Interoperable slide microscopy viewer and annotation tool for imaging data science and computational pathology.

Area of Science:

• Digital pathology
• Biomedical imaging
• Computational pathology

Background:

• Slide microscopy imaging data exchange is hindered by poor standardization and interoperability.
• This lack of standardization impacts scientific reproducibility and clinical integration of innovations.
• Existing systems lack seamless integration for complex slide imaging data.

Purpose of the Study:

• Introduce Slim, an open-source, web-based slide microscopy viewer.
• Enable interoperability using the Digital Imaging and Communications in Medicine (DICOM) standard.
• Facilitate visualization and analysis of diverse microscopy data.

Main Methods:

• Developed Slim, a web-based viewer implementing the DICOM standard.
• Integrated Slim with the NCI Imaging Data Commons.
• Utilized DICOMweb services for data access and visualization.
• Demonstrated interactive visualization of brightfield and immunofluorescence microscopy images.

Main Results:

• Slim provides interoperability with existing medical imaging systems via DICOM.
• Successfully visualized diverse microscopy data from The Cancer Genome Atlas and Human Tissue Atlas Network.
• Enabled standardized image annotation collection for machine learning.
• Facilitated visual interpretation of machine learning model inference results.

Conclusions:

• Slim enhances the exchange and analysis of slide microscopy imaging data.
• The DICOM standard implementation in Slim promotes interoperability and reproducibility.
• Slim supports machine learning model development and validation in pathology.