Abstract:

Objective

To construct and verify a predictive model for evaluating the pain risk of patients with liver cancer after transcatheter hepatic arterial chemoembolization （TACE）.

Methods

The clinical data of 212 patients with liver cancer who received TACE in a first-class hospital in Guangzhou from January 2022 to December 2023 were collected retrospectively.These patients were divided into pain group and pain-free group according to whether they experienced pain after operation.Logistic regression method was used to screen the key factors causing pain after TACE， and a nomogram was created to predict the pain risk.Bootstrap technology is used to verify the prediction model internally， and C- statistics and calibration curve are used to evaluate its prediction effect.

Results

Among the 212 patients with liver cancer treated by TACE， 74 patients had postoperative pain， and the incidence of pain was 34.9%.It was found that patients with a history of liver cirrhosis had higher independent risk factors for pain after TACE （P＜0.05）， while the use of painkillers before and during the operation significantly reduced the risk of postoperative pain （P＜0.05）.Based on this， a prediction model of postoperative pain risk of patients with liver cancer was established， and the formula was Logit（P） = 62.39 + 1.676× ［liver cirrhosis］ - 1.643×［analgesic used before operation］ - 1.293×［analgesic used during operation］.The C statistic of this model is 0.729 （95%CI： 0.661～0.795）， the sensitivity is 45.8%， and the specificity is 70.0%.The calibration curve and Brier score verify the good fitting degree of the model.

Conclusion

In this study， a pain prediction model for patients with liver cancer after TACE was established， which showed good discrimination and calibration， and provided a reliable tool for clinic experience.

Key words: Liver Cancer, Transcatheter Arterial Chemoembolization, Pain, Risk Prediction Model