Decade-long profile of imaging biomarker use in ophthalmic clinical trials

PURPOSE. The purpose of this study was to investigate the use of imaging biomarkers in published clinical trials (CTs) in ophthalmology and its eventual changes during the past 10 years. METHODS. We sampled from published CTs in the fields of cornea, retina, and glaucoma between 2005–2006 and 2015–2016. Data collected included year of publication, phase, subspecialty, location, compliance with Consolidated Standards for Reporting Trials, impact factor, presence and use of imaging biomarkers (diagnostic, prognostic and predictive; primary and secondary surrogate endpoints), and use of centralized reading centers. RESULTS. We included 652 articles for analysis, equally distributed in three timeframes (2005– 2006, 2010–2011, and 2015–2016), mainly reporting phase IV CTs and trials on procedures (42.2% and 35.4%, respectively). Imaging biomarkers were included in 46.3% of the analyzed CTs and their use significantly increased over time (P < 0.05). Optical coherence tomography was the most frequently used device (27.7%), whereas diagnostic biomarkers and secondary surrogate endpoints were the most frequent biomarker types (19.5% and 22.5%, respectively). Early-phase CTs showed an increase in the use of biomarkers for patient selection and stratification over time (P < 0.05), but not in the use of imaging surrogate endpoints (P = 0.90). Only 3 of 59 (5.1%) of phase III CTs included primary surrogate imaging endpoints, whereas secondary surrogate imaging endpoints were present in 50.8% of these trials (P < 0.001). Retinal CTs had the highest prevalence for each type of imaging biomarker (P < 0.001). Reading centers were used in 52 of 302 CTs (17.2%), with no significant time-related increase. CONCLUSIONS. Imaging biomarkers are increasingly used in published CTs in ophthalmology. Additional efforts, including centralized reading centers, are needed to improve their validation and use, allowing a wider use of these tools as primary surrogate endpoints in phase III CTs.