OBJECTIVES : Entrectinib is a tyrosine kinases inhibitor (TKI) targeting ROS1-fusion proteins; in NSCLC, it inhibits tyrosine kinases encoded by the gene ROS1. Most recent clinical data in conjunction with Italian clinical practice inputs are used in a previously developed international AUC model in order to estimate its cost-effectiveness in the treatment of ROS1+ NSCLC relative to crizotinib, the TKI currently approved in Italy.

METHODS : Proportion of patients within each health state are based on the progression-free survival (PFS) and overall survival (OS) curves, estimated by parametric extrapolation of observed data from the pooled entrectinib trials; given the lack of head-to-head studies, a Matching-Adjusted Indirect Comparison (MAIC) has been carried out to estimate hazard ratios vs. crizotinib. Economical inputs on oral drug costs, adverse events management, and supportive care are based on national literature data and current tariffs. Drugs are costed based on protocol/label planned dose and observed length of treatment. Deterministic and probabilistic sensitivity analyses assess the impact of uncertainty surrounding the variables estimates.

RESULTS : The model estimates that entrectinib, when compared to crizotinib, induces a survival gain of 1.4 life years in the post-progression phase. The associated costs increase by about 115K €, mainly due to acquisition costs of entrectinib, longer TKI treatment duration and improvement of life expectancy, leading to an ICER around 90K €/LY.

CONCLUSIONS : Entrectinib increases expected survival in ROS1+ NSCLC patients with respect to crizotinib. ROS1-positive NSCLC may be considered as a rare disease, making such an ICER potentially acceptable considering its effectiveness and safety profile in a developed country as Italy. However, these results should be interpreted cautiously: limitations are due to the ineluctable indirect comparison approach and poor knowledge on the natural history of the ROS1-NSCLC population that make the analysis subject to considerable uncertainty.