Magnetic Resonance Imaging Assessing the Correlation of Components and Prognosis in Myxoid Liposarcoma

Abstract

Abstract Objective: To analyse the components of Myxoid liposarcoma(MLS) by Magnetic Resonance Imaging(MRI) features and explore the correlation with prognosis. Methods: Twenty MLS patients who underwent MRI were retrospectively enrolled. Tumor components was analysed by MRI features and the diagnostic efficacy of prognosis was presented as sensitivity and specificity. Patients were categorized into good and poor prognosis according to the postoperative follow-up. Results: The proportion (mean±SD) of non-fatty/non-myxoid components in the good and poor prognosis groups were （18.18±11.68）% and （66.67±15.00）% respectively(P<0.001). The proportion of mucus composition also had significantly different between the two groups((73.64±13.62)% vs (30.56±16.09)%, P<0.001). Good prognosis group has higher ADC value (mean±SD: (1.66±0.23)mm2/s) and lower ADC low signal ratio(median±interquartile range: (5.00±10.00)%) in non-fat/non-mucinous areas than poor((1.21±0.41)mm2/s, (20.00±39.00)%; P=0.006，P=0.001). The prognosis between the component ratio <25% and> 50% of both non-fat/non-mucinous and mucus composition were statistically significant(P=0.001，P=0.008). The proportion of non-fat/non-mucus component had the highest diagnostic efficacy, with a sensitivity, specificity and accuracy of >90%(100%，90.09%，90%), AUC=0.99. Conclusion: Imaging features were closely related to the histological components of MLS. Diffusion-restricted areas in non-fatty/non-mucinous components may suggest dense areas of round cell. Using MRI features to assess the components of MSL have certain implications for prognostic prediction，especially before preoperative puncture biopsy.