Difference between revisions of "Documentation/Nightly/Modules/dPetBrainQuantification"

From Slicer Wiki
Jump to: navigation, search
Line 9: Line 9:
 
{{documentation/{{documentation/version}}/module-introduction-row}}
 
{{documentation/{{documentation/version}}/module-introduction-row}}
 
<br>
 
<br>
 +
This module allows the quantization of dynamic PET (dPET) brain scans. It is optimized to work with FDG tracer.
 +
 +
It is possible to segment relevant blood pools, derive pTAC curves with methods based on Chen [1], and perform voxel or region based Patlak analysis [2]
 +
 
Author: Martín Bertran and Natalia Martínez (Facultad de Ingeniería, Udelar, Uruguay)<br>
 
Author: Martín Bertran and Natalia Martínez (Facultad de Ingeniería, Udelar, Uruguay)<br>
 
Contributor: Guillermo Carbajal, Álvaro Gómez (Facultad de Ingeniería, Udelar, Uruguay)<br>
 
Contributor: Guillermo Carbajal, Álvaro Gómez (Facultad de Ingeniería, Udelar, Uruguay)<br>
Line 15: Line 19:
  
 
<!-- ---------------------------- -->
 
<!-- ---------------------------- -->
{{documentation/{{documentation/version}}/module-section|Module Description}}
+
{{documentation/{{documentation/version}}/module-section|Module description}}
The purpose of this module is to provide an interface for analysis and quantification of Brain dynamic PET studies. In it simplest form, it takes either a 4D DICOM, or a (Nifiti + Sif) study and performs the Patlak analysis. K-Map estimation is computed using an automated pTAC extraction algorithm. The user can optionally select the region of interest over which the Patlak estimation will take place, and supply its own estimated or extracted pTAC using a .csv input. Ptac curves can be automatically extracted using either IDIF (Image Derived Input Functions) techniques or Hunter PBIF (Population Based Input Functions) techniques. Supplying blood samples is optional for IDIF techniques. For IDIF methods, the required carotid segmentation can be done either automatically or user assisted by providing a ROI.
+
 
 +
This module provides a tool for processing dynamic PET brain scans with FDG tracer into Patlak-Gedde [2] parametric K-Maps. It attempts to minimize the required user input, as well as additional information other than the dPET scan.
 +
 
 +
Processing is divided into 3 blocks:
 +
 
 +
*'''Vascular segmentation:''' Allows the segmentation of vasculature and adjacent tissue from the dPET scan (Necessary for implemented IDIF pTAC extraction methods). Segmentation can either be done automatically or through a user defined ROI.
 +
*'''pTAC estimation:'''  Estimates pTAC from the dPET scan by attempting to correct for Partial Volume Effects. Estimation can be done without additional information, but the module allows the incorporation of additional late blood samples. Additionally, the Hunter [3] PBIF method is also implemented.
 +
*'''K-Map Estimation:''' The module allows the creation of a scalar volume with the parametric K-Map. Estimation can be done voxel by voxel or over user defined regions.
 +
 
 +
The most usual “end to end” workflow involves loading the dPET brain scan, and obtaining a K-map volume. However, the processing blocks can be used independently, each allowing external inputs, the relevant output variables can also be saved and stored.
 +
 
 +
 
 +
Figure 1 shows a basic illustration of the processing blocks.
 +
 
 +
[[File:Fig1.png|800px|Illustration of the dPET processing blocks]]
 +
 
 +
Illustration of the dPET processing blocks
 +
 
 +
<!-- ---------------------------- -->
 +
{{documentation/{{documentation/version}}/module-section|Panel 1 : PET dynamic study and information.}}
 +
 
 +
This panel allows the loading of the dPET scan, and optional blood samples.
 +
Figure 2 shows a screencap of the available options in this panel.
 +
 
 +
[[File:Fig2.png|800px|Figure 2. Screencap: PET dynamic study and information panel]]
 +
 
