Original Article Evaluation of risk assessment tools for breast cancer screening in Chinese population

Junyu Zhao, Xiuling Song, Ling Leng, Huanjun Wang, Lin Liao, Jianjun Dong
2017 Int J Clin Exp Med   unpublished
There are many models for screening breast cancer, but not all of them were suitable for all women worldwide including the most famous Gail model. There is no independent risk prediction model in China and no verification of the Gail model in Chinese female population. To evaluate the Gail model for screening breast cancer in Chinese population and compare it with the health risk appraisal (HRA) model. Methods: A total of 3030 Chinese females between 45-70 years were randomly selected for five
more » ... selected for five years follow-up from 2008 to 2014. Gail model and HRA model were compared based on the screening database for breast cancer screening in Chinese female population. Other potential risk factors were also analyzed. Results: The sensitivity of the Gail model was 5.0% and specificity was 97.1%. For the HRA model, the sensitivity was 70.0% and the specificity was 60.6%. The AUC of the Gail model was 0.542 (95% CI, 0.426, 0.658) and the Youden index was 2.1%. The AUC and Youden index of the HRA model was more reliable than the Gail model (AUC, 0.734 (95% CI, 0.643, 0.825); Youden index, 30.6%). In addition , non-menopausal women had a higher risk of developing breast cancer than menopausal women (P=0.000). The age of menarche was statistically associated with the risk of developing breast cancer (P=0.000). Conclusions: The HRA model is more suitable for Chinese females than the classic risk assessment tool. However, the HRA model needs to be updated to increase its sensitivity and specificity.