Soft-tissue fat tumours: differentiating malignant from benign using proton density fat fraction quantification MRI

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Lipomas, atypical lipomatous tumors (ALT), and liposarcomas are soft tissue tumors with overlapping imaging features.
• Accurate differentiation is crucial for appropriate clinical management and treatment planning.

Purpose of the Study:

• To assess the utility of proton density fat fraction (PDFF) in differentiating lipoma, ALT, and liposarcoma.
• To explore the use of magnetic resonance imaging (MRI)-based fatty acid composition (FAC) and 3D texture analysis for distinguishing ALT from non-classical lipomas.

Main Methods:

• Prospective study involving 41 patients (20 women, 21 men) with lipomas, ALTs, and myxoid liposarcomas.
• Quantitative MRI techniques including spoiled gradient multi-echo imaging for PDFF and FAC analysis.
• Support vector machine algorithm and 3D texture analysis were employed for differentiation.

Main Results:

• Significant differences in mean PDFF were observed between liposarcomas and both ALT and lipomas (p=0.0001).
• PDFF also showed a significant difference between ALT and lipomas (p=0.021), with optimal thresholds identified.
• Texture analysis failed to differentiate ALT from non-classical lipomas, but FAC unsaturation degree showed a significant difference (p=0.013).

Conclusions:

• PDFF measurement is a valuable adjunct to standard MRI for differentiating liposarcomas from lipomas and ALTs.
• Current PDFF, FAC, and texture analysis methods are insufficient to confidently distinguish between ALT and lipomas not solely composed of fat.