Related Experiment Video
Updated: Sep 10, 2025
High-Speed Ultraviolet Photoacoustic Microscopy for Histological Imaging with Virtual-Staining assisted by Deep Learning
Published on: April 28, 2022
Enhancing frozen histological section images using permanent-section-guided deep learning with nuclei attention
Elad Yoshai1, Gil Goldinger2, Tatiana Kogan3
1School of Biomedical Engineering, Tel Aviv University, Tel Aviv, Israel.
Abstract:
In histological pathology, frozen sections are often used for rapid diagnosis during surgeries, as they can be produced within minutes. However, they suffer from artifacts and often lack crucial diagnostic details, particularly within the cell nuclei region. Permanent sections, on the other hand, contain more diagnostic detail but require a time-intensive preparation process. Here, we present a generative deep learning approach to enhance frozen section images by leveraging guidance from permanent sections. Our method places a strong emphasis on the nuclei region, which contains critical information in both frozen and permanent sections. Importantly, our approach avoids generating artificial data in blank regions, ensuring that the network only enhances existing features without introducing potentially unreliable information. We achieve this through a segmented attention network, incorporating nuclei-segmented images during training and adding an additional loss function to refine the nuclei details in the generated permanent images. We validated our method across various tissues, including kidney, breast, and colon. This approach significantly improves histological efficiency and diagnostic accuracy, enhancing frozen section images within seconds, and seamlessly integrating into existing laboratory workflows.
Insights
Generative deep learning enhances frozen section pathology images for faster surgical diagnosis. The AI focuses on cell nuclei, improving diagnostic detail without creating artificial data.
Area of Science:
- Histological pathology
- Medical imaging
- Artificial intelligence
Background:
- Frozen sections offer rapid intraoperative diagnosis but often lack nuclear detail and contain artifacts.
- Permanent sections provide superior diagnostic detail but require lengthy preparation.
- Accurate histological diagnosis is critical for surgical decision-making.
Purpose of the Study:
- To develop a generative deep learning method for enhancing frozen section images using permanent sections as guidance.
- To improve the diagnostic quality of frozen sections, particularly in the critical cell nuclei region.
- To accelerate the histological diagnosis workflow during surgery.
Main Methods:
- A segmented attention network was employed, leveraging nuclei-segmented images during training.
- A novel loss function was incorporated to refine nuclear details in the enhanced images.
- The approach was validated on diverse tissue types, including kidney, breast, and colon samples.
Main Results:
- The deep learning method successfully enhanced frozen section images, improving clarity and diagnostic detail.
- Enhancement focused on critical regions like cell nuclei, preserving existing features without generating artificial data.
- The process significantly improved histological efficiency, providing enhanced images within seconds.
Conclusions:
- This generative deep learning approach offers a significant advancement in histological image enhancement.
- The method improves diagnostic accuracy by enhancing crucial details in frozen sections.
- The technique seamlessly integrates into existing laboratory workflows, accelerating intraoperative diagnosis.
