Variability of manual ciliary muscle segmentation in optical coherence tomography images

Ciliary muscle and anterior segment characteristics in pre-presbyopic adults with Down syndrome

PURPOSE

Previous research has shown that accommodation deficits are common in individuals with Down syndrome (DS), but the origin and mechanisms behind these deficits are still unknown. The aim of this study was to investigate the characteristics of different ocular structures involved in accommodation, in particular the ciliary muscle (CM), in a population of individuals with DS to further understand this deficit and its mechanisms.

METHODS

Thirty-two volunteer participants of pre-presbyopic age with (n = 16) and without DS (n = 16) were recruited. Temporal and nasal images of the CM were acquired using anterior segment optical coherence tomography (AS-OCT) while participants fixated an eccentrically located target. Analysis of CM parameters was undertaken using validated semi-automated software. Axial length, anterior chamber depth, lens thickness and corneal curvature were obtained with the Topcon Aladdin Optical Biometer and Corneal Topographer. Non-cycloplegic refractive error and accommodative ability were obtained with an open-field autorefractor and dynamic retinoscopy, respectively. Independent t-tests were conducted to determine differences in CM and other anterior segment parameters between participants with and without DS.

RESULTS

No significant differences were found in the CM parameters studied between participants with and without DS (p > 0.05). In contrast, significant differences were found in visual acuity (p < 0.001), accommodative response (p < 0.001) and corneal curvature (K1 p = 0.003 and K2 p < 0.001) between participants with and without DS.

CONCLUSIONS

Despite having poorer accommodation, pre-presbyopic adults with DS do not have a different CM morphology to that found in typically developing adults. These findings suggest that the accommodative deficit found in this population is not due to a mechanical deficit of the CM.

CMS-NET: deep learning algorithm to segment and quantify the ciliary muscle in swept-source optical coherence tomography images

Background

The ciliary muscle plays a role in changing the shape of the crystalline lens to maintain the clear retinal image during near work. Studying the dynamic changes of the ciliary muscle during accommodation is necessary for understanding the mechanism of presbyopia. Optical coherence tomography (OCT) has been frequently used to image the ciliary muscle and its changes during accommodation in vivo. However, the segmentation process is cumbersome and time-consuming due to the large image data sets and the impact of low imaging quality.

Objectives

This study aimed to establish a fully automatic method for segmenting and quantifying the ciliary muscle on the basis of optical coherence tomography (OCT) images.

Design

A perspective cross-sectional study.

Methods

In this study, 3500 signed images were used to develop a deep learning system. A novel deep learning algorithm was created from the widely used U-net and a full-resolution residual network to realize automatic segmentation and quantification of the ciliary muscle. Finally, the algorithm-predicted results and manual annotation were compared.

Results

For segmentation performed by the system, the total mean pixel value difference (PVD) was 1.12, and the Dice coefficient, intersection over union (IoU), and sensitivity values were 93.8%, 88.7%, and 93.9%, respectively. The performance of the system was comparable with that of experienced specialists. The system could also successfully segment ciliary muscle images and quantify ciliary muscle thickness changes during accommodation.

Conclusion

We developed an automatic segmentation framework for the ciliary muscle that can be used to analyze the morphological parameters of the ciliary muscle and its dynamic changes during accommodation.

Performance of the Deep Neural Network Ciloctunet, Integrated with Open-Source Software for Ciliary Muscle Segmentation in Anterior Segment OCT Images, Is on Par with Experienced Examiners

