This study aims to investigate a novel Centre-Adaptive Multi-Modal Fusion Framework (Cad-MMFF) that integrates clinical, ultrasound, and MRI data to improve the detection of clinically significant prostate cancer (csPCa), reduce unnecessary biopsies, and optimize MRI utilization.
Ethnically Chinese men aged ≥ 50
Suspicion of PCa [elevated PSA (>4-ng/ml) or abnormal DRE/PHI] and indicated for prostate biopsy
Obtained clinical consent (for prospectively enrolled patients only)
Prior biopsy/treatments for PCa
History of genitourinary cancer
Poor image quality
Incomplete imaging scans/clinical data/pathological results
The entire cohort of this study will be targeted at patients who are suspected of PCa and indicated for ultrasound-guided and/or MRI-guided targeted prostate biopsy. Brightness-modulated ultrasound data (longitudinal and transverse views with full prostate coverage), pathological outcomes (Gleason score, ISUP GG), pre-biopsy multi-parametric MRI scans, PI-RADS scores, patients demographics and clinical data, will be acquired from January 2023 to August 2026 retrospectively, and from September 2026 to March 2028 prospectively, from the existing database of the Picture Archiving and Communication System (PACS). Informed consent from subjects will be sought only for prospectively enrolled patients, while there will be no interventions involved given its nature of retrospective analysis
Cohort
Prospective
Not listed
Area under Receiver-Operating-Curve of using the Centre-adaptive multi-modal fusion framework (Cad-MMFF)
The number of unnecessary biopsy rate decreased of using the Centre-adaptive multi-modal fusion framework (Cad-MMFF)
Proportion of MRI safely saved of using the Centre-adaptive multi-modal fusion framework (Cad-MMFF)
Net-Benefit of using the Centre-adaptive multi-modal fusion framework (Cad-MMFF)
By Decision Curve analysis
Click on the lead sponsor to view all of their trials.
Chinese University of Hong Kong
Lead sponsor
The Hong Kong Polytechnic University
Collaborator
This trial is not recruiting at the moment. You can still explore other options:
