NAMIC Wiki - User contributions [en]

DBP3:Utah:RegSegPipeline

2011-08-29T19:23:03Z

Dillonl: /* Current Progress */

[[DBP3:Utah| back to DBP3 home]]
= The CARMA DBP: MRI-based study and treatment of atrial fibrillation =
== Pilot Studies on a Registration & Segmentation Pipeline & Workflow ==
Alex Zaitsev, Dominik Meier, Ron Kikinis

=== Background ===
By Rob MacLeod, PI AFib DBP

The Utah group carries out analysis of MRI scans from all patients coming to our clinical with symptoms of Atrial Fibrillation (AF) and continues to conduct regular MRI scans for patients once they undergo treatment by our physicians. The result is over 1000 MRI scans per year that we process using a set of heterogeneous tools. One goal of our collaboration with NAMIC is to explore new tools and capabilities to consolidate and streamline this processing pipeline. The goals of the process vary slightly depending on the timing of the imaging relative to patient care but all involve some combination of image correction, region of interest definition, segmentation of the left atrium (in some cases also right atrium and esophagus), and quantitative evaluation of enhancement in the images, due either to pre-existing fibrosis or post-treatment scar. In many cases, registration is also necessary to compare multiple images from the same patients over time and to compare multiple patients at the same stage of disease progression and treatment.

For more information, please see the following references:

R.S. Oakes, T.J. Badger, E.G. Kholmovski, N. Akoum, N.S. Burgon, E.N. Fish, J.J. Blauer, S.N. Rao, E.V. DiBella, N.M. Segerson, M. Daccarett, J. Windfelder, C.J. McGann, D. Parker, R.S. MacLeod, N.F. Marrouche. “Detection and quantification of left atrial structural remodeling with delayed-enhancement magnetic resonance imaging in patients with atrial fibrillation,” In Circulation, Vol. 119, No. 13, pp. 1758--1767. 2009. http://www.sci.utah.edu/publications/oaks09/Oaks_Circ2009.pdf

C.J. McGann, E.G. Kholmovski, R.S. Oakes, J.J. Blauer, M. Daccarett, N. Segerson, K.J. Airey, N. Akoum, E. Fish, T.J. Badger, E.V. DiBella, D. Parker, R.S. MacLeod, N.F. Marrouche. “New Magnetic Resonance Imaging-Based Method for Defining the Extent of Left Atrial Wall Injury After the Ablation of Atrial Fibrillation,” In Journal of the American College of Cardiology, Vol. 52, No. 15, pp. 1263--1271. Oct 7, 2008. http://www.sci.utah.edu/publications/mcgann08/mcg2008.pdf

C. Mahnkopf, T.J. Badger, N.S. Burgon, M. Daccarett, T.S. Haslam, C.T. Badger, C.J. McGann, N. Akoum, E. Kholmovski, R.S. Macleod, N.F. Marrouche. “Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheter ablation,” In Heart Rhythm, Vol. 7, No. 10, pp. 1475--1481. 2010. http://www.sci.utah.edu/publications/mahnkopf10/mahnkopf2010.pdf

We are in the process of making a small set of anonymized test data available for general use and more extensive deidentified data for exclusive use with NAMIC. Please contact me (macleod@sci.utah.edu) for more details.

=== Main processing pipeline ===
To facilitate the workflow, we can place all the automated steps at the beginning and cluster interactive elements at the end. Exception is the cropping step required as input for nonrigid registration.<br>
1. N4 bias field correction for the MRI (surface coils):
::*'''Input''': MRI_pre and MRI_post, each run separately with the same parameters below; '''Output:''' MRI_pre_n4 and MRI_post_n4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:N4ITKBiasFieldCorrection-Documentation-3.6 N4 ITK]; '''Parameters:''' convergence: 1e-5, iterations: 50,40,30,20, shrink factor: 3
::*''Comments:'' a run on entire image gives some benefit that '''may''' be improved with masking: again the dominant intensity dropoff from the surface coil occurs along the chest wall and ribcage. Even if that is not the structure of interest, it is the low-freq. variation the bias correction algorithm is searching for, and masking that out can be counter-productive: via masking we may end up with a smoother image, but the intensity variations removed were not caused by the coil but are actually true signal.
2. affine registration MRA ->MRI (both pre and post)
::*'''Input''': MRI_pre_n4 (fixed) and MRA_pre ; '''Output:''' Xf1_pre_MRA-MRI.tfm"
::*'''Input''': MRI_post_n4 (fixed) and MRA_post ; '''Output:''' Xf2_post_MRA-MRI.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' no initialization; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' from the example dataset (P1) we infer that the two scans tend to have little initial alignment, hence initialization steps are not recommended. We choose registration parameters such that the chance of the registration making the alignment worse is minimized. Decide in a subsequent review/QC step if we keep the transform or the original pose. The MRA contains the same FOV and has surrounding structures (liver, chest, spine etc) visible also, despite lower intensities. A global affine is thus not necessarily going to benefit from masking the heart, unless the relative motion of the heart becomes the dominant reason for misalignment. We tried masking with both BrainsFit and RobustMultires modules. Both failed to provide better alignment with masking.
3. registration post -> pre: phase 1: AFFINE
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' Xf3_pre-post_affine.tfm
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' DOF: rigid+similarity+affine (possibly rigid only); initialization: Moments align ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' done on the entire image. The surrounding structures are useful in constraining the solution transform and should provide a more robust behavior. Cropping down to the cardiac only ROI is deferred to the nonrigid registration in phase 2. The inferior-superior FOV can differ, e.g. how much of the liver is included. If the two exams differ significantly (>30%) in that content, the above affine could fail and a prior cropping step is then suggested to better match image content before registration. Resampling to isotropic voxel size at this stage is also advantageous but will generate very large files > 100MB due to the full FOV.
4. Cropping to VOI:
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' MRI_pre_n4_cropped, MRI_post_n4_cropped
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 Crop Volume]
::*''Comments:'' the [http://www.slicer.org/slicerWiki/index.php/Modules:VolumeRendering-Documentation-3.6 volume rendering module] may help in obtaining a good cropping ROI. Because of the high contrast, the MRA provides a good source for volume rendering.
5. registration post -> pre: phase 2: nonrigid / BSPLINE
::*'''Input''': MRI_pre_n4_cropped (fixed) and MRI_post_n4_cropped (moving) ; '''Output:''' Xf4_pre-post_BSpline.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*'''Parameters:''' DOF: BSpline only; initialization: Xf3_pre-post_affine.tfm ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' crop to volume of interest: [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 CropVolume module], use ''Resample to isotropic voxelsize'' option.
6. resample MRA: apply above BSpline to the pre MRA
::*'''Input Volume''': MRA_post_n4_cropped
::*'''Input Transform''': above Xf4_pre-post_BSpline.tfm; '''Output Volume''': MRA_post_n4_cropped_Xf4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:ResampleScalarVectorDWIVolume-Documentation-3.6 ResampleScalarVectorDWIVolume]
::*'''Parameters:''' output-to-input box checked, interpolation: linear
7. ROI definition (manual box ROI or automated via atlas)
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
8. segmentation of LA from MRA -> inner wall
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:RobustStatisticsSeg-Documentation-3.6 Robust Statistics module] or [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding] for thresholding within [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*''Comments:'' as a dynamic image the MRA contains significant spread and likely requires interactive segmentation/thresholding to yield a satisfactory LA volume. For validation/visualization, use the :[http://www.slicer.org/slicerWiki/index.php/Modules:Volumes-Documentation-3.6 Volumes] thresholding option within ''Display'' tab, use ''iron'' colormap & low alpha setting to check for ventricular wall borders. (see figures below)
9. LA wall segmentation
::*'''Module used:''' ''Atrium Cardiac Wall Segmentation'' (Extension Module, no documentation yet, authors: Yi Gao, Behnood Gholami, Allen Tannenbaum)
10. segmentation of enhancement
::*'''Module:''' [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding]
::*'''alternative Module:''' [http://wiki.na-mic.org/Wiki/index.php/Projects:AblationScarSegmentation AblationScarSegmentation]
::*''Comments:'; operate only on ROI within LA wall. Based on proper intensity statistics. An atlas-based set of intensity distributions may be more meaningful here than a simple Otsu, because both amount and location of enhancement is unknown and can in theory be 0.

