A decision tree was specified for patients with a confirmed diagnosis of non-small-cell lung cancer. Six strategies were defined from relevant combinations of mediastinoscopy, endoscopic or endobronchial ultrasound with needle aspiration, and combined positron emission tomography-computed tomography with F18-fluorodeoxyglucose. Patients without distant metastases and central or contralateral nodal involvement (N2/N3) were considered to be candidates for surgical resection. Diagnostic accuracies were informed from literature reviews, prevalence and survival from the Danish Lung Cancer Registry, and procedure costs from national average tariffs. All parameters were specified probabilistically to determine the joint decision uncertainty. The cost-effectiveness analysis was based on the net present value of expected costs and life years accrued over a time horizon of 5 years.<h4 class=""h4"">Resultsh4>
At threshold values of around 30,000 for cost-effectiveness, it was found to be cost-effective to send all patients to positron emission tomography-computed tomography with confirmation of positive findings on nodal involvement by endobronchial ultrasound. This result appeared robust in deterministic sensitivity analysis. The expected value of perfect information was estimated at 52 per patient, indicating that further research might be worthwhile.<h4 class=""h4"">Conclusionsh4>
The policy recommendation is to make combined positron emission tomography-computed tomography and endobronchial ultrasound available for supplemental staging of patients with non-small-cell lung cancer. The effects of alternative strategies on patients¡¯ quality of life, however, should be examined in future studies.