Related Experiment Video
Updated: Jul 13, 2025
08:08
Detection and Isolation of Cancer in Prostate Biopsies Using Stimulated Raman Histology and Artificial Intelligence
Published on: June 10, 2025
69
Prostate volume analysis in image registration for prostate cancer care: a verification study
Jessica M Bugeja1, Georges Mehawed2,3,4,5, Matthew J Roberts2,3,4,6,7
1Australian e-Health Research Centre, Commonwealth Scientific and Industrial Research Organisation, Health and Biosecurity, Herston, Australia. Jess.Bugeja@csiro.au.
Physical and Engineering Sciences in Medicine
|October 11, 2023
Summary
Image registration techniques for prostate cancer care are compared. Automated non-rigid registration showed superior accuracy, but semi-automated rigid registration is a feasible alternative for clinical use.
Area of Science:
- Medical Imaging
- Radiology
- Urology
Background:
- Combined magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) offer diagnostic and surgical planning benefits in prostate cancer care.
- Limited availability of PET-MRI necessitates exploring alternative image fusion techniques.
- Image fusion of separately acquired Prostate Specific Membrane Antigen PET/CT and MRI is a potential clinical solution.
Purpose of the Study:
- To compare the accuracy of different computed tomography (CT) and MRI registration algorithms for urological prostate cancer care.
- To evaluate semi-automated rigid, automated rigid, and automated non-rigid registration techniques.
- To assess the clinical feasibility of these registration methods.
Main Methods:
- Paired whole-pelvis MR and CT scan data from 20 patients were analyzed.
- Manual prostate clinical target volume (CTV) contours were created on both MR and CT images.
- Semi-automated rigid, automated rigid, and automated non-rigid registration algorithms were applied to align MR and CT data.
- Accuracy was quantified using Dice Similarity Index (DSI), 95% Hausdorff distance (95%HD), and average surface distance (ASD).
Main Results:
- The automated non-rigid registration technique demonstrated significantly superior performance over both automated rigid and semi-automated rigid methods (p < 0.001).
- Automated rigid registration also showed significantly improved accuracy compared to semi-automated rigid registration (p < 0.001).
- All tested registration techniques exhibited sufficient accuracy for potential clinical application.
Conclusions:
- Automated non-rigid registration offers the highest accuracy for aligning MR and CT images in prostate cancer care.
- Semi-automated rigid registration presents a quick, feasible, and accessible option for immediate clinical use by urologists and radiation oncologists.
- These image registration methods support enhanced diagnosis, surgical planning, and intra-operative guidance in prostate cancer management.