<gallery perrow="4" widths="200px">
Image:DBP3_AFib_LV_overlay_1.jpg| Example contrast MRI with thresholded MRA as color overlay. Registration validation is not trivial: we can in theory use the position of the blood in the MRA relative to the LV wall as registration check, but the MRA itself does not provide a clear boundary at the lower intensities.
Image:DBP3_AFib_LA_overlay_2.jpg|Example contrast MRI with thresholded MRA as color overlay and areas of enhancement marked
Image:DBP3_histogramMRA.png|Histogram of MRA intensities: green = entire FOV, red: cardiac structures only (excl. aorta), which make only 6% of the total image content and thus are unlikely to disturb a registration based on the entire image.
Image:DBP3_Biasfield_masked-unmasked.png|Bias field correction results (N4ITK) with and without masking of the structures of interest. Masking causes a different correction not related to coil sensitivity but rather local image content. Masking is not recommended at this point.

</gallery>

=== Example Cases ===
* [[Media:AFib_Example_P1_Data.zip|Example Case P1 : pre-post N4 bias corregtion, registration, LA segmentation]]

=== Current Progress ===
As we proceed and make progress with the pipeline we have simplified it by consolidating many of the steps with minimal user input. Users are now prompted to provide a pre and post set of MRI data. Once provided, a button labeled 'MRI Intensity Correction' is enabled. When the user clicks the button, Slicer begins the pipeline by using N4 ITK MRI bias correction immediately followed by a whole-image affine registration. Once completed, the user is then prompted to crop the images to the desired ROI. After the ROI is defined a non-rigid registration is performed yielding registered pre and post MRI images and the corresponding transformation matrix. The following image gallery shows the progress and current state of the GUI.

<gallery perrow="2" widths="550px" heights="300px">
Image:PipelineStart.png|Registration Segmentation Pipeline input selection - raw images.
Image:PipelineStep1.png|N4 ITK MRI bias correction completed.
</gallery>

DBP3:Utah:RegSegPipeline

2011-08-29T19:21:35Z

Dillonl: /* Current Progress */

[[DBP3:Utah| back to DBP3 home]]
= The CARMA DBP: MRI-based study and treatment of atrial fibrillation =
== Pilot Studies on a Registration & Segmentation Pipeline & Workflow ==
Alex Zaitsev, Dominik Meier, Ron Kikinis

=== Background ===
By Rob MacLeod, PI AFib DBP

The Utah group carries out analysis of MRI scans from all patients coming to our clinical with symptoms of Atrial Fibrillation (AF) and continues to conduct regular MRI scans for patients once they undergo treatment by our physicians. The result is over 1000 MRI scans per year that we process using a set of heterogeneous tools. One goal of our collaboration with NAMIC is to explore new tools and capabilities to consolidate and streamline this processing pipeline. The goals of the process vary slightly depending on the timing of the imaging relative to patient care but all involve some combination of image correction, region of interest definition, segmentation of the left atrium (in some cases also right atrium and esophagus), and quantitative evaluation of enhancement in the images, due either to pre-existing fibrosis or post-treatment scar. In many cases, registration is also necessary to compare multiple images from the same patients over time and to compare multiple patients at the same stage of disease progression and treatment.

For more information, please see the following references:

R.S. Oakes, T.J. Badger, E.G. Kholmovski, N. Akoum, N.S. Burgon, E.N. Fish, J.J. Blauer, S.N. Rao, E.V. DiBella, N.M. Segerson, M. Daccarett, J. Windfelder, C.J. McGann, D. Parker, R.S. MacLeod, N.F. Marrouche. “Detection and quantification of left atrial structural remodeling with delayed-enhancement magnetic resonance imaging in patients with atrial fibrillation,” In Circulation, Vol. 119, No. 13, pp. 1758--1767. 2009. http://www.sci.utah.edu/publications/oaks09/Oaks_Circ2009.pdf

C.J. McGann, E.G. Kholmovski, R.S. Oakes, J.J. Blauer, M. Daccarett, N. Segerson, K.J. Airey, N. Akoum, E. Fish, T.J. Badger, E.V. DiBella, D. Parker, R.S. MacLeod, N.F. Marrouche. “New Magnetic Resonance Imaging-Based Method for Defining the Extent of Left Atrial Wall Injury After the Ablation of Atrial Fibrillation,” In Journal of the American College of Cardiology, Vol. 52, No. 15, pp. 1263--1271. Oct 7, 2008. http://www.sci.utah.edu/publications/mcgann08/mcg2008.pdf

C. Mahnkopf, T.J. Badger, N.S. Burgon, M. Daccarett, T.S. Haslam, C.T. Badger, C.J. McGann, N. Akoum, E. Kholmovski, R.S. Macleod, N.F. Marrouche. “Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheter ablation,” In Heart Rhythm, Vol. 7, No. 10, pp. 1475--1481. 2010. http://www.sci.utah.edu/publications/mahnkopf10/mahnkopf2010.pdf

We are in the process of making a small set of anonymized test data available for general use and more extensive deidentified data for exclusive use with NAMIC. Please contact me (macleod@sci.utah.edu) for more details.

