Markers for hepatocellular cancer include the best and worst of cancer detection. Although hepatocellular cancer is relatively infrequent compared to other cancers in the western world, HCC has a very high incidence in parts of Asia and Africa. It is estimated to be one of the most common cancer worldwide. High risk factors for HCC include previous hepatitis B infection, heavy alcohol consumption, cirrhosis, and aflatoxin exposure. Alpha fetoprotein may be the best human cancer marker that appears in the serum, but levels of this marker are often not elevated until the tumor is beyond surgical treatment. No other serum or tissue marker is particularly useful. Screening of high-risk populations in China has detected previously undiagnosed HCC in 1,000 of 5 million individuals tested and has led to an increase in survival from 5.5 to 61.6% with surgical resection over those who are later diagnosed with HCC without screening. Elevations of AFP due to yolk sac tumors may be differentiated from those due to HCC on the basis of Concanavalin A reactivity. Immunodetection using radiolabeled anti-AFP and immunoscintigraphy have given inconsistent results that are not as sensitive as ultrasonography in detecting HCC in the liver. Various enzymes, isoenzymes, and other markers may be useful as adjuncts to diagnosis in selected cases, but are not generally as good as AFP alone. If a patient has an AFP-producing tumor, the serum levels of AFP provide an excellent means of monitoring its progression. If the serum AFP levels drop to normal and stay there, cure is almost certain. If, however, the serum AFP level does not fall at the normal catabolic rate after therapy, or subsequently rises, regrowth of metastases are indicated. Immunotherapy using anti-AFP has not been shown to induce remission, but experimental studies indicate that drug-conjugated anti-AFP is effective in inhibiting growth of AFP-producing tumors. Clinical trials using drug-conjugated anti-AFP are now underway. Monoclonal antibodies have not yet identified the "antigens" useful for the diagnosis or treatment of HCC, but epitopes identified by monoclonal antibodies have been studied experimentally in rats which indicate multiple cellular lineages to HCC in cases of experimental chemically induced hepatocarcinoma.