Intravoxel incoherent motion diffusion-weighted imaging for ruling out high-risk esophageal varices in cirrhosis.

Purpose: To explore the role of liver- and spleen-based Intravoxel Incoherent Motion Diffusion-weighted Imaging (IVIM) imaging in ruling out high-risk esophageal varices (HREVs) in patients with cirrhosis and compare it with other noninvasive tests (NITs).

Methods: Patients with liver cirrhosis who underwent paired Esophagogastroduod-enoscopy (EGD) and upper abdominal IVIM imaging between January 2015 and December 2021 were retrospectively enrolled. Logistic regression analyses were performed to determine the potential predictors of HREVs. The diagnostic performance was estimated with the area under the receiver operating characteristic curve (AUROC). We used the 95% sensitivity to determine the cutoff for ruling out HREVs and compared the spared EGD rate of IVIM imaging with that of NITs. A sensitivity analysis was carried out to assess the impact of the prevalence of HREVs on the negative predictive value (NPV) at given cutoff values.

Results: A total of 244 patients were included, and HREVs were found in 54.1% (132/244) of the patients. The parameters f_S, D*_S and D*_L independently predicted HREVs. Compared with NITs, f_S (AUROC: 0.90) performed significantly better (all p < 0.05), and D*_S (AUROC: 0.74) and D*_L (AUROC: 0.73) performed similarly. f_S (cutoff ≤ 16.1%) had the highest spared EGD rate (27.0%), with a missed HREVs rate of less than 5%, and showed excellent interobserver repeatability. The NPV of f_S remained above 90% as the prevalence of HREVs increased from 5 to 50%.

Conclusion: Splenic microcirculatory perfusion parameters (f_S and D*_S) are potential noninvasive markers for ruling out HREVs and avoiding unnecessary EGDs.