=== Main processing pipeline ===
To facilitate the workflow, we can place all the automated steps at the beginning and cluster interactive elements at the end. Exception is the cropping step required as input for nonrigid registration.<br>
1. N4 bias field correction for the MRI (surface coils):
::*'''Input''': MRI_pre and MRI_post, each run separately with the same parameters below; '''Output:''' MRI_pre_n4 and MRI_post_n4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:N4ITKBiasFieldCorrection-Documentation-3.6 N4 ITK]; '''Parameters:''' convergence: 1e-5, iterations: 50,40,30,20, shrink factor: 3
::*''Comments:'' a run on entire image gives some benefit that '''may''' be improved with masking: again the dominant intensity dropoff from the surface coil occurs along the chest wall and ribcage. Even if that is not the structure of interest, it is the low-freq. variation the bias correction algorithm is searching for, and masking that out can be counter-productive: via masking we may end up with a smoother image, but the intensity variations removed were not caused by the coil but are actually true signal.
2. affine registration MRA ->MRI (both pre and post)
::*'''Input''': MRI_pre_n4 (fixed) and MRA_pre ; '''Output:''' Xf1_pre_MRA-MRI.tfm"
::*'''Input''': MRI_post_n4 (fixed) and MRA_post ; '''Output:''' Xf2_post_MRA-MRI.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' no initialization; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' from the example dataset (P1) we infer that the two scans tend to have little initial alignment, hence initialization steps are not recommended. We choose registration parameters such that the chance of the registration making the alignment worse is minimized. Decide in a subsequent review/QC step if we keep the transform or the original pose. The MRA contains the same FOV and has surrounding structures (liver, chest, spine etc) visible also, despite lower intensities. A global affine is thus not necessarily going to benefit from masking the heart, unless the relative motion of the heart becomes the dominant reason for misalignment. We tried masking with both BrainsFit and RobustMultires modules. Both failed to provide better alignment with masking.
3. registration post -> pre: phase 1: AFFINE
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' Xf3_pre-post_affine.tfm
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' DOF: rigid+similarity+affine (possibly rigid only); initialization: Moments align ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' done on the entire image. The surrounding structures are useful in constraining the solution transform and should provide a more robust behavior. Cropping down to the cardiac only ROI is deferred to the nonrigid registration in phase 2. The inferior-superior FOV can differ, e.g. how much of the liver is included. If the two exams differ significantly (>30%) in that content, the above affine could fail and a prior cropping step is then suggested to better match image content before registration. Resampling to isotropic voxel size at this stage is also advantageous but will generate very large files > 100MB due to the full FOV.
4. Cropping to VOI:
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' MRI_pre_n4_cropped, MRI_post_n4_cropped
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 Crop Volume]
::*''Comments:'' the [http://www.slicer.org/slicerWiki/index.php/Modules:VolumeRendering-Documentation-3.6 volume rendering module] may help in obtaining a good cropping ROI. Because of the high contrast, the MRA provides a good source for volume rendering.
5. registration post -> pre: phase 2: nonrigid / BSPLINE
::*'''Input''': MRI_pre_n4_cropped (fixed) and MRI_post_n4_cropped (moving) ; '''Output:''' Xf4_pre-post_BSpline.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*'''Parameters:''' DOF: BSpline only; initialization: Xf3_pre-post_affine.tfm ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' crop to volume of interest: [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 CropVolume module], use ''Resample to isotropic voxelsize'' option.
6. resample MRA: apply above BSpline to the pre MRA
::*'''Input Volume''': MRA_post_n4_cropped
::*'''Input Transform''': above Xf4_pre-post_BSpline.tfm; '''Output Volume''': MRA_post_n4_cropped_Xf4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:ResampleScalarVectorDWIVolume-Documentation-3.6 ResampleScalarVectorDWIVolume]
::*'''Parameters:''' output-to-input box checked, interpolation: linear
7. ROI definition (manual box ROI or automated via atlas)
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
8. segmentation of LA from MRA -> inner wall
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:RobustStatisticsSeg-Documentation-3.6 Robust Statistics module] or [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding] for thresholding within [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*''Comments:'' as a dynamic image the MRA contains significant spread and likely requires interactive segmentation/thresholding to yield a satisfactory LA volume. For validation/visualization, use the :[http://www.slicer.org/slicerWiki/index.php/Modules:Volumes-Documentation-3.6 Volumes] thresholding option within ''Display'' tab, use ''iron'' colormap & low alpha setting to check for ventricular wall borders. (see figures below)
9. LA wall segmentation
::*'''Module used:''' ''Atrium Cardiac Wall Segmentation'' (Extension Module, no documentation yet, authors: Yi Gao, Behnood Gholami, Allen Tannenbaum)
10. segmentation of enhancement
::*'''Module:''' [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding]
::*'''alternative Module:''' [http://wiki.na-mic.org/Wiki/index.php/Projects:AblationScarSegmentation AblationScarSegmentation]
::*''Comments:'; operate only on ROI within LA wall. Based on proper intensity statistics. An atlas-based set of intensity distributions may be more meaningful here than a simple Otsu, because both amount and location of enhancement is unknown and can in theory be 0.

<gallery perrow="4" widths="200px">
Image:DBP3_AFib_LV_overlay_1.jpg| Example contrast MRI with thresholded MRA as color overlay. Registration validation is not trivial: we can in theory use the position of the blood in the MRA relative to the LV wall as registration check, but the MRA itself does not provide a clear boundary at the lower intensities.
Image:DBP3_AFib_LA_overlay_2.jpg|Example contrast MRI with thresholded MRA as color overlay and areas of enhancement marked
Image:DBP3_histogramMRA.png|Histogram of MRA intensities: green = entire FOV, red: cardiac structures only (excl. aorta), which make only 6% of the total image content and thus are unlikely to disturb a registration based on the entire image.
Image:DBP3_Biasfield_masked-unmasked.png|Bias field correction results (N4ITK) with and without masking of the structures of interest. Masking causes a different correction not related to coil sensitivity but rather local image content. Masking is not recommended at this point.

</gallery>

=== Example Cases ===
* [[Media:AFib_Example_P1_Data.zip|Example Case P1 : pre-post N4 bias corregtion, registration, LA segmentation]]

=== Current Progress ===
As we proceed and make progress with the pipeline we have simplified it by consolidating many of the steps with minimal user input. Users are now prompted to provide a pre and post set of MRI data. Once provided, a button labeled 'MRI Intensity Correction' is enabled. When the user clicks the button, Slicer begins the pipeline by using N4 ITK MRI bias correction immediately followed by a whole-image affine registration. Once completed, the user is then prompted to crop the images to the desired ROI. After the ROI is defined a non-rigid registration is performed yielding registered pre and post MRI images and the corresponding transformation matrix. The following image gallery shows the progress and current state of the GUI.

<gallery perrow="2">
Image:PipelineStart.png|Registration Segmentation Pipeline input selection - raw images.
Image:PipelineStep1.png|N4 ITK MRI bias correction completed.
</gallery>

DBP3:Utah:RegSegPipeline

2011-08-29T19:21:13Z

Dillonl: /* Current Progress */

[[DBP3:Utah| back to DBP3 home]]
= The CARMA DBP: MRI-based study and treatment of atrial fibrillation =
== Pilot Studies on a Registration & Segmentation Pipeline & Workflow ==
Alex Zaitsev, Dominik Meier, Ron Kikinis

=== Background ===
By Rob MacLeod, PI AFib DBP

The Utah group carries out analysis of MRI scans from all patients coming to our clinical with symptoms of Atrial Fibrillation (AF) and continues to conduct regular MRI scans for patients once they undergo treatment by our physicians. The result is over 1000 MRI scans per year that we process using a set of heterogeneous tools. One goal of our collaboration with NAMIC is to explore new tools and capabilities to consolidate and streamline this processing pipeline. The goals of the process vary slightly depending on the timing of the imaging relative to patient care but all involve some combination of image correction, region of interest definition, segmentation of the left atrium (in some cases also right atrium and esophagus), and quantitative evaluation of enhancement in the images, due either to pre-existing fibrosis or post-treatment scar. In many cases, registration is also necessary to compare multiple images from the same patients over time and to compare multiple patients at the same stage of disease progression and treatment.

