The article discusses how RetinAI addresses the challenges faced in conducting clinical studies in the field of Ophthalmology. It highlights the low rate of drug approvals, high costs, and the difficulty in establishing relevant endpoints as specific challenges in this domain. RetinAI's data management platform, Discovery®, is introduced as a solution that tackles these challenges by enhancing clinical study workflows, reducing time and costs, and generating endpoint insights, in real-time.
In our two previous articles, we first presented the main challenges related to conducting clinical studies, and then discussed how AI-enabled technologies and effective data management can tackle these challenges.
Here, we will explain how RetinAI’s solutions can tackle the challenges of conducting clinical studies in the Ophthalmology space, by enhancing clinical study workflows with benefits from a reduction in time and costs. When dealing with clinical studies in Ophthalmology, it is important to mention that:
- on average only one drug per year gets into the market1
- Ophthalmology has the second highest average per-study costs across phases one, two and three2
- another important challenge is related to endpoints. Take nAMD for example, most relevant endpoints (imaging for example) have been surrogate measures and were not considered as primary outcome endpoints by health authorities. Indeed, anatomical endpoints derived from OCT parameters are well-defined but they do not directly measure how a patient survives, feels or functions, which is what the definition of an endpoint requires. As a consequence, it is hard to address unmet medical needs and measure successful outcomes in a study with the endpoints currently available.
With this premise, we will now re-address the three categories of clinical study challenges (administrative, human and data related), and how RetinAI’s Discovery® for Clinical Studies, can tackle them.
With eCRF management functionality, Discovery can digitalize and centralize all clinical and imaging findings, across multiple study stakeholders, depending on their role and access permissions for reviewing, completing case report forms or analyzing the data. This feature with Discovery significantly reduces manual steps and manual data transfers, and ensures that valuable clinical data is insured, or in other words is centralized and accessed in a secured environment, speeding up submissions for drug approval.
When it comes to patient assessment and enrollment, Discovery is capable of real time evaluation of enrollment criteria and defining study population by identifying threshold endpoint measures for patient subgroups.
RetinAI’s AI models can be used to automatically identify and measure retinal biomarkers, retinal thickness and fluid volume to identify subgroups of patients related to disease progression and treatment outcomes, accelerating clinical trial design and enrollment.
As Discovery centralizes data collection in a single platform, the increased visibility and real time monitoring, in turn, enable better coordination of the study team, together with participative research and teleophthalmology.
Clinical studies’ stakeholders that have access to the platform, indeed, can securely share information, monitor enrollment, ensure the quality of data collected and overall study progress from anywhere, at anytime. This real time monitoring and reviewing opportunity results in earlier and faster GO/NO GO decisions that can shorten time and costs of the clinical study.
Managing, standardizing, aggregating and analyzing huge amounts of complex data, such as medical imaging data, is Discovery's powerhouse.
Discovery has already proven to be able to aggregate clinical study data and RWE to analyze patient biomarkers and treatment response at scale, accelerating the standard analysis workflow in specific cases, reducing timelines of six months to six weeks.
Shortening the clinical study duration is possible thanks to the cloud set up of the platform and is beneficial for the clinical study involved. Contrary to the traditional study set up that is generally sequence-based across steps, when the data are in a secure, cloud-based environment, some of the steps can be done simultaneously, or in parallel: for example, as soon as data is uploaded to the platform, this data is automatically deployed to reviewers for review, so data collection and review happen in tandem.
If this cloud setup is combined with AI models, the potential of reducing clinical study’s duration and enabling interim analysis and GO/NO GO decisions earlier and faster is even greater. This is demonstrated in the Razorbill study, where Discovery is being used for both data collection, data management and analysis.
Related to the cloud, privacy and security matters are of utmost importance. Discovery allows controlled access to data, as each individual user has a personal login where access is limited to just the data that he/she uploaded or that has been shared with them, in a secured and anonymous way.
Lots still needs to be done for rendering clinical studies more cost and time efficient but we are working everyday with this purpose in mind, by developing new AI models and continuously improving our data management capabilities.
1) Emmett Cunningham. Ophthalmology Innovation Summit AAO (2019). https://ois.net/wp-content/uploads/2019/10/ETCunningham-OIS@AAO-2019-Year-in-Review-EXPANED-FINAL-10.18.2019.pdf
2) Sertkaya, A. et al., Examination of Clinical Trial Costs and Barriers for Drug Development (2014). ASPE https://aspe.hhs.gov/reports/examination-clinical-trial-costs-barriers-drug-development-0