 +
Screencap: PET dynamic study and information panel
 +
 
 +
*'''Import DICOM Study:'''Provides an alternative one button access to the DICOM browser.
 +
*'''Nifti + SIF input directory:''' The dPET study can also be imported from a 3D Nifti stack and a SIF frametime definition file. This encapsulates a call to the multivolume importer module.
 +
*'''Input multivolume:''' The multivolume node to be processed must be selected from the 3D SLICER scene.
 +
*'''Import acquired blood samples:''' Optional input that allows the loading of aquired blood samples from a CSV file.  The format is as follows:
 +
time,value,time,value,...
 +
Time is given in seconds, value in Bq/ml.
 +
 
 +
<!-- ---------------------------- -->
 +
{{documentation/{{documentation/version}}/module-section|Panel 2: Visualization options.}}
 +
 
 +
 
 +
A screencap of the panel can be seen in figure 3.
 +
 
 +
This panel has the following options.
 +
 
 +
[[File:Fig3.png|800px|Figure 3. Visualization panel screencap. The segmented brain mask is displayed over the 14th frame of the imported dPET study.]]
 +
 
 +
Visualization panel screencap. The segmented brain mask is displayed over the 14th frame of the imported dPET study.
 +
 
 +
*'''Display frame selector:'''Slidebar that allows to change the displayed frame from the multivolume in the background.
 +
*'''Display Brain Mask:''' Renders the automatic Brain Mask segmentation in the foreground. This is the default region used for voxel based Patlak analysis.
 +
 
 +
<!-- ---------------------------- -->
 +
{{documentation/{{documentation/version}}/module-section|Panel 3 : Carotid Segmentation Options.}}
 +
 
 +
This panel presents the available segmentation options.
 +
Two segmentation methods are available:
 +
*Automatic segmentation (See figure 4).
 +
*ROI assisted segmentation(See figure 5).
 +
 
 +
[[File:Fig4.png|800px|Figure 4. Screencap of the Carotid segmentation options panel. The automatically segmented vascular regions are displayed on the viewer window.]]
 +
 
 +
Screencap of the Carotid segmentation options panel. The automatically segmented vascular regions are displayed on the viewer window.
 +
 
 +
[[File:Fig5.png|800px|Figure 5. Screencap of the Carotid segmentation options panel. The ROI assisted segmented vascular regions and the input ROI are displayed on the viewer window.]]
 +
 
 +
Screencap of the Carotid segmentation options panel. The ROI assisted segmented vascular regions and the input ROI are displayed on the viewer window.
 +
 
 +
 
 +
*'''Type :Automatic Carotid segmentation:''' Automatically segments vasculature within the image.
 +
 
 +
*'''Type :Segmentation with manual ROI over Carotids:''' Requires a user input ROI node over the carotid region. This node must be set as input in the '''Input ROI''' option.
 +
If possible, the ROI should contain approximately equal ammounts of tissue and vasculature.
 +
 
 +
*'''Choose a frame for better segmentation:''' Allows the user to choose the frame that shows the best contrast between vasculature and tissue. By default this option is disabled, it is of use for certain tracers.
 +
 
 +
*'''Get chart with tissue and carotid mean activity:''' Displays in scene a chart with the mean activity curves of segmented tissue and vasculature. Both TACs remain in scene, which allows for later exporting as a mcsv file.
 +
 
 +
*'''Output volume:''' Allows saving the segmented vasculature as a scalar volume node.
 +
 
 +
*'''Display Carotid Segmentation''' Executes the segmentation algorithm with the selected options.
 +
 
 +
<!-- ---------------------------- -->
 +
{{documentation/{{documentation/version}}/module-section|Panel 4:pTAC estimation Options.}}
 +
 
 +
This panel presents the available pTAC estimation options.
 +
 
 +
#'''Type: IDIF pTAC estimation.'''
 +
This options estimates the plasma time activity curve (pTAC) from the segmented vasculature and adjacent tissue.
 +
 
 +
By default, the initial estimator used is the 5% hottest voxels.
 +
 
 +
Available options are shown in figure 6, and are as follows:
 +
 
 +
[[File:Fig6.png|800px|Figure 6. Screencap of pTAC estimation options using IDIF. The resulting estimators are shown in the viewer window.]]
 +
 
 +
Screencap of pTAC estimation options using IDIF. The resulting estimators are shown in the viewer window.
 +
 
