CTSC:ARRA.051110

Back to CTSC:ARRA supplement

Agenda

• mi2b2 installation update
• IRB update
• Review long term goals for mi2b2 clinical research deliverables

Harvard Catalyst Medical Informatics group Meeting Minutes May 11, 2010

In attendance:

• Shawn Murphy
• Yong Gao
• Randy Gollub
• Chris Herrick
• Valerie Humblet
• Steve Pieper
• Charles McGow
• Kathy Andriole (phone)
• Wendy Plesniack (phone)
• Dan Marcus (phone)
• Paul Lamonica
• Karl Helmer (phone)
• Alex Zeitsev
• Mark Anderson

Meeting Minutes

• Paul and Chris started the mi2b2 STAR-D installation process but ran into a few problems including a hard drive crash. Before the crash, they did get it to log into the real Children's PACS. In this meeting they discovered a misconception about how the system is configured. That was part of the problem that caused the crash. Paul will write up the installation instructions on this wiki page. The http site for the XP Windows only software is http://i2b2.slicer.org/downloads/ . The username and password to download the zip file were agreed upon. If you need this information, contact any of the meeting attendees to ask them to give it to you. It will not be put on the wiki pages.
• Server hardware site update:
  • (BIDMC) Jesse is working to schedule a meeting between himself, Chris H and BIDMC Radiology IT.
  • (BWH) Kathy computer's is still in the box but she can be ready to install as soon as CHB site is working.
  • (MGH) Darren ?

• Shawn continues to work with each site to get all the right people added to correct site i2b2 IRB protocols. Kathy and Darren are now added to the Partners wide one. Chris, Bill and Paul are in the process of being added to CHB, Jesse and Chris to BIDMC. Chris finished his CITI training. Bill and Paul still are working on their CITI training.

• Shawn reviewed the value of having some clinical research projects (such as that worked on last week on traumatic brain injury) that would use the mi2b2 infrastructure that would be developed in parallel with the mi2b2 development. A first step after the scientific questions are outlined might be to query and retrieve just a few scans at each site to test feasibility (e.g. access to images, naming conventions of scans, scan quality, etc.).
• More discussion followed regarding details of potential future projects. People suggested the kinds of bioinformatic data to request (Demographics, co-morbid diseases, concurrent medications) and for the TBI study that all kinds of scans: MRI, DTI, CT might be useful. Also, image study information: sequence of studies, time between traumatic event and image acquisition. Karl notes that there are lots of new methods in development for image analysis of Diffusion images for TBI. 3D gradient recalled sequence with T2* weighting MRI (need to save raw dicom if wanted to own analysis) is used to measure susceptibility weighting (SWI). Standard scanners have SWI analysis embedded image acquisition software. These specialized scans are already in the hands of research scientists.
• Steve Pieper suggested using the data to address questions such as:
  • Which imaging studies (down to the level of acquisition parameters e.g slice thickness) contributed more to successful diagnosis and treatment outcomes?
  • What novel and/or specific image acquisition or analysis methods provide quantifiable clinical benefit over current standard practices? Provides a way to translate medical imaging research developments into guidelines for clinical practice.

Randy suggested partnering with the QI and Radiology decision maker types to design this as a parallel study. We should look for new mi2b2 collaborators from the fields of healthcare outcomes and economics. Verify that our methods will support this cadre of research scientists.
Other opportunities ripe for harvesting (low hanging fruit) in terms of what clinically standard practice acquisitions support advanced image analysis methods:

1. Vascular CT angiography (e.g. cerebral vascular anyeurism morphology as predictors of outcome)
2. Body mass index from MR (e.g. obesity studies)
3. Orthopedics
4. Osteoporosis (e.g. in hip fracture patients look at bone density measures)

On going Action Items

1. Finish amendments to each site i2b2 IRB to add mi2b2 study personnel