The role of diffusion-weighted MRI (DWI) in the differentiation of benign from malignant skeletal lesions of the pelvis.

PURPOSE:To evaluate the utility of Diffusion-weighted MRI in the differentiation of benign from malignant skeletal lesions of the pelvis. MATERIALS AND METHODS:In this retrospective study 33 patients with indeterminate skeletal lesions of the pelvis were evaluated with DWI. Minimum, mean, maximum ADC-values of the skeletal lesions were measured followed by qualitative assessment of DWI. All patients underwent histological confirmation using CT-guided biopsy or surgical resection. The histology of the skeletal lesions was correlated with the findings on DWI. RESULTS:There were 13 malignant lesions and 20 benign lesions. The mean, minimum and maximum ADC values (×10-6mm2/s) for benign skeletal lesions was higher than the mean ADC-values for malignant lesions (1422.2 vs 1263.7; 780.4 vs 771.8; 1969.6 vs 1676.8 respectively). These differences were however not statistically significant (P-values=0.29; 0.94; 0.149 respectively). The sensitivity, specificity, positive predictive value and negative predictive value for qualitative assessment of Diffusion-weighted MRI in the differentiation of benign from malignant skeletal lesions were: 53.9%, 85%, 70%, 73.9% respectively. Qualitative assessment of DWI (restricted diffusion versus non-restricted diffusion) allowed differentiation of benign from malignant skeletal lesions (P-value=0.0259). CONCLUSIONS:Qualitative assessment of DWI may aid in the differentiation of benign skeletal lesions from malignant skeletal lesions of the pelvis. Although DWI has a low sensitivity in the distinction of the two disease entities, it may be a useful adjunct due to its relatively high specificity. This is of particular importance in lymphoma where biopsy may only show chronic inflammatory cells and hence may be false negative.