Segmentation of elastographic images using a coarse-to-fine active contour model

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Image Processing

Background:

• Accurate delineation of radiofrequency-ablation-induced coagulation (thermal lesion) boundaries is crucial but challenging with conventional imaging.
• Elastography offers a novel method for visualizing thermal coagulations by imaging local strain after compression.

Purpose of the Study:

• To present an automated segmentation approach for thermal coagulations in 3D elastographic data.
• To enable rapid acquisition of both area and volume information for thermal lesions.

Main Methods:

• A coarse-to-fine active contour initialization method.
• A gradient vector flow active contour model for deformable contour optimization.
• Incorporation of prior knowledge regarding the geometry of thermal coagulations.

Main Results:

• The automated algorithm demonstrated performance comparable to manual delineation by medical physicists on elastograms (r = 0.99 for volumes).
• Results from the automated algorithm on elastograms showed strong agreement with manual delineation on pathology images (r = 0.96).
• Successful application of the algorithm on in vivo elastograms was achieved.

Conclusions:

• The developed automated segmentation approach provides rapid and accurate quantification of thermal lesion volumes and areas from 3D elastography.
• This method shows significant potential for clinical application in monitoring radiofrequency ablation treatments.
• The algorithm's accuracy is validated against expert manual delineations on both elastographic and pathology images.