Enlisting artificial intelligence to fight cancer raises the hope of winning the prolonged war against this despised disease
“Cancer is an umbrella term for thousands of different types of conditions, yet treatment offered today is often generic and does not consider the need for differing therapies for different people,” says Professor Toby Walsh, a leading expert in artificial intelligence (AI).
“However, with AI, all of us can have access to the best experts on the planet to get the best diagnosis and treatment.”
Major technology companies such as Alphabet’s Verily and Google’s DeepMind, alongside a slew of startups, are using cognitive computing to fight cancer by building tools that essentially sort and accumulate medical knowledge and data on a scale that is impossible for humans alone to achieve.
Using machines to encapsulate the knowledge of physicians and experts, and to interpret data better than the specialists, can create a new understanding of cancer to provide better diagnosis and treatment outcomes.
For example, opportunities provided by genome sequencing can be unlocked with AI. The cost of sequencing someone’s genome, which is the unique arrangement of their DNA, is falling and can now be done in 24 hours for $1,000.
Analysing a patients’ DNA enables doctors to understand the type of cancer they have and, importantly, to act on the molecular code of the disease for more accurate treatment.
But there is no standardised approach to handling the high-volume, high-quality data produced from genome sequencing. It can differ from one hospital to another.
Swiss startup SOPHiA Genetics has developed AI that takes raw genome data and studies it to decipher the molecular profile of a person’s cancer to find more suitable and personalised treatment options.
“This technology allows doctors to understand what is driving the cancer and to tackle it with the medicine which best treats the cause of the molecular event that is no longer working,” says the company’s chief executive Jurgi Camblong. The more data analysed by the AI, the more it learns and the more powerful it becomes.
By 2020 the company, which currently has partnerships with five UK healthcare institutions, including Oxford University Hospital, wants to start collecting data on treatment outcomes, so it can identify how a specific strain of cancer responds to different treatments to determine which is most successful.
This article also appeared in The Times newspaper on Thursday the 6th of December.