AI can screen bowel cancer patients who won't respond to new NHS drug

Researchers in the UK and Ireland have developed an artificial intelligence (AI) based tool to predict which patients with advanced stage bowel cancer are unlikely to respond to bevacizumab or drugs to slow the growth of the cancer. The findings are important because the drug is only available on the NHS, but its benefits are only felt by a small proportion of patients.

Citing a report by The Guardian, Tuesday, April 14, this method was developed by researchers from the Institute of Cancer Research in London and RCSI University of Medicine and Health Sciences in Dublin. The goal is to avoid patients from treatment that is likely to be ineffective, but still carries serious side effects.

The problem is indeed not small. Almost 10,000 cases of advanced stage colon cancer are found each year in the UK. Diagnosis is also increasing in the young adult group. Colon cancer is the second highest cause of death after lung cancer. If detected early, the chance of survival can reach 98 percent. However, in the advanced stage, the five-year survival rate can drop to 10 percent.

In the study, researchers tracked 117 European bowel cancer patients undergoing chemotherapy and bevacizumab, a drug approved by the National Health Service (NHS), the UK's public health service, in December. The drug works by inhibiting a protein that tumors need to grow. However, only a small proportion of patients actually benefit, while the risks are quite heavy, ranging from blood clots to gastrointestinal disorders.

Still according to The Guardian report, researchers use an AI tool called PhenMap to read complex tumor genetic data. From there, they found patterns of patient response to drugs, as well as identifying groups of patients with the same gene mutations who are at high risk of negative reactions.

Professor of stratification and precision medicine at the Institute of Cancer Research, Anguraj Sadanandam, said the choice of therapy for patients is very limited when the cancer has spread. "However, we know that the majority of patients will not benefit from this drug, which means thousands of people can face unpleasant side effects in vain," he said.

He added that the AI method helps read patterns that are difficult for humans to capture. "In this study, we show that this method allows us to identify patients who are least likely to respond to treatment with bevacizumab," Sadanandam said as quoted by The Guardian.

The researchers say the findings still need to be tested in a larger group of patients. They also want to see if a similar approach could be used for other types of cancer.

This finding suggests that treatment is no longer about what drugs are available, but who is likely to be helped by the drug.