Related Experiment Videos
Computer-aided intrapatient comparison of brain SPECT images: the gray-level normalization issue applied to children
Catherine Pérault1, Dimitri Papathanassiou, Hubert Wampach
1Nuclear Medicine and Biophysics Unit, Jean Godinot Institute, Reims, France. catherine.perault@reims.fnclcc.fr
Insights
Automated comparison of brain SPECT images requires careful gray-level normalization. Simple, robust scaling methods are recommended for ictal-interictal comparisons, avoiding image maximum as a reference.
Area of Science:
- Medical Imaging
- Nuclear Medicine
- Epilepsy Research
Background:
- Intrapatient comparison of brain SPECT images is crucial for diagnosing conditions like epilepsy.
- Automated tools can improve efficiency and consistency in image analysis.
- Gray-level normalization is a critical step that can significantly impact comparison results.
Purpose of the Study:
- To develop and evaluate an automated tool for intrapatient comparison of brain SPECT images.
- To assess the impact of different gray-level normalization methods on image comparison.
- To determine the most reliable normalization strategy for ictal-interictal SPECT image analysis.
Main Methods:
- Acquired ictal and interictal (99m)Tc-ethyl cysteinate dimer SPECT images from 6 children with partial epilepsy.
- Performed 3D rigid geometric ictal-to-interictal image registration using correlation coefficient and ratio uniformity criteria.
- Applied various 1- and 2-parameter linear gray-level normalization methods, including scaling based on image statistics and linear regression.
- Assessed registration validity and normalization plausibility using scatterplots, twin displays, and focused numeric comparisons.
Main Results:
- Satisfactory geometric registration was achieved for most patients.
- Most normalization methods yielded similar subtraction images for 5 out of 6 patients.
- The maxima ratio normalization method produced noticeably different and, in one case, incorrect results.
- Robust 1-parameter scaling methods generally performed better than 2-parameter methods for dissimilar images.
Conclusions:
- The choice of gray-level normalization method significantly impacts the clinical interpretation of intrapatient brain SPECT images.
- Simple, robust scaling methods are recommended for ictal-interictal SPECT image comparisons.
- Image maximum should not be used as a reference for normalization, and additive constants are generally not required.
Unlabelled:
A tool was developed for automated intrapatient comparison of brain SPECT images, with specific emphasis on gray-level normalization.
Methods:
Ictal and interictal (99m)Tc-ethyl cysteinate dimer SPECT images were acquired for 6 children with partial epilepsy (age range, 2-10 y). For each patient, 3-dimensional rigid geometric ictal-to-interictal image registration optimizing different classic criteria (correlation coefficient, ratio uniformity) in a multiscale translation-rotation 6-parameter space was first performed. Gray-level normalization was then performed with different methods, using a 1- or 2-parameter linear model. In the 1-parameter case, the scaling factor was equal to the interictal-to-ictal ratio of the maximum, mean, or median values calculated within different reference volumes (whole brain or cerebellum) or obtained by linear regression between ictal and interictal counts in the brain or by maximizing a robust criterion, the number of deterministic sign changes in the subtraction images. In the 2-parameter case, the scaling factor and additive constant were estimated using these last 2 methods. For each patient, registration validity and normalization plausibility were assessed by considering the correlation scatterplot together with the different normalization lines and by comparing interictal and registered normalized ictal images using a twin display (with isocontours) in the 3 orthogonal planes. Three-dimensional volumes of interest could be selected on coupled interictal-subtraction images for further focused numeric comparison.
Results:
After a satisfactory and stable geometric registration with both criteria, the different normalization methods led to similar subtraction images for 5 of 6 patients, except the maxima ratio, which gave noticeably different results in 2 patients. For the remaining patient, with highly dissimilar ictal-interictal images, the maxima ratio normalization was obviously wrong and the other 1-parameter methods probably better depicted the data than did the 2-parameter methods.
Conclusion:
When comparing intrapatient brain SPECT images, one should be aware of the potential impact of the gray-level normalization method on clinical interpretation. For ictal-interictal images, simple robust scaling should be recommended. In particular, image maximum should generally not be considered a valid reference, and no additive constant is needed in the linear gray-level normalization model.