Related Experiment Video
Updated: Oct 28, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Multi-view convolutional neural networks for automated ocular structure and tumor segmentation in retinoblastoma
Victor I J Strijbis1,2, Christiaan M de Bloeme3, Robin W Jansen3
1Department of Radiology and Nuclear Medicine, Cancer Center Amsterdam, Amsterdam UMC, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands. v.strijbis@amsterdamumc.nl.
Abstract:
In retinoblastoma, accurate segmentation of ocular structure and tumor tissue is important when working towards personalized treatment. This retrospective study serves to evaluate the performance of multi-view convolutional neural networks (MV-CNNs) for automated eye and tumor segmentation on MRI in retinoblastoma patients. Forty retinoblastoma and 20 healthy-eyes from 30 patients were included in a train/test (N = 29 retinoblastoma-, 17 healthy-eyes) and independent validation (N = 11 retinoblastoma-, 3 healthy-eyes) set. Imaging was done using 3.0 T Fast Imaging Employing Steady-state Acquisition (FIESTA), T2-weighted and contrast-enhanced T1-weighted sequences. Sclera, vitreous humour, lens, retinal detachment and tumor were manually delineated on FIESTA images to serve as a reference standard. Volumetric and spatial performance were assessed by calculating intra-class correlation (ICC) and dice similarity coefficient (DSC). Additionally, the effects of multi-scale, sequences and data augmentation were explored. Optimal performance was obtained by using a three-level pyramid MV-CNN with FIESTA, T2 and T1c sequences and data augmentation. Eye and tumor volumetric ICC were 0.997 and 0.996, respectively. Median [Interquartile range] DSC for eye, sclera, vitreous, lens, retinal detachment and tumor were 0.965 [0.950-0.975], 0.847 [0.782-0.893], 0.975 [0.930-0.986], 0.909 [0.847-0.951], 0.828 [0.458-0.962] and 0.914 [0.852-0.958], respectively. MV-CNN can be used to obtain accurate ocular structure and tumor segmentations in retinoblastoma.
Insights
Multi-view convolutional neural networks (MV-CNNs) accurately segment ocular structures and tumors in retinoblastoma patients using MRI. This automated approach aids personalized treatment by providing precise eye and tumor segmentation.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Accurate segmentation of ocular structures and tumor tissue is crucial for personalized treatment in retinoblastoma.
- Current segmentation methods may lack the precision required for complex ocular anatomy and tumor delineation.
Purpose of the Study:
- To evaluate the performance of multi-view convolutional neural networks (MV-CNNs) for automated segmentation of ocular structures and tumors in retinoblastoma patients using MRI.
- To assess the impact of different imaging sequences and data augmentation techniques on segmentation accuracy.
Main Methods:
- A retrospective study utilizing MRI data from 40 retinoblastoma and 20 healthy eyes.
- Manual delineation of ocular structures and tumors on Fast Imaging Employing Steady-state Acquisition (FIESTA) images served as the reference standard.
- Multi-view convolutional neural networks (MV-CNNs) were trained and validated, with performance assessed using intra-class correlation (ICC) and dice similarity coefficient (DSC).
Main Results:
- Optimal MV-CNN performance was achieved using FIESTA, T2, and T1c sequences with data augmentation.
- High volumetric intra-class correlation coefficients (ICCs) were obtained for eyes (0.997) and tumors (0.996).
- Median dice similarity coefficients (DSCs) for ocular structures ranged from 0.828 (retinal detachment) to 0.975 (vitreous humour), with tumors achieving a median DSC of 0.914.
Conclusions:
- Multi-view convolutional neural networks (MV-CNNs) demonstrate high accuracy in segmenting ocular structures and tumors in retinoblastoma patients.
- This automated segmentation approach shows significant potential for improving personalized treatment strategies in retinoblastoma care.
- The study highlights the effectiveness of combining multiple MRI sequences and data augmentation for robust segmentation performance.
