Grid based evaluation of a liver segmentation method for contrast enhanced abdominal MRI

Area of Science:

• Radiology
• Medical Imaging
• Computer-Aided Diagnosis

Background:

• Advancements in MRI contrast agents offer new diagnostic and therapeutic planning possibilities.
• Specialized image analysis techniques are necessary for interpreting these new MRI contrast agents.
• Gadolinium-ethoxybenzyl-diethylenetriaminepentaacetic acid (Gd-EOB) enhanced MRI requires precise liver segmentation.

Purpose of the Study:

• To validate and optimize a novel, fully automatic liver segmentation method for Gd-EOB enhanced MRI.
• To leverage an academic research grid for extensive parameter scanning and evaluation.
• To present the implementation and initial results of the liver segmentation technique.

Main Methods:

• Utilizing an academic research grid for computational analysis.
• Implementing a fully automatic method for liver segmentation in MRI.
• Performing extensive parameter scans and validation against expert reference segmentations.

Main Results:

• Demonstration of the implementation layout for the liver segmentation method.
• Presentation of preliminary results from the validation and optimization process.
• Insights into the production phase and grid exploitation for Healthgrid applications.

Conclusions:

• The developed automatic liver segmentation method shows promise for Gd-EOB enhanced MRI.
• Academic research grids are valuable tools for optimizing medical image analysis techniques.
• Experiences gained inform the broader use of research grids in healthcare applications.