PURPOSE: The purpose was to validate diffusion-weighted imaging (DWI) in the assessment of hepatic metastases compared with superparamagnetic iron oxide (SPIO)-enhanced magnetic resonance imaging. MATERIALS AND METHODS: For 21 consecutive patients with 160 metastases from extrahepatic malignancy and 25 benign focal lesions, two radiologists evaluated four separate review sessions (I, SPIO-enhanced T2-weighted images; II, precontrast DWI; III, SPIO-enhanced T2-weighted images and precontrast DWI; IV, SPIO-enhanced T2-weighted images plus precontrast and SPIO-enhanced DWI) and assigned confidence levels using a five-grade scale for each hepatic lesion. RESULTS: The A(z) values after receiver operating characteristic curve analysis for Reader 1 and Reader 2 were 0.80 and 0.75 on session I, 0.91 and 0.91 on session II, 0.97 and 0.96 on session III and 0.96 and 0.96 on session IV, respectively. The A(z) value of session II was significantly higher than that of session I (Reader 1, P=.004; Reader 2, P<.001), and that of session III was significantly higher than that of session I (P<.001 for each reader) or session II (Reader 1, P=.004; Reader 2, P=.003). Although there was no significant difference of A(z) value between session III and session IV (Reader 1, P=.231; Reader 2, P=.878), the sensitivity improved for session IV compared with that for session III (Reader 1, P=.031; Reader 2, P=.039). CONCLUSION: In the assessment of hepatic metastases, DWI can provide more accurate information than can SPIO-enhanced images. Diagnostic accuracy can be increased even more through the combination of both techniques.
PURPOSE: The purpose was to validate diffusion-weighted imaging (DWI) in the assessment of hepatic metastases compared with superparamagnetic iron oxide (SPIO)-enhanced magnetic resonance imaging. MATERIALS AND METHODS: For 21 consecutive patients with 160 metastases from extrahepatic malignancy and 25 benign focal lesions, two radiologists evaluated four separate review sessions (I, SPIO-enhanced T2-weighted images; II, precontrast DWI; III, SPIO-enhanced T2-weighted images and precontrast DWI; IV, SPIO-enhanced T2-weighted images plus precontrast and SPIO-enhanced DWI) and assigned confidence levels using a five-grade scale for each hepatic lesion. RESULTS: The A(z) values after receiver operating characteristic curve analysis for Reader 1 and Reader 2 were 0.80 and 0.75 on session I, 0.91 and 0.91 on session II, 0.97 and 0.96 on session III and 0.96 and 0.96 on session IV, respectively. The A(z) value of session II was significantly higher than that of session I (Reader 1, P=.004; Reader 2, P<.001), and that of session III was significantly higher than that of session I (P<.001 for each reader) or session II (Reader 1, P=.004; Reader 2, P=.003). Although there was no significant difference of A(z) value between session III and session IV (Reader 1, P=.231; Reader 2, P=.878), the sensitivity improved for session IV compared with that for session III (Reader 1, P=.031; Reader 2, P=.039). CONCLUSION: In the assessment of hepatic metastases, DWI can provide more accurate information than can SPIO-enhanced images. Diagnostic accuracy can be increased even more through the combination of both techniques.