 +
 
 +
*'''Use previous carotid segmentation:'''
 +
This checkbox allows the previous carotid segmentation to be used for pTAC estimation. Else, a new segmentation will be run using the currently selected segmentation options.
 +
 
 +
*'''Use venous blood samples:'''This checkbox allows the use of imported venous samples for pTAC estimation.
 +
 
 +
*'''Fit with Hunter tail (FDG) when one or no sample is provided:'''
 +
Allows the use of a Hunter based initial estimator for pTAC extraction.
 +
 +
*'''Get chart with estimated pTAC:''' Displays in scene a chart with the estimated pTAC. The pTAC remains in scene, which allows for later exporting as a mcsv file.
 +
 
 +
 
 +
 
 +
 
 +
 
 +
 
 +
#'''Type: PBIF Hunter pTAC estimation with venous samples.'''
  
<!-- ---------------------------- -->
+
This estimation is based on Hunter et al [3] for FDG based on populational data. It requires at least one user input late venous sample. Figure 7 shows a screencap of the options panel.
{{documentation/{{documentation/version}}/module-section|Use Cases}}
 
  
'''Use Case 1: Simplest usage'''
+
[[File:Fig7.png|800px|Figure 7. Screencap of the PBIF pTAC estimation panel. The obtained estimation is displayed in scene.]]
  
Import a PET dynamic study using either DICOM or Nifti + SIF, a .csv file containing venous blood samples is optional. Select "apply K map estimation" which returns a scalar volume containing the K parameters from patlak estimation
+
Screencap of the PBIF pTAC estimation panel. The obtained estimation is displayed in scene.
  
'''Use Case 2 : Patlack Estimation using Hunter (PBIF) pTAC estimation'''
+
This panel contains:
  
Import a dynamic PET study and a .csv file containing at least one venous blood sample. In the pTAC parameters panel choose "PBIF Hunter pTAC estimation with venous samples" Input the patients lean weight and inyected dosage. Optionally you can display the estimated pTAC by clicking the "get pTAC estimation button" Click "apply K map estimation" to get the parametric K map.
+
*'''Dosage injected in Mbq:''' Input injected tracer dosage in MBq.
  
'''Use Case 3 : Patlack Estimation using Chen based IDIF pTAC estimation'''
+
*'''Lean weight Kg:''' Input patient's lean weigth.
  
Import a dynamic PET study and optionally a .csv file containing at least one venous blood sample. In the pTAC estimation options panel select the type "IDIF pTAC estimation". If no blood sample file is loaded or the "Use venous blood sample" tick box is not checked the estimation will use the hot voxels of the carotid segmentation, more details below. You can display the estimated pTAC by clicking the "get pTAC estimation button" Click "apply K map estimation" to get the parametric K map.
+
*'''Get chart with estimated pTAC:''' Displays a graph in scene with the estimated pTAC. This pTAC remains in scene, which allows for later exporting as a mcsv file.
  
'''Use Case 4 : Patlack Estimation using pTAC input file'''
+
<!-- ---------------------------- -->
 +
{{documentation/{{documentation/version}}/module-section|Panel 5:K-Map Estimation options.}}
  
Import a dynamic PET study From the K-Map estimation options menu select "Load pTAC estimation from .csv" this file should contain the pTAC values evaluated at the endtime of each frame. If not, a warning will be issued and interpolation will be attempted Select Apply selected K-Map estimation
+
This panel presents the available options for a Gedde-Patlak analysis. A screencap of the panel can be seen in Figure 8
  
<!-- ---------------------------- -->
+
[[File:Fig8.png|800px|Figure 8. Screencap of the K-Map estimation panel. The estimated K-Map volume is overlaid over the dPET scan.]]
{{documentation/{{documentation/version}}/module-section|Tutorials}}
 
N/A
 
  
<!-- ---------------------------- -->
+
Screencap of the K-Map estimation panel. The estimated K-Map volume is overlaid over the dPET scan.
{{documentation/{{documentation/version}}/module-section|Panels and their use}}
 
  
* '''PET dynamic Study and Information:''' Input Parameters
+
*'''pTAC Input options:''' The Patlak analysis can be applied with either the last estimated pTAC or a .csv input external pTAC.
** '''Import DICOM Study:''' Alternative access to the DICOM browser. Used to import DICOM studies into the slicer scene.
 