For more information, please see the following references:

R.S. Oakes, T.J. Badger, E.G. Kholmovski, N. Akoum, N.S. Burgon, E.N. Fish, J.J. Blauer, S.N. Rao, E.V. DiBella, N.M. Segerson, M. Daccarett, J. Windfelder, C.J. McGann, D. Parker, R.S. MacLeod, N.F. Marrouche. “Detection and quantification of left atrial structural remodeling with delayed-enhancement magnetic resonance imaging in patients with atrial fibrillation,” In Circulation, Vol. 119, No. 13, pp. 1758--1767. 2009. http://www.sci.utah.edu/publications/oaks09/Oaks_Circ2009.pdf

C.J. McGann, E.G. Kholmovski, R.S. Oakes, J.J. Blauer, M. Daccarett, N. Segerson, K.J. Airey, N. Akoum, E. Fish, T.J. Badger, E.V. DiBella, D. Parker, R.S. MacLeod, N.F. Marrouche. “New Magnetic Resonance Imaging-Based Method for Defining the Extent of Left Atrial Wall Injury After the Ablation of Atrial Fibrillation,” In Journal of the American College of Cardiology, Vol. 52, No. 15, pp. 1263--1271. Oct 7, 2008. http://www.sci.utah.edu/publications/mcgann08/mcg2008.pdf

C. Mahnkopf, T.J. Badger, N.S. Burgon, M. Daccarett, T.S. Haslam, C.T. Badger, C.J. McGann, N. Akoum, E. Kholmovski, R.S. Macleod, N.F. Marrouche. “Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheter ablation,” In Heart Rhythm, Vol. 7, No. 10, pp. 1475--1481. 2010. http://www.sci.utah.edu/publications/mahnkopf10/mahnkopf2010.pdf

We are in the process of making a small set of anonymized test data available for general use and more extensive deidentified data for exclusive use with NAMIC. Please contact me (macleod@sci.utah.edu) for more details.

=== Main processing pipeline ===
To facilitate the workflow, we can place all the automated steps at the beginning and cluster interactive elements at the end. Exception is the cropping step required as input for nonrigid registration.<br>
1. N4 bias field correction for the MRI (surface coils):
::*'''Input''': MRI_pre and MRI_post, each run separately with the same parameters below; '''Output:''' MRI_pre_n4 and MRI_post_n4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:N4ITKBiasFieldCorrection-Documentation-3.6 N4 ITK]; '''Parameters:''' convergence: 1e-5, iterations: 50,40,30,20, shrink factor: 3
::*''Comments:'' a run on entire image gives some benefit that '''may''' be improved with masking: again the dominant intensity dropoff from the surface coil occurs along the chest wall and ribcage. Even if that is not the structure of interest, it is the low-freq. variation the bias correction algorithm is searching for, and masking that out can be counter-productive: via masking we may end up with a smoother image, but the intensity variations removed were not caused by the coil but are actually true signal.
2. affine registration MRA ->MRI (both pre and post)
::*'''Input''': MRI_pre_n4 (fixed) and MRA_pre ; '''Output:''' Xf1_pre_MRA-MRI.tfm"
::*'''Input''': MRI_post_n4 (fixed) and MRA_post ; '''Output:''' Xf2_post_MRA-MRI.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' no initialization; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' from the example dataset (P1) we infer that the two scans tend to have little initial alignment, hence initialization steps are not recommended. We choose registration parameters such that the chance of the registration making the alignment worse is minimized. Decide in a subsequent review/QC step if we keep the transform or the original pose. The MRA contains the same FOV and has surrounding structures (liver, chest, spine etc) visible also, despite lower intensities. A global affine is thus not necessarily going to benefit from masking the heart, unless the relative motion of the heart becomes the dominant reason for misalignment. We tried masking with both BrainsFit and RobustMultires modules. Both failed to provide better alignment with masking.
3. registration post -> pre: phase 1: AFFINE
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' Xf3_pre-post_affine.tfm
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' DOF: rigid+similarity+affine (possibly rigid only); initialization: Moments align ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' done on the entire image. The surrounding structures are useful in constraining the solution transform and should provide a more robust behavior. Cropping down to the cardiac only ROI is deferred to the nonrigid registration in phase 2. The inferior-superior FOV can differ, e.g. how much of the liver is included. If the two exams differ significantly (>30%) in that content, the above affine could fail and a prior cropping step is then suggested to better match image content before registration. Resampling to isotropic voxel size at this stage is also advantageous but will generate very large files > 100MB due to the full FOV.
4. Cropping to VOI:
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' MRI_pre_n4_cropped, MRI_post_n4_cropped
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 Crop Volume]
::*''Comments:'' the [http://www.slicer.org/slicerWiki/index.php/Modules:VolumeRendering-Documentation-3.6 volume rendering module] may help in obtaining a good cropping ROI. Because of the high contrast, the MRA provides a good source for volume rendering.
5. registration post -> pre: phase 2: nonrigid / BSPLINE
::*'''Input''': MRI_pre_n4_cropped (fixed) and MRI_post_n4_cropped (moving) ; '''Output:''' Xf4_pre-post_BSpline.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*'''Parameters:''' DOF: BSpline only; initialization: Xf3_pre-post_affine.tfm ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' crop to volume of interest: [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 CropVolume module], use ''Resample to isotropic voxelsize'' option.
6. resample MRA: apply above BSpline to the pre MRA
::*'''Input Volume''': MRA_post_n4_cropped
::*'''Input Transform''': above Xf4_pre-post_BSpline.tfm; '''Output Volume''': MRA_post_n4_cropped_Xf4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:ResampleScalarVectorDWIVolume-Documentation-3.6 ResampleScalarVectorDWIVolume]
::*'''Parameters:''' output-to-input box checked, interpolation: linear
7. ROI definition (manual box ROI or automated via atlas)
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
8. segmentation of LA from MRA -> inner wall
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:RobustStatisticsSeg-Documentation-3.6 Robust Statistics module] or [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding] for thresholding within [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*''Comments:'' as a dynamic image the MRA contains significant spread and likely requires interactive segmentation/thresholding to yield a satisfactory LA volume. For validation/visualization, use the :[http://www.slicer.org/slicerWiki/index.php/Modules:Volumes-Documentation-3.6 Volumes] thresholding option within ''Display'' tab, use ''iron'' colormap & low alpha setting to check for ventricular wall borders. (see figures below)
9. LA wall segmentation
::*'''Module used:''' ''Atrium Cardiac Wall Segmentation'' (Extension Module, no documentation yet, authors: Yi Gao, Behnood Gholami, Allen Tannenbaum)
10. segmentation of enhancement
::*'''Module:''' [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding]
::*'''alternative Module:''' [http://wiki.na-mic.org/Wiki/index.php/Projects:AblationScarSegmentation AblationScarSegmentation]
::*''Comments:'; operate only on ROI within LA wall. Based on proper intensity statistics. An atlas-based set of intensity distributions may be more meaningful here than a simple Otsu, because both amount and location of enhancement is unknown and can in theory be 0.

