Hepatic Arterial Stenosis Assessed with Doppler US after Liver Transplantation: Frequent False-Positive Diagnoses with Tardus Parvus Waveform and Value of Adding Optimal Peak Systolic Velocity Cutoff

Abstract

To evaluate the utility of the tardus parvus waveform of the hepatic artery at Doppler ultrasonography (US) in the diagnosis of hepatic arterial stenosis in liver transplant (LT) recipients and determine whether the accuracy of such a diagnosis is enhanced by including an optimal peak systolic velocity (PSV) cutoff. This retrospective study was institutional review board approved; the requirement for informed consent was waived. The authors identified 361 LT recipients (267 male, 94 female) who underwent Doppler US and either computed tomography (CT) or angiography, with an interval between these examinations of less than 1 week. At Doppler US, tardus parvus pattern was defined as a waveform with a resistive index (RI) of less than 0.5 and a systolic acceleration time longer than 0.08 second. At CT or angiography, patients were assigned to the hepatic arterial stenosis (≥50% vessel narrowing) or nonstenosis group. The capability of the tardus parvus pattern to facilitate the diagnosis of hepatic arterial stenosis was calculated. The difference in PSV between the true- and false-positive tardus parvus patterns was evaluated. Receiver operating characteristic (ROC) analysis was performed to determine the optimal cutoff PSV for diagnosing hepatic arterial stenosis. The capability of the tardus parvus pattern and an optimal PSV cutoff in the diagnosis of hepatic arterial stenosis was determined. Sixty transplant recipients had the tardus parvus pattern at Doppler US. The sensitivity, specificity, and positive predictive value (PPV) of the tardus parvus pattern were 72% (23 of 32 LT recipients), 88.8% (292 of 329 LT recipients), and 38% (23 of 60 LT recipients), respectively. The false-positive rate was 11.2% (37 of 329 LT recipients). ROC analysis revealed an optimal PSV cutoff of less than or equal to 48 cm/sec for diagnosing hepatic arterial stenosis. The combination of the tardus parvus pattern and a PSV cutoff of less than or equal to 48 cm/sec improved specificity to 99.1% (326 of 329 LT recipients) and the PPV to 88% (22 of 25 LT recipients), thereby reducing the false-positive rate to 1% (three of 329 LT recipients) while slightly decreasing the sensitivity to 69% (22 of 32 LT recipients). Use of the tardus parvus waveform of the hepatic artery resulted in a low PPV and a high false-positive rate. However, the combination of the tardus parvus pattern and an optimal PSV cutoff greatly improved the PPV and reduced the false-positive rate in the diagnosis of hepatic arterial stenosis.