Much of the recent attention on artificial intelligence has focused on the potential threats it poses. But for life scientists, it represents a powerful new tool for improving the diagnosis and treatment of cancer.
Louisiana health care now finds itself at the cusp of a new era, with AI powering molecular-level cancer therapies, detection and diagnostics. Some of the breakthroughs have already taken place. Others are developing. All of them lead to the promised land of improved outcomes and patient care.
“We’re moving toward what we call ‘precision medicine,’” said Joe Ramos, director and chief executive officer of the Louisiana Cancer Research Center.
Precision medicine uses a patient’s genomic, environmental and lifestyle information to guide treatment. LSU Associate Professor of Computational Biology Michal Brylinksi and his team built an AI-powered cancer drug discovery engine that could formulate molecular-driven treatments for each patient.
Brylinski’s team trained its AI to recognize the connections between specific cancers and drugs that control the production of kinases, proteins that accelerate cancer growth.
That’s important because not all cancers are created equal
“Even when you get to the level of the specific subtype of breast cancer,” Ramos said, “not every triple negative breast cancer is the same from patient to patient.”
Brylinksi’s drug discovery engine -- CancerOmics.net -- provides a clearer picture of how a specific cancer will respond to a particular drug or combination of drugs.
Ramos, who came to LCRC in June 2022, said CancerOmics.net’s ability to match the drug to the patient is one of five major areas where artificial intelligence is improving cancer treatment. The others are:
Ramos said LCRC, which was founded in downtown New Orleans in 2002, wants to encourage its researchers to focus on these five major areas, and also plans to recruit more scientists to continue to add specialized expertise. The center currently hosts over 200 researchers from LSU Health New Orleans, Tulane University, Xavier University and Ochsner Health.
Yaguang Xi, a professor and vice chair for Research at LSU Health New Orleans’ Department of Genetics, is among them. Xi is collaborating with scientists working with AI applications to match drugs to the proteins controlling colorectal cancer.
Dr. Hong-Wen Deng, a professor and director at Tulane Integrated Institute of Data and Health Science, trained his artificial intelligence to determine whether a particular polyp or growth is colorectal cancer and if so, what stage of the disease it is.
Deng published a paper in the journal Nature Communications that concludes AI can detect colorectal cancer as well as or better than pathologists.
Ramos said Deng’s approach could apply to other forms of cancer – distinguishing, for example, between a breast cancer lesion and a benign growth.
“We could have sort of an artificial intelligence pathologist who could also take a look at these things. You can imagine that it reduces the amount of time the (human) pathologist might need to look,” Ramos said.
Although the AI diagnostics are powerful, Ramos said it will be some time before medicine moves to entirely artificial intelligence-driven diagnostics.
It’s more likely that AI would be used to take the first pass, with the pathologist determining whether the AI finding was correct, he said. This approach would reduce costs and improve reliability: the AI could catch things the pathologist might miss, he said, without removing the essential human element entirely from the process.