Glynn RJ, et al. - Since the Gail model and the model developed by Tyrer and Cuzick are two questionnaire-based approaches with demonstrated ability to predict breast cancer development in the general population, researchers compared the calibration, discrimination and net reclassification of these models using questionnaire data sent every 2 years to 76,922 Nurses' Health Study participants between 1980 and 2006, with 4,384 invasive breast cancer incidents identified by 2008. The Tyrer-Cuzick model had both overall and age-specific comparisons of slightly higher C-statistics than the Gail model. Net reclassification assessment did not favor either model. The Rosner-Colditz model had better calibration and discrimination in the one-third sample than the other two models. Investigators suggested that both the Gail and Tyrer-Cuzick models had some ability to discriminate cases and non-cases of breast cancer, but their fit model has limitations.