# Construction of nomogram model based on contrast-enhanced ultrasound parameters to predict the degree of pathological differentiation of hepatocellular carcinoma

## TL;DR

This study builds a model using ultrasound data to predict how aggressive liver cancer is, helping doctors make better diagnoses.

## Contribution

A novel nomogram model using CEUS parameters to predict HCC pathological differentiation was developed and validated.

## Key findings

- The model achieved an AUC of 0.831 in the training set and 0.811 in the testing set.
- Three CEUS parameters (mTTI, FT, and lesion size) were identified as significant predictors.
- The model provides clinical utility for non-invasive HCC differentiation assessment.

## Abstract

To predict the degree of pathological differentiation of hepatocellular carcinoma (HCC) by quantitative analysis the correlation between the perfusion parameters of contrast-enhanced ultrasound (CEUS) and the pathological grades of HCC using VueBox® software.

We enrolled 189 patients who underwent CEUS and liver biopsy at our hospital from July 2019 to September 2024 and were pathologically confirmed with primary HCC. The Edmondson-Steiner pathological classification system was used as the gold standard for dividing the patients into the low-grade and high-grade groups. The patients were randomly divided into training set and testing set in a ratio of 7:3, in which the parameters of the training set were analyzed by univariate analysis and then stepwise regression to construct the prediction model, and the diagnostic efficacy of the validation model was evaluated by discrimination, calibration, and clinical applicability.

A total of 189 patients with primary hepatocellular carcinoma were enrolled, including 118 patients in the low-grade group and 71 patients in the high-grade group; they were randomly divided into training set of 128 patients and testing set of 61 patients. The prediction model was constructed by logistic regression in the training set, and the final model included three variables: mTTI, FT, and maximum diameter of a single lesion, resulting in the equation was
Y=−2.360+1.674X1+1.019X2+0.753X3(2)+1.570X3(3)
.The area under the ROC curve (AUC) of the training set was 0.831, with a sensitivity of 82.0% and a specificity of 79.5%; the area under the ROC curve (AUC) of the testing set was 0.811, with a sensitivity of 81.0% and a specificity of 70.0%.

The regression model constructed by combining multiple parameters can effectively improve the diagnostic performance of CEUS in predicting the pathological differentiation grade of HCC, thus providing a clinical basis and empirical support for the use of CEUS as a diagnostic imaging method for this disease.