Primary Soft-Tissue Lymphomas: MRI Features Help Discriminate From Other Soft-Tissue Tumors

Abstract

MRI presentation of extra-nodal soft-tissue lymphomas (STLs) is scarcely reported and lacks of comparison with other soft-tissue tumors (STTs) including sarcomas (STS). Yet, suggesting this diagnosis on MRI would considerably reduce diagnostic intervals. Our aim was to investigate if conventional MRI could discriminate STLs from other STTs. MRIs of STL patients were compared with those of patients addressed to a sarcoma reference center for the diagnosis of a STT. MRI characteristics depicting the tumor (size, signal, habitats, shape, surrounding tissues) were reported. Uni- and multivariate associations with STL diagnosis were evaluated in the entire cohort, and in the subgroups of benign and malignant STTs patients. Diagnostic performances of MRI features combinations were tested. We included 39 patients with STLs (median age: 69 years) and 368 patients with other STTs (122 benign STTs and 246 STS; median age: 58 years). Six MRI features were independent predictors of STL compared to all other STTs: intermediate SI on T1-WI, homogeneous enhancement (without necrotic areas), no blood signal, no fibrotic signal, no peritumoral enhancement and lack of abnormal intra- and peritumoral vasculature (p-value range: <0.0001-0.0163). Their simultaneous presence had a sensitivity of 0.88 (0.71-0.96) and a specificity of 0.88 (0.84-0.91). Other relevant MRI features were: no fat signal to discriminate against STS (p=0.0409), the infiltrative growth pattern and the vessel and nerve encasement to discriminate against benign STTs (p=0.0016 and 0.0011, respectively). Our research demonstrates that conventional MRI can help discriminating STLs from other STTs. Indeed, radiologists can help suggesting the possible diagnosis of STL, which could speed-up the subsequent proper histopathological analysis in light of MRI findings.