Simple Clinical Prognostic Model for Hepatocellular Carcinoma in Developing Countries and Its Validation

Abstract

Purpose: More than 80% of hepatocellular carcinomas (HCCs) worldwide occur in developing countries, especially in Asia. It often presents at an advanced stage beyond treatment. In this circumstance, a simple prognostic model is useful. Previous prognostic models require radiologic and laboratory investigations that are not readily available in developing countries. Our aim is to formulate and then validate a simple clinical prognostic model for HCC in an Asian population using only clinical parameters and with serum alpha-fetoprotein (AFP) as the sole laboratory test. Patients and Methods: Cox regression modeling was performed on several clinical parameters and serum AFP level in 397 patients with HCC who received only supportive care in Singapore. A later group of 324 HCC patients from an Asia-Pacific–wide randomized trial was then used to validate the model. Results: Ascites, physical performance status, and serum AFP were independently predictive of survival. Cox analysis yielded a simple score based on these three variables that categorizes patients into low-, medium-, and high-risk groups with 6-month survivals of 43%, 21%, and 5%, respectively. The prospective validation data provided corresponding estimates of 33%, 15%, and 3% and give confirmation of the utility of the simple model. Conclusion: We have formulated and prospectively validated a simple prognostic score for untreated HCC that only requires a clinical evaluation for ascites and physical performance status and measurement of serum AFP. This simple model is particularly apt for developing country circumstances and can also be used to select patients for treatment trials.