Developing a novel predictive model for identifying risk factors associated with being lost to follow-up among high-risk patients for recurrence following radical resection of hepatocellular carcinoma: the first report.

Abstract

Background: Follow-up is essential especially for patients who are at a high risk of recurrence after radical resection of hepatocellular carcinoma (HCC). The aim of this study was to develop a predictive model aimed at identifying the risk factors associated with being lost to follow-up (LTFU) in high-risk patients for recurrence following radical resection of HCC.

Methods: The retrospective study was conducted at our institution between October 2018 to May 2023. The patients who underwent radical liver resection for HCC and had high-risk factors for recurrence were categorized into an LTFU group and a control group. Multivariate logistic regression analysis was utilized to determine risk factors and construct a nomogram predictive model.

Results: A total of 352 patients were included and subsequently classified into two distinct groups: the LTFU group (n = 123, 34.94%) and the control group (n = 229, 65.06%). Logistic regression analysis was then conducted to explore the potential associations between various factors and the occurrence of LTFU. The findings identified several independent risk factors for LTFU, including smoking (odds ratio, OR = 1.823, 95% confidence interval, CI 1.086-3.060, p = 0.023); residing more than 200 km away from the hospital (OR = 1.857, 95% CI 1.105-3.121, p = 0.019); having an unstable profession (OR = 1.918, 95% CI 1.112-3.311, p = 0.019); and lacking medical insurance (OR = 5.921, 95% CI 1.747-20.071, p = 0.004); the presence of liver cirrhosis (OR = 2.161, 95% CI 1.153-4.048, p = 0.016); an operation time less than 240 min (OR = 2.138, 95% CI 1.240-3.688, p = 0.006); and the absence of postoperative adjuvant therapy (OR = 2.641, 95% CI 1.504-4.637, p = 0.001). Based on these seven significant factors, a main effects model was established, designated as the Wei-LTFU model, which achieved an area under the curve value of 0.744 (95% CI 0.691-0.798) in predicting the likelihood of LTFU.

Conclusion: A main effects model, namely the Wei-LTFU model, incorporating the seven significant factors was formulated to predict the likelihood of LTFU occurrence, ultimately aiming to assist healthcare workers in developing effective strategies to improve follow-up outcomes for patients.