81) for EMZL, 30 months (IQR 16, 52) for SMZL, and 18 months (IQR 14, 23) for NMZL. In univariable analysis, predictors of HT were Stage III/IV (HR=3.4, 95% CI 1.36-8.53, p=0.009), elevated LDH (HR=2.98, 95% CI 1.25-7.12, p=0.014), and presence of ≥2 extranodal sites (HR=2.69, 95% CI 1.22-5.94, p=0.014). In multivariable analysis, elevated LDH was the only non-composite predictor of HT (HR=2.65, 95% CI 1.08-6.52, p=0.034). MALT-IPI score 2-3 (HR=3.61, 95% CI 1.06-12.25, p=0.040) and FLIPI score 3-5 (HR=2.19, 95% CI 1.22-8.22, p=0.018) also predicted HT. We analyzed OS after HT in 32 cases of tMZL (21 tEMZL, 7 tSMZL, 4 tNMZL). At HT, median age was 68 years (range 39-89); 24 (80%) patients presented with stage III/IV disease, 11 (58%) with elevated LDH, 12 (55%) with hemoglobin <12g/dL, 3 (14%) with platelet count <100x10 9/L and 9 (39%) with biopsy-proven bone marrow involvement. Twenty-six (81%) were treated with a systemic therapy, mainly CHOP-backbone therapy (R-CHOP n=13, CHOP n=4, R-EPOCH n=3, lenalidomide-R-CHOP n=1). Among these patients, the overall response rate was 85% (n=22/26). Three of the 4 non-responders died within 1 year. At last follow-up, 18/32 (56%) patients had died, and 6/13 (46%) alive patients had relapsed (MZL 3/6, DLBCL 3/6). After HT, 2-, 4- and 10-year OS were 76%, 55% and 55%, respectively. In univariable analysis, the only predictor of OS at HT was age ≥70 (HR 3.57 ,95% CI 1.34-9.48, p=0.011). Conclusion: In this North American prospective cohort of mainly non-Hispanic White patients, we found a low cumulative incidence of HT in MZL of 2.9% and 4.5% at 5- and 10-year, which was consistent with the Finnish findings (Kalashnikov et al., 2023). At HT, age was associated with OS and there was a high response rate to CHOP-backbone therapies.
hematology