PURPOSE

We aimed to develop models for predicting overall survival (OS) and progression-free survival (PFS) of patients with primary hepatocellular carcinoma (HCC).

METHODS

Clinicopathological characteristics and laboratory information of patients were collected. We retrospectively analyzed presurgical data of 216 patients with primary HCC. The random forest and least absolute shrinkage and selection operator regression models were used to select features. We established prognostic models for predicting OS and PFS of primary liver cancer using ultrasonic imaging as well as clinical and pathological features. Accuracy of the models was evaluated using area under the curve, C index, and calibration curves, whereas their clinical application value was assessed using decision curve analysis.

RESULTS

Models for predicting OS and PFS were established based on ultrasonic imaging accessible features. The models showed excellent accuracy and prognosis prediction of OS and PFS in patients with primary HCC.

CONCLUSION

The established models based on factors such as aspartate aminotransferase platelet ratio index, Child-Turcotte-Pugh grade, tumor grade, hepatitis B virus-DNA, the intensity of ultrasound enhancement at the portal stage, lymphocyte/monocyte ratio, portal hypertension, gender, stage, the beginning time of ultrasonic contrast, and the total grade of ultrasonic enhancement can effectively predict OS and PFS of primary HCC.