** '''Nifti + SIF input directory:''' Allows importing studies in a Nifti + SIF format. This option is recommended to avoid DICOM overflow errorrs.
 
** '''Input Multivolume:''' Select the input multivolume (Imported from either the DICOM browser or the Nifti + SIF input directory)
 
** '''Import acquired blood samples:''' Allows the importing of acquired samples through a .csv file. Optional, although some methods require it.
 
  
* '''Visualization Options:'''
+
*'''Region of interest Input options:''' Allows the user to either perform a voxel based Patlak estimation over the automatically generated brain mask ('''Automatic Mask Generation'''). Or to perform region based estimation over either a user defined ROI ('''Input ROI''') or a labelmap ('''Input Labelmap''')
**'''Display frame selector:''' Slider that displays the selected frame into the scene.
 
**'''Display Brain Mask :''' Shows the obtained Brain Mask. It separates brain from background, used for internal calculations.
 
  
* '''pTAC Estimation Options:''' The estimation of the K-map requires the plasma activity curve, currently, two options for its estimation are provided.
 
** '''Type: IDIF pTAC estimation :''' Image derived input function estimation using Chen's approach (ref C). It requires carotid segmentation, which is also provided.
 
*** '''Use previous carotid segmentation :''' Check if you want to use a previously acquired carotid segmentation map. Otherwise, segmentation will be done using the currently selected parameters in the "Carotid Segmentation Options" panel.
 
*** ''' Use venous blood samples :''' If a venous sample file has been loaded, it is possible to use it to aid in the estimation. Otherwise, a sample-less estimation will take place.
 
**'''Type: PBIF Hunter pTAC estimation with venous samples:''' The pTAC is extracted according to population based data curves taken from Hunter (ref H)
 
*** '''Dosage inyected in MBq :''' Patient inyected dosage in MBq
 
*** '''Lean Weight in Kg :''' Patient Lean weight in Kg
 
** '''Chart estimated pTAC :''' Displays estimated pTAC in a chart
 
** '''get pTAC estimation :''' Executes the pTAC estimation with currently selected options
 
** ''' Write estimated pTAC to .csv file :''' Allows the saving of the estimated pTAC into a .CSV file.
 
* '''Carotid Segmentation Options :''' Carotid segmentation is required for the IDIF pTAC estimation. Two methods are implemented, one is fully automatic and the other requires a user input ROI over the carotid region. Both estimation methods return a segmentation containing the detected blood vessels and its surrounding tissue
 
** '''Type : Automatic Carotid Segmentation:''' Default selection, the algorithm searchs for voxels that are both highly similar to a rough estimate of the expected pTAC and that posses sufficient signal intensity in the initial time frames. For more details about the methdo, see CAR
 
** '''Type : ROI Assisted Segmentaion : This option requires a user-input ROI over the carotid region'''
 
*** '''Input ROI :''' User input ROI
 
** '''Choose a frame for better segmentation:''' Forces the segmentation to use a particular frame for segmentation. Useful for tracers that reach rapid equilibrium like C-11 Flumazenil, where the automatic method may present issues.
 
** '''Apply Connectivity Filter:''' Returns the two largest classified clusters. For automatic segmentation, it automatically ignores the upper half of the brain. Usually not needed
 
  
 
<!-- ---------------------------- -->
 
<!-- ---------------------------- -->
Line 77: Line 168:
 
<!-- ---------------------------- -->
 
<!-- ---------------------------- -->
 
{{documentation/{{documentation/version}}/module-section|References}}
 
{{documentation/{{documentation/version}}/module-section|References}}
N/A
+
 
 +
[1]Chen K, Bandy D, Reiman E, Huang SC, Lawson M, Feng D,
 +
Yun LS, Palant A (1998) Noninvasive quantification of
 +
the cerebral metabolic rate for glucose using positron
 +
emission tomography, 18F-fluoro-2-deoxyglucose, the
 +
Patlak method, and an image-derived input function.
 +
J Cereb Blood Flow Metab 18:716–23
 +
 
 +
[2] Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab. 1983; 3: 1-7.
 +
 
 +
[3] Hunter, G. J., Hamberg, L. M., Alpert, N. M., Choi, N. C., & Fischman, A. J. (1996). Simplified measurement of deoxyglucose utilization rate. Journal of nuclear medicine: official publication, Society of Nuclear Medicine, 37(6), 950-955.
  
