PURPOSE: To compare the diagnostic accuracies of diffusion-weighted imaging (DWI), T2-weighted imaging (T2WI), and the combination of both sequences in discriminating benign from malignant focal liver lesions (FLLs). MATERIALS AND METHODS: In all, 166 patients with 269 FLLs (153 benign and 116 malignant) were retrospectively evaluated. Two abdominal readers visually assessed the DWI, T2WI, and the combined (DWI+T2WI) image sets in an independent and blinded manner. The diagnostic abilities of each image set in discriminating the benign from the malignant FLLs set were compared using a binary logistic regression model. Pathologic results, consensus reading, and follow-up imaging were used as the reference standard. RESULTS: The overall characterization accuracy in all lesions of the combined set (80.3%) was significantly higher than those of the T2WI set (68.8%) and DWI set (73.2%) (combined vs. T2WI, P < 0.001; combined vs. DWI, P = 0.001), while there was no significant difference between the T2WI and DWI sets (P = 0.058). All image sets were more accurate in the characterization of malignant FLLs than of benign FLLs (P < 0.001). CONCLUSION: T2WI and DWI are complementary in discriminating benign from malignant FLLs; their combination improves diagnostic confidence.
PURPOSE: To compare the diagnostic accuracies of diffusion-weighted imaging (DWI), T2-weighted imaging (T2WI), and the combination of both sequences in discriminating benign from malignant focal liver lesions (FLLs). MATERIALS AND METHODS: In all, 166 patients with 269 FLLs (153 benign and 116 malignant) were retrospectively evaluated. Two abdominal readers visually assessed the DWI, T2WI, and the combined (DWI+T2WI) image sets in an independent and blinded manner. The diagnostic abilities of each image set in discriminating the benign from the malignant FLLs set were compared using a binary logistic regression model. Pathologic results, consensus reading, and follow-up imaging were used as the reference standard. RESULTS: The overall characterization accuracy in all lesions of the combined set (80.3%) was significantly higher than those of the T2WI set (68.8%) and DWI set (73.2%) (combined vs. T2WI, P < 0.001; combined vs. DWI, P = 0.001), while there was no significant difference between the T2WI and DWI sets (P = 0.058). All image sets were more accurate in the characterization of malignant FLLs than of benign FLLs (P < 0.001). CONCLUSION: T2WI and DWI are complementary in discriminating benign from malignant FLLs; their combination improves diagnostic confidence.
Authors: Maike Bode; Shuo Zhang; Mark N Terwolbeck; Caroline Molavi Tabrizi; Masami Yoneyama; Nils A Kraemer; Christiane K Kuhl; Alexandra Barabasch Journal: Sci Rep Date: 2022-07-21 Impact factor: 4.996