Type:
Poster Presentation
Keywords:
Cardiac, Computer applications, MR, Image manipulation / Reconstruction, Diagnostic procedure, Computer Applications-3D, Segmentation, Inflammation
Authors:
O. Aparina, O. Stukalova, D. Parkhomenko; Moscow/RU
Results
Factors of poor quality LGE MR-images of atria:
|
Solutions:
|
Thin walls
|
High resolution scan – use of 3D acquisition; Special cardiac cycle ECG synchronization
|
Motion in respiratory cycle
|
Respiratory synchronization
|
Motion in cardiac cycle
|
Special cardiac cycle ECG synchronization
|
Fat
|
Fat saturation
|
Acquisition Time
|
Use fast imaging technologies
|
Factors of mistakes in atrial LGE MRI image analysis:
|
Solutions:
|
Image artifacts
|
Development new MR-pulse sequences for image aquisition
|
Thin walls
|
Performing manual,
semi-automated or automated atrial segmentation
Performing Automatic scar and fibrosis detection
|
Low size of wall scar and fibrotic lesions
|
Diffuse distribution of fibrosis
|
Low reproducibility of lesions detection
|
Different automated algorithms perform analysis of the black and white images of atrial wall to detect scar and\or fibrosis.
These programs work with the series atrial wall images after manual,
semi-automated or automated contouring.
Analysis of pixel intensity histogram,
pixel location and relations of pixel distribution can be analyzed.
The automated definition of pathological sites in atrial walls may base on different approaches:
- Experts evaluation and feature analysis:
- Intensity of pathological region
- Location of pathological region
- Shape & Size of pathological region
- Correlation with electro-anatomic map.
This approach defines the sites of fibrosis on 3D reconstruction of MR-image fitted to the sites of pathological electrical activity on 3D electro-anatomic map.
- Gaussian modeling and N standard deviation threshold.
Performs analysis of histogram of pixel intensity by fitting it to Gaussian distribution curves and/or Gaussian mix.
It helps to define the sites of fibrosis at a lower intensity threshold (3SD),
sites of scar at higher intensity threshold (5-6SD) or full width at half maximum.
The method is often used to perform the analysis of ventricle myocardium.
- Based on non-Gaussian modeling (Rayleigh,
Levi,
Caushi curves etc.).
Some researchers consider that histograms of pixel intensities may be better fitted to non-Gaussian curves and then analyzed another mathematic methods.
Clinical reasons to evaluate atrial scar/fibrosis.
Defining the structural remodeling in the future will give a new opportunities for decision making in management patients with AF.
Recent studies showed that 3D models of atrial fibrosis and scar is a helpful tool in planning invasive treatment techniques.
Quantitative assessment of preablation fibrosis and postablation scar predicts the results of interventional treatment and need for repeat procedures.
Use LGE MRI of mediastinum helps to assess the another organs damage after ablation procedure (esophagus,
heart fat pads with nerve ganglia).
In the future it may provide information for choosing strategy rhythm control or rate control taking in consideration degree of fibrosis in atrial walls.
It may be the method of choice to evaluate the structural remodeling progression and antifibrotic drugs effects.