Impact of clinical and molecular predictors of benefit from erlotinib in advanced non-small cell lung cancer on cost-effectiveness

Abstract

6531 Background: Previously we have shown that the incremental cost effectiveness ratio (ICER) of erlotinib, in unselected patients with advanced non-small cell lung cancer (NSCLC) after chemotherapy failure, was $95,869 (2007 CAD), per life year gained (LYG), using original patient data from the NCIC CTG BR.21 trial of erlotinib versus placebo in advanced NSCLC. Here, we investigate the value of combining clinical and molecular predictors of outcome to identify more, and less cost- effective patient subgroups for treatment. Methods: Cost-effectiveness ranges of interest were identified a priori: group A ICER <$50,000/LYG, B $50,000- 75,000/LYG and C >$200,000/LYG CAD. Significant clinical and molecular predictors of outcome from the BR.21 trial were assessed to identify groups within these thresholds of cost-effectiveness, and to characterize patient characteristics of each group that had >50 patients. Due to small numbers, EGFR mutation status was not included. Results: The subgroup with lowest ICER ($22,564/LYG; n=52) consisted of patients with adenocarcinoma, never smoker status who had received one prior chemotherapy regimen. By contrast, past and present smokers with wild-type EGFR and K-ras gene mutation had the highest ICER for erlotinib treatment ($3,862,962/LYG; n=52). The presence of one of the following variables, never smoking status, high EGFR copy number or EGFR protein positive tumours (immunohistochemistry), predicted for lower ICERs in groups A and B, irrespective of histological subtype. Female gender was associated with higher ICERs in all three groups. Conclusions: Incorporation of molecular and clinical predictors into economic analyses has the potential to differentiate cost-effective versus cost-ineffective subgroups for treatment. We identified subsets of patients (e.g. smoking history, non-adenocarcinoma histology) who could be treated within a reasonable range of cost effectiveness. While subset analyses must be interpreted with caution, examining different methodological approaches in the economic analysis of novel agents is warranted and a parallel investigation using classification and regression tree methodology is underway. Author Disclosure Employment or Leadership Consultant or Advisory Role Stock Ownership Honoraria Research Expert Testimony Other Remuneration Roche and OSIP Roche, Roche and OSIP