Anterior segment optical coherence tomography (AS-OCT), being non-invasive and well-tolerated, is the method of choice for an in vivo investigation of ciliary muscle morphology and function. The analysis requires the segmentation of the ciliary muscle, which is, when performed manually, both time-consuming and prone to examiner bias. Here, we present a convolutional neural network trained for the automatic segmentation of the ciliary muscle in AS-OCT images. Ciloctunet is based on the Freiburg U-net and was trained and validated using 1244 manually segmented OCT images from two previous studies. An accuracy of 97.5% for the validation dataset was achieved. Ciloctunet's performance was evaluated by replicating the findings of a third study with 180 images as the test data. The replication demonstrated that Ciloctunet performed on par with two experienced examiners. The intersection-over-union index (0.84) of the ciliary muscle thickness profiles between Ciloctunet and an experienced examiner was the same as between the two examiners. The mean absolute error between the ciliary muscle thickness profiles of Ciloctunet and the two examiners (35.16 µm and 45.86 µm) was comparable to the one between the examiners (34.99 µm). A statistically significant effect of the segmentation type on the derived biometric parameters was found for the ciliary muscle area but not for the selective thickness reading ("perpendicular axis"). Both the inter-rater and the intra-rater reliability of Ciloctunet were good to excellent. Ciloctunet avoids time-consuming manual segmentation, thus enabling the analysis of large numbers of images of ample study cohorts while avoiding possible examiner biases. Ciloctunet is available as open-source.

Automated segmentation of the ciliary muscle in OCT images using fully convolutional networks

Quantifying shape changes in the ciliary muscle during accommodation is essential in understanding the potential role of the ciliary muscle in presbyopia. The ciliary muscle can be imaged in-vivo using OCT but quantifying the ciliary muscle shape from these images has been challenging both due to the low contrast of the images at the apex of the ciliary muscle and the tedious work of segmenting the ciliary muscle shape. We present an automatic-segmentation tool for OCT images of the ciliary muscle using fully convolutional networks. A study using a dataset of 1,039 images shows that the trained fully convolutional network can successfully segment ciliary muscle images and quantify ciliary muscle thickness changes during accommodation. The study also shows that EfficientNet outperforms other current backbones of the literature.

In vivo measurement of the attenuation coefficient of the sclera and ciliary muscle

We acquired 1325 nm OCT images of the sclera and ciliary muscle of human subjects. The attenuation coefficients of the sclera and ciliary muscle were determined from a curve fit of the average intensity profile of about 100 A-lines in a region of interest after correction for the effect of beam geometry, using a single scattering model. The average scleral attenuation coefficient was 4.13 ± 1.42 mm-1 with an age-related decrease that was near the threshold for statistical significance (p = 0.053). The average ciliary muscle attenuation coefficient was 1.72 ± 0.88 mm-1, but this value may be an underestimation due to contributions from multiple scattering. Overall, the results suggest that inter-individual variations in scleral attenuation contribute to variability in the quality of transscleral OCT images of the ciliary muscle and the outcome of transscleral laser therapies.

Review of the application of the open-source software CilOCT for semi-automatic segmentation and analysis of the ciliary muscle in OCT images

Presbyopia and myopia research shows a growing interest in ciliary muscle biometry using optical coherence tomography (OCT). Until now, segmentation of the ciliary muscle is often performed manually using either custom-developed programs or image processing software. Here we present a novel software for semi-automatic segmentation of the ciliary muscle. It provides direct import of OCT images in DICOM format, a standardized procedure for segmentation, image distortion correction, the export of anatomical ciliary muscle landmarks, like ciliary muscle apex and scleral spur, as well as a continuous thickness profile of the ciliary muscle as a novel way of analysis. All processing steps are stored as XML files, fostering documentation and reproducibility of research through the possibility of replicating the analysis. Additionally, CilOCT supports batch processing for the automated analysis of large numbers of images and the respective data export to tabulated text files based on the stored XML files. CilOCT was successfully applied in several studies and their results will be summarized in this paper.

Ciliary muscle thickness profiles derived from optical coherence tomography images

The purpose of this study was to provide an in-depth analysis of the ciliary muscle's (CM) morphological changes during accommodation by evaluating CM thickness (CMT) profiles. The CM of 15 near-emmetropic subjects (age 20-39) was imaged via optical coherence tomography (OCT) during far (0 D) and near vision (3 D). A custom-made Java-based program was used for semi-automatic CM segmentation and thickness measurements. CMT profiles were generated to determine regions of the largest shape changes. The results revealed on average a thinning within the first 0.25 mm and a thickening from 0.36 to 1.48 mm posterior to scleral spur when accommodating from 0 to 3 D. In contrast to previous analyses, this method offers pixel-wise reconstruction of CM shapes and quantification of accommodative change across the entire muscle boundary.