Clinical trial recruitment and AI: Easing the burden of GA studies

(Image Credit: AdobeStock/ALL YOU NEED studio)

Reviewed by Dominic J. Williamson

Artificial intelligence (AI) is finding its niche in medical research, with more applications being explored. A team of researchers in the United Kingdom (UK) are exploring AI to simplify appropriate patient recruitment in trials focused on geographic atrophy (GA) associated with age-related macular degeneration (AMD).

First authors and PhD candidates Dominic J. Williamson and Robbert R. Struyven point out that developing new treatments is expensive, can take decades, and relies heavily on data from clinical trials to demonstrate the safety and efficacy of the new treatments.¹

“Today, challenges in recruiting participants stand as a significant obstacle in clinical trials.^2,3 Time taken to fully recruit a trial adds significantly to the cost and can lead to delays, with up to 86% of trials not finishing on schedule,”^4,5 the researchers write.

Williamson and Struyven are from the National Institute for Health and Care Research Biomedical Research Centre for Ophthalmology
at Moorfields Eye Hospital National Health Service Foundation Trust in London, UK, and the Centre for Medical Imaging Computing at University College London.

In light of this, Williamson, Struyven, and their colleagues conducted a cross-sectional study to explore the role of AI in clinical trial recruitment of individuals with GA. The investigators used a retrospective data set from the
INSIGHT Health Data Research Hub at Moorfields Eye Hospital that included 306,651 patients (602,826 eyes) with suspected retinal disease who underwent optical coherence tomography (OCT) imaging between January 1, 2008, and April 10, 2023.

The investigators recount that they trained a deep learning model on OCT scans to identify patients potentially eligible for GA trials using AI-generated segmentations of retinal tissue. This method was compared with a keyword-based electronic health record (EHR) search for accuracy. They also carried out clinical validation with fundus autofluorescence (FAF) images to calculate the positive predictive value (PPV) of this approach by comparing AI predictions with expert assessments.

The main outcome measure was the PPV of AI for identifying trial-eligible patients. The secondary outcome was the intraclass correlation between the GA areas segmented on FAF by experts and AI-segmented OCT scans.

AI performance

The authors report, “The AI system shortlisted a larger number of eligible patients with greater precision (1139[;] PPV, 63%; 95% CI, 54%-71%) compared [with] the electronic health records search (693[;] PPV, 40%; 95% CI, 39%-42%).” When AI and the EHR approach were combined, a total of 604 eligible patients with a PPV of 86% (95% CI, 79%-92%) were identified.

The intraclass correlation of GA area segmented on FAF vs AI-segmented area on OCT was 0.77 (95% CI, 0.68-0.84) for eligible cases. The AI also adjusts to the distinct imaging criteria from several clinical trials, generating tailored shortlists ranging from 438 to 1817 patients, the investigators report.

The investigators conclude that AI has the potential to streamline the clinical trial process, first by identifying patients with greater ease and later in other stages of the clinical trial process.

The researchers noted that they demonstrated that an AI tool can facilitate clinical trial recruitment in AMD.

“Given the intense clinical trial activity currently underway, we believe that implementing this AI-driven strategy will offer a scalable solution to the recruitment challenges in GA trials,” they concluded. “Our future efforts will concentrate on assessing the real-world robustness of this AI solution, ensuring it performs well with routinely collected images of varying quality, protocols, [and] equipment and across diverse populations.”

They also believe that, in addition to prescreening patients, AI may enable site feasibility assessments, data-driven protocol design, and cost reduction. Once treatments are available, similar AI systems could also be used to identify individuals who may benefit from treatment, they hypothesize.

References:

1. Williamson DJ, Struyven RR, Antaki F, et al. Artificial intelligence to facilitate clinical trial recruitment in age-related macular degeneration. Ophthalmol Sci. Published online June 19, 2024. doi:10.1016/j.xops.2024.100566

2. Parke DW II. RCTs: the gold standard’s future. American Academy of Ophthalmology. February 1, 2019. Accessed November 11, 2023. https://www.aao.org/eyenet/article/rcts-the-gold-standards-future

3. Treweek S, Lockhart P, Pitkethly M, et al. Methods to improve recruitment to randomised controlled trials: Cochrane systematic review and meta-analysis. BMJ Open. 2013;3(2):e002360. doi:10.1136/bmjopen-2012-002360

4. Sertkaya A, Birkenbach A, Berlind A, Eyraud J; Eastern Research Group Inc. Examination of Clinical Trial Costs and Barriers for Drug Development. US Department of Health and Human Services. July 24, 2014. Accessed April 9, 2023. https://aspe.hhs.gov/reports/examination-clinical-trial-costs-barriers-drug-development-0

5. Huang GD, Bull J, Johnston McKee K, Mahon E, Harper B, Roberts JN; CTTI Recruitment Project Team. Clinical trials recruitment planning: a proposed framework from the Clinical Trials Transformation Initiative. Contemp Clin Trials. 2018;66:74-79. doi:10.1016/j.cct.2018.01.003