RETURN TO ABSTRACT LISTING


Objective:
Diffusion weighted imaging (DWI) has been shown to be useful for differentiating benign from malignant tumors and for tumor grading. The aim was to determine whether apparent diffusion coefficient (ADC) or conventional MRI allows for differentiation of malignant extremity soft tissue tumors.

Materials and Methods:
A consecutive series of histology-proven, malignant soft tissue tumors in adult patients with conventional MRI and DWI were included. Post-operative cases were excluded. Conventional MRI evaluation included T1W/T2W signal alterations and heterogeneity, peri-tumoral edema, necrosis or cystic changes, internal hemorrhage, and the maximum longitudinal dimension. ADC mean and minimum using a free hand region of interest (ROI) of the whole tumor and darkest ADC area of the tumor were recorded. A final diagnosis prediction was also recorded based on all imaging findings, blinded to the final histologic diagnosis. Kruskal-Wallis and Fisher’s Exact Test were used to determine associations and the significance between tumor subtypes. Intraclass correlation (ICC) and kappa calculations were used, specifically for the ADC values and each reader’s diagnosis prediction.

Results:
A total of 78 soft tissue sarcomas of the upper and lower extremities were tested, including 2 chondrosarcomas, 13 myxofibrosarcomas, 5 leiomyosarcomas, 11 liposarcomas, 4 malignant peripheral nerve sheath tumors, 20 undifferentiated pleomorphic soft tissue sarcomas, 12 synovial sarcomas, 2 histiocytomas, 1 angiosarcoma, 1 epithelioid sarcoma, 5 lymphomas, and 2 spindle cell sarcomas. On MRI, only T1 enhancement and hemorrhage were determined to be different across the samples (p = 0.034 and p = <0.001), respectively. More notably, liposarcomas consistently exhibited absence of hemorrhage (9/11, 81.8%) while undifferentiated pleomorphic sarcomas consistently showed hemorrhage (18/20, 90%). Liposarcomas were shown to be heterogenous and hyperintense on T1W when compared to a tumor not classified as a liposarcoma (p = 0.046 and p = 0.010), respectively. Myxoid tumors were similarly more heterogenous on T2W images when compared to nonmyxoid tumors (p = 0.044). Lymphomas registered lower ADC values in all four dependent, continuous variables for both readers (p < 0.05). Consistently lower mean, whole tumor ADC values were demonstrated for synovial sarcomas (p < 0.05). The inter-reader agreement was good to excellent for all four continuous measurements (ADC values) (ICC= 0.65 - 0.98). The inter-reader agreement was poor with respect to arriving at the same final diagnosis given the same case (ICC= 0.31) based on conventional MRI and DWI.

Conclusion:
DWI and ADC currently offers limited insight in being able to differentiate between malignant soft tissue malignancies, apart from lymphomas and synovial sarcomas, while preserving good to excellent inter-reader agreement in ADC measurements.