Prostate Cancer Risk Calculator Apps in a Taiwanese Population Cohort: Validation Study

Area of Science:

• Urology
• Digital Health
• Oncology

Background:

• Mobile health (mHealth) apps are increasingly used for patient care and clinical decision-making.
• Prostate cancer (PCa) risk calculator apps, such as Rotterdam and Coral, aid in assessing PCa and high-grade PCa risks.
• External validation is crucial as PCa epidemiology varies across populations.

Purpose of the Study:

• To validate the Rotterdam and Coral PCa risk calculator mobile apps in a Taiwanese cohort.
• To assess the accuracy of these apps using both prebiopsy and postprostatectomy pathology outcomes.
• To evaluate the apps' performance in predicting high-grade PCa (Gleason score ≥7).

Main Methods:

• Retrospective analysis of 1134 male patients undergoing transrectal ultrasound prostate biopsies (2012-2018).
• Calculation of PCa and high-grade PCa probabilities using Rotterdam and Coral apps.
• Comparison of app predictions with biopsy and prostatectomy results, including calibration (Hosmer-Lemeshow test) and discrimination (AUC).

Main Results:

• The Rotterdam app showed good calibration for high-grade PCa detection in the biopsy cohort.
• Rotterdam app demonstrated significantly better discrimination for both PCa (AUC: 0.779) and high-grade PCa (AUC: 0.862) compared to Coral.
• In the smaller prostatectomy cohort, both apps showed similar predictive capacity for high-grade PCa.

Conclusions:

• The Rotterdam and Coral apps are applicable to the Taiwanese population with acceptable accuracy.
• The Rotterdam app is superior to the Coral app for predicting PCa and high-grade PCa in this cohort.
• The Rotterdam app is a potential tool for improving PCa and high-grade PCa detection among Taiwanese men.