Detection of hepatocellular carcinoma by Gd-EOB-DTPA-enhanced liver MRI: comparison with triple phase 64 detector row helical CT.

Abstract Purpose To compare the diagnostic performance of Gd-EOB-DTPA-enhanced MRI with that of triple phase 64-MDCT in the detection of hepatocellular carcinoma (HCC). Patients and methods Thirty-four patients with 52 surgically proven lesions underwent Gd-EOB-DTPA-enhanced MRI and triple phase 64-MDCT. Two observers independently evaluated MR and CT imaging on a lesion-by-lesion basis. Sensitivity, positive and negative predictive values and reproducibility were evaluated. The diagnostic accuracy of each modality was assessed with alternative-free response receiver operating characteristic (ROC) analysis. Results Both observers showed higher sensitivity in detecting lesions with MRI compared to CT, however, only the difference between the two imaging techniques for observer 2 was significant ( P  = 0.034). For lesions 1 cm or smaller, MRI and CT showed equal sensitivity (both 62.5%) with one observer, and MRI proved superior to CT with the other observer (MRI 75% vs. CT 56.3%), but the latter difference was not significant ( P  = 0.083). The difference in positive and negative predictive value between the two imaging techniques for each observer was not significant ( P  > 0.05). The areas under the ROC curve for each observer were 0.843 and 0.861 for MRI vs. 0.800 and 0.833 for CT and the differences were not significant. Reproducibility was higher using MRI for both observers, but the result was not significant (MRI 32/33 vs. CT 29/33, P  = 0.083). Conclusion Gd-EOB-DTPA-enhanced MRI tended to show higher diagnostic accuracy, sensitivity and reproducibility compared to triple phase 64-MDCT in the detection of hepatocellular carcinoma, however statistical significance was not achieved.