Cancer treatment created in 30 days and survival rate prediction achieved with AI- edited by Aishee Biswas

A potential breakthrough in  cancer treatment has been made using artificial intelligence (AI) .  A study published in the journal Chemical Science reveals that researchers from the University of Toronto and Insilico Medicine have developed a new treatment for hepatocellular carcinoma (HCC), the most common form of liver cancer, using an AI drug discovery platform called Pharma.AI. The researchers applied AlphaFold, an AI-powered protein structure database, to the platform to uncover a previously unknown treatment pathway for cancer. They also developed a "novel hit molecule" that can bind to the target without assistance. The AI system was able to create the treatment in just 30 days and can predict a patient's survival rate.

 

The potential drug was developed within a span of just 30 days  after selecting the target and synthesizing only seven compounds.

Further rounds of compound generation led to the discovery of a more potent hit molecule. However, before it can be widely used, any potential drug would have to undergo clinical trials.

 

It was stated that while the world was fascinated with generative AI's advances in art and language, their algorithms using generative AI managed to create potent inhibitors of a target with an AlphaFold-derived structure. AI is quickly revolutionizing the drug discovery and development process, as traditional trial and error methods are slow, costly, and limit exploration opportunities. Michael Levitt, a Nobel Prize laureate in chemistry, said that this paper is further proof of AI's potential to transform the drug discovery process with improved speed, efficiency, and accuracy. Combining AlphaFold's predictive power with Insilico Medicine's Pharma.AI platform's target and drug-design capability could signal a new era of AI-powered drug discovery.

 

The potential of AI in healthcare goes beyond the advancements in individual AI applications. According to Alan Aspuru-Guzik, a professor of chemistry and computer science at U of T's Faculty of Arts & Science, AI developments can revolutionize healthcare by expanding the range of diseases that can be targeted with generative models that use an AI-derived protein. He also hinted at the possibility of self-driving labs, which would push the boundaries even further. Another study published in JAMA Network Open showed that an AI system developed by scientists at the University of British Columbia and BC Cancer could predict cancer patient survival rates using doctors' notes.

 

Using natural language processing (NLP), a type of AI that can comprehend intricate human language,  this model is capable of examining doctors' notes following an initial consultation  and recognizing distinct characteristics of each patient. It successfully forecasted survival rates at six months, 36 months, and 60 months with an accuracy rate of over 80%. In contrast to previous models that could only apply to certain cancer types, this model can determine rates for all cancers.