Skip to main content

Product Highlights

Flywheel Offers Powerful Support for Digital Pathology Research and AI Development


Digital pathology research often involves tedious and time-consuming digital manipulation, analysis and annotation. With features from Flywheel, these steps can be simplified and automated to seamlessly manage multi-modal and multi-center digital pathology research projects.

Flywheel’s powerful digital pathology platform

Flywheel’s tools allow researchers to:

  • Convert virtual slides, or whole slide images (WSIs), to DICOM
  • Store batches of imaged tissues (tissue microarrays, or TMAs)
  • View and annotate digital pathology images using powerful new features

Flywheel offers ready-to-use plugin applications called Gears that automate data pre-processing workflows and analytic pipelines to standardize digital pathology research data and prepare it for analysis. Our container-based approach facilitates reproducible computation, handling large processing workloads with elastic compute scaling, and tracking inputs, versions and outputs. Some of our new Gears that automate routine tasks in digital pathology research include “WSI-to-DICOM” and “Process-TMA.”

Viewing Digital Pathology in Flywheel

Flywheel enables whole slide image viewing in addition to curation, search, processing, and sharing.

The WSI-to-DICOM Gear helps users to convert slide images to DICOM for inspection and annotation in our native image viewer. The Gear uses OpenSlide, a C library familiar to many researchers that provides a simple interface to read WSIs. OpenSlide supports a variety of the file formats used by different vendors, ensuring maximum functionality and flexibility to Flywheel users when converting to files such as DICOM.

A key feature of the new “Process-TMA” Gear is its ability to ingest TMA data from digital pathology systems and add it as metadata to images, connecting them to unique subjects and specific annotations. It can split data from multiple patients in tissue microarrays into useful subcomponents, maximizing the data’s research potential.

User-friendly features accelerate digital pathology research, viewing and labeling

Flywheel features also allow users to zoom in, measure, annotate, and draw regions of interest (ROIs) on digital pathology images, easily and intuitively. These functions are integrated seamlessly into the Flywheel Viewer, our custom version of the popular Open Health Imaging Foundation (OHIF) viewer that allows users to choose data and metadata fields, label images, review selected outputs, and export for manual external analysis.

Develop AI in digital pathology research

Now, Flywheel users can more easily develop tissue-based machine learning using Flywheel.

Flywheel’s research data management solution allows data scientists to automate pre-processing, standardize quality control with a quality assurance rule framework, customize computational pipelines, and manage training and learning data sets for AI development with full provenance and regulatory compliance.

Flywheel can also be used to manage digital pathology multi-reader studies, where readers at separate institutions have controlled access to a Flywheel project to follow study worklists, make annotations and ROIs, and fill quantitative summary forms. The annotations, labels, and ROI data can be searched to develop test and training data sets for AI.

Finally, testing digital pathology-based algorithms can be streamlined with Flywheel’s ability to elastically scale large processing workloads, deploy models in clinical settings, and capture algorithm versions and outputs in the platform for further iteration. Flywheel offers an open REST API framework to offer researchers the flexibility to work with digital pathology research data in their machine learning platform of choice.

With these tools and features, Flywheel offers users even more ways to accelerate AI in  digital pathology and slide imaging, tissue-based research, and drug development.

Request a demo to learn how Flywheel can accelerate your digital pathology research, AI development, and clinical trials.