Background and aim: The incidence of hepatocellular carcinoma is increased in patients with cirrhosis. Therefore, surveillance for detection of small tumors has been proposed. The aim of this study was to determine the clinical and economical effects of screening for small hepatocellular carcinoma in Western patients with Child-Pugh class A cirrhosis.
Methods: Based on a decision analysis model representing the natural history of cirrhosis and the continuing risk of developing cancer, we compared a strategy of performing ultrasound and alpha-fetoprotein dosage every 6 months with a strategy of seeking tumors only if they are clinically suspected. In both strategies, partial hepatectomy was performed for patients with compensated cirrhosis and diagnosed with resectable tumors. We did not consider orthotopic liver transplantation as a therapeutic option. Data were drawn from MEDLINE search.
Results: For most patients seen in the daily practice, screening provides negligible benefits in life expectancy (< 3 months), even when the incidence of cancer is high (6% per year), and despite our choice of consistent biases in favor of screening. The cost-effectiveness ratios of systematic surveillance range between $48,000 and $284,000 for each additional life-year gained, more than other common medical practices. However, for a minority of patients with a predicted cirrhosis-related survival rate above 80% at 5 years (the "ideal" candidates) screening may increase mean life expectancy by 3 to 9 months depending on age, cancer incidence (1.5% to 6% per year), and survival rate after surgery (40% to 60% at 3 years). In this clinical setting, the cost-effectiveness ratios range between $26,000 and $55,000 for each additional life-year gained.
Conclusions: For most patients with cirrhosis seen in the daily practice, biannual screening to detect symptomless tumors accessible to surgical resection provides negligible benefit in life expectancy. In addition, the cost-effectiveness ratios incurred by this strategy is more important than that of many current medical practices. On the other hand, for well-targeted patients with the longest reported cirrhosis-related survival rate, screening may substantially increase mean life expectancy, at lower costs. Careful selection of these patients with a favorable cirrhosis-related prognosis requires further studies.