Comparison of the tolerability and safety of hedgehog inhibitors in real-life: A cohort of 330 patients with locally advanced basal cell carcinoma.

Abstract

9583 Background: Two hedgehog pathway inhibitors (HHIs) have been approved for the treatment of locally advanced basal cell carcinoma (laBCC): vismodegib and sonidegib. Efficacy of both seem similar, even though no head-to-head comparison has been performed. However, adverse events (AEs) in pivotal trials seemed less frequent with a later onset with sonidegib. CARADERM is a French national database created in 2013 to improve the management of rare skin tumors, including laBCC. The objective of our study was to compare the safety profile of HHIs in this real-life cohort. Methods: LaBCC patients from the CARADERM database were reviewed. Patients who started either vismodegib or sonidegib at least one year before analysis were included. Type and grade of AEs were collected when available. Cumulative incidence of the first occurrence of adverse events was estimated, with treatment discontinuation as a competing event. Results: In the total cohort of 452 laBCC patients, 330 met the inclusion criteria: 280 (85%) treated with vismodegib and 50 (15%) with sonidegib. The median follow-up was 22.3 months. Clinical characteristics (including age, gender, performance status, localization and tumor size) were similar for both groups. The cumulative incidence of the first AE was significantly lower with sonidegib (43.5%, 95% confidence interval (95%CI) = 29.0-57.2 at 12 months) than with vismodegib (63.5%, 95%CI = 57.5-68.9 at 12 months, p=0.0014) (Table 1). For vismodegib treated patients, 191 (68%) experienced at least one AE. The most frequent were cramps (n=123, 44%), dysgeusia (n=123, 44%) and alopecia (n=88, 31%). For sonidegib treated patients, 22 (44%) experienced at least one AE. The most frequent were cramps (n=8, 16%), alopecia (n=7, 14%) and dysgeusia (n=6, 12%). Major AEs seemed to be less frequent and to appear later with sonidegib, with a significant difference for cramps (p=0.007) and dysgeusia (p=0.001), but not significant for alopecia (p=0.059). Conclusions: We present here the largest real-life comparison of HHIs in a real-life cohort of laBCC patients. The cumulative incidence of the first AE was significantly lower with sonidegib. Even though the range of AEs was similar with both HHIs, dysgeusia and cramps were less frequent with sonidegib. These data highlight the difference in terms of tolerance of both HHIs, with a later onset and lower frequencies of AEs with sonidegib. However, though both groups were clinically comparable, vismodegib was overrepresented compared to current prescriptions of HHIs, and further analyses are mandatory to confirm these data. Cumulative incidence of first adverse event at 3, 6 and 12 months. 3 months [95CI] 6 months [95CI] 12 months [95CI] Sonidegib 16.3 % [7.5 ; 28.0] 26.8 % [15.2 ; 39.8] 43.5 % [29.0 ; 57.2] Vismodegib 31.6 % [26.2 ; 37.2] 55.5 % [49.5 ; 61.2] 63.5 % [57.5 ; 68.9] 95CI = 95% confidence interval.