Screening tests for hepatocellular carcinoma

HCC meets most of the criteria established for a cancer surveillance program. It is a common health care problem, a target population is identified, effective treatments are available, recall strategies are in place, and surveillance of cirrhotics is cost effective. The main problem with surveillance for HCC is the lack of acceptable accuracy of the current available screening test in the population studied. As reviewed in this article, the current available literature on AFP is fraught with limitations. Most of the studies were not powered to assess the performance of markers as surveillance tests for HCC. The performance of the AFP assay such as reproducibility was seldom described. Analyses of covariates such as demographics and etiology of liver disease were rarely performed. As a consequence, most new biomarkers for HCC have remained stagnant in phase 1 or 2 studies even though some were first reported as potential HCC markers more than 15 years ago. The prospective studies reviewed evaluated AFP at the time of entry into the study and not around the time of HCC development, which is critical to know the performance of the marker in surveillance. Another limitation of the data is the lack of outcomes to eliminate the possibility of lead time bias with screening with AFP. This is likely related to the lack of randomized control trials to compare markers. Two recent independent studies assessed the quality of AFP as a screening test for HCC. They all indicate the lack of data regarding the effectiveness of AFP for the surveillance of HCC for many of the reasons outlined above. Although surveillance of cirrhotics with AFP and ultrasound is recommended, shown to be cost effective, and widely performed, the current data regarding the performance of AFP as a surveillance test for HCC are poor. Several markers, including glypican-3, DCP, AFP-L3, insulin growth factor-1, and human hepatocyte growth factor, appear to be promising, and should be further evaluated in properly designed phase 2 studies to determine their ability to diagnosed early stage HCC, followed by phase 3 studies that will retrospectively determine if they can detect preclinical disease. If the results hold up, phase 4 studies will prospectively assess their ability to detect asymptomatic HCC. For progress to be made in the validation of current or new markers, several issues need to be addressed. One is identifying the target population for which the surveillance test will be applied. It is unlikely that a single marker will be effective for the surveillance of HCC in all etiologies. Therefore, as the marker(s) progresses through the phases of validation careful attention to the design of the study should be made regarding the eligibility of patients, that is, cirrhotics or HBsAg carriers, viral versus nonviral etiologies, attention to ethnicities, and the focus on early stage HCC. Second, there is an urgent need for a widely accepted prognostic staging system for HCC so there is homogeneity in the studies throughout the world, especially at phases 1 and 2. The five phases used of biomarker development help to evaluate the published literature and to organize biomarker development so that precious specimen resources and funding can be allocated in a rational manner. This structured approach will be important, as many new biomarkers are likely to be discovered based on gene microarray, proteomics, and tumor immunology during the next decade.