 
<!-- ---------------------------- -->
 
<!-- ---------------------------- -->

Revision as of 15:57, 12 February 2015

Home < Documentation < Nightly < Modules < dPetBrainQuantification


For the latest Slicer documentation, visit the read-the-docs.


Introduction and Acknowledgements


This module allows the quantization of dynamic PET (dPET) brain scans. It is optimized to work with FDG tracer.

It is possible to segment relevant blood pools, derive pTAC curves with methods based on Chen [1], and perform voxel or region based Patlak analysis [2]

Author: Martín Bertran and Natalia Martínez (Facultad de Ingeniería, Udelar, Uruguay)
Contributor: Guillermo Carbajal, Álvaro Gómez (Facultad de Ingeniería, Udelar, Uruguay)
Contact: Guillermo Carbajal, <email>carbajal@fing.edu.uy</email>

Module description

This module provides a tool for processing dynamic PET brain scans with FDG tracer into Patlak-Gedde [2] parametric K-Maps. It attempts to minimize the required user input, as well as additional information other than the dPET scan.

Processing is divided into 3 blocks:

  • Vascular segmentation: Allows the segmentation of vasculature and adjacent tissue from the dPET scan (Necessary for implemented IDIF pTAC extraction methods). Segmentation can either be done automatically or through a user defined ROI.
  • pTAC estimation: Estimates pTAC from the dPET scan by attempting to correct for Partial Volume Effects. Estimation can be done without additional information, but the module allows the incorporation of additional late blood samples. Additionally, the Hunter [3] PBIF method is also implemented.
  • K-Map Estimation: The module allows the creation of a scalar volume with the parametric K-Map. Estimation can be done voxel by voxel or over user defined regions.

The most usual “end to end” workflow involves loading the dPET brain scan, and obtaining a K-map volume. However, the processing blocks can be used independently, each allowing external inputs, the relevant output variables can also be saved and stored.


Figure 1 shows a basic illustration of the processing blocks.

Illustration of the dPET processing blocks

Illustration of the dPET processing blocks

Panel 1 : PET dynamic study and information.

This panel allows the loading of the dPET scan, and optional blood samples. Figure 2 shows a screencap of the available options in this panel.

Figure 2. Screencap: PET dynamic study and information panel

Screencap: PET dynamic study and information panel

  • Import DICOM Study:Provides an alternative one button access to the DICOM browser.
  • Nifti + SIF input directory: The dPET study can also be imported from a 3D Nifti stack and a SIF frametime definition file. This encapsulates a call to the multivolume importer module.
  • Input multivolume: The multivolume node to be processed must be selected from the 3D SLICER scene.
  • Import acquired blood samples: Optional input that allows the loading of aquired blood samples from a CSV file. The format is as follows:
time,value,time,value,...

Time is given in seconds, value in Bq/ml.

Panel 2: Visualization options.

A screencap of the panel can be seen in figure 3.

This panel has the following options.

Figure 3. Visualization panel screencap. The segmented brain mask is displayed over the 14th frame of the imported dPET study.

Visualization panel screencap. The segmented brain mask is displayed over the 14th frame of the imported dPET study.

  • Display frame selector:Slidebar that allows to change the displayed frame from the multivolume in the background.
  • Display Brain Mask: Renders the automatic Brain Mask segmentation in the foreground. This is the default region used for voxel based Patlak analysis.

Panel 3 : Carotid Segmentation Options.

This panel presents the available segmentation options. Two segmentation methods are available:

  • Automatic segmentation (See figure 4).
  • ROI assisted segmentation(See figure 5).

Figure 4. Screencap of the Carotid segmentation options panel. The automatically segmented vascular regions are displayed on the viewer window.

Screencap of the Carotid segmentation options panel. The automatically segmented vascular regions are displayed on the viewer window.

Figure 5. Screencap of the Carotid segmentation options panel. The ROI assisted segmented vascular regions and the input ROI are displayed on the viewer window.

