Deformable image registration between pathological images and MR image via an optical macro image

Area of Science:

• Medical Imaging
• Pathology
• Image Registration

Background:

• Computed tomography (CT) and magnetic resonance (MR) imaging are vital for human body visualization.
• Definitive pathological diagnosis often requires invasive biopsy or resection for tissue evaluation.
• Bridging the gap between tissue information and imaging signals is crucial for advanced diagnostics.

Purpose of the Study:

• To develop a registration scheme connecting physical tissue (PT) images of specimens with 3D MR images.
• To establish a method for analyzing the relationship between tissue characteristics and MR image signals.
• To enhance diagnostic capabilities by integrating detailed tissue information with non-invasive imaging.

Main Methods:

• A registration scheme was proposed using optical macro (OM) images as a reference.
• Two registration processes were developed to derive conversion parameters between PT and MR images.
• The method was validated using a resected human brain specimen.

Main Results:

• Aligned PT images closely matched the shape of the OM image.
• Extracted cross-sectional MR images were also similar to the OM image.
• The method achieved high accuracy with TREs of 0.56±0.39mm and 0.87±0.42mm, enabling accurate region searching between PT and MR images.

Conclusions:

• The proposed method successfully aligns PT images with 3D MR images.
• It enables the analysis of relationships between physical tissue properties and MR signals.
• This technique holds potential for improving pathological diagnosis and understanding tissue-MR signal correlations.