The team – from international institutions including Imperial College London and the University of Leuven – say the system could allow medics to classify a tumour as cancerous before surgery, and could potentially improve outcomes for patients. In a study looking at the tumours of 4,500 women, the model predicted cancer risk with 98 per cent accuracy.
Ovarian cancer affects approximately 230,000 women globally – around 7,100 in the UK. Women aged over 50 are most commonly affected, and symptoms include abdominal pain and persistent bloating. The disease is diagnosed through a combination of scans and blood tests.
In the new study, which was conducted with over 4,500 patients from across Europe the researchers used an existing system of ovarian tumour classification called the Simple Rules.
These act like a checklist for healthcare professionals performing an ultrasound scan of the tumour. They highlight ten features – such as size of the growth or whether it has smooth or rough edges – and classify it into one of three categories: cancerous, non-cancerous (benign) or inconclusive. This helps medics decide whether or not to operate.
The team behind the current study wanted to see if the Simple Rules system could be made more detailed – and identify the percentage cancer risk for each individual patient. They also wanted to reduce the number of tumours dubbed inconclusive.
After analysing the tumours of over 4,500 women over the course of ten years, the team created a mathematical model that weights each of the features in the Simple Rules, and calculates the risk of a tumour being cancerous.
The study found the model predicted cancer risk with 98 per cent accuracy. The system maintained its accuracy even without using results from the blood test currently used to help diagnose the condition, called CA125.
Professor Tom Bourne, one of the senior authors on the study, from the Department of Surgery and Cancer at Imperial said: “This system enables us to accurately calculate a cancer risk for each individual patient with ovarian growths, before they undergo surgery. Using this information we can then discuss with patients the best course of treatment.
“Crucially, providing an accurate cancer risk from an ultrasound scan may enable patients with a high cancer risk to see a specialist – and see them sooner. Some estimates suggest less than half of women with ovarian cancer in the UK are operated on by a specialist gynaecological cancer surgeon, despite the outcome being significantly better for women when this happens. If we can get the right women to the right surgeon we can improve outcomes for women with this disease. On the other hand we can also avoid women with benign cysts having surgery they may not need.”
The authors argue that the Simple Rules system is more accurate than the current system also in use in the UK, called the Risk of Malignancy Index, and hope it will be adopted more widely.