Purpose Primary breast lymphoma (PBL) is a group of lymphoproliferative diseases with obvious histological and clinical heterogeneity. Because of no extensive extranodal involvement and the usual confinement of the disease to the breast ± regional nodes, the application of international prognostic index (IPI) in PBL is limited. The purpose of the study is to investigate the prognostic value of PET-derived metabolic parameters and establish a prognostic model based on the PET parameters and clinical characteristics to classify patients with PBL at diagnosis into different risk categories.

Methods In this study, we retrospectively investigated 94 PBL cases treated at Tianjin medical university cancer institute and hospital between 2005 and 2021. The inclusion criteria for this study were that the diagnosis of PBL was based on the Wiseman-Liao criteria adopted in 1972. The disease is limited to unilateral or bilateral breast ± regional lymph nodes. There was no previous history of extramammary lymphoma, and the patient at diagnosis was not accompanied by widespread lymphoma at the same time.

Results A total of 94 patients with PBL were included in this study. The median age of patients at diagnosis was 52 years (range, 16-87 years). Right breast is more likely to be involved than left breast (62.8% vs 29.8%). In addition, seven patients had bilateral breast involvement at the time of diagnosis. Six patients (6.7%) had a poor performance status (ECOG≥2). At the time of diagnosis, 73.0% of patients had normal lactate dehydrogenase; 61.7% of patients have increased β2-microglobulin.

Of the 94 patients, 10 patients underwent only surgery without chemotherapy or radiotherapy. Most patients (89.4%) received chemotherapy, the most common of which was CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) or CHOP-like regimens. About 77.4% of patients received rituximab combined with chemotherapy. About 47.6% of patients received consolidation radiotherapy of the involved breast with or without lymph nodes in the ipsilateral area. With a median follow-up of 72 months, the 5-year OS and PFS rates were 61.9% and 72.3%, respectively.

We designed a targeted sequencing panel for PBL including 45 mutated genes described in the literatures. After sequencing 42 samples from the 35 patients included in the study, the recurrently mutated genes were CD79B (in 13/31 samples; 41.9%), KMT2D, MYD88 (11/31; 35.5%), CARD11 (8/31; 25.8%), PIM1 (6/31; 19.4%), and CREBBP (4/31; 12.9%). MYD88/CD79B-mutated (MCD) genotype accounted for 22.6% of PBL.

In the whole population, the median TMTV0 was 87 cm3 (range: 3-312 cm3). The ROC analysis showed that the optimal TMTV cut-off value for estimating PFS and OS was 90 cm3. The areas under the curves (AUCs) were 0.767 for PFS and 0.719 for OS. The sensitivities and specificities were 73.6% and 77.2% for PFS, and 71.3% and 73.2% for OS respectively. The prognosis of patients with high TMTV0 (TMVT0 > 90 cm3, n = 44) was significantly worse, with 5-year PFS and OS of 44.5% and 62.3%, respectively, while those with low TMTV0 (TMVT0 ≤ 90 cm3, n = 50) were 74.8% and 84.9% (P = .005 for PFS and P = .017 for OS). In multivariate analysis, TMTV0 also remained a more reliable predictor of survival than other prognostic factors including TLG (TMTV0: P < .001 for PFS and OS; TLG: P = .003 and P = .002 for PFS and OS individually). In this study, we also demonstrated that baseline high serum β2-microglobulin level and ECOG performance status predicted both shorter PFS and shorter OS.

We then constructed a new prognostic model of PBL based on serum β2-microglobulin level, ECOG performance status in combination with TMTV0. According to whether there are adverse factors, we divide patients into four risk categories. The prognosis of the four groups was significantly different, with 5-year PFS of 86.9%, 62.5%, 38.6 and 19.3% (P = .001) and 5-year OS of 92.4%, 70.6%, 52.1% and 29.4 (P = .013), respectively.

Conclusion In summary, this is the first study to stratify the prognosis of PBL by integrating the quantitative functional parameters of PET/CT, β2-microglobulin and ECOG performance score. Our research shows that TMTV0 is the most robust predictive index of outcome. Combined with β2-microglobulin and ECOG score, it can be used to accurately predict the prognosis of PBL patients. For patients with a high risk of progression or relapse, we may need more active treatment options to improve the prognosis of patients.

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution