Related Experiment Video
Updated: Aug 13, 2026
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Deep learning-assisted detection of lymph node metastases in bladder cancer
Adam Gorczynski1,2, Balazs Tarján3, Noora Neittaanmäki4,5
1Department of Clinical Pathology, Sahlgrenska University Hospital, Gothenburg, Region Västra Götaland, Sweden. adam.gorczynski@vgregion.se.
Diagnostic Pathology
|August 12, 2026
Summary
An AI model assists pathologists in detecting bladder cancer lymph node metastases, significantly reducing review time. This technology can support routine histopathological assessment, even in institutions with limited case volumes.
Area of Science:
- Oncology
- Pathology
- Artificial Intelligence
Background:
- Bladder cancer necessitates radical cystectomy and lymph node dissection for prognosis and treatment.
- Histopathological evaluation of lymph nodes is critical but labor-intensive and time-consuming.
Purpose of the Study:
- To develop a supervised deep learning model for detecting nodal metastases in bladder cancer.
- To evaluate the AI model's impact on pathologists' assessment efficiency.
Main Methods:
- A deep learning model was trained on 100 whole slide images with 3437 pixelwise annotations.
- The AI model's performance was validated on 50 regions by AI and two specialist pathologists.
- Pathologists reviewed slides with and without AI assistance to measure review times.
Main Results:
- The AI model achieved 100% sensitivity and 60% specificity in detecting metastases.
- AI assistance reduced median review time for pathologists from 9s to 4.2s and 12s to 4s.
- The AI model effectively assisted in identifying lymph node metastases.
Conclusions:
- A locally trained AI model can effectively aid pathologists in detecting bladder cancer lymph node metastases.
- AI significantly reduces histopathological review time, enhancing efficiency.
- This approach supports routine pathological assessment, particularly in institutions with limited case volumes.