Related Experiment Video
Updated: Aug 12, 2026
An R-Based Landscape Validation of a Competing Risk Model
Published on: September 16, 2022
Validation and Clinical Utility of a Prediction Model for the Risk of Upstaging to Invasive Breast Cancer After a
Claudia J C Meurs1,2, Sara van Bekkum3, Joost van Rosmalen4,5
1Department of HTSR, University of Twente, Enschede, The Netherlands. c.j.c.meurs@utwente.nl.
Background:
This study aimed to validate the DCIS-upstage model, a previously developed model to predict the risk of upstaging to invasive breast cancer in patients with biopsy-proven ductal carcinoma in situ (DCIS) in a more recent cohort and to assess the model's clinical utility.
Methods:
The model was validated in a registry cohort (n = 2269) and in an institution cohort (n = 302). A calibration plot was made, followed by a decision curve analysis (DCA). The model's area under the curve (AUC) was compared with the AUC of another published model and with the AUCs of new models using the risk factors of the DCIS-upstage model and additional risk factors.
Results:
The DCIS-upstage model had an AUC of 0.67 at development; in the validation, the AUC was 0.65 in the registry cohort and 0.73 in the institution cohort. The DCA showed that the model has clinical utility. The other published model had an AUC of 0.66 in the institution cohort. Adding risk factors to the DCIS-upstage model slightly increased the AUC.
Conclusions:
The DCIS-upstage prediction model is valid in other cohorts. The model has clinical utility and may be used to select patients with biopsy-proven DCIS for sentinel lymph node biopsy.
