A semiautomatic tool for prostate segmentation in radiotherapy treatment planning

Area of Science:

• Medical Imaging
• Radiotherapy
• Computational Anatomy

Background:

• Accurate prostate delineation is crucial but time-consuming in radiotherapy planning.
• Existing methods require significant manual effort, impacting treatment efficiency.
• The proposed approach aims to streamline 3D prostate modeling for improved workflow.

Purpose of the Study:

• To develop and evaluate an intuitive 3D modeling method for the prostate using slice-wise best-fitting ellipses.
• To assess the accuracy and reproducibility of the proposed elliptical modeling technique.
• To explore the potential time-saving benefits of this automated approach.

Main Methods:

• A novel method for 3D prostate modeling based on fitting ellipses to cross-sectional image slices.
• Initialization using control points, adaptable to various imaging modalities assuming smooth, elliptical cross-sections.
• Development of a prior shape model from 23 patients' data and evaluation on 10 test patients using T1-weighted MRI scans.

Main Results:

• The best-fitting ellipses (BFE) method achieved high accuracy: Volume overlap (Vo) of 0.96 ± 0.01, Accuracy (Ac) of 0.92 ± 0.01, and Hausdorff distance (HD) of 1.6 ± 0.27 mm.
• The mean best-fitting ellipses (MBFE) method showed slightly lower but still significant results: Vo (0.9 ± 0.02), Ac (0.81 ± 0.03), and HD (4.05 ± 1.3) mm.
• Moderate improvements were observed for MBFE after applying the Monte Carlo Markov Chain (MCMC) method.

Conclusions:

• The best-fitting ellipses method shows significant potential for robust and reproducible 3D prostate modeling in radiotherapy.
• The technique offers a promising avenue for reducing the time required for target volume delineation.
• Further validation on larger patient cohorts is suggested to confirm time-saving benefits.