Artificial intelligence (AI) is no longer advancing in baby steps. In fact, it’s starting to leap forward and sprint toward the future. Its recent applications in medicine have been remarkable, especially as a diagnostic tool. For instance, AI can detect melanoma — a form of skin cancer — in its early stages, allowing quick response and increased survival rates — up to 98 per cent! That’s a big deal, because melanoma claims as many as 65,000 lives a year, the World Health Organization

reports. AI is now also able to diagnose diabetic retinopathy (DR) — a form of blindness caused by diabetes. This  program, developed by the research team at Google Brain, has passed the first tests and is as precise and efficient as highly-skilled and experienced ophthalmologists, Forbes reports. With almost 40 per cent of Americans affected by DR, and nearly 415 million patients worldwide diagnosed with diabetes who are at risk of DR, this affects too many people. These are amazing advances, but perhaps the most exciting examples of AI in medicine can be found in pharmacology.

Nothing personal, but AI is just better in chemistry

ITProPortal’s research shows that 40 per cent of respondents from the pharmaceutical industry confirmed that their organisations had already deployed AI, and that there have been no problems so far. This only makes sense. In a field in which data analysis demands more than the human mind can provide, the ability to find patterns and discover meaning in large data sets places a premium on the strengths of artificial intelligence. It’s all possible thanks to the ability of AI to process huge amounts of patient data, analyse medical images, and conduct automated genetic profiling. This personal, rather than one-size-fits-all, approach to care finds the right combination of treatment options for each patient.

But to bring personalised medicine to the masses, we need new, highly individualised drugs — and that means more R&D. Molecule synthesis is a painstaking process of research and analysis. Traditionally, it involved hours and hours spent studying chemical reactions and combinations, then carefully analysing them to plan out a synthesis of the new molecule. Even then, the work is just beginning, as the synthesised chemical must then endure a series of tests and trials to see whether it performs as planned. But, those long days in the lab are behind us now as a new, smart algorithm yields the same results in a fraction of the time. Mark Waller and his colleagues from Shanghai University have designed a program that can “map out a six-step synthesis for an Alzheimer’s drug intermediate” in a mere 5.4 seconds. And it’s not just fast — it’s smart, too. The team conducted an experiment involving 45 organic chemists, asking them to choose between two synthetic pathways for nine different molecules. For each of the nine molecules, one of the pathways was designed by chemists, while Waller’s AI designed the other. To everyone’s surprise, 57 per cent of the participants preferred the AI’s pathway formulation. Clearly, then, AI is central to the future of pharmaceutical research.

The pharmaceutical Industry open its gates to AI

The power of AI hasn’t gone unnoticed, and the world’s biggest pharmaceutical names have decided to turn to artificial intelligence to speed up drug discovery. Developing an effective drug takes time, and AI can greatly improve the process by providing precise data about molecule combinations. This could significantly reduce the time spent synthesising effective medicine by allowing AI to explore these tricky molecular combinations rather than running expensive, lengthy laboratory tests. Some of the biggest names in the industry are already embracing AI because of that. For instance, GlaxoSmithKline is planning to invest $43 million in the field, and Merck & Co, Johnson & Johnson, and Sanofi, true giants in the industry, are also keen on exploring the potential of AI and its ability to “streamline the drug discovery process”, Business Insider reports.

According to the latest studies from Weill Cornell Medicine researchers, AI can help with cancer patients, finding the best possible drug combinations for each patient, saving research time and enabling quicker treatment. Again, this is possible because artificial intelligence can easily work with large data sets, finding meaning where people — even the smartest people — can only see numbers. This talent is critical in the fight against cancer, where we need to study combinations of different medications that quickly add up to mind-boggling numbers. For instance, a hundred different drugs can yield as many as 5,000 unique combinations. And the greater the initial number of drugs, and the more complex the combinations, the more impractical and impossible human analysis becomes. That’s where AI is really proving its value.

Of course, this is just a hint of what AI could achieve as its capabilities grow with every new breakthrough in the field. As AI gets smarter, it’ll become easier to use it in the best possible way – to once and for all erase the word ‘incurable’ from medical records.

Bibliography

The above guest blog is courtesy of Richard Van Hooijdonk.