MR imaging of fat-containing tumours: The distinction between lipoma and liposarcoma

• Main Authors: S. Jaovisidha, Y. Suvikapakornkul, P. Woratanarat, T. Subhadrabandhu, A. Nartthanarung, P. Siriwongpairat
• Other Authors: Mahidol University
• Format: Review
• Published: 2018
• Subjects: Medicine
• Online Access: https://repository.li.mahidol.ac.th/handle/123456789/29679

Introduction: This study aimed to retrospectively review the magnetic resonance (MR) imaging features of the lipomatous tumour in order to differentiate between lipoma and liposarcoma. Methods: The MR images of 38 patients (24 female and 14 male with a mean age 48 years) in a consecutive five-year period, who had histologically verified lipoma (n is 29) and liposarcoma (n is 9), were retrospectively reviewed. The images were assessed for the number, site, size and margin of the lesions, as well as the signal intensity (homogenous, bright signal on T1-weighted [T1W] image, bright signal suppressed on T1W with fat-suppression image, bright signal on T2-weighted with fatsuppression image), the internal architecture (thin/thick septum, cystic change), the effect on the surrounding tissue (oedema, neurovascular involvement) and the enhancement pattern. Results: A partially ill-defined margin, neurovascular involvement, enhancing thick/nodular septum and a partially bright signal intensity on T1W images were statistically significant MR imaging features that favoured a diagnosis of liposarcoma (p-value is less than 0.0001). Male gender, an internal cystic change and surrounding soft tissue oedema increased the risk of liposarcoma approximately 2.8, 3.5 and 3.5 times, respectively, compared with the reference group (lipoma), but this was not a statistically significant finding. Thick/nodular septum was signif icantly associated with liposarcoma compared with lipoma (odds ratio 69.3, 95 percent confidence interval 5.2-3184.8, p-value is less than 0.0001). Conclusion: Statistically significant MR imaging features that favour a diagnosis of liposarcoma included a par tially ill-defined margin, neurovascular involvement, enhancing thick/ nodular septum, and a partially bright signal intensity on T1W images. The most statistically significant predictor of liposarcoma was thick/ nodular septum.