The use of mutual information in registration of CT and MRI datasets post permanent implant

1
The use of mutual information in registration of
CT and MRIdatasets post permanent implant

• By Patrick W. McLaughlin, Vrinda Narayana, Marc
  Kessler, Daniel McShan, Sara Troyer, Lon Marsh,
  George Hixson, Peter L. Roberson
• Presentation by Jon Faust
• March 31st, 2005

2
Introduction

• Current American Brachytherapy Society guidelines
  for postimplant dosimetric analysis recommend
  CT-based imaging.
• The weakness of this method is poor definition of
  prostate volume by CT imaging relative to MRI or
  ultrasound imaging. This is especially true in
  the postimplant state, when significant
  anatomical distortion is present due to implanted
  radioactive sources (seeds).

3
Background What is CT?

• Physically, X-rays can traverse a cross-section
  of an object along straight lines, be attenuated
  by the object, and detected outside it.
• During CT scanning, the cross-section is probed
  with X-rays from various directions, attenuated
  signals are recorded and converted to projections
  of the linear attenuation coefficient
  distribution of the cross-section

4
Sample CT image
5
Background What is MRI?

• The MRI machine applies an RF pulse that is
  specific only to hydrogen
• The pulse causes the protons in that area to
  absorb the energy required to make them spin, in
  a different direction.
• The specific frequency of resonance is calculated
  based on the particular tissue being imaged and
  the strength of the main magnetic field.

Photo courtesy NASA
6
Sample MRI Image
7
Why dont we just use MRI?

• Registration remains necessary because no single
  imaging modality provides optimal definition of
  both prostate and seeds.
• While some pulse sequences of MRI do allow better
  definition of seeds, those same pulse sequences
  do not define the prostate volume.

8
Why is CT and MRI registration important?

• More detailed information
• Both seeds and volume can be determined

9
Methods and Materials

• 5 patients who underwent a permanent implant for
  prostate carcinoma were studied
• Two weeks post implant an axial CT, sagittal and
  coronal MRI were taken
• MRI to CT datasets were performed by mutual
  information

10
Cropping

• Necessary to avoid registration being influenced
  by structures beyond the region of interest
• Overlaid using mutual information to form an image

11
Translation

• The MRI axial dataset was aligned to the MRI
  sagittal data set through MI (TMsMa). The MRI
  axial dataset was aligned to the MRI coronal data
  set through MI (TMcMa). The CT axial dataset was
  aligned to the MRI axial data set through MI
  (TMaCt).
• The net rotation-translation matrix for the
  sagittal dataset was obtained by matrix
  multiplication of the sagittal MRI adjustment
  matrix with the axial MRI adjustment matrix
  (TMaCt x TMsMa). The net rotation-translation
  matrix for the coronal datasets was obtained by
  matrix multiplication of the coronal MRI
  adjustment matrix with the axial MRI adjustment
  matrix (TMaCt x TMsMa).

12
Results

• Successfully matched CT and MRI datasets, and
  were able to overlay the CT scan over the MRI
  scan
• The root-mean square (RMS) random errors were
  1.4mm and 0.5o
• Using cropping to focus the registration in the
  region of interest minimizes the impact of
  anatomical differences

13
Window-wash technique for registration evaluation
14
Seed-to-seed registration

• There is a detectable difference in the optimal
  registration using seed position only.
• Differences in registration averaged 1.5 and 1.8
  mm per dimension.
• RMS differences averaged 3.6 mm.

15
Conclusion

• MRI registration can be accomplished rapidly and
  accurately, overcoming the prior substantial time
  investment required.
• MRI-based postimplant dosimetry improves the
  accuracy of prostate dose calculation, improves
  feedback for treatment planning strategies, and
  improves normal tissue dose and toxicity
  analysis.

16
Summary

• Rapid registration of MRI and CT data sets can be
  accomplished through use of mutual information
  software.
• Cropping and translation were effective tools in
  registration of CT and MRI
• Were able to successfully register MRI and CT

17
References

• Nag S, Bice W, deWyngaert K, et al. The American
  Brachytherapy Society recommendations for
  permanent prostate brachytherapy postimplant
  dosimetric analysis. Int J Radiat Oncol Biol Phys
  200046 221230.
• Narayana V, Roberson PL, Winfield RJ, et al.
  Impact of ultrasound and computed tomography
  prostate volume registration on evaluation of
  permanent prostate implant. Int J Radiat Oncol
  Biol Phys 199739 341346.
• Yu Y, Waterman FM, Suntharalingam N, et al.
  Limitations of the minimum peripheral dose as a
  parameter for dose specification in permanent
  125I prostate implants. Int J Radiat Oncol Biol
  Phys 1996 34717725.

18
References

• Algan O, Hanks GE, Shaer AH. Localization of the
  prostatic apex for radiation treatment planning.
  Int J Radiat Oncol Biol Phys 1995 33925930.
• Roach M, Faillace-Akazawa P, Malfatti C, et al.
  Prostate volumes defined by magnetic resonance
  imaging and computerized tomographic scans for
  three-dimensional conformal radiotherapy. Radiat
  Oncol Biol Phys 19963510111018.
• Milosevic M, Voruganti S, Blend R, et al.Magnetic
  resonance imaging (MRI) for localization of
  prostatic apex Comparison to computed tomography
  (CT) and urethrography. Radiother Oncol
  199847277 284.