Precision medicine and AI: Made for each other

As precision medicine transforms disease treatment into a patient-by-patient art and science, AI is poised to help quickly identify or even predict genetic mutations, pointing the way to highly targeted therapies.

Eventually algorithms may completely displace current genetic testing processes, which can take weeks to complete and apply to individual patients.

And genetics is just one avenue of precision medicine in which AI may change the game. Big data is increasingly available for training cancer-busting algorithms in genomics, proteomics, epigenomics and others.

So noted the authors of an article running in the Journal of Translational Medicine.

“Digital biomarkers present a big opportunity to measure clinical endpoints in a remote, objective and unbiased manner,” wrote Attila Seyhan, PhD, of Brown University and Claudio Carini, MD, PhD, of King's College London. “AI, machine learning algorithms, computational biology and digital biomarkers will offer an opportunity to translate new data into actionable information, allowing earlier diagnosis and precise treatment options.”

Seyhan and Carini gave several examples of advances already achieved. In one, scientists retrained an off-the-shelf Google deep learning algorithm to identify the most common types of lung cancers. The tool had 97% accuracy and even identified altered genes driving abnormal cell growth.

“The genetic changes identified by this study often cause the abnormal growth seen in cancer and they can change a cell’s shape and interactions with its surroundings, providing visual clues for automated analysis,” wrote Seyhan and Carini.

Summarizing their observations, the authors placed AI’s role in precision medicine within the broader context of patient-centric cancer care.

Drug development, for example, “is a challenging long process with many obstacles (in) the way,” they wrote. “Though several strategies have been proposed to tackle this issue, there is a general consensus that a better use of biomarkers, ‘omics’ data, AI and machine learning will accelerate the implementation of a new medical practice that will depart from the widely spread concept ‘one drug fits all.’”

The journal is offering the article in full for free.