Vendor sets plan to gather diverse clinical data for AI

An Australian company is offering a creative solution to building the datasets needed to build artificial intelligence products.

The approach seeks participation from clinics around the world to contribute information and share in the proceeds that result from the end-products.

The concept is being pitched by Presagen, which bills itself as an online platform for clinics to safely and privately connect global medical data with artificial intelligence to co-create scalable and unbiased medical products.

The Adelaide, Australia-based company unveiled the proposal at the HLTH conference in Las Vegas this week. Under the plan, the company says it wants to crowdsource globally diverse datasets to build AI medical products. More data will result in AI products that are robust, scalable and unbiased, company executives contend.

And clinics can then benefit from their participation, says Michelle Perugini, founder and CEO of Presagen. “The financial value that is created from economies of scale can then be shared among contributing clinics via royalties,” she adds. “This enables clinics—particularly small and medium sized clinics—to benefit and unlock the value of their data with AI without wearing the technical or commercial cost and risk.”

Perugini-Michelle-CROP.jpg
Michelle Perugin

The company calls the initiative AI Open Projects, and it’s hoping to kick off initiatives in radiology, ophthalmology (retinal issues) and fertility treatment.

Drawing data from a variety of sources could address two major concerns with AI. Many times, algorithms are based on a limited number of large datasets that are not necessarily diverse. Thus, expanding the pool of data on which AI findings are based.

"To build AI products that solve global problems, you need a global dataset which is diverse and represents different types of people and clinical settings,” Perugini contends. “This is challenging because data privacy laws can prevent private medical data leaving the country of origin. As a result, many focus on building AI from local datasets that are not diverse, creating AI that will be biased and simply will not scale."

Presagen's first AI Open Project in radiology will focus on the detection of lung cancer, in partnership with John MacLean, MD, a clinician with experience in radiology, surgical pathology and general practice, and founder of Doclink, a Sydney-based company.

For reprint and licensing requests for this article, click here.