<gallery perrow="4" widths="200px">
Image:DBP3_AFib_LV_overlay_1.jpg| Example contrast MRI with thresholded MRA as color overlay. Registration validation is not trivial: we can in theory use the position of the blood in the MRA relative to the LV wall as registration check, but the MRA itself does not provide a clear boundary at the lower intensities.
Image:DBP3_AFib_LA_overlay_2.jpg|Example contrast MRI with thresholded MRA as color overlay and areas of enhancement marked
Image:DBP3_histogramMRA.png|Histogram of MRA intensities: green = entire FOV, red: cardiac structures only (excl. aorta), which make only 6% of the total image content and thus are unlikely to disturb a registration based on the entire image.
Image:DBP3_Biasfield_masked-unmasked.png|Bias field correction results (N4ITK) with and without masking of the structures of interest. Masking causes a different correction not related to coil sensitivity but rather local image content. Masking is not recommended at this point.

</gallery>

=== Example Cases ===
* [[Media:AFib_Example_P1_Data.zip|Example Case P1 : pre-post N4 bias corregtion, registration, LA segmentation]]

=== Current Progress ===
As we proceed and make progress with the pipeline we have simplified it by consolidating many of the steps with minimal user input. Users are now prompted to provide a pre and post set of MRI data. Once provided, a button labeled 'MRI Intensity Correction' is enabled. When the user clicks the button, Slicer begins the pipeline by using N4 ITK MRI bias correction immediately followed by a whole-image affine registration. Once completed, the user is then prompted to crop the images to the desired ROI. After the ROI is defined a non-rigid registration is performed yielding registered pre and post MRI images and the corresponding transformation matrix. The following image gallery shows the progress and current state of the GUI.

<gallery perrow="2" widths="300px">
Image:PipelineStart.png|Registration Segmentation Pipeline input selection - raw images.
Image:PipelineStep1.png|N4 ITK MRI bias correction completed.
</gallery>

DBP3:Utah:RegSegPipeline

2011-08-29T19:20:17Z

Dillonl:

[[DBP3:Utah| back to DBP3 home]]
= The CARMA DBP: MRI-based study and treatment of atrial fibrillation =
== Pilot Studies on a Registration & Segmentation Pipeline & Workflow ==
Alex Zaitsev, Dominik Meier, Ron Kikinis

=== Background ===
By Rob MacLeod, PI AFib DBP

The Utah group carries out analysis of MRI scans from all patients coming to our clinical with symptoms of Atrial Fibrillation (AF) and continues to conduct regular MRI scans for patients once they undergo treatment by our physicians. The result is over 1000 MRI scans per year that we process using a set of heterogeneous tools. One goal of our collaboration with NAMIC is to explore new tools and capabilities to consolidate and streamline this processing pipeline. The goals of the process vary slightly depending on the timing of the imaging relative to patient care but all involve some combination of image correction, region of interest definition, segmentation of the left atrium (in some cases also right atrium and esophagus), and quantitative evaluation of enhancement in the images, due either to pre-existing fibrosis or post-treatment scar. In many cases, registration is also necessary to compare multiple images from the same patients over time and to compare multiple patients at the same stage of disease progression and treatment.

For more information, please see the following references:

R.S. Oakes, T.J. Badger, E.G. Kholmovski, N. Akoum, N.S. Burgon, E.N. Fish, J.J. Blauer, S.N. Rao, E.V. DiBella, N.M. Segerson, M. Daccarett, J. Windfelder, C.J. McGann, D. Parker, R.S. MacLeod, N.F. Marrouche. “Detection and quantification of left atrial structural remodeling with delayed-enhancement magnetic resonance imaging in patients with atrial fibrillation,” In Circulation, Vol. 119, No. 13, pp. 1758--1767. 2009. http://www.sci.utah.edu/publications/oaks09/Oaks_Circ2009.pdf

C.J. McGann, E.G. Kholmovski, R.S. Oakes, J.J. Blauer, M. Daccarett, N. Segerson, K.J. Airey, N. Akoum, E. Fish, T.J. Badger, E.V. DiBella, D. Parker, R.S. MacLeod, N.F. Marrouche. “New Magnetic Resonance Imaging-Based Method for Defining the Extent of Left Atrial Wall Injury After the Ablation of Atrial Fibrillation,” In Journal of the American College of Cardiology, Vol. 52, No. 15, pp. 1263--1271. Oct 7, 2008. http://www.sci.utah.edu/publications/mcgann08/mcg2008.pdf

C. Mahnkopf, T.J. Badger, N.S. Burgon, M. Daccarett, T.S. Haslam, C.T. Badger, C.J. McGann, N. Akoum, E. Kholmovski, R.S. Macleod, N.F. Marrouche. “Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheter ablation,” In Heart Rhythm, Vol. 7, No. 10, pp. 1475--1481. 2010. http://www.sci.utah.edu/publications/mahnkopf10/mahnkopf2010.pdf

We are in the process of making a small set of anonymized test data available for general use and more extensive deidentified data for exclusive use with NAMIC. Please contact me (macleod@sci.utah.edu) for more details.

