Table 4.

Change in age-adjusted AUC for breast cancer by different risk models among postmenopausal women not currently using menopausal hormone replacement therapy.

Gail 5-year riskaTyrer–Cuzick 5-year riskbCAD2Y 2-year riskc
ModelN cases/controlsAUC (SE)P valuedP valueeN cases/controlsAUC (SE)P valuedP valueeN cases/controlsAUC (SE)P valued
Risk model only403/84255.1 (1.8)403/84260.4 (1.7)250/82968.8 (1.9)
Risk model + DHEA285/60657.1 (2.1)0.091285/60660.7 (2.1)0.092174/59567.4 (2.2)0.355
Risk model + DHEAS380/78857.2 (1.8)0.008380/78861.7 (1.8)0.007236/77667.8 (2.0)0.023
Risk model + mammographic density397/83157.2 (1.8)0.014397/83161.5 (1.7)0.006
Risk model + DHEA and mammographic density281/59857.6 (2.1)0.0310.086281/59861.3 (2.1)0.0180.083
Risk model + DHEAS and mammographic density374/77758.4 (1.8)0.0020.006374/77761.9 (1.8)0.0010.007
  • aGail model included risk factors of age, age at menarche, age at first live birth, number of previous breast biopsies, atypical hyperplasia, and first-degree family history of breast cancer.

  • bTyrer–Cuzick model included risk factors of age, age at menarche, age at first child, menopause, height, weight, MHT, hyperplasia, atypical hyperplasia, lobular cancer in situ, and first-/second-degree family history of breast cancer.

  • cCAD2Y risk model included age, menopausal status, BMI, current use of MHT, breast cancer in family, percent mammographic breast density, mammographic density difference (absolute difference between right and left breasts), microcalcification (absolute difference between right and left breasts), and interaction between mammographic density and masses.

  • dP value indicates difference in prediction by model with only risk score on the basis of log-likelihood ratio test.

  • eP value indicates difference in prediction by model with risk score and mammographic density compared with prediction by risk score together with mammographic density and DHEA or DHEAS, on the basis of log-likelihood ratio test.