NA-MIC Project Weeks

Back to Projects List

Exploring Ultrasound Multi-frame Image Storage Structured Report support for Slicer/OHIF

Key Investigators

• Deepa Krishnaswamy (BWH, USA)
• Andrey Fedorov (Brigham and Women's Hospital, USA)
• Chris P. Bridge (Massachusetts General Hospital, USA)
• Matt McCormick (Fideus Labs, USA)
• Dave Dinh (BWH, USA)
• Tina Kapur (BWH, USA)

Github Repository

• SlicerUltrasound
• create_LUS_SR
• lus-dashboard

Funding Source(s)

• R01EB035679
• R21EB034075

Project Description

Our team has built a 3D Slicer extension for ultrasound anonymization, annotation, and adjudication (https://github.com/SlicerUltrasound/SlicerUltrasound). We are currently trying to port the functionality of these tools to OHIF in order to make the processing more streamlined for clinicians and avoid the downloading of additional tools.

However, we currently store all of our annotations as JSON. This can be problematic as JSON can be prone to misinterpretation; we had to write specialized software to load these annotations into Slicer, and we are missing much of the metadata to associate the file with the referenced image.

In this project, I want to see if using DICOM Structured Reports (SR) will help our use case. Having our files as SRs would help us to utilize the cloud and retain all important metadata. But it might also make processing more complicated. Therefore, the goal of this week is to see how much effort is needed to add functionality to Slicer/OHIF for Ultrasound Multi-frame Image Storage.

Objective

1. Create support for Ultrasound Multi-frame Image Storage in 3D Slicer and OHIF

Approach and Plan

1. Create a DICOM Structured Report for our annotations using Ultrasound Multi-frame Image Storage (both pleura line and B-line points/lines, and manual annotations like pleural percentage)
2. Create these SRs using both highdicom and itk-wasm (as suggested by Matt, better for OHIF)
3. Add functionality to 3D Slicer core for loading, parsing, and viewing these SRs.
4. Make sure that the SR can be loaded into the official OHIF viewer.
5. Add functionality to the OHIF usAnnotation extension to save the annotations as an SR
6. Deploy OHIF with usAnnotation to connect to the cloud (Google Healthcare DICOM datastore)

Progress and Next Steps

1. Converted our JSON annotation file to a DICOM SR
2. Confirmed the DICOM SR can load in OHIF
3. Added functionality to 3D Slicer to load, parse, and view these DICOM SRs
4. Next week - will make a PR to Slicer
5. WIP - deploying OHIF connected to a Google Cloud Platform DICOM datastore, with the usAnnotation plugin support (Thank you, Martin and Andrey, for your help!)

Illustrations

Current workflow:

Successfully loaded the converted annotations in OHIF!

Successfully loaded the converted annotations in Slicer!

Interactive dashboard to keep track of one type of annotations:

Background and References

Google Colab notebook for creating the Structured Report

GitHub code for the dashboard

No response

Funding

This work is supported by NIH R01EB035679 and R21EB034075.