End-to-end prostate cancer detection in bpMRI via 3D CNNs: Effects of attention mechanisms, clinical priori and decoupled false positive reduction.
Anindo Saha, Matin Hosseinzadeh, Henkjan Huisman
https://pubmed.ncbi.nlm.nih.gov/34245943/Actively Recruiting
Led by Guangxi Medical University · Updated on 2026-05-29
3000
Participants Needed
1
Research Sites
26 weeks
Total Duration
Researchers are evaluating a deep learning model that combines multimodal imaging and digital pathology to improve the prediction of clinically significant prostate cancer. This observational study includes patients suspected of having prostate cancer who undergo standard preoperative multiparametric MRI, transrectal ultrasound, and subsequent biopsy or prostate removal. The goal is to reduce unnecessary biopsies and overdiagnosis, helping to optimize diagnosis and treatment strategies. Participants receive standard-of-care imaging and pathology procedures, including multiparametric MRI and transrectal ultrasound, followed by prostate biopsy or radical prostatectomy as needed. High-quality imaging and pathology data are aligned using advanced spatial registration techniques. A transformer-based deep learning model analyzes these combined data to predict cancer risk and support clinical decisions. The study includes both retrospective and prospective patient data from multiple centers. During the study, researchers collect clinical information such as age, body mass index, prostate-specific antigen levels, and pathology scores. The model undergoes validation through internal and external datasets to assess its accuracy using the area under the receiver operating characteristic curve AUC. Participants data are monitored without additional treatments or interventions. The study aims to improve early diagnosis and risk assessment for prostate cancer while minimizing unnecessary procedures.
CONDITIONS
Multimodal Imaging and Digital Pathology for Prostate Cancer Prediction
C
Caigou Shi, MD
Explore thousands of other clinical trials that might be a better match.
Sign up to get personalized trial recommendations delivered to your inbox.
Already have an account? Log in here
Guangxi Medical University
Lead Sponsor
You may qualify if you...
You will not qualify if you...
Duration - 2 to 4 weeks
Participants are screened for eligibility to participate in the trial.
1 visit (in-person)
Duration - Up to several weeks depending on clinical scheduling
Participants undergo standard preoperative multiparametric MRI and transrectal ultrasound examinations, followed by prostate biopsy or radical prostatectomy as part of their routine clinical care.
1 to 3 visits depending on procedures
Duration - Until study completion
Participants who undergo routine care are observed as clinical data, imaging, and pathology records are collected and analyzed to develop and validate the predictive model for prostate cancer.
No additional study visits; observation based on routine care
Study Type
OBSERVATIONAL
Masking
N/A
Allocation
N/A
Model
N/A
Primary Purpose
N/A
Number of Arms
0
Have more questions? Get in touch with our team for quick support
Anindo Saha, Matin Hosseinzadeh, Henkjan Huisman
https://pubmed.ncbi.nlm.nih.gov/34245943/Frank-Jan H Drost, Daniël F Osses, Daan Nieboer...
https://pubmed.ncbi.nlm.nih.gov/31022301/Hongmei Zeng, Wanqing Chen, Rongshou Zheng...
https://pubmed.ncbi.nlm.nih.gov/29653628/Freddie Bray, Mathieu Laversanne, Hyuna Sung...
https://pubmed.ncbi.nlm.nih.gov/38572751/Jasper J Twilt, Anindo Saha, Joeran S Bosma...
https://pubmed.ncbi.nlm.nih.gov/40512493/Sébastien Molière, Dimitri Hamzaoui, Guillaume Ploussard...
https://pubmed.ncbi.nlm.nih.gov/39547898/Jonathan I Epstein, Mahul B Amin, Victor E Reuter...
https://pubmed.ncbi.nlm.nih.gov/28177964/Young Joon Lee, Hyong Woo Moon, Moon Hyung Choi...
https://pubmed.ncbi.nlm.nih.gov/40067105/Mirabela Rusu, Hassan Jahanandish, Sulaiman Vesal...
https://pubmed.ncbi.nlm.nih.gov/39880746/Lizhi Shao, Chao Liang, Ye Yan...
https://pubmed.ncbi.nlm.nih.gov/40897909/Guangxi Medical University
Lead Sponsor