Made-in-Canada software helps better predict cancerous lesions in CT scans
Published Wednesday, September 4, 2013 5:17PM EDT
A team of Canadian researchers led by the Terry Fox Research Institute has developed new software that can accurately identify whether a lesion on the lung of a high-risk individual -- as seen on a CT scan -- is benign or malignant nine times out of ten.
To develop the technology, Dr. Stephen Lam, of the B.C. Cancer Agency, said his team looked at data sets of current and former smokers who had nodules on their lungs. They were able to predict whether a spot was cancerous with an accuracy rate of greater than 90 per cent.
The tool, when combined with a low-dose CT scan, will help doctors make decisions and potentially save the patient from unnecessary testing, Lam said.
"What this means is that we can reduce the number of repeat CTs -- or even biopsy -- in a great proportion of people. Maybe reduce the proportion of people that require follow-up within the first year to less than 10 per cent."
The researchers say the finding is important because Canada, along with many other countries, are considering whether to start screening current and ex-smokers at high risk of developing lung cancer.
But they want to avoid the “overdiagnosis” of these patients, as additional testing shows that 98 per cent of the spots detected by CT scans end up being non-malignant.
At the moment, doctors measure the risk of a lesion or nodule developing into cancer based on nodule size.
However, the Canadian-developed software includes other parameters identified as risk factors, such as: old age, female sex, family history, emphysema, nodule location, nodule type, lower nodule count and the presence of sharp or needle-like points on the nodule.
Dr. Heidi Roberts, a radiologist at Toronto's Princess Margaret Cancer Centre who helped author the study, said the software will help eliminate risks associated with unnecessary testing.
"It makes testing more efficient, it's reducing the number of invasive tests you would have to do and it decreases the harm to the person who is entering the lung-cancer screening," she told CTV News.
"Almost everybody will have at least one lung nodule and if we have an easy, fast reliable method to non-invasively say this nodule is harmful or harmless, it will help us a lot in implementing lung-cancer screening."
In the study, which was published Wednesday in the New England Journal of Medicine, a total of 12,029 lung cancer nodules were observed on the CT scans of 2,961 current and former smokers between the ages of 50 to 75.
The CT scans were obtained from two separate studies, where participants were followed for an average of three and eight-and-a-half years respectively.
Lam hopes the tool will help to establish Canadian guidelines on lung-cancer screening with CT scans. Provincial health programs would then have to approve the scanning programs.
"We hope it will change in the near future," he said.
While there are still gaps regarding the use of CT screening, researchers in Ontario have already developed recommendations for screening populations at high risk of lung cancer.
Their recommendations are published in the October issue of the International Association for the Study of Lung Cancer's journal, the Journal of Thoracic Oncology. They include:
- Screening for lung cancer with a low-dose CT is recommended in high-risk populations, defined as people aged 55 to 74 with a minimum smoking history of 30 "pack years" or more. (Pack years are determined by multiplying the number of packs smoked each day by the number of years a person has smoked.)
- People at high risk for lung cancer should commence screening with an initial low-dose CT scan. That is to be followed by annual screens for two consecutive years, and then once every two years after each negative scan.
Roberts said these guidelines represent where the software tool may be able to play a role.
"This tool is easy to implement for anybody who is reading CT scans," she said. "If this is available that will help everybody out in the community."
In another study published Wednesday, it was found that low-dose CT scans are more sensitive in detecting early-stage lung cancers than chest radiography.
Using data from the U.S. National Lung Screening Trial, the study found that when compared to radiography, two annual lung-cancer screenings using low-dose CT scans resulted in a greater number of early-stage lung cancers being diagnosed.
Positive screening results were observed in 27.9 and 16.8 per cent of participants in the low-dose CT group during the first and second screenings respectively, compared to 6.2 per cent and 5 per cent of participants in the radiography group.
The study found that low-dose CT scans also resulted in a decrease in the number of advanced-stage cancers being diagnosed when compared to radiography.
A total of 53,454 participants enrolled in the trial between 2002 to 2004, with participants being tracked until 2009.
Vancouver resident and study participant Chris Douglas has little doubt that getting an early CT scan saved his life.
Douglas, 65, smoked for 45 years before finally quitting in 2008. In 2011 after getting tested doctors discovered a nodule on his lung and decided to remove it.
"I don't think it's an exaggeration to say that they saved my life," he said.
Now Douglas is getting CT scans every year as a cancer survivor and has converted to a healthy lifestyle by exercising and not smoking.
The TFRI-led study comes just two years after the 2011 U.S. National Lung Screening trial found that CT screening can reduce lung cancer mortality by up to 20 per cent.
With a report from CTV medical specialist Avis Favaro and producer Elizabeth St. Philip