=== Main processing pipeline ===
To facilitate the workflow, we can place all the automated steps at the beginning and cluster interactive elements at the end. Exception is the cropping step required as input for nonrigid registration.<br>
1. N4 bias field correction for the MRI (surface coils):
::*'''Input''': MRI_pre and MRI_post, each run separately with the same parameters below; '''Output:''' MRI_pre_n4 and MRI_post_n4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:N4ITKBiasFieldCorrection-Documentation-3.6 N4 ITK]; '''Parameters:''' convergence: 1e-5, iterations: 50,40,30,20, shrink factor: 3
::*''Comments:'' a run on entire image gives some benefit that '''may''' be improved with masking: again the dominant intensity dropoff from the surface coil occurs along the chest wall and ribcage. Even if that is not the structure of interest, it is the low-freq. variation the bias correction algorithm is searching for, and masking that out can be counter-productive: via masking we may end up with a smoother image, but the intensity variations removed were not caused by the coil but are actually true signal.
2. affine registration MRA ->MRI (both pre and post)
::*'''Input''': MRI_pre_n4 (fixed) and MRA_pre ; '''Output:''' Xf1_pre_MRA-MRI.tfm"
::*'''Input''': MRI_post_n4 (fixed) and MRA_post ; '''Output:''' Xf2_post_MRA-MRI.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' no initialization; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' from the example dataset (P1) we infer that the two scans tend to have little initial alignment, hence initialization steps are not recommended. We choose registration parameters such that the chance of the registration making the alignment worse is minimized. Decide in a subsequent review/QC step if we keep the transform or the original pose. The MRA contains the same FOV and has surrounding structures (liver, chest, spine etc) visible also, despite lower intensities. A global affine is thus not necessarily going to benefit from masking the heart, unless the relative motion of the heart becomes the dominant reason for misalignment. We tried masking with both BrainsFit and RobustMultires modules. Both failed to provide better alignment with masking.
3. registration post -> pre: phase 1: AFFINE
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' Xf3_pre-post_affine.tfm
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit]
::*'''Parameters:''' DOF: rigid+similarity+affine (possibly rigid only); initialization: Moments align ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' done on the entire image. The surrounding structures are useful in constraining the solution transform and should provide a more robust behavior. Cropping down to the cardiac only ROI is deferred to the nonrigid registration in phase 2. The inferior-superior FOV can differ, e.g. how much of the liver is included. If the two exams differ significantly (>30%) in that content, the above affine could fail and a prior cropping step is then suggested to better match image content before registration. Resampling to isotropic voxel size at this stage is also advantageous but will generate very large files > 100MB due to the full FOV.
4. Cropping to VOI:
::*'''Input''': MRI_pre_n4 (fixed) and MRI_post_n4 (moving) ; '''Output:''' MRI_pre_n4_cropped, MRI_post_n4_cropped
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 Crop Volume]
::*''Comments:'' the [http://www.slicer.org/slicerWiki/index.php/Modules:VolumeRendering-Documentation-3.6 volume rendering module] may help in obtaining a good cropping ROI. Because of the high contrast, the MRA provides a good source for volume rendering.
5. registration post -> pre: phase 2: nonrigid / BSPLINE
::*'''Input''': MRI_pre_n4_cropped (fixed) and MRI_post_n4_cropped (moving) ; '''Output:''' Xf4_pre-post_BSpline.tfm"
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*'''Parameters:''' DOF: BSpline only; initialization: Xf3_pre-post_affine.tfm ; samples: 200k, convergence: 1e-5, iterations:
::*''Comments:'' crop to volume of interest: [http://www.slicer.org/slicerWiki/index.php/Modules:CropVolume-Documentation-3.6 CropVolume module], use ''Resample to isotropic voxelsize'' option.
6. resample MRA: apply above BSpline to the pre MRA
::*'''Input Volume''': MRA_post_n4_cropped
::*'''Input Transform''': above Xf4_pre-post_BSpline.tfm; '''Output Volume''': MRA_post_n4_cropped_Xf4
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:ResampleScalarVectorDWIVolume-Documentation-3.6 ResampleScalarVectorDWIVolume]
::*'''Parameters:''' output-to-input box checked, interpolation: linear
7. ROI definition (manual box ROI or automated via atlas)
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:BRAINSFit BRAINSfit] or [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
8. segmentation of LA from MRA -> inner wall
::*'''Module used:''' [http://www.slicer.org/slicerWiki/index.php/Modules:RobustStatisticsSeg-Documentation-3.6 Robust Statistics module] or [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding] for thresholding within [http://www.slicer.org/slicerWiki/index.php/Modules:CMTK CMTK]
::*''Comments:'' as a dynamic image the MRA contains significant spread and likely requires interactive segmentation/thresholding to yield a satisfactory LA volume. For validation/visualization, use the :[http://www.slicer.org/slicerWiki/index.php/Modules:Volumes-Documentation-3.6 Volumes] thresholding option within ''Display'' tab, use ''iron'' colormap & low alpha setting to check for ventricular wall borders. (see figures below)
9. LA wall segmentation
::*'''Module used:''' ''Atrium Cardiac Wall Segmentation'' (Extension Module, no documentation yet, authors: Yi Gao, Behnood Gholami, Allen Tannenbaum)
10. segmentation of enhancement
::*'''Module:''' [http://www.slicer.org/slicerWiki/index.php/Modules:Editor-Documentation-3.6 Editor: thresholding]
::*'''alternative Module:''' [http://wiki.na-mic.org/Wiki/index.php/Projects:AblationScarSegmentation AblationScarSegmentation]
::*''Comments:'; operate only on ROI within LA wall. Based on proper intensity statistics. An atlas-based set of intensity distributions may be more meaningful here than a simple Otsu, because both amount and location of enhancement is unknown and can in theory be 0.

<gallery perrow="4" widths="200px">
Image:DBP3_AFib_LV_overlay_1.jpg| Example contrast MRI with thresholded MRA as color overlay. Registration validation is not trivial: we can in theory use the position of the blood in the MRA relative to the LV wall as registration check, but the MRA itself does not provide a clear boundary at the lower intensities.
Image:DBP3_AFib_LA_overlay_2.jpg|Example contrast MRI with thresholded MRA as color overlay and areas of enhancement marked
Image:DBP3_histogramMRA.png|Histogram of MRA intensities: green = entire FOV, red: cardiac structures only (excl. aorta), which make only 6% of the total image content and thus are unlikely to disturb a registration based on the entire image.
Image:DBP3_Biasfield_masked-unmasked.png|Bias field correction results (N4ITK) with and without masking of the structures of interest. Masking causes a different correction not related to coil sensitivity but rather local image content. Masking is not recommended at this point.

</gallery>

=== Example Cases ===
* [[Media:AFib_Example_P1_Data.zip|Example Case P1 : pre-post N4 bias corregtion, registration, LA segmentation]]

=== Current Progress ===
As we proceed and make progress with the pipeline we have simplified it by consolidating many of the steps with minimal user input. Users are now prompted to provide a pre and post set of MRI data. Once provided, a button labeled 'MRI Intensity Correction' is enabled. When the user clicks the button, Slicer begins the pipeline by using N4 ITK MRI bias correction immediately followed by a whole-image affine registration. Once completed, the user is then prompted to crop the images to the desired ROI. After the ROI is defined a non-rigid registration is performed yielding registered pre and post MRI images and the corresponding transformation matrix. The following image gallery shows the progress and current state of the GUI.

<gallery perrow="4" widths="200px">
Image:PipelineStart.png|Registration Segmentation Pipeline input selection.
Image:PipelineStage1.png|N4 ITK MRI bias correction completed.
</gallery>

File:PipelineStep1.png

2011-08-29T19:16:59Z

Dillonl:

File:PipelineStart.png

2011-08-29T19:16:17Z

Dillonl:

DBP3:Utah:RegSegPipeline

2011-08-29T19:15:57Z

Dillonl: /* Current Progress */

DBP3:Utah:RegSegPipeline

2011-08-29T19:13:28Z

Dillonl:

File:01 3mo s5 ManualIso 2.png

2011-05-25T22:17:01Z

Dillonl: uploaded a new version of "File:01 3mo s5 ManualIso 2.png"

File:01 3mo s5 ManualIso.png

2011-05-25T22:16:48Z

Dillonl: uploaded a new version of "File:01 3mo s5 ManualIso.png"

File:01 3mo s5 AutoIso 2.png

2011-05-25T22:16:31Z

Dillonl: uploaded a new version of "File:01 3mo s5 AutoIso 2.png"

File:01 3mo s5 AutoIso.png

2011-05-25T22:16:04Z

Dillonl: uploaded a new version of "File:01 3mo s5 AutoIso.png"

File:01 3mo s5 Auto Endo Manual slice.png

2011-05-25T18:01:43Z

Dillonl: uploaded a new version of "File:01 3mo s5 Auto Endo Manual slice.png"

File:01 3mo s4 Auto Endo Manual slice.png

2011-05-25T18:00:12Z

Dillonl: uploaded a new version of "File:01 3mo s4 Auto Endo Manual slice.png"

File:01 3mo s1 Auto Endo Manual slice.png

2011-05-25T17:59:36Z

Dillonl: uploaded a new version of "File:01 3mo s1 Auto Endo Manual slice.png"

File:01 3mo s1 Auto Endo 4PixDilate slice.png

2011-05-25T17:57:55Z

Dillonl: uploaded a new version of "File:01 3mo s1 Auto Endo 4PixDilate slice.png"

DBP3:Utah:AutoWallSeg

2011-04-22T00:53:33Z

Dillonl: /* Automatic Segmentation */

=Automatic Segmentation=

==Automatic Wall Segmentation from GA Tech==

GA Tech produced a slicer module to automatically segment the left atrial wall, given the original data AND the endo (blood pool) segmentation. Below is some evaluation of those segmentations.

