102 - Accepted Sonographic Findings to Predict Clinical TIPS Malfunction: Time to Reassess Normal?

Transjugular Intrahepatic Portosystemic Shunt (TIPS) creation is used to manage complications of cirrhosis and portal hypertension. Advances in stent design have improved portal decompression control through smaller initial stent diameters, but this may present challenges in imaging follow-up in predicting shunt dysfunction. This study evaluated the appropriateness of current imaging guidelines and their implications for imaging protocols post-TIPS.

Materials and Methods:

187 adult patients (78 females, mean age 57.5 ± 10.8 years) with TIPS creation between January 2018 and December 2022 were reviewed. Clinical and imaging data were analyzed at 1, 3, 6, 12, 24, 36, and 48 months. Variable follow-up resulted in 367 appointments evaluated as separate data points. Clinical end points were recurrent esophageal variceal bleeding, need for paracentesis, or significant recurrent ascites on physical exam. Historical cutoff values for TIPS velocities (90-190 cm/s) and portal vein velocities ( > 30 cm/s) were evaluated for their ability to predict clinical dysfunction of the TIPS using binomial GEE regression.

Results:

Alcoholic liver disease (29.9%, n = 56) was the primary cause of liver disease; primary indication for TIPS was ascites (48.7%, n = 91). During placement of the 8-10 mm controlled expansion stent grafts, 150 (80.2%) TIPS were dilated to 8mm, 21 (11.2%) to 9mm, and 16 (8.6%) to 10mm. 71 patients (38.0%) had > 1 reintervention, resulting in 118 unique reinterventions. Among 367 follow-up observations, 231 (62.9%) experienced clinical normal TIPS function, and 136 (37.1%) experienced clinical TIPS malfunction as defined.

Using currently accepted imaging guidelines, post-TIPS duplex demonstrated a sensitivity of 88.0% (95% CI: 79.6-93.9%) and a specificity of 33.6% (95% CI: 25.8-42.0%) in diagnosing clinical TIPS malfunction. The positive predictive value was 46.6% (95% CI: 39.0-54.3%), and the negative predictive value was 81.0% (95% CI: 68.6-90.1%). False positive and negative error rates were 66.4% (95% CI: 58.0-74.2%) and 12.0% (95% CI: 6.1-20.4%), respectively. Mid-TIPS velocity was associated with clinical abnormality across all stent diameters (Wald χ²(1) = 7.20, P = 0.007).

Conclusion:

Accepted post-TIPS imaging values are highly sensitive but also with a high false positive rate and low specificity in patients with controlled expansion stent grafts. These findings suggest that patients may be inaccurately radiologically diagnosed with shunt dysfunction.