Slicer:StevePieper RonKikinis VisitToDartmouth 051906

From Slicer Wiki
Jump to: navigation, search
Home < Slicer:StevePieper RonKikinis VisitToDartmouth 051906

Logistics

  • Date: May 19 2006
  • Location: Dartmouth-Hitchcock Medical Center
    • Take I-93 North to I-89 North. Take Exit 18 for Lebanon, NH and turn right onto Route 120, toward Hanover. Follow Rt. 120 for about 3 miles to the entrance to DHMC. The entrance is across from Centerra Park. There are two left turn lanes to get to the hospital – it is easiest if you stay in the one on the right. Make your left turn and at the traffic light, make a right. There is a parking lot called Lot 20 on your left. Visitors can park anywhere in that lot and wait at one of the small glass shelters for a shuttle bus to take you across the street to DHMC.
    • Exit the bus and enter the hospital through the door on your left (Cancer Center Entrance). Make a right turn and walk down to the small café on your right. To your left there is a hallway and an elevator, which you should take to the 5th floor. Exit the elevator and make a right turn. Desk 5D is on your left. The desk staff will be aware of your arrival and can page myself or Andy to let us know you are here.
  • Hotel: Fireside Inn and Suites, West Lebanon, NH (1-800-962-3198)
    • Continue on I-89N, take exit 20. Left off the exit, stay in the left lane. A little way up on the left there is an airport sign, take that left. Take the first road on your left and follow it around to your right. There should be a Fireside Inn and Suites sign pointing you to the hotel. It is located behind LaValley's Hardware.

Questions? Please contact Heather Pixley(heather.pixley@dartmouth.edu).

Proposed Schedule

  • 9am - 10am:
  • segmentation (SP, JW, AS, Heather Wishart)
  • discuss application in Slicer for FLAIRS WMHIs (SP, AS, HW, John MacDonald)
  • review ALICE package and autosegment and "nudgeball" features
  • 10am - 11am:
  • DTI, eddy current correction, 32 direction data (Steve Pieper, John West, Andy Saykin, Ron Kikinis)
  • 11am - noon:
  • review experience with Slicer relative to Analyze (Howie Cleavinger) and BRAINS (Bob Roth)
  • demonstration of brains- basic processing steps.
  • Noon: Lunch, Aud A
  • Meet with Dartmouth team working on NAMIC related projects
  • 1pm: BRAINS and Slicer Hands-on Session and Demonstrations
  • demonstration of brains- basic processing steps. If there is time, will do similar steps in Slicer (SP, AS, Tara McHugh, Bob Roth, Ron Kikinis)
  • Comparison of hippocampal ROI (one side) in brains and Slicer using one case (SP, TM, AS, RK, others?)
  • 2:30pm: Review action Items for Slicer and projects
  • 3:00: Tour of the Dartmouth Advanced Imaging Center
  • 3:30-4:30pm: Administrative Meeting (Ron and Andy)

Meeting Highlights & Notes

  • 9am - 10:30am: Segmentation - Attendees (Steve Pieper, Heather Wishart, John West, John MacDonald, Tara McHugh)
    • Discussed EM Local Segment and how this would apply to MS Lesion Segmentation
      • Discussed creation of Lesion, Gray, White, CSF, and BRAIN apriori (probability) maps and atlases
      • Discussed advanced features of module including creation of sub classes and super classes as well as incorporation of prior probability maps
      • Went over histogram displays and how this could be used to improve training module on different classes
      • Discussed possible improvements to program such as reseting of learned points
    • Action Items
      • Create probability maps from current MS data
      • Use new maps to improve use of EM Local Segment module
      • Use segmentation tool to segment lesions from a group of MS cases and compare to manual segmentation (gold standard) of same lesions
      • Possibly schedule visit from Kilian Pohl to further advance our use of the module
  • 10:30am - 11:30am: DTI - Attendees (SP, JW, TM)
    • Showed Steve new 32 direction DTI data from 3T Philips Scanner
    • Eddy current corrected data vs. same data without eddy current correction
      • FA maps in corrected data were much cleaner and had higher FA values
      • Discussed current eddy current correction scheme (affine registration) used at Dartmouth (developed by Konstantinos Arfanakis, IIT)
    • Gordon Kindlmann's eddy current correction tool in Slicer
      • How to incorporate into our processing pipeline
      • How to run from Slicer command window
      • Discussed testing methods to optimize this for our data
    • Addressed issues with data processing
      • Learned how to view each individual diffusion direction in Slicer for QC of eddy current correction
      • Discussed 'gradient overplus' used in Philips DTI protocol and how this might affect processing of data
      • Discussed Philips PAR/REC (raw data) format and how to incorporate into slicer
    • Action Items
      • Get sample data uploaded (eddy current corrected and not corrected) for Gordon to test with his correction scheme
      • Work with PAR/REC data and determine how to incorporate into slicer (coordinate system issues)
      • Determine how gradient directions for 'gradient overplus' data affect processing of DTI data and come up with correct processing scheme. Strategy: Process data with Philips tools and compare to Slicer processing output of same data
  • 2:00pm - 5:30pm: ROI Editing - Attendees (Ron Kikinis, Andy Saykin, SP, TM)
    • Demonstrated tracing left hippocampus in the brains program (TM)
    • SPL_PNL brain atlas discussed and shown to be a useful teaching tool (RK)
    • Instruction on using Scene options and color codes and how to change them with labelmaps
    • Transformed original spgr volume, problems; see Notes (RK, SP)
    • Demonstrated tracing left hippocampus in Slicer 2.7-dev (TM)
    • Discussed issue of crosshair misalignment:
      • Crosshairs are only aligned on 2 of 3 planes at any one time using the mode for the crosshairs to follow the mouse
      • Looked at IJK mode for the voxel coordinates, unfortunately an underlying issue with Slicer 2 code. (SP)

Background and Preparation

Agenda

  • Demonstration and comparison of manual tracing methods
  • Overview and discussion of progress in fMRI and DTI processing
  • Further develop strategic plan for publication of findings

Attendees

  1. Ron Kikinis
  2. Steve Pieper
  3. Andy Saykin
  4. Tara McHugh
  5. Bob Roth
  6. Laura Flashman
  7. Thomas McAllister
  8. Heather Wishart
  9. Nancy Koven
  10. Jo Cara Pendergrass
  11. Howie Cleavinger
  12. John West
  13. John MacDonald
  14. Heather Pixley

Notes from Editing Mode

  • Looked at IJK mode for the voxel coordinates. There is always one plane out of 3 mis-aligned by 1 - 2 voxels. This is likely due to the original SLICER program and subsequent development.(Tara McHugh)
  • Attempt at transform 12X original spgr scan to lessen pixelation of screen. Using 0.5 x 0.5 x 0.5 was too much. I will test isometric voxel measures at 0.7^3 to see if this can be supported. This might lessen subjective issues of GM-WM boundaries for some neuronal areas.(Tara McHugh) 5/23/06 Missing a step in the transform, so not able to replicate.TM