===Visual===

(click on thumbnails to view full size)

====Slice====

{| class="wikitable"
|-
! Auto, Endo, and Manaul Wall (01 3mo s5)
! Auto, Endo, and Manual Wall (01 3mo s4)
! Auto, Endo, and Manual Wall (01 3mo s1)
! Auto, Endo, and 4 Pixel Dilation from Endo (01 3mo s1)
|-
| [[File:01 3mo s5 Auto Endo Manual slice.png|thumb]]
| [[File:01 3mo s4 Auto Endo Manual slice.png|thumb]]
| [[File:01 3mo s1 Auto Endo Manual slice.png|thumb]]
| [[File:01 3mo s1 Auto Endo 4PixDilate slice.png|thumb]]
|-
| This image shows how the automatic wall segmentation creates a complete surface around the endo - while a manual segmentor would cut off the wall around the veins (although the location this cutoff decision is arbitrary).
| This image shows where the automatic segmentation surrounds a vein, while the manual segmentor decided to not surround the vein.
| This image shows some island artifacts that may be the result of the process being done in 3D (leaking from the slice above) - note that the data below the islands is lighter.
| Here we show a 4 pixel dilation from the endo compared to the automatic segmentation. This also image shows more islands.
|-
|}

====Isosurface====

Note how the automatic segmentation creates a smooth/complete (mostly) surface on the top side of the atrium, while the bottom has gaps similar to the manual.

{| class="wikitable"
|-
! Auto Wall Segmenation Isosurface (01 3mo s5)
! Manual Wall Segmentation Isosurface (01 3mo s5)
|-
| [[File:01 3mo s5 AutoIso.png|thumb]]
| [[File:01 3mo s5 ManualIso.png|thumb]]
|-
| [[File:01 3mo s5 AutoIso 2.png|thumb]]
| [[File:01 3mo s5 ManualIso 2.png|thumb]]
|-
|}

===Comparison Statistics===

====Inter Auto Seg Comparison====

Before getting the automatic wall segmentation code we had done a small study on our manual segmentations (see [[LA Segmentation reproducibility - intra and inter observer]]), this involved 64 total Endo + Wall Segmentations (4 different segmentors segmenting the same scan, for 16 scans).

To compare them we used the STAPLE algorithm to create a "ground truth" and then calculated overlap and accuracy, among other things, according to that ground truth.

We did the same analysis, but on the automatic wall segmentations (each derived from the different manual endo segmentations), and compared it to the original manual segmentation data.

Below are boxplots of the two data sets, for both overlap and accuracy.

For two sets S1 and S2, overlap is computed: (2 * ||S1 ^ S2|| ) / ( ||S1|| + ||S2|| )

For two sets S1 and S2, accuracy is computed: (||S1 ^ S2|| + ||!S1 ^ !S2|| ) / (||S1|| + ||S2||)

(For brevity: ^ is intersection, U is union, and !S is the complement of S)

[[File:AutoSegOverlap.png]]

[[File:AutoSegAccuracy.png]]

As the graphs show the automatic wall segmentations have much better overlap and accuracy with their STAPLE produced ground truths.

However the most these results seem to say is that the automatic algorithm is more consistent than manual segmentations. We hope to run a new comparison to be able to really evaluate the performance of the automatic segmentation.

A side note is that the manual segmentations are actually very consistent when compared to the automatic segmentations - one would expect the gap to be larger.

Original data: [[File:Autoseg.xlsx]], [[File:Manual Compare Results.xlsx]]

====Auto to Manual Overlap Comparison====

In this comparison we took each automatic segmentation produced using a manual endo segmentation and compared it with the corresponding manual wall segmentation.

The box plot below is for all 64 automatic-manual segmentation pairs.

[[File:AugoSeg Overlap pairs.png]]

Original data: [[File:Autoseg pairs.xlsx]]

DBP3:Utah:AutoWallSeg

2011-04-22T00:52:25Z

Dillonl: /* Automatic Segmentation */

=Automatic Segmentation=

==Automatic Wall Segmentation from GA Tech==

GA Tech produced a slicer module to automatically segment the left atrial wall, given the original data AND the endo (blood pool) segmentation. Below is some evaluation of those segmentations.

{| class="wikitable"
|-
! Auto, Endo, and Manaul Wall (01 3mo s5)
! Auto, Endo, and Manual Wall (01 3mo s4)
! Auto, Endo, and Manual Wall (01 3mo s1)
! Auto, Endo, and 4 Pixel Dilation from Endo (01 3mo s1)
|-
| [[File:01 3mo s5 Auto Endo Manual slice.png|thumb]]
| [[File:01 3mo s4 Auto Endo Manual slice.png|thumb]]
| [[File:01 3mo s1 Auto Endo Manual slice.png|thumb]]
| [[File:01 3mo s1 Auto Endo 4PixDilate slice.png|thumb]]
|-
| This image shows how the automatic wall segmentation creates a complete surface around the endo - while a manual segmentor would cut off the wall around the veins (although the location this cutoff decision is arbitrary).
| This image shows where the automatic segmentation surrounds a vein, while the manual segmentor decided to not surround the vein.
| This image shows some island artifacts that may be the result of the process being done in 3D (leaking from the slice above) - note that the data below the islands is lighter.
| Here we show a 4 pixel dilation from the endo compared to the automatic segmentation. This also image shows more islands.
|-
|}

====Isosurface====

Note how the automatic segmentation creates a smooth/complete (mostly) surface on the top side of the atrium, while the bottom has gaps similar to the manual.

{| class="wikitable"
|-
! Auto Wall Segmenation Isosurface (01 3mo s5)
! Manual Wall Segmentation Isosurface (01 3mo s5)
|-
| [[File:01 3mo s5 AutoIso.png|thumb]]
| [[File:01 3mo s5 ManualIso.png|thumb]]
|-
| [[File:01 3mo s5 AutoIso 2.png|thumb]]
| [[File:01 3mo s5 ManualIso 2.png|thumb]]
|-
|}

===Comparison Statistics===

====Inter Auto Seg Comparison====

Before getting the automatic wall segmentation code we had done a small study on our manual segmentations (see [[LA Segmentation reproducibility - intra and inter observer]]), this involved 64 total Endo + Wall Segmentations (4 different segmentors segmenting the same scan, for 16 scans).

To compare them we used the STAPLE algorithm to create a "ground truth" and then calculated overlap and accuracy, among other things, according to that ground truth.

We did the same analysis, but on the automatic wall segmentations (each derived from the different manual endo segmentations), and compared it to the original manual segmentation data.

