Evidenced by the numerous innovations that have arisen lately, artificial intelligence has seemingly boundless potential to help improve human life. Uses for the technology span a vast array of applications, including many within the healthcare sector. Matching medical professionals to understaffed hospitals and helping doctors develop patient-specific treatment plans may very well be just the tip of the iceberg.
Dallas-based MedCognetics is using AI to take cancer detection to the next level. The company’s medical imaging tech works to pinpoint early signs of breast cancer based on an extensively trained machine learning model. Its system is designed to assist radiologists as they conduct mammograms and review the scans’ findings.
MedCognetics was founded in 2019, right before the pandemic spread rampantly across the U.S. and healthcare workers became swamped with an influx of patients. These days, healthcare systems across the country are still seeing lingering effects of this event. Medical staff, such as radiologists, are often fewer in number, have heavier workloads and work longer hours to keep up with patient demand, often leading to burnout.
Helping ensure these individuals can still provide patients with the services they need, MedCognetics provides digital health software that can be deployed on-premises via the cloud or through the web. MedCognetics’ QmTRIAGE solution works to detect the earliest manifestations of cancer, according to the company.
A key focus of MedCognetics work isn’t just accurately spotting signs of breast cancer but infallibly identifying it in patients spanning all ethnicities, according to the company. It does so by training its machine learning model in cities with diverse patient populations.
“We’re dealing with what is called AI bias. A high-level way to look at it is AI systems are like a child, and every time you teach it something that’s what its reference is,” Ron Nag, CEO and co-founder of MedCognetics, told Built In. “This means that if you train with, for example, breast imaging data from a location like Beverly Hills, where the population has access to good, consistent healthcare and is sustained by a healthy, supportive lifestyle, these patients are not representative of the people that [live] in Sub-Saharan Africa, Kampala [or] Malaysia. ... It has been already proven that these kinds of systems will miss [most] of the time if they’re not trained at the very beginning.”
In training its AI system to be unbiased, MedCognetics is simultaneously working to reduce disparities that people of color have historically experienced in the healthcare space. MedCognetics wants to help improve healthcare, prevention and diagnoses and help diverse groups see better health outcomes overall, the company said in a statement.
The company has garnered national recognition for its tech solution that aims to make medical imaging more efficient and reliable. The National Institutes of Health, or NIH, established the Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity program to close gaps in the AI and machine learning field caused by a lack of diversity in the sector’s researchers and data. The program is known by the shorthand AIM-AHEAD.
MedCognetics received a $750,000 grant from AIM-AHEAD earlier this year. The money will help further MedCognetics’ research and achieve its mission of providing unbiased AI-enabled healthcare services around the world, Nag said in a January statement.
MedCognetics’s tech is intended to be a decision support tool for physicians. Rather than having to manually discern potential areas of concern for over a hundred cases a day, QmTRIAGE pre-processes those cases and arranges them according to which results contain “highly suspicious” findings, according to Nag. This way, physicians can quickly decide if these readings are genuinely concerning and prioritize them as such.
Having recently gained 510(k) clearance from the U.S. Food and Drug Administration to implement its tech, MedCognetics is one step closer to achieving its goals. In recent years, the FDA has heightened its standards for AI medical devices in an effort to edge out biased systems, according to Nag. When applying for the certification, MedCognetics presented its data findings to the FDA to prove QmTRIAGE’s capabilities.
“The FDA [clearance] and the NIH grant that we got really lends itself to the solemn belief that if not [for] eradicating cancer, you can use this kind of tech [to turn cancer] into a chronic illness or a manageable illness,” Nag said. “I think cancer can get there if people thoughtfully address the problem, not just make point solutions, and that’s our point.”
As MedCognetics continues to pursue its work of reducing bias in medical imaging, the healthtech company plans to expand its cancer detection capabilities into additional parts of the body, like lung cancer for instance, Nag said. It also wants to include other modalities for deploying its tech. Already offering cloud and on-premise versions of QmTRIAGE, MedCognetics will build out a web portal for institutions that don’t have an imaging solution on-site.