Coregistration and Correlation of MRI and MALDIIMS data

1
Co-registration and Correlation of MRI and
MALDI-IMS data

• Zhengyu Yang

2
Overview

• Project Aim
• Integrate MALDI-IMS information and MRI
  information to better interpret both modalities
• Project Plan
• Co-register MRI images, MALDI-IMS images and
  blockface images
• Total water content measured by MR ??the total
  protein content in tissue using MALDI-IMS
• (Short term)
• Relaxomics
• ( Long term)

3
MRI

• High resolution MRI shows exquisite tissue
  structure and pathology
• Contrast reflects interaction of water with
  tissue macromolecules and cell boundaries

4
MRI relaxation

• We lack adequate models of tissue-protein
  interactions to quantitatively explain the
  variations seen in MR images or the changes
  measured in various pathological states in terms
  of specific underlying macromolecular changes.

5
MALDI-IMS

• Matrix-assisted Laser Desorption/Ionization
• A powerful technique for identifying changes in
  proteomic profile that vary spatially
• A soft ionization technique used in mass
  spectrometry
• Analysis biomolecules and large
• organic molecules
• Similar to electrospray ionization

6
MALDI-IMS

• Ionization is triggered by a laser beam
• A matrix is used to protect the biomolecule from
  being destroyed by direct laser beam and to
  facilitate vaporization and ionization.
• Ionized particles are available for traditional
  mass analysis (i.e. TOF)

7
MALDI-IMS

• Ionization is triggered by a laser beam
• A matrix is used to protect the biomolecule from
  being destroyed by direct laser beam and to
  facilitate vaporization and ionization.
• Ionized particles are available for traditional
  mass analysis (i.e. TOF)

8
MALDI-IMS

• Advantages
• Softer ionization process
• (capable of producing protonated molecular ions
  for small proteins)
• Good sensitivity (into the picomole range)
• High mass accuracy
• Co-registered with conventional histological
  images

9
Co-registration of MRI and MALDI- IMS

• Accurate co-registration of MALDI-IMS to MRI data
  will increase the information provided by each
  modality
• In addition, integration of the different
  techniques will provide a unique ability to
  correlate different types of information in order
  to enhance the interpretation of both modalities

10
Preliminary Results

• The areas of high protein concentration agreed
  well with contrast variations in MR and blockface
  volumes

11
Our hypothesize

• However, quantitative relationships are still
  unknown
• relaxation within tissue ?? variations in its
  protein contents
• Voxel-voxel variations in water content are
  correlated with total protein concentration

12
Current work

• To further investigate these relationships, we
  have designed a study to quantitatively correlate
  total water content measured by MR with total
  protein content in tissue using MALDI-IMS.

13
Current work

• MRI data collection and processing
• Data Acquisition
• Fresh calf liver sealed in a 10mm Eppendorf tube
• MR aquisiton _at_ 7 Tesla ( MR spectrometer at
  varian)
• Spin-echo sequence, Multi-slice
• TR 333-4000ms
• TE 12-33ms

14
Current work

• Data Processing
• M0, T1 and T2 can be estimated by nonlinearly fit

(From left to right) M0 , T1 and T2 maps
generated from scanner results of calf liver.
Thickness 0.5. Gap 0. Matrix 256x256

15
Future work

• MALDI-IMS data collection and processing
• Samples will be sectioned and imaged via
  MALDI-IMS
• Images will be co-registered with MRI images
• Explore how voxel-voxel variation in MRI data
  correlated to MALDI-IMS data

16
Future work

• The Model
• Suppose within a voxel there is a total mass of
  water distributed between bound and free
  fractions, fb and ff 1-fb
• fb is the fraction of bound/total water
• In fast exchanges,

17
Future work

• In any case,
• So that
• Assuming and are constants

18
Disscusion

• Hope to show that the voxel-by-voxel relationship
  of total protein content measured by MALDI-IMS
  can be quantitatively correlated with M0 measures
  from MR.
• Expect to see minor spatial variations in this
  concentration going from voxel to voxel related
  to local deviations in the tissue composition
• Hypothesize a similar phenomena will be observed
  when the same tissue sample is imaged using NMR
• Hope to build quantitative model to describe the
  correlation between the spatially variations in
  NMR and MALDI-IMS data

19
Acknowledgement

• Tuhin K. Sinha
• John C. Gore
• Richard Baheza
• Jeffrey James Luci
• Don Nolting
• Wellington Pham
• Jarrod M. True

Funding NIH-NIBIB R01-EB000214, R01-CA109106,
T32-EB001628