Purpose: Primary mediastinal large B-cell lymphoma (PML) has clinicopathologic features distinct from those of other diffuse large-cell lymphomas. However, the optimal treatment for this tumor is evolving, and in particular, the role of radiation therapy remains undefined. We conducted a retrospective review to evaluate the role of radiation therapy in this disease. Methods and Materials: The medical records of 40 consecutive patients with Ann Arbor Stage I or II PML treated at our institution from January 1980 to December 1995 were reviewed. There were 18 patients with Stage I disease and 22 patients with Stage II disease; 62.5% were women and 37.5% were men. The median age was 32.4 Sears (range, 17-74 years). The tumor scores were 0 in 1 patient, I in 5 patients, II in 13 patients, III in 7 patients, IV in 4 patients, and unknown in 10 patients. The International Prognostic Index (IPI) was 0 in 10 patients, I in 26 patients, II in 2 patients, and unknown in 2 patients. All patients were treated with doxorubicin-based chemotherapy, and 35 patients received radiation therapy. For most patients who received radiation therapy, an involved field or a modified-mantle field was used, and a dose of 40 Gy in 20 fractions or 39.6 Gy in 22 fractions was administered. Univariate analysis was performed to identify prognostic factors. Results: The median follow-up in surviving patients was 56 months (range, 19-194 months). The actuarial 5-year relapse-free survival (RFS) rate and overall survival (OS) rate for all patients were 67% and 72%, respectively. Thirty-five patients achieved a complete response; 32 of these patients received radiation therapy. The patterns of failure for the complete responders were as follows: locoregional failure alone for 1 patient (at the margin of the radiation field); distant failure alone for 5 patients; and both locoregional (in-field) and distant failure for 1 patient. There were no failures after 2.5 years. None of the 5 patients who never achieved a complete response had local control, and all died,vith disease. Only 2 of the 5 completed the planned course of radiation therapy; both had massive mediastinal disease. There was no treatment-related death from the initial chemotherapy or radiation therapy. One patient developed a second malignancy (sarcoma) within the radiation field after 13 years. The tumor score was a significant predictor of RFS (p = 0.016) and OS (p = 0.006), but the IPI did not prove to be a significant predictor. Conclusion: We recommend consolidative radiation therapy in view of the excellent local control and the lack of Significant toxicity. Modified mantle or involved field appears to be an adequate volume, and 39.6-40 Gy appears to be an adequate dose. The tumor score is a significant prognostic factor. (C) 2000 Elsevier Science Inc.
