Detecting a cancer can be personalized down to the individual tumor, scientists say.
The changing nature of cancer, which is often drug-resistant, is one of the biggest problems in their treatment. However, this AI may allow doctors to predict how cancer cells will mutate, and thus enable patients to receive personalized treatment earlier.
The technology, known as Revolver (re-evolution cancer) that detects patterns in DNA mutations that can predict future drug resistance in tumors.
To create Revolver, the researchers analyzed the 768 samples of tumors from 178 patients with lung cancer, breast cancer, kidney and intestine.
Lead author Dr Andrea Sottoriva from the Institute of cancer research in London said: “With this tool we hope to remove one of the trump cards of cancer – the fact that it develops unpredictably, without our knowledge of what will happen next”.
Chief Executive officer of ICR, Professor Paul Worker added: “This new approach with the use of AI can allow to personalize the treatment in more detail and at an earlier stage, adapting it to the characteristics of each individual tumor.”
Revolver can also be used to predict if in the future patients will increase the resistance.
Professor Workman said: “If we can predict how it will develop the tumor, treatment can be changed before adaptation and drug resistance will ever happen, putting us a step ahead of cancer.”
This means that doctors will be able to predict how it will look this tumor in the future.
This new technology uses AI may have many implications for cancer treatment, not only in drug resistance but also in preventing relapse. Scientists also found a link between certain DNA sequences repeated tumor mutations and survival of the patient.
This suggests that recurring patterns of DNA mutations can be used as an indicator of treatment prognosis, helping to shape future treatments.
Dr Sottoriva said: “We have developed a powerful tool AI – it can make predictions about future steps in the evolution of tumors on the basis of certain patterns of mutation, which until now remained hidden in complex datasets. Giving a glimpse into the future, we could use this tool AI for intervention at an earlier stage, predicting the next step cancer.”