The aim was to build a prognostic model to stratify patients at diagnosis into different risk categories. We investigated the prognostic value of functional PET parameters and clinical features in 64 primary breast lymphoma (PBL) patients. With a median follow-up of 60 months, 5-year progression-free survival (PFS) and overall survival (OS) was 62.5% and 73.4%. In multivariate analysis, baseline total metabolic tumor volume (TMTV0) and β2-microglobulin remained more reliable predictors of survival than other prognostic factors. The optimal TMTV0 cut-off value was 90 cm . Among 29 patients with high TMTV0, 5-year PFS and OS were 44.8% and 62.1%, respectively, while 5-year PFS and OS of 35 patients with low TMTV0 were 74.3% and 85.7%, respectively. TMTV0 combined with β2-microglobulin identified three groups with very different prognosis, including low-risk group with low TMTV0 and β2-microglobulin≤normal (n = 30), intermediate-risk group with high TMTV0 or β2-microglobulin>normal (n = 20), and high-risk group with high TMTV0 and β2-microglobulin>normal (n = 14). In the three groups, 5-year PFS rates were 80%, 55% and 28.6% (P = 0.003), and 5-year OS rates were 90%, 65% and 50% (P = 0.023) respectively. We established a new prognostic model through TMTV0 and β2-microglobulin, and can divide PBL at diagnosis into different risk categories. This article is protected by copyright. All rights reserved.This article is protected by copyright. All rights reserved.