Purpose: Positron Emission Tomography has gained an established place in the work-up of patients with Non-Small Cell Lung Cancer (NSCLC). The Scottish Health Technology Board Report (2002) suggests its use could only be supported on cost effectiveness grounds, for patients whose CT scans showed no hilar node spread. This conclusion was sensitive to assumptions about patient utilities for PET-determined health states, real data for which are not available in the literature. Our purpose was to directly elicit patient utilities for PET determined health states.

Methods: Patients referred for the determination of surgical candidacy were recruited and interviewed by a research nurse, prior to PET. Using a series of cards that described the experiences of patients receiving specific forms of treatment and their outcomes, patients were asked to indicate, on a Visual Analogue Scale, how they valued the avoidance of a needless thoracotomy, when compared to experiencing surgery or other treatments with or without the possibility of cure, states contingent upon the accuracy of their PET scan results. WHO performance status and quality of life were evaluated, the latter using the EORTC-Lung instrument.

Results: For our initial 25 patients, the VAS score (with 95% confidence limits) for being a “false positive”, when the PET result might deny potentially curative surgery was 12 [4, 21] but the difference between this state and the VAS scores for the “false negative state” (that would subject them to futile thoracotomy) was not significant (24 [12, 37]). The VAS scores for being “true negative” and “true positive” were 90 [84, 97] and 72 [60, 84] respectively.

Conclusions: The only two randomized trials of PET guided management have produced contrary results, one showing a significant reduction in rates of futile thoracotomy and the other showing no significant difference. Our results suggest that directly eliciting patient valuations of possible PET-determined outcomes would be worth incorporating into future cost-effectiveness analyses, but future work should also assess the patient's valuations of these states after actually experiencing the outcomes.