British researchers have developed a new model to predict the 10-year risk of breast cancer
London Aug 27: A team of British researchers has developed a new model that reliably predicts a woman’s likelihood of developing breast cancer and then dying within a decade.
The study, published in The Lancet Digital Health, analyzed anonymized data from 11.6 million women ages 20 to 90 from 2000 to 2020.
Not all of these women had a previous history of breast cancer, a precancerous condition called ductal carcinoma in situ, or DCIS.
Identifying women most at risk of developing the deadly cancer could improve screening. These women may be called upon to start screening early, called for more frequent exams, or seen with different types of imaging.
Such a personalized approach could lead to lower breast cancer mortality while avoiding unnecessary screening of women at lower risk. And researchers at the University of Oxford said that women at higher risk of developing fatal cancer could also consider treatments that try to prevent the development of breast cancer.
“This is an important new study that has the potential to introduce a new approach to screening. Risk-based strategies can provide a better balance of benefits and harms in breast cancer screening, allowing more personalized information for women to help improve decision-making. Risk-based approaches can,” she said. “Also helping to use health service resources more efficiently by targeting interventions to those most likely to benefit.”
The researchers tested four different modeling techniques for predicting the risk of dying from breast cancer.
There were two more traditional models based on statistics, and two that used machine learning, a form of artificial intelligence.
All forms included the same types of data, such as the woman’s age, weight, smoking history, family history of breast cancer, and use of hormone therapy (HRT).
The models were evaluated for their ability to accurately predict risk in general, and across a variety of groups of women, such as different age groups and ethnic backgrounds.
A technique called “internal and external cross-validation” was used. This involves dividing the dataset into structurally different parts, in this case, by region and time period, to understand how well the model has moved to different settings.
The results showed that one of the statistical models, which was developed using the Competitive Risk Regression, was the best overall. It more accurately predicted which women would develop breast cancer and die from it within 10 years.
The machine learning models were less accurate, especially for different ethnic groups of women. “If further studies confirm the accuracy of this new model, it could be used to identify women at high risk of fatal breast cancer who might benefit from improved screening and preventative treatment,” the team said.
Disclaimer: This story has not been edited by the WBSETCL team and is auto-generated from syndicated feed.