Improving diagnostic accuracy in cirrhotic patients: a comprehensive multicenter analysis of contrast-enhanced ultrasound-guided biopsy in a large cohort of 2056 patients.

Abstract

Purpose: To investigate the factors influencing ultrasound-guided liver biopsy accuracy and develop a decision model for managing biopsy results.

Materials and methods: This prospective study enrolled 2056 adult patients with focal hepatic lesions from nine Chinese hospitals. Diagnostic accuracy was calculated, and variables were analyzed using multivariate logistic regression. A prediction model was constructed using chi-square automatic interaction detection.

Results: This study enrolled 2056 participants (1297 men, 759 women; mean age, 67.8 ± 10.8 years) with 2056 biopsied lesions (mean ± standard deviation: 4.0 ± 2.8 cm; range 0.7-17.7 cm), and 94.2% were accurately diagnosed. Two to three biopsy passes were significantly more accurate than one pass (95.1% vs. 87.3%, P < 0.001) and comparable to the accuracy of > 3 passes (95.1% vs. 93.2%, P = 0.408). Independent predictors of accurate diagnoses included absence of cirrhosis [2.428 (1.457-3.741), P < 0.001], CEUS guidance [1.899 (1.288-2.801), P = 0.001], and number of biopsy passes > 1 [1.775 (1.103-2.855), P = 0.018]. The predictive decision tree model demonstrated that for cirrhotic patients who underwent CEUS-guided biopsy, the probability of an accurate diagnosis increased from 88.3% (when US-guided biopsy was used) to 93.3%. Conversely, in noncirrhotic patients undergoing US-guided biopsy, a diagnostic accuracy probability of 96.4% was observed. The overall prediction accuracy of the model was 94.4%.

Conclusion: CEUS-guided biopsy enhances diagnostic accuracy in cirrhotic patients, whereas US-guided biopsy is highly accurate in noncirrhotic patients. Two to three needle punctures are sufficient for 95% accuracy, with no additional improvement from more punctures.