Tumour markers can aid in areas such as diagnosis, surveillance of recurrence, staging and prognosis. This article focuses on 2 tumour markers, alpha-fetoprotein (AFP) and beta-human chorionic gonadotropin (HCG). These tumour markers have been examined for their utility as prognostic indicators in 2 different manners. First, the marker level at diagnosis has been studied to determine if it is prognostic of outcome presumably because of its relation to tumour bulk or to the biological nature of the tumour. A more recent trend has been to investigate tumour marker decline. The finding of a delayed rate of decline suggests a poorer response of the malignancy to chemotherapy. The major focus of the article will be on marker decline of AFP and HCG as prognostic tools in peripheral and central nervous system (CNS) germ cell tumours (GCTs) and hepatic tumours (hepatoblastoma and hepatocellular carcinoma). The articles reviewed here suggest that HCG and AFP can correlate with survival if examined in specific ways, and could potentially be used to tailor treatment for individual patients. One group of authors presents data on patients with GCTs suggesting that satisfactory marker regression is an independent prognostic factor for survival. In a study of hepatoblastoma, data demonstrate that both the magnitude and rate of decline are associated with survival. Marker decline studies in hepatocellular carcinoma do not exist and marker levels at diagnosis do not appear to have a role in potential therapeutic changes. However, data on fucosylated subtype of AFP, Lens culinaris agglutinin A reactive AFP, has shown prognostic significance in hepatocellular carcinoma. The data for CNS GCTs are limited and studies examining serial cerebrospinal fluid HCG/AFP are ongoing. In some diseases, issues relating to timing of marker sampling when examining marker decline need to be studied in greater detail. Hopefully, marker decline studies can be duplicated in the other diseases, to document a potential role in determining outcome. Further studies are needed to test the ability to alter therapy in attempts to improve survival while decreasing toxicity to patients.