Documentation/4.4/Modules/IntensityDifferenceMetric

From Slicer Wiki
Jump to: navigation, search
Home < Documentation < 4.4 < Modules < IntensityDifferenceMetric


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



Introduction and Acknowledgements

This work is supported by NA-MIC, NAC, NCIGT, and the Slicer Community. This work is partially supported by Brain Science Foundation and NIH U01 CA151261.
Author: Andrey Fedorov, Kilian Pohl, Peter Black, Ron Kikinis, SPL
Contact: Andrey Fedorov <email>fedorov@bwh.harvard.edu</email>

NA-MIC  
Brain Science Foundation  
NCIGT  

Module Description

Intensity Difference Metric can be used to quantify the differences between the two images. The assumptions made by this metric are that:

  • the images are aligned and the major source of difference is due to the anatomical changes of the structures
  • the changes are small
  • the structure of interest is hyperintensive in the baseline volume, and its segmentation is available
  • increase in signal intensity between the two analyzed scans corresponds to "growth"
Left to right: (1) baseline volume with the outline of the structure of interest; (2) followup volume co-registered with the baseline, with the same outline; (3) baseline volume with the voxels categorized as growth highlighted in red.


Use Cases

Most frequently used for these scenarios:

  • This metric is used in the last step of processing by ChangeTracker.
  • The metric can be used on its own, as long as the assumptions listed above are considered.

Tutorials

See ChangeTracker documentation.

Panels and their use

The logic used in this module is the following. It is assumed that the changes are happening near the boundary of the object, which should be defined by a segmentation in the baseline volume. Therefore, it is also assumed that the difference in intensity for the voxels inside the object are due to noise.

The processing steps are first to estimate the noise level using the signal difference in the inner portion of the object, and then categorize the voxels near the boundary as "growth" or "shrinkage" based on whether they are located inside or outside the voxel, and whether signal increases or decreases in the followup volume compared to baseline.

For a more formal description of the method take a look at the papers in the References section.

  • Parameters
    • Sensitivity affects how much intensity difference should be observed at a given voxel to be considered as "changing". Lower sensitivity will lead to larger number of voxels marked as "changing".
    • Changing band size defines the margin around the segmentation boundary where intensity differences will be considered and categorized as "growth" or "shrinkage". Larger number will lead to more voxels considered and generally larger number of changing voxels reported.
  • IO
    • Baseline volume
    • Baseline segmentation volume
    • Followup volume
    • Output volume: will contain the result of change quantification. Voxels marked with red color correspond to "growth", and green correspond to "shrinkage".
  • Report
    • Report file name is an optional parameter. If specified, it will contain formatted report that summarizes the changes contained in the output volume. This file is used by ChangeTracker to report the results.
IntensityDifferenceMetric parameters panel


Similar Modules

References

  • Konukoglu, E., Wells, W. M., Novellas, S., Ayache, N., Kikinis, R., Black, P. M., & Pohl, K. M. (2008). Monitoring slowly evolving tumors. 2008 5th IEEE International Symposium on Biomedical Imaging: From Nano to Macro (pp. 812-815). IEEE. doi:10.1109/ISBI.2008.4541120 URL
  • Pohl, K. M., Konukoglu, E., Novellas, S., Ayache, N., Fedorov, A., Talos, I.-F., Golby, A., et al. (2011). A new metric for detecting change in slowly evolving brain tumors: validation in meningioma patients. Neurosurgery, 68(1 Suppl Operative), 225-33. doi:10.1227/NEU.0b013e31820783d5 URL

Information for Developers