Primary Mediastinal (Thymic) Large B-cell Lymphoma: The Working Diagnosis vs WHO Classification Criteria

Abstract

Primary mediastinal (thymic) large B-cell lymphoma (PMBCL) is an uncommon but distinct entity in the WHO classification of hematologic malignancies. Accurately diagnosing PMBCL is challenging as it is a clinicopathological entity. There are no pathognomonic features. Many characteristics overlap with diffuse large B-cell lymphoma NOS, nodular sclerosis Hodgkin lymphoma, and so-called “grey-zone” lymphomas. Thus, it is likely that patients are occasionally misdiagnosed as having PMBCL. Accurate diagnosis is critical as it impacts systemic therapy selection and receipt of radiation therapy (RT). We sought to investigate the fidelity between a working diagnosis of PMBCL at our institution and the clinicopathological criteria established within the 2017 WHO classification. Patients treated for stage I-II PMBCL from 1998-2018 at our institution were retrospectively reviewed for clinical and pathological conformity with WHO criteria in collaboration with a hematopathologist who reviewed archival tissue. Patients were characterized as definitely having PMBCL if they had all of the following: an anterior mediastinal mass +/- regional lymph node involvement, no extranodal disease, B-cell antigen expression, and supportive feature(s)- female gender, young age, bulky primary tumor, CD30 weakly positive, compartmentalizing alveolar fibrosis, lack of surface immunoglobulin expression, and MUM1 or CD23 positivity. Patients without supportive features were characterized as equivocal. PMBCL was excluded if they had any of the following: no anterior mediastinal mass, distant lymph node involvement or extranodal disease, lack of B cell antigen expression, or EBV positivity. 63 patients were evaluated. 71% were female. Median age- 35 years (range, 20-73). Stage 1- 51%; stage II- 49%. Median mediastinal mass size- 10 cm (range, 3-20). 48% of patients received R-CHOP, 32% R-EPOCH, and 21% received other chemotherapy regimens; 79% received rituximab; 83% of patients received 6 chemotherapy cycles. 86% of patients had a post chemotherapy PET/gallium response assessment (78% CR; 22% PR). 64% of patients received RT with PR patients more likely to receive RT (83% vs 62%, p = 0.16). Median follow up was 6 years. Overall, 92% of patients met WHO criteria for PMBCL; 5% (n = 3) did not meet criteria, all due to a lack of an anterior mediastinal mass. 3% were equivocal- grey zone lymphoma (n = 1) and lack of supportive features (n = 1). 5-year PFS and OS was 67% (95% CI 54-77%) and 81% (95% CI 68-89%), respectively. Failure to receive rituximab was the only predictor associated with inferior PFS on univariate analysis, HR 2.79 (1.15, 6.78). Despite the complexity of the clinicopathological criteria of PMBCL, most (92%) patients who were treated for stage I-II PMBCL at our institution appear to have been accurately diagnosed. Treatment regimens varied significantly given changes in patterns of care over the time period. The importance of rituximab was confirmed.