The digital journey towards personalized cancer therapy
Tailored medical treatment is tantalisingly close now that medical practitioners are able to collect, analyze and learn from petabytes of individual patient data. The Institute of Cancer Research’s head of data science reveals what this will mean for cancer treatment.
As consumers, we’ve become accustomed to online experiences that are dynamically tailored to our browsing and buying habits. YouTube and Amazon can take a good guess at what we’d like to watch, read or listen to next. And, as a result of mass customization initiatives like ‘Build Your Own BMW’ or ‘mi adidas’, a custom car or a pair of sports shoes can be created that are truly unique to us by drawing on thousands or even millions of permutations.
But can such personalization now be extended beyond the realm of the consumer to provide medical treatment that is specifically customized to the idiosyncrasies of an individual and their condition?
Doctors treat millions of people for diseases such as cancer every year, says Dr Bissan Al-Lazikani, head of data science at The Institute of Cancer Research (ICR), but in the absence of a deep knowledge of each patient’s unique profile, the drugs and therapies used to counteract the disease are often blunt instruments. “Every patient will have a slightly different cancer, different conditions and drugs they are taking. No two people are the same, so why should we assume any two people will respond to the same treatment in the same way?” she says.
Personalized diagnosis and treatment is the ultimate aim, and the growing ability to gather vast amounts of data through personalized techniques such as genome sequencing, coupled with the capability of processing it through big data analytics and machine learning, are making this a reality. “We want to be a lot smarter at tailoring therapy, giving the right treatment to the right patients at the right time. And doing that is a data science problem,” says Al-Lazikani.
As one of the world’s most influential cancer research organizations the ICR, along with partner the Royal Marsden Hospital, is already collecting many petabytes of patient data. To make sense of the information ICR has created two data platforms: canSAR, a public database integrating multidisciplinary data designed to help researchers across the globe get faster answers to their drug discovery queries; and Knowledge Hub, an in-house platform designed to provide the big data analysis and prediction capability to tailor treatments to individual patients.
Using the Knowledge Hub, Al-Lazikani and her team are bringing together clusters of patient data based on their genetics, and are already demonstrating that they can dramatically improve the chances of successful treatment. An example of this is the team’s work with prostate cancer clinicians to help them determine the amount of radiation treatment each patient should receive. Currently, patients are given an equal amount of radiation, but this is too much for some and perhaps not enough for others — both scenarios come with their own set of consequences. She says: “Using the Knowledge Hub we’ve integrated all the patient data, represented it graphically and started to build machine-learning predictors of toxicity. We’re already getting really amazing results — by combining signatures that are coming from the genetics, from the lab and from the physics data, you can start predicting which patients are more likely to get toxicity than others. So in future we will be able to really tailor the radiation to each patient.”
Alongside numerous other personalized medical initiatives, such advances give Al-Lazikani an optimistic view of what is a rapidly evolving field. “I absolutely believe that we are on track to see genuinely individual, tailored therapy for patients in about 10 years’ time. We just need to hold our nerve and not be afraid to innovate. We have the brainpower, we have the data. We just need to go ahead and do it.”
- Photography: Ben Gold