AI is FOUR TIMES more effective at predicting an ovarian cancer sufferer’s risk of dying than existing CT scans
- Assesses a tumour’s size, structure and DNA to predict a patient’s prognosis
- Could even help doctors choose the best course of treatment for the individual
- Analyses how similar growths responded to chemo or surgery in the past
AI is four times more effective at predicting an ovarian cancer sufferer’s risk of death than existing CT scans, research suggests.
A study found an artificial intelligence model that assesses a tumour’s structure and genetic make-up better predicts a patient’s prognosis than methods currently used by doctors.
The technology could even help doctors choose the best course of treatment for an individual patient by analysing how similar growths have responded to chemo or surgery in the past.
Researchers claim AI has the ‘potential to transform the way healthcare is delivered and improve patient outcomes’.
AI is four times more effective at predicting an ovarian cancer sufferer’s risk of death than existing CT scans used by doctors, research suggests (stock)
The research was carried out by Imperial College London and led by Professor Eric Aboagye, from the faculty of medicine, department of surgery & cancer.
‘The long-term survival rates for patients with advanced ovarian cancer are poor despite the advancements made in cancer treatments,’ Professor Aboagye said.
‘There is an urgent need to find new ways to treat the disease.
‘Our technology is able to give clinicians more detailed and accurate information on the how patients are likely to respond to different treatments, which could enable them to make better and more targeted treatment decisions.’
Ovarian cancer affects around 7,300 new women every year in the UK, Target Ovarian Cancer statistics show.
And in the US, 22,530 women are diagnosed with the disease annually, according to the American Cancer Society.
Although 90 per cent of sufferers live five years or more if the cancer is diagnosed during its earliest stage, vague symptoms – including bloating, loss of appetite and abdominal pain – mean it often goes unnoticed.
To test whether AI could assist ovarian-cancer management, the researchers input CT scan and tissue sample data into a mathematical software called TEXLab.
This information was taken from 354 patients who battled the disease between 2004 and 2015.
Ovarian cancer is currently diagnosed via a blood test thats look for a substance called CA125, followed by a CT scan that shows details of the tumour.
This helps doctors know how far the disease has spread and determines the type of treatment patients should receive.
However, the scan do not give a detailed insight into a patient’s likely outcomes or the most effective treatment.
TEXLab examined the tumours’ structure, shape, size and DNA – four characteristics that significantly influence patient survival and give an indication of their prognosis.
The patients were then given a score, known as Radiomic Prognostic Vector (RPV), which indicates how severe their disease is.
Results – published in the journal Nature Communications – were compared against blood tests and existing prognostic scores used by doctors.
Not only was the software four times more accurate when it came to assessing patient survival, the team also discovered just five per cent of patients with high RPV scores had a survival rate of less than two years.
High RPV was also associated with poor response to chemotherapy and surgery.
This suggests RPV could be used to predict how patients will respond to treatments.
Co-author Professor Andrea Rockall, clinical chair of radiology, added: ‘Artificial intelligence has the potential to transform the way healthcare is delivered and improve patient outcomes.
‘Our software is an example of this and we hope it can be used as a tool to help clinicians with how to best manage and treat patients with ovarian cancer.’
Larger studies are due to be carried out to determine how accurately TEXLab can predict the outcome of surgery and drug therapies for individual patients.