Fat-suppressed T2-weighted MR imaging of hepatocellular carcinoma and metastases: comparison of conventional spin-echo, fast spin-echo, and echoplanar pulse sequences.

The purpose of our study was to compare the diagnostic accuracy of fat-suppressed T2-weighted magnetic resonance (MR) images obtained with conventional spin-echo (SE), respiratory-triggered fast SE, and breath-hold multishot SE echoplanar (EP) sequences for the detection of hepatocellular carcinoma and metastases. Images obtained with the three sequences in 17 patients (15 with cirrhosis) with 31 hepatocellular carcinomas and 14 patients with 45 metastases were retrospectively analyzed. Image review was conducted on a segment-by-segment basis; in all, 248 liver segments were reviewed separately and independently for detection of solid, malignant lesions. Diagnostic accuracy was evaluated with receiver-operating-characteristic (ROC) curve analysis. Diagnostic accuracy for hepatocellular carcinoma as determined by ROC curve analysis was better by a statistically significant amount for conventional SE (Az = 0.95) images when compared with respiratory-triggered fast SE (Az = 0.83, P < 0.05) and breath-hold multishot SE EP (Az = 0.80, P < 0.05) images. Conventional SE MR sequences should not be replaced with respiratory-triggered fast SE or breath-hold multishot SE EP sequences for T2-weighted MR imaging of patients with hepatocellular carcinoma in cirrhosis, unless sufficient contrast-enhanced dynamic MR imaging is subsequently performed.