Preoperative identification of hepatocellular carcinoma from focal liver lesions ≤ 20 mm in high-risk patients using clinical and contrast-enhanced ultrasound features

Abstract

Objective

We aimed to develop and validate a prediction model to identify HCC in focal liver lesions (FLLs) ≤20 mm among patients at risk for HCC based on clinical and contrast-enhanced ultrasound (CEUS) features.

Methods

Between January 2022 and July 2023, 386 patients (mean age 58 ± 11 years; 277 male) at risk for HCC with FLLs ≤20 mm and clinical and preoperative CEUS data from three centers were retrospectively enrolled. Three prediction models based on clinical data (Cli-M), CEUS features (CEUS-M), and combined clinical and CEUS features (Com-M) were constructed using the training cohort (187 patients). Their predictive performance was evaluated using the area under the receiver operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA) in the internal and external validation cohorts. All patients were reclassified using the American College of Radiology CEUS Liver Imaging Reporting and Data System (CEUS LI-RADS) and combined with the best-performing model (modified LI-RADS).

Results

The AUCs of Com-M were 0.873–0.951 in the training, internal, and external validation cohorts, which were higher than those of Cli-M (0.749–0.795, all P < 0.05) and CEUS-M (0.848–0.899, all P < 0.05). The sensitivity of LR-5 of modified LI-RADS was significantly improved from 83.1 % to 88.9 % (p＜0.001) in the training, internal and external validation cohort while there was no statistical different on its specificity (82.6 %-94.7 % vs 95.7 %-97.6 %., p = 0.162–0.650).

Conclusions

The model based on clinical and CEUS features can help identify HCC in FLLs ≤ 20 mm in high-risk patients.