Treatment outcome and prognostic factors for primary mediastinal (thymic) B-cell lymphoma: a multicenter study of 106 patients.

ABSTRACT

PURPOSE: To define clinicopathologic features, response to treatment, and prognostic factors of primary mediastinal B-cell lymphoma (MBL), a CD20+ tumor recognized as a distinct entity among non-Hodgkin's lymphomas (NHL). PATIENTS AND METHODS: One hundred six patients presented with disease confined to thorax (86%), bulky mediastinum (73%), superior vena cava syndrome (47%), and contiguous infiltration (57%). Ninety-nine received doxorubicin-containing chemotherapy (CHT). RESULTS: Thirty-five of 99 patients were primarily CHT-resistant, and 64 responded 23 achieved complete response (CR) and 41 achieved response with residual mediastinal abnormality. Seventy-seven percent of responders received mediastinal radiotherapy (RT). Of 64 responders, 18 (28%) relapsed none of 23 CR patients and 18 of 41 (44%) with residual mediastinal abnormality. Relapse-free survival rate of responders was 71% at 3 years. Actuarial 3-year survival rate was 52% for all patients and 82% for responders. Predictive factors of poor outcome were identified by logistic regression; Cox survival analysis was performed on death and relapse. Pericardial effusion (P < .001) and Eastern Cooperative Oncology Group (ECOG) performance status > or = 2 (P = .009) predicted nonresponse (NR) and affected survival. Less than partial midway response to CHT predicted NR to subsequent therapies. Bulky disease was related to persistent mediastinal abnormality and risk of relapse (P = .025). CONCLUSION: MBL is an aggressive NHL with unique clinicopathologic aspects, often refractory to current CHT designed for high-grade NHL. Poor performance status and pericardial effusion predict NR and poor survival. Inadequate response after the first courses of front-line CHT predicts failure of subsequent treatment. Responders with bulky mediastinum or residual mediastinal abnormality after CHT are at risk of relapse. These factors should help to select high-risk patients for intensive treatments.