Screencap of the Carotid segmentation options panel. The ROI assisted segmented vascular regions and the input ROI are displayed on the viewer window.


  • Type :Automatic Carotid segmentation: Automatically segments vasculature within the image.
  • Type :Segmentation with manual ROI over Carotids: Requires a user input ROI node over the carotid region. This node must be set as input in the Input ROI option.

If possible, the ROI should contain approximately equal ammounts of tissue and vasculature.

  • Choose a frame for better segmentation: Allows the user to choose the frame that shows the best contrast between vasculature and tissue. By default this option is disabled, it is of use for certain tracers.
  • Get chart with tissue and carotid mean activity: Displays in scene a chart with the mean activity curves of segmented tissue and vasculature. Both TACs remain in scene, which allows for later exporting as a mcsv file.
  • Output volume: Allows saving the segmented vasculature as a scalar volume node.
  • Display Carotid Segmentation Executes the segmentation algorithm with the selected options.

Panel 4:pTAC estimation Options.

This panel presents the available pTAC estimation options.

  1. Type: IDIF pTAC estimation.

This options estimates the plasma time activity curve (pTAC) from the segmented vasculature and adjacent tissue.

By default, the initial estimator used is the 5% hottest voxels.

Available options are shown in figure 6, and are as follows:

Figure 6. Screencap of pTAC estimation options using IDIF. The resulting estimators are shown in the viewer window.

Screencap of pTAC estimation options using IDIF. The resulting estimators are shown in the viewer window.


  • Use previous carotid segmentation:

This checkbox allows the previous carotid segmentation to be used for pTAC estimation. Else, a new segmentation will be run using the currently selected segmentation options.

  • Use venous blood samples:This checkbox allows the use of imported venous samples for pTAC estimation.
  • Fit with Hunter tail (FDG) when one or no sample is provided:

Allows the use of a Hunter based initial estimator for pTAC extraction.

  • Get chart with estimated pTAC: Displays in scene a chart with the estimated pTAC. The pTAC remains in scene, which allows for later exporting as a mcsv file.




  1. Type: PBIF Hunter pTAC estimation with venous samples.

This estimation is based on Hunter et al [3] for FDG based on populational data. It requires at least one user input late venous sample. Figure 7 shows a screencap of the options panel.

Figure 7. Screencap of the PBIF pTAC estimation panel. The obtained estimation is displayed in scene.

Screencap of the PBIF pTAC estimation panel. The obtained estimation is displayed in scene.

This panel contains:

  • Dosage injected in Mbq: Input injected tracer dosage in MBq.
  • Lean weight Kg: Input patient's lean weigth.
  • Get chart with estimated pTAC: Displays a graph in scene with the estimated pTAC. This pTAC remains in scene, which allows for later exporting as a mcsv file.

Panel 5:K-Map Estimation options.

This panel presents the available options for a Gedde-Patlak analysis. A screencap of the panel can be seen in Figure 8

Figure 8. Screencap of the K-Map estimation panel. The estimated K-Map volume is overlaid over the dPET scan.

Screencap of the K-Map estimation panel. The estimated K-Map volume is overlaid over the dPET scan.

  • pTAC Input options: The Patlak analysis can be applied with either the last estimated pTAC or a .csv input external pTAC.
  • Region of interest Input options: Allows the user to either perform a voxel based Patlak estimation over the automatically generated brain mask (Automatic Mask Generation). Or to perform region based estimation over either a user defined ROI (Input ROI) or a labelmap (Input Labelmap)


Similar Modules

N/A

References

[1]Chen K, Bandy D, Reiman E, Huang SC, Lawson M, Feng D, Yun LS, Palant A (1998) Noninvasive quantification of the cerebral metabolic rate for glucose using positron emission tomography, 18F-fluoro-2-deoxyglucose, the Patlak method, and an image-derived input function. J Cereb Blood Flow Metab 18:716–23

[2] Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab. 1983; 3: 1-7.

[3] Hunter, G. J., Hamberg, L. M., Alpert, N. M., Choi, N. C., & Fischman, A. J. (1996). Simplified measurement of deoxyglucose utilization rate. Journal of nuclear medicine: official publication, Society of Nuclear Medicine, 37(6), 950-955.

Information for Developers