A comparison of US- versus MR-based 3-D Prostate Shapes Using Radial Basis Function Interpolation and Statistical...

Area of Science:

• Medical Imaging
• Radiology
• Oncology

Background:

• Accurate prostate segmentation is crucial for localized prostate cancer treatment planning, particularly for brachytherapy.
• Current pre-implant planning predominantly relies on ultrasound (US) imaging.
• Integrating magnetic resonance (MR) imaging offers potential for enhanced segmentation accuracy.

Purpose of the Study:

• To compare 3D prostate segmentations derived from US and MR imaging data.
• To investigate the interchangeability of US and MR modalities for automatic contour segmentation in prostate brachytherapy.
• To lay the groundwork for coupled US-MR segmentation strategies.

Main Methods:

• Developed a 3D prostate segmentation approach using radial basis function interpolation and a 3D point distribution model.
• Employed modified principal axis transformation for rigid registration of US and MR images.
• Utilized statistical shape models for cross-modality geometric and shape comparisons.

Main Results:

• Presented a novel scheme for comparing 3D segmented prostate shapes across US and MR modalities.
• Conducted geometric comparisons of prostate volumes and dimensions between US and MR.
• Evaluated the interchangeability of US and MR for automatic contour segmentation.

Conclusions:

• The study provides a foundational step towards coupled US-MR segmentation for prostate brachytherapy.
• Findings suggest potential for improved pre-implant planning by integrating both imaging modalities.
• Further research can build upon these methods for enhanced treatment strategies.