Purpose: To assess whether dynamic contrast-enhanced (DCE) MRI timing bolus data from routine clinical examinations can be postprocessed to obtain hepatic perfusion parameters for diagnosing cirrhosis.
Materials and methods: We retrospectively identified 57 patients (22 with cirrhosis and 35 without cirrhosis) who underwent abdominal MRI, which included a low-dose (2 mL gadodiamide) timing bolus using a volumetric spoiled gradient echo T1-weighted sequence through the abdomen. Using a dual-input single-compartment model, the following perfusion parameters were measured: arterial, portal, and total blood flow; arterial fraction; mean transit time; and distribution volume. Those parameters were compared between patients with and without cirrhosis using t-tests. Receiver operating characteristic (ROC) curve analysis was used to identify the perfusion parameters that can best predict the presence of cirrhosis.
Results: The hepatic arterial fraction, arterial flow, and distribution volume in patients with cirrhosis (27.7 +/- 8.3%, 44.8 +/- 14.1 mL/minute/100 g, and 16.3 +/- 4.5%, respectively) were significantly higher than those without cirrhosis (18.7 +/- 4.4%, 28.5 +/- 11.7 mL/minute/100 g, and 14.0 +/- 4.2%, respectively; P < 0.05 for all). ROC analysis showed arterial fraction as the best predictor of cirrhosis, with sensitivity of 73% and specificity of 86%.
Conclusion: Timing bolus DCE MR images from routine examinations can be postprocessed to yield potentially useful hepatic perfusion parameters.