Accuracy Validation of an Automated Method for Prostate Segmentation in Magnetic Resonance Imaging
Related Experiment Video
Updated: Mar 5, 2026
A Cognitive Fusion-guided Prostate Biopsy Using Multiparametric Magnetic Resonance Imaging and Transrectal Ultrasound
Published on: March 21, 2025
Maysam Shahedi1,2,3, Derek W Cool4,5, Glenn S Bauman6,7,8
1Baines Imaging Research Laboratory, London Regional Cancer Program, A3-123A, 790 Commissioners Rd E, London, ON, N6A 4L6, Canada. mshahedi@uwo.ca.
Abstract:
Three dimensional (3D) manual segmentation of the prostate on magnetic resonance imaging (MRI) is a laborious and time-consuming task that is subject to inter-observer variability. In this study, we developed a fully automatic segmentation algorithm for T2-weighted endorectal prostate MRI and evaluated its accuracy within different regions of interest using a set of complementary error metrics. Our dataset contained 42 T2-weighted endorectal MRI from prostate cancer patients. The prostate was manually segmented by one observer on all of the images and by two other observers on a subset of 10 images. The algorithm first coarsely localizes the prostate in the image using a template matching technique. Then, it defines the prostate surface using learned shape and appearance information from a set of training images. To evaluate the algorithm, we assessed the error metric values in the context of measured inter-observer variability and compared performance to that of our previously published semi-automatic approach. The automatic algorithm needed an average execution time of ∼60 s to segment the prostate in 3D. When compared to a single-observer reference standard, the automatic algorithm has an average mean absolute distance of 2.8 mm, Dice similarity coefficient of 82%, recall of 82%, precision of 84%, and volume difference of 0.5 cm3 in the mid-gland. Concordant with other studies, accuracy was highest in the mid-gland and lower in the apex and base. Loss of accuracy with respect to the semi-automatic algorithm was less than the measured inter-observer variability in manual segmentation for the same task.
Insights
A new automatic algorithm accurately segments the prostate on MRI scans in about 60 seconds. This automated prostate segmentation reduces variability compared to manual methods.
Area of Science:
- Medical Imaging
- Radiology
- Computational Anatomy
Background:
- Manual 3D prostate segmentation on MRI is time-consuming and prone to inter-observer variability.
- Accurate prostate segmentation is crucial for cancer diagnosis and treatment planning.
Purpose of the Study:
- To develop and evaluate a fully automatic segmentation algorithm for T2-weighted endorectal prostate MRI.
- To assess the algorithm's accuracy across different prostate regions and compare it to manual segmentation variability.
Main Methods:
- Developed a fully automatic algorithm using template matching for initial localization and learned shape/appearance models for surface definition.
- Evaluated the algorithm on 42 T2-weighted endorectal prostate MRI datasets.
- Compared algorithm performance against manual segmentations from multiple observers and a previous semi-automatic method.
Main Results:
- The automatic algorithm achieved an average segmentation time of approximately 60 seconds.
- In the mid-gland, accuracy metrics included a mean absolute distance of 2.8 mm, Dice similarity of 82%, recall of 82%, precision of 84%, and volume difference of 0.5 cm³.
- Accuracy was highest in the mid-gland, decreasing towards the apex and base.
Conclusions:
- The developed automatic algorithm provides accurate and efficient 3D prostate segmentation on MRI.
- The algorithm's performance loss compared to the semi-automatic method was less than the inter-observer variability of manual segmentation.
- This automated approach has the potential to reduce the labor and variability associated with manual prostate MRI segmentation.
