Background: The recommended surveillance strategy for oesophageal adenocarcinoma may prevent as few as 50% of cancer deaths. Tissue biomarkers have been proposed to identify high-risk patients.
Aim: To determine performance characteristics of an ideal biomarker, or panel of biomarkers, that would make its use more cost-effective than the current surveillance strategy.
Methods: We created a Markov model using data from published literature, and performed a cost-utility analysis. The population consisted of 50-year-old Caucasian men with gastro-oesophageal reflux, who were monitored until age 80. We examined strategies of observation only, current practice (dysplasia-guided surveillance), surveillance every 3 months for patients with a positive biomarker (biomarker-guided surveillance), and oesophagectomy immediately for a positive biomarker (biomarker-guided oesophagectomy). The primary outcome was the threshold cost and performance characteristics needed for a biomarker to be more cost-effective than current practice.
Results: Regardless of the cost, the biomarker needs to be at least 95% specific for biomarker-guided oesophagectomy to be cost-effective. For biomarker-guided surveillance to be cost-effective, a $100 biomarker could be 80% sensitive and specific.
Conclusions: Biomarkers predicting the development of oesophageal adenocarcinoma would need to be fairly accurate and inexpensive to be cost-effective. These results should guide the development of biomarkers for oesophageal adenocarcinoma.