All Hands Meeting 2004

Title: Planned GCRC-BIRN Collaboration at BWH

Jonathan Sacks, PhD

Quantitative perfusion analysis

Methods - MRI protocol

• Absolute myocardial blood flows (MBF) (ml/min/g) at rest and during dipyridamole vasodilation were quantified by validated Fermi model deconvolution.

• Rest myocardial blood flow corrected for rate pressure product

• 6 segments at each endomyocardial and epimyocardial levels in each slice location

• Perfusion reserve (PR) = MBF during vasodilation MBF at rest

Abnormal quantitative FPMP = Segmental endomyocardial PR significantly lower than matching epimyocardial PR by paired student t-test (p < 0.05)

Scenario One: Cardiac MR

• Cardiac research is being done at GCRC

Under G. Adler M.D., PH.D., R. Kwong M.D. and others

• The study uses Cardiac MR to visualize blood perfusing through the heart. The purpose of the study is to determine whether treatment with an aldosterone receptor antagonist will improve myocardial perfusion in patients with diabetes mellitus.

• The image of the heart is divided into segments at each step in a time series. This enables blood perfusion to be visualized through time on each segment

Scenario One: Cardiac MR

• The Surgical Planning Lab (SPL) at BWH has over many years developed a set of tools for visualizing and analyzing MR images of the human brain and other organs. SPL is a participating site for both Morphometry and Function BIRN.

• These techniques provide superior resolution of MR images and many of them are automatic or semi-automatic

• If the techniques developed at SPL are applied to the study of blood perfusion in the cardiac study, very substantial gains in resolution and efficiency of computation are realized in comparison with current Cardiac MR

Scenario One: Cardiac MR

• One of the key requirements for tracking perfusion through a time series is registration of elements of the series with a baseline (the ‘first’ time slice)

• SPL has developed several registration techniques which are so accurate that it will be possible to track blood perfusion through a single voxel

• Segmentation using level sets, followed by ICPRegistration using non-linear warpingSpatio-temporal filtering

Scenario One: Cardiac MR

Alignment of anatomical models, MRI data and functional data from the CARTO system. Depicted here is simulated CARTO data and deidentified preoperative MRI data where the red dots show the position of the CARTO data points. The initial misaligned data (left) and the registration

Scenario One: Cardiac MR

Figure 12. Ibrowser controller showing image intervals and controls for indexing and animating.

nameorder

visibilitycopy

deletehold

viewer FGviewer BG

animation & viewing

manual indexing

Figure 1. Generalized visualization and morphometric tools supporting FBIRN FLAC

QA spatialprocessing

activationdetection

datainspection morphometric

interactive,multi-modalityvisualization

report

Interactive visualization tools

DB

morphometricfunctionalgeneticclinicaldemographic …

validation,reliability testing,populationpooling

activation detector suite

…

Interactive query for knowledge discovery

Scenario One: Cardiac MR

fMRI AMPA Potentiator Study Design

Baseline

LY451395 (0.2 mg BID)

or Placebo

Scan 1 Scan 2 Scan 3

Week 0 Week 5 Week 9

ScreeningVisit

-7 to -9days

LY451395 (1.0 mg BID)

or Placebo

fMRI Block Design Paradigm

NovelFace-Name

Pairs

RepeatedFace-Name

Pairs

NovelFace-Name

Pairs

RepeatedFace-Name

Pairs

40s 25s5s 40s 25s 40s 40s25s 5s

+ + ++ +

Scenario 2: Alzheimer Study

LEAM fMRI

Thirteen AD subjects completed study with all three adequate fMRI scan sessions• Eight subjects on MGH Siemens 3T (Trio)• Five subjects on BWH GE 3T (long bore)

fMRI Block Design: Test-Retest Reliability in AD patient

Differences in prefrontal activation in LY451395 vs. placebo

Session 3 > Session 1 for Drug (n=4) > Placebo (n=9)

GCRC: Gail Adler, MD, PhD

Reisa Sperling

Ray Kwong, MD

Joe Ronda, MS

Surgical Planning Lab, BWH:

Ron Kikinis, MD

Steve Pieper, PhD

C.F. Westin, PhD

Wendy Plesniak, PhD

Acknowledgements