To investigate the role of multimodal ultrasound features, including ultrasound, contrast-enhanced ultrasound, and endoscopic ultrasound, in predicting imaging remission and acute exacerbation of IgG4-related pancreatitis following steroid therapy.

This study included 84 patients with biopsy-confirmed IgG4-related pancreatitis with complete imaging data in the First Affiliated Hospital of Zhejiang University School of Medicine from January 2018 to January 2024. Baseline clinical information and features of ultrasound, contrast-enhanced ultrasound, and endoscopic ultrasound (including pancreatic parenchymal changes, biliary ductal changes, and enhancement patterns) were recorded. Patients were followed for a median duration of 18 months via imaging and medical records to document remission and acute exacerbation post-steroid treatment. The baseline clinical and sonographic features were compared between patients who achieved remission and those who did not, and between patients who experienced acute exacerbation and those who did not, to investigate the predictive value of multimodal ultrasound for these outcomes. Logistic regression models were constructed to predict imaging remission and acute exacerbation.

Among the 84 patients, 43 achieved imaging remission, and 15 had acute exacerbation during the follow-up. Compared with the non-remission group, the imaging remission group had higher proportions of patients with elevated baseline amylase, bile duct stenosis, formation of pseudocysts, and heterogeneous enhancement (22/43 vs 12/41, 25/43 vs 10/41, 9/43 vs 1/41, and 19/43 vs 9/41, respectively), with statistically significant differences between the two groups (χ²=4.176, 9.836, 6.843, and 4.669, P=0.048, 0.002, 0.009, and 0.031, respectively). Logistic regression analysis was applied to construct a model for predicting imaging remission: Imaging remission =0.553×whether baseline amylase was elevated +1.779×whether the bile duct was stenotic +3.280×whether there was a pseudocyst -0.800×whether there was homogeneous enhancement -0.701. The area under the curve (AUC) of this model for predicting imaging remission was 0.81, with a sensitivity of 83% and specificity of 57%. Compared with the non-exacerbation group, the acute exacerbation group had higher proportions of patients with elevated baseline amylase, bile duct stent implantation, and hypo-enhancement (10/15 vs 24/69, 9/15 vs 8/69, and 11/15 vs 30/69, respectively), with statistically significant differences between the two groups (χ²=5.199, 17.885, and 4.395, P=0.040, <0.001, and =0.036, respectively). Logistic regression was applied to construct a model for predicting acute exacerbation: Acute exacerbation =0.589× whether baseline amylase was elevated +2.332× whether there was a biliary stent + 1.488×whether there was hypo-enhancement -3.442. The AUC of this model for predicting acute exacerbation was 0.83, with a sensitivity and specificity of 93% and 51%, respectively.

Multimodal sonograms can provide comprehensive lesion characterization and effectively predict imaging remission and acute exacerbation of IgG4-related pancreatitis following steroid treatment.