Below are boxplots of the two data sets, for both overlap and accuracy.

For two sets S1 and S2, overlap is computed: (2 * ||S1 ^ S2|| ) / ( ||S1|| + ||S2|| )

For two sets S1 and S2, accuracy is computed: (||S1 ^ S2|| + ||!S1 ^ !S2|| ) / (||S1|| + ||S2||)

(For brevity: ^ is intersection, U is union, and !S is the complement of S)

[[File:AutoSegOverlap.png]]

[[File:AutoSegAccuracy.png]]

As the graphs show the automatic wall segmentations have much better overlap and accuracy with their STAPLE produced ground truths.

However the most these results seem to say is that the automatic algorithm is more consistent than manual segmentations. We hope to run a new comparison to be able to really evaluate the performance of the automatic segmentation.

A side note is that the manual segmentations are actually very consistent when compared to the automatic segmentations - one would expect the gap to be larger.

Original data: [[File:Autoseg.xlsx]], [[File:Manual Compare Results.xlsx]]

====Auto to Manual Overlap Comparison====

In this comparison we took each automatic segmentation produced using a manual endo segmentation and compared it with the corresponding manual wall segmentation.

The box plot below is for all 64 automatic-manual segmentation pairs.

[[File:AugoSeg Overlap pairs.png]]

Original data: [[File:Autoseg pairs.xlsx]]

File:Autoseg pairs.xlsx

2011-04-22T00:51:57Z

Dillonl:

File:Manual Compare Results.xlsx

2011-04-22T00:51:15Z

Dillonl:

File:Autoseg.xlsx

2011-04-22T00:50:54Z

Dillonl:

File:AugoSeg Overlap pairs.png

2011-04-22T00:50:12Z

Dillonl:

File:AutoSegAccuracy.png

2011-04-22T00:49:42Z

Dillonl:

File:AutoSegOverlap.png

2011-04-22T00:49:15Z

Dillonl:

File:01 3mo s4 Auto Endo Manual slice.png

2011-04-22T00:47:56Z

Dillonl: uploaded a new version of "File:01 3mo s4 Auto Endo Manual slice.png"

File:01 3mo s5 ManualIso 2.png

2011-04-22T00:47:20Z

Dillonl:

File:01 3mo s5 AutoIso 2.png

2011-04-22T00:47:02Z

Dillonl:

File:01 3mo s5 ManualIso.png

2011-04-22T00:46:37Z

Dillonl:

File:01 3mo s5 AutoIso.png

2011-04-22T00:46:13Z

Dillonl:

DBP3:Utah:AutoWallSeg

2011-04-22T00:21:37Z

Dillonl:

File:01 3mo s1 Auto Endo 4PixDilate slice.png

2011-04-21T23:56:09Z

Dillonl:

File:01 3mo s1 Auto Endo Manual slice.png

2011-04-21T23:55:27Z

Dillonl: uploaded a new version of "File:01 3mo s1 Auto Endo Manual slice.png"

File:01 3mo s1 Auto Endo Manual slice.png

2011-04-21T23:54:59Z

Dillonl:

File:01 3mo s4 Auto Endo Manual slice.png

2011-04-21T23:54:33Z

Dillonl:

File:01 3mo s5 Auto Endo Manual slice.png

2011-04-21T23:52:51Z

Dillonl:

DBP3:Utah

2011-04-21T23:48:36Z

Dillonl: /* Activites In Progress */

=== The CARMA DBP: MRI-based study and treatment of atrial fibrillation ===
Tissue remodeling of the atrial wall is the hallmark of Atrial Fibrillation
(AF), a progressive cardiac disease that develops over time (months to
years). The mechanisms of of this transformation are only partially understood,
but the current scientific focus on tissue remodeling and its putative role in
AF suggests an MRI image-based approach to the study of AF.

The Comprehensive Arrhythmia Research and MAnagement (CARMA) Center at the
University of Utah is a world leader in the rapidly emerging field of
MRI-managed evaluation and ablation of AF. Other groups have begun to recognize
the potential of this approach and to investigate and validate some of our
findings. Still others are developing new refinements of the MRI technique
driven by the specific needs of this application domain.

Preliminary investigation by CARMA has identified image processing and analysis
as the rate-limiting step to the development of MRI-based therapies. Novel
forms of MRI can be used to evaluate new patients, predict success before
ablation, analyze outcomes post-ablation, and guide repeat ablations. Such
MRI-based therapies, however, urgently require advanced tools and software to
support efficient workflows and accelerate the quantification and analysis of
images. The CARMA-NAMIC DBP project will address some of these needs with the
development of algorithms and tools for the automated segmentation of heart
structures and the MRI-based evaluation of AF progression and recovery.

This remainder of this wiki page describes ongoing CARMA-NAMIC activities and
specific plans. For more background information, references, and description of
the specific scientific and clinical aims of this project,
please see the project web page: http://www.na-mic.org/pages/DBP:Atrial_Fibrillation .

=== DBP Specific Aims ===
* Develop and validate image-based longitudinal diagnostic indices for AF
* Develop automatic segmentation methods for the atrium and adjacent structures
* Develop an AF scoring scheme to evaluate disease progression and recovery from therapy

*

[[Image:TV1.jpg|400px|The CARMA Center's Utah classification for Atrial Fibrillation staging involves segmentation of the left atrial wall from MRI, followed by quantification of enhanced vs. non-enhanced voxels in the wall.]]

(IMAGE ABOVE) The CARMA Center's Utah classification for Atrial Fibrillation staging involves segmentation of the left atrial wall from MRI, followed by quantification of enhanced vs. non-enhanced voxels in the wall.

=== Activites In Progress ===
* Begin technical collaboration with GA Tech
** Construct test dataset for GA Tech -- identify, anonymize, transmit
** Define NAMIC project week activities and goals
** [[DBP3:Utah:AutoWallSeg|Automatic Wall Segmentation]]
* CARMA team learning/testing of Slicer tools
** Image registration capabilities for longitudinal analysis
*[[DBP3:Utah:RegSegPipeline|Registration & Segmentation pipeline pilots & workflow development (Dominik Meier, Alex Zaitsev)]]

=== Completed Activities ===

=== Engineering Wishlist ===
* Some extend support for DICOM I/O (ITK V.4 may take care of this)
* Tools for deformable image registration (example applications exist in both clinical image management and in experiments) .)
* Tools for image registration, in general, is a big need that we have. Again, examples exist for both clinical and experimental data.
* GUI support for "steerable" segmentation methods like level-set methods, active contours, etc. (e.g. GA Tech's work on left atrium segmentation).
** Workflow
** Interaction widgets
*** Placing fudicials, guides, constraints, walls, sources, sinks, etc.
* "Show me a histogram, let me pick a range from the histogram" and provide means for histogram processing to suggest thresholds.

===Sample Data===
*Two Subjects Two Time Points: Data is [[File:CardiacMRIs.zip|here]]

DBP3:Utah:AutoWallSeg

2011-04-21T23:47:55Z

Dillonl: Created page with '=Automatic Segmentation= ==Automatic Wall Segmentation from GA Tech== GA Tech produced a slicer module to automatically segment the left atrial wall, given the original data AN…'