Cancer researchers at the Georgia Cancer Center hope the use of AI can help improve patient care as well as assist in developing new treatments.
As a graduate student, Dr. Ravindra Kolhe spent countless hours poring over pages upon pages of data. Now through the use of AI, scientists can access information quicker than they could in the past, aiding them in their research.
“In closed system-based research, we are teaching computers to become smarter and smarter, helping us get research done faster. What took years or decades is now taking months or years,” said Kolhe, the Medical College of Georgia at Augusta University’s professor and chair of the Department of Pathology.
Kolhe likens the use of machine learning and algorithms to the evolution of the iPhone. The first iPhone had substantially less capability than the current iteration.
“We do a good bit of things with AI not only research but in diagnosis that’s where my expertise is. There’s a lot more discussion on early detection and diagnosis of cancer by using algorithms which are based on machine learning,” he said.
Machines have years of images from mammograms to tissue samples under the microscope to cull data. There are samples from people as well as animal models. AI can examine how molecules interact as well as analyzing genomes and biomarkers to provide the diagnosis and suggest what types of treatments would provide the best outcomes for the patient. That’s only part of the type of data available for the AI to analyze.
Not only does AI have the potential for speed, but Kolhe said it can help with accuracy as well.
One study AI is already at work in is funded through Paceline.
Previous article: Paceline grants awarded Aug. 2, 2023
Cancer affects more than just the patient’s physical health; it reaches into the person’s whole life. This new study focuses on the toxicity of cancer to a person’s finances. AI can analyze that type of data and pull it together much quicker than a person can.
While there are benefits to this rapidly growing technology, Kolhe said that AI still must be proven. There can be side effects and downsides.
“I think this is very early in the research, where we want to make sure we validate these systems, and we are 100 percent sure about these machines and how they calculate some of the decisions they are making behind the scenes. This is where most of the research goes to validate some of the hypothesis. Unless it’s 99% accurate, we will take time to get it into clinical use. Most if not all of this is, at the moment, in the research stage,” he said.
Kolhe believes that human interaction is necessary alongside the use of AI.
“No, AI is not going to replace pathologists, but the pathologists who don’t use AI will be replaced by pathologists who use AI,” he said.
Charmain Z. Brackett, the publisher of Augusta Good News and Inspiring: Women of Augusta, has covered Augusta’s news for 36 years. Reach her at charmain@augustagoodnews.com. Sign up for the newsletter here.
