“Virtual cancer patient” predicts how breast cancer patients respond to treatment

A computer generated “virtual cancer patient” can predict how patients with advanced breast cancer respond to treatment with 70 per cent accuracy, scientists reveal at the NCRI Cancer Conference in Birmingham today.

The team from Nottingham City Hospital, in collaboration with researchers at the Institute for Medical Biomathematics in Israel, undertook a pilot study on 33 patients with advanced breast cancer that had spread to the liver, lymph nodes or lungs. They used the cyber-patient, based on advanced mathematical models, to find out which drug out of two would work best in each patient, based on certain characteristics of their cancer, such as the size of their tumours and how fast they were growing.

In this retrospective study, part funded by Cancer Research UK, the Optimata “virtual cancer patient” (OVP) model accurately predicted how around 70 per cent of the patients responded to their treatment. In the future, technology like this could help doctors tailor treatment more accurately to ensure every patient receives the most appropriate therapy to treat their particular disease.

The two chemotherapy drugs they compared were called docetaxel and doxorubicin – these can be used on their own to treat a number of cancers but can have different effects in different people. Looking at drugs that work best on their own, rather than in combination with other drugs, provided the researchers with the most clear cut data. Patients with advanced cancer were chosen for this pilot study because they are most likely to suffer serious side effects, such as fatigue and sickness, due to the volume of anti-cancer drugs they receive.

Dr Abhik Mukherjee from Nottingham City Hospital who worked on the study, said: “Every cancer is slightly different and every patient will respond to treatment differently. We wanted to find a way to predict how patients would respond to a particular drug in order to limit their side effects and give them the best chance of beating their disease.”

The OVP was “trained” using clinical data from real patients. The team programmed the model to look at how the drugs affected the growth of the cancer, how the drugs behaved in the body and how the cancer cells responded to the drugs. Once the model had been fully “trained” they compared the predictions of the OVP programme with the actual response of patients to the treatment to test the effectiveness of the technology.

Dr Stephen Chan, who also worked on the study, added: “We found the computer programme accurately predicted how the patients responded to treatment in around 70 per cent of cases. However this was a pilot study in a small number of patients, so now we want to fine tune the model to improve its accuracy and test it in a larger study. We also want to see how it works when we use combinations of drugs and whether the model can predict if a patient will suffer other side effects in response to the treatment.”

Kate Law, director of clinical trials at Cancer Research UK, said: “This was a very interesting early study that could potentially have a big impact on how cancer patients are treated in the future. Tailoring treatments to individual patients will ensure the best possible outcome for every patient. This is a hugely important area of cancer research, so we look forward to seeing how this technology performs in a larger trial.”

ENDS

For media enquiries please contact Emma Gilgunn-Jones on 020 7061 8300 or, out-of-hours, the duty press officer on 07050 264 059.