Assessment of the cut-off value of quantitative liver-portal vein contrast ratio in the hepatobiliary phase of liver MRI.

AIM To calculate the quantitative liver–portal vein contrast ratio (Q-LPC) cut-off value based on tumour detectability by using receiver operating characteristic (ROC) curves. MATERIALS AND METHODS Seventy-four patients with tumours (46 men and 28 women; age, 71 ± 8.1 years), who underwent liver magnetic resonance imaging (MRI) using gadolinium–ethoxybenzyl–diethylenetriamine pentaacetic acid (Gd–EOB–DTPA) were enrolled. Some patients were found to have multiple tumours. In total, 102 tumour images were evaluated for quantitative liver–spleen contrast ratio (Q-LSC) and Q-LPC 10 minutes after the administration of Gd–EOB–DTPA. Q-LPC and Q-LSC were compared to assess the cut-off values and usefulness. The ROC curve was evaluated using the method for continuously distributed test results, with a free scale of 50 mm. A score of ≥30 out of 50 points was considered good. Cut-off values of Q-LPC and Q-LSC were then calculated. The areas under the ROC curve (AUCs) were also examined and compared. RESULTS The AUC–ROC for Q-LPC was 0.858 (95% confidence interval [CI], 0.783–0.933). The cut-off value was determined to be at 1.462. Sensitivity was 0.747, and specificity was 0.852 at the cut-off value. The AUC–ROC for Q-LSC was 0.710 (95% CI, 0.597–0.822). The cut-off value was at 1.543, the sensitivity was 0.560, and the specificity was 0.778 at the cut-off value. A significant difference was noted between the AUCs (p=0.0016). CONCLUSION Q-LPC can be used for hepatobiliary phase MRI evaluation.