University of Arizona researchers combined label-free multiphoton microscopy with neural networks to accurately classify pancreatic neuroendocrine neoplasms in tissue samples.
Key Details
- 1Multiphoton microscopy (MPM) was used to image pancreatic neuroendocrine neoplasm (PNEN) samples without labeling.
- 2Researchers trained both traditional machine learning and four convolutional neural networks (CNNs) on these images.
- 3CNNs achieved classification accuracies ranging from 90.8% to 96.4%, outperforming the ML algorithm’s 80.6%.
- 4Analysis showed key features included collagen content and image texture metrics.
- 5The approach is faster than traditional pathology and was validated across samples from multiple biorepositories.
- 6Publication: Biophotonics Discovery, October 2, 2025, DOI: 10.1117/1.BIOS.2.4.045001.
Why It Matters
This study demonstrates the potential of advanced imaging and AI for real-time, highly accurate tumor detection, which could transform intraoperative pathology and improve surgical outcomes in oncology. Faster, label-free diagnosis may reduce delays and error in tumor identification during surgery.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.