Figure 1 from Assessing the Value of Incorporating a Polygenic Risk Score with Nongenetic Factors for Predicting Breast Cancer Diagnosis in the UK Biobank
Collister JA., Liu X., Littlejohns TJ., Cuzick J., Clifton L., Hunter DJ.
<p>Calibration plots of 10-year observed versus predicted risk by risk decile, for the Tyrer–Cuzick and Gail models, plotting the observed proportion of women with breast cancer by 10 years of follow-up against the predicted 10-year risk from the models. Top row, predicted risks in training dataset from Tyrer–Cuzick and Gail models. Middle row, predicted risks in test dataset from Tyrer–Cuzick and Gail models, calibrated using the training dataset. Bottom row, predicted risks in test dataset from Cox model containing Tyrer–Cuzick or Gail model and PRS<sub>BC</sub>, developed in training dataset.</p>