Validation of Risk Assessment Models for Predicting the Incidence of Breast Cancer in Korean Women.
10.4048/jbc.2014.17.3.226
- Author:
Jun Won MIN
1
;
Myung Chul CHANG
;
Hae Kyung LEE
;
Min Hee HUR
;
Dong Young NOH
;
Jung Han YOON
;
Yongsik JUNG
;
Jung Hyun YANG
Author Information
1. Department of Surgery, Dankook University College of Medicine, Cheonan, Korea. changmc@dankook.ac.kr
- Publication Type:Validation Studies ; Original Article
- Keywords:
Breast neoplasms;
Risk assessment;
Validation studies
- MeSH:
Area Under Curve;
Asian Americans;
Asian Continental Ancestry Group;
Breast Neoplasms*;
Female;
Hope;
Humans;
Incidence*;
Korea;
Mass Screening;
Mortality;
Risk Assessment*;
ROC Curve
- From:Journal of Breast Cancer
2014;17(3):226-235
- CountryRepublic of Korea
- Language:English
-
Abstract:
PURPOSE: The Gail model is one of the most widely used tools to assess the risk of breast cancer. However, it is known to overestimate breast cancer risk for Asian women. Here, we validate the Gail model and the Korean model using Korean data, and subsequently update and revalidate the Korean model using recent data. METHODS: We validated the modified Gail model (model 2), Asian American Gail model, and a previous Korean model using screening patient data collected between January 1999 and July 2004. The occurrence of breast cancer was confirmed by matching the resident registration number with data from the Korean Breast Cancer Registration Program. The expected-to-observed (E/O) ratio was used to validate the reliability of the program, and receiver operating characteristics curve analysis was used to evaluate the program's discriminatory power. There has been a rapid increase in the incidence of breast cancer in Korea, and we updated and revalidated the Korean model using incidence and mortality rate data from recent years. RESULTS: Among 40,229 patients who were included in the validation, 161 patients were confirmed to have developed breast cancer within 5 years of screening. The E/O ratios and 95% confidence intervals (CI) were 2.46 (2.10-2.87) for the modified Gail model and 1.29 (1.11-1.51) for the Asian American Gail model. The E/O ratio and 95% CI for the Korean model was 0.50 (0.43-0.59). For the updated Korean model, the E/O ratio and 95% CI were 0.85 (0.73-1.00). In the discriminatory power, the area under curve and 95% CI of the modified Gail model, Asian American Gail model, Korean model and updated Korean model were 0.547 (0.500-0.594), 0.543 (0.495-0.590), 0.509 (0.463-0.556), and 0.558 (0.511-0.605), respectively. CONCLUSION: The updated Korean model shows a better performance than the other three models. It is hoped that this study can provide the basis for a clinical risk assessment program and a future prospective study of breast cancer prevention.
