If successfully adopted, the researchers say their new measures could substantially improve screening, reduce mortality and reduce anxiety around screening. The study was published in the International Journal of Cancer.
“These measures could revolutionize mammographic screening at little extra cost, as they simply use computer programs,” Lead researcher and University of Melbourne Professor John Hopper said in a press release.
“The new measures could also be combined with other risk factors collected at screening, such as family history and lifestyle factors, to provide an even stronger and holistic picture of a woman’s risk," he said.
Using computer software and artificial intelligence, the system measures mammograms for density, brightness, called cirrocumulus, and cirrus, which analyzes the texture of a mammogram. Using these two measures of the cloud-like features of the images, cirrocumulus and cirrus, the researchers substantially improved risk prediction beyond that of all other known risk factors.
The study involved 944 women who had been screened and diagnosed with breast cancer in three studies compared to 2639 matched controls or women who were screened and not diagnosed with cancer. Using their computer program and algorithms the mammogram images of the women were analyzed and compared.
The results showed that the predicted risk using the researchers’ mammogram risk gradient was twice as predictive as other methods of risk prediction.
Currently only about 55 percent of Australian women aged 50-74 get breast cancer screens. The researchers conclude that knowing that screening could also give an accurate risk prediction would encourage more women to get screened. Women with high risk could be monitored more closely and free women with low risk from unnecessary anxiety.
“The new measures could also be combined with other risk factors collected at screening, such as family history and lifestyle factors, to provide an even stronger and holistic picture of a woman’s risk," he said.
Using computer software and artificial intelligence, the system measures mammograms for density, brightness, called cirrocumulus, and cirrus, which analyzes the texture of a mammogram. Using these two measures of the cloud-like features of the images, cirrocumulus and cirrus, the researchers substantially improved risk prediction beyond that of all other known risk factors.
The study involved 944 women who had been screened and diagnosed with breast cancer in three studies compared to 2639 matched controls or women who were screened and not diagnosed with cancer. Using their computer program and algorithms the mammogram images of the women were analyzed and compared.
The results showed that the predicted risk using the researchers’ mammogram risk gradient was twice as predictive as other methods of risk prediction.
Currently only about 55 percent of Australian women aged 50-74 get breast cancer screens. The researchers conclude that knowing that screening could also give an accurate risk prediction would encourage more women to get screened. Women with high risk could be monitored more closely and free women with low risk from unnecessary anxiety.
Sources: University of Melbourne press release and the International Journal of Cancer
No comments:
Post a Comment