Multimodal Imaging and Digital Pathology for Prostate Cancer Prediction

Официальное название

A Multicenter Study of a Deep Learning Model Based on Spatial Registration of Multimodal Imaging and Digital Pathology for Predicting Clinically Significant Prostate Cancer

Обзор

This is a multicenter observational study. A deep learning model integrated with multimodal imaging and digital pathology spatial registration is built based on preoperative multiparametric magnetic resonance imaging, transrectal ultrasound and postoperative digital pathological whole slide images. The study is designed to achieve accurate prediction of clinically significant prostate cancer and non-invasive risk stratification. Unnecessary prostate biopsy and overdiagnosis can be reduced to support the optimization of clinical diagnosis and treatment strategies.

Подробное описание

This prospective and retrospective multicenter observational study enrolls patients with suspected prostate cancer who receive standardized preoperative multiparametric magnetic resonance imaging, transrectal ultrasound examination, followed by prostate biopsy or radical prostatectomy. Complete clinical data including age, BMI, prostate specific antigen indicators, PI-RADS v2.1 scores, Gleason score and ISUP grading are collected from all eligible participants.

Biomechanically constrained non-rigid spatial registration technique is applied to achieve precise alignment between preoperative multimodal images and postoperative digital pathological whole slide images using high-quality multicenter datasets. A transformer-based multimodal deep learning fusion model is developed to analyze correlations between macroscopic imaging features and microscopic pathological heterogeneity, thereby establishing an interpretable artificial intelligence framework for clinically significant prostate cancer prediction.

Comprehensive model validation is conducted via internal cross-validation, external multicenter independent verification and international public datasets. Decision curve analysis and clinical impact curve are applied to assess clinical applicability. The model serves as an intelligent auxiliary tool to refine biopsy strategies, avoid redundant puncture and excessive treatment, and facilitate early precise diagnosis and risk stratification of prostate cancer.

Вмешательства

• Другое No Intervention: Observational Cohort
  This is an observational study. No new treatment, drug, device, or procedure is being administered to participants. Only standard-of-care clinical data, imaging, and pathology records are collected and analyzed.

Первичные конечные точки

• Area Under the Receiver Operating Characteristic Curve (AUC) for predicting clinically significant prostate cancer (csPCa) [Срок оценки: Baseline (at the time of imaging/pathology data collection)]

Критерии участия

Критерии включения

• Subjects who are scheduled to undergo or have undergone prostate biopsy or radical prostatectomy.
• Subjects who have completed standard-of-care preoperative multiparametric MRI (mpMRI) and transrectal ultrasound (TRUS) examinations.
• Subjects with complete pathological diagnosis results available.
• Age between 40 and 90 years.
• Able and willing to provide written informed consent (for prospective cohort participants only).

Критерии исключения

• Prior history of pelvic radiation therapy or radical prostatectomy.
• Incomplete or poor-quality mpMRI or TRUS images (e.g., motion artifacts, insufficient sequences).
• Concurrent other primary malignant tumors.
• Severe systemic diseases that may affect the evaluation of the prostate.
• Subjects with incomplete clinical or pathological data.
• Contraindications to MRI examination (e.g., incompatible metallic implants, severe claustrophobia).

Критерии приведены из реестра в оригинале (на английском). Окончательную оценку соответствия проводит исследовательский центр.

Здоровые добровольцы: Нет

Дизайн исследования

Модель наблюдения

Когортное

Центры проведения

Китай · 1 центр

• Liuzhou People's Hospital Affiliated to Guangxi Medical University — Liuzhou

Публикации

• Shao L, Liang C, Yan Y, Zhu H, Jiang X, Bao M, Zang P, Huang X, Zhou H, Nie P, Wang L, Li J, Zhang S, Ren S. An MRI-pathology foundation model for noninvasive diagnosis and grading of prostate cancer. Nat Cancer. 2025 Oct;6(10):1621-1637. doi: 10.1038/s43018-025-01041-x. Epub 2025 Sep 2. PMID 40897909
• Rusu M, Jahanandish H, Vesal S, Li CX, Bhattacharya I, Venkataraman R, Zhou SR, Kornberg Z, Sommer ER, Khandwala YS, Hockman L, Zhou Z, Choi MH, Ghanouni P, Fan RE, Sonn GA. ProCUSNet: Prostate Cancer Detection on B-mode Transrectal Ultrasound Using Artificial Intelligence for Targeting During Prostate Biopsies. Eur Urol Oncol. 2025 Apr;8(2):477-485. doi: 10.1016/j.euo.2024.12.012. Epub 2025 Jan 2 PMID 39880746
• Saha A, Hosseinzadeh M, Huisman H. End-to-end prostate cancer detection in bpMRI via 3D CNNs: Effects of attention mechanisms, clinical priori and decoupled false positive reduction. Med Image Anal. 2021 Oct;73:102155. doi: 10.1016/j.media.2021.102155. Epub 2021 Jun 29. PMID 34245943
• Lee YJ, Moon HW, Choi MH, Eun Jung S, Park YH, Lee JY, Kim DH, Eun Rha S, Kim SH, Lee KW, Choi YJ, Lee YS, Lee W, Lee S, Grimm R, von Busch H, Han D, Lou B, Kamen A. MRI-based Deep Learning Algorithm for Assisting Clinically Significant Prostate Cancer Detection: A Bicenter Prospective Study. Radiology. 2025 Mar;314(3):e232788. doi: 10.1148/radiol.232788. PMID 40067105
• Twilt JJ, Saha A, Bosma JS, Padhani AR, Bonekamp D, Giannarini G, van den Bergh R, Kasivisvanathan V, Obuchowski N, Yakar D, Elschot M, Veltman J, Futterer J, Huisman H, de Rooij M; PI-CAI Consortium. AI-Assisted vs Unassisted Identification of Prostate Cancer in Magnetic Resonance Images. JAMA Netw Open. 2025 Jun 2;8(6):e2515672. doi: 10.1001/jamanetworkopen.2025.15672. PMID 40512493
• Goel S, Shoag JE, Gross MD, Al Hussein Al Awamlh B, Robinson B, Khani F, Baltich Nelson B, Margolis DJ, Hu JC. Concordance Between Biopsy and Radical Prostatectomy Pathology in the Era of Targeted Biopsy: A Systematic Review and Meta-analysis. Eur Urol Oncol. 2020 Feb;3(1):10-20. doi: 10.1016/j.euo.2019.08.001. Epub 2019 Sep 4. PMID 31492650
• Pham THN, Schulze-Hagen MF, Rahnama'i MS. Targeted multiparametric magnetic resonance imaging/transrectal ultrasound-guided (mpMRI/TRUS) fusion prostate biopsy versus systematic random prostate biopsy: A comparative real-life study. Cancer Rep (Hoboken). 2024 Feb;7(2):e1962. doi: 10.1002/cnr2.1962. Epub 2024 Jan 12. PMID 38217298
• Drost FH, Osses DF, Nieboer D, Steyerberg EW, Bangma CH, Roobol MJ, Schoots IG. Prostate MRI, with or without MRI-targeted biopsy, and systematic biopsy for detecting prostate cancer. Cochrane Database Syst Rev. 2019 Apr 25;4(4):CD012663. doi: 10.1002/14651858.CD012663.pub2. PMID 31022301

Идентификаторы

NCT: NCT07614256 · GXMU-PCa-AI-KY20260016

Первоисточники (государственные реестры)

• ClinicalTrials.gov (NIH, США)

Открыть это исследование на ClinicalTrials.gov ↗