A team of researchers on the University of Oxford, led by the Nuffield Department of Primary Care Health Sciences, have developed a latest model that reliably predicts a lady’s likelihood of developing after which dying of breast cancer inside a decade.
The study, published today within the Lancet Digital Health, analyzed anonymized data from 11.6 million women aged 20-90 from 2000 to 2020. All of those women had no prior history of breast cancer, or the precancerous condition called ‘ductal carcinoma in situ’, or DCIS.
Breast cancer screening is important but has challenges. While it reduces breast cancer deaths, it sometimes detects tumors that are usually not harmful (‘overdiagnosis’), which results in unnecessary treatments. This not only harms some women, but additionally causes unnecessary costs to the NHS. For each 10,000 UK women aged 50 years invited to breast screening for the following 20 years, 43 breast cancer deaths are prevented by screening, but 129 women can be ‘overdiagnosed’.(https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3693450/).
‘Risk-based screening’ goals to personalize screening based on a person’s risk, to maximise the advantages and minimize the downsides of such screening. Tailoring screening programs on the premise of individual risks was recently highlighted as an avenue for further improvement in screening strategy by Prof Chris Whitty (https://www.gresham.ac.uk/watch-now/medical-screening). Currently, in risk-based breast screening, most models of risk work by estimating a lady’s risk of a breast cancer diagnosis. Nonetheless, not all breast cancers are fatal, and we all know that the danger of being diagnosed doesn’t at all times align well with the danger of dying from breast cancer once diagnosed.
The brand new model developed by the team works to predict a lady’s 10-year combined risk of developing after which dying from breast cancer. Identifying women at the very best risk of deadly cancers could improve screening. These women might be invited to start out screening earlier, be invited for more frequent screenings, or be screened with several types of imaging. Such a personalised approach could further lower breast cancer deaths while avoiding unnecessary screening for lower-risk women. Women at higher risk for developing a deadly cancer is also considered for treatments that try to stop breast cancers developing.
Professor Julia Hippisley-Cox, Professor of General Practice and Epidemiology and senior writer from the Nuffield Department of Primary Care Health Sciences on the University of Oxford, said: ‘That is a very important latest study which potentially offers a latest approach to screening. Risk-based strategies could offer a greater balance of advantages and harms in breast cancer screening, enabling more personalized information for girls to assist improve decision-making. Risk-based approaches may help make more efficient use of health service resources by targeting interventions to those most definitely to profit. We thank the numerous 1000’s of GPs who’ve contributed anonymized data to the QResearch database without which this research wouldn’t have been possible.’
The researchers tested 4 different modeling techniques to predict breast cancer mortality risk. Two were more traditional statistical-based models and two used machine learning, a type of artificial intelligence. All models included the identical varieties of data, like a lady’s age, weight, history of smoking, family history of breast cancer, and use of hormone therapy (HRT).
The models were evaluated for his or her ability to predict risk accurately overall and across a various range of groups of ladies, reminiscent of from different ethnic backgrounds and age groups. A method called ‘internal-external cross-validation’ was used. This involves splitting the dataset into structurally different parts, on this case, by region and time period, to grasp how well the model might transport into different settings.
The outcomes showed that one statistical model, developed using ‘competing risks regression’ performed the most effective overall. It most accurately predicted which women will develop and die from breast cancer inside 10 years. The machine learning models were less accurate, especially for various ethnic groups of ladies.
Dr Ashley Kieran Clift, first writer and Clinical Research Fellow on the Nuffield Department of Primary Care Health Sciences, University of Oxford, said: ‘Funded by Cancer Research UK and making the most of the scale and richness of the QResearch database with its linked data sources on the University of Oxford, we were capable of explore different approaches to develop a tool that may be helpful for brand new, risk-based public health strategies.
If further studies confirm the accuracy of this latest model, it might be used to discover women at high risk of deadly breast cancers who may profit from improved screening and preventative treatments.’
This paper took a latest approach and asked, “can we predict which women are at highest risk of developing a cancer that may kill them?” We could use that information to higher goal screening and even for prevention strategies to those that stand to profit probably the most.
Further evaluation of the competing risks model should include assessment of the models in a special setting, reminiscent of one other dataset from the UK or abroad.’
Professor Stavros Petrou, Co-Writer and Health Economics Lead within the Nuffield Department of Primary Care Health Sciences, University of Oxford
Source:
Journal reference:
Clift, A. K., et al. (2023) Predicting 10-year breast cancer mortality risk in the final female population in England: a model development and validation study. The Lancet Digital Health. doi.org/10.1016/S2589-7500(23)00113-9.