Accuracy of generative deep learning model for macular anatomy prediction from optical coherence tomography images in macular hole surgery
Ontology highlight
ABSTRACT:
SUBMITTER: Kwon H
PROVIDER: S-EPMC10959933 | biostudies-literature | 2024 Jan
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:To evaluate the automated determination of the center of an idiopathic macular hole (MH) by using swept-source optical coherence tomography (OCT) images with new macro-based algorithms in ImageJ and to compare the difference between the MH center measurements obtained automatically and manually. This cross-sectional study included 39 eyes of 39 elderly individuals (22 women, 17 men) with stage 3 and 4 MH. The MH center was automatically determined using the ImageJ macro. The foveal center was also manually identified by two masked examiners using horizontal and vertical serial B-scan OCT angiography images. The mean age was 68.8 ± 8.3 years. After adjusting for the effect of magnification, the mean distance between the MH center determined manually by Examiner 1 and that determined automatically was 15.5 ± 9.9 µm. The mean distance between the two manually determined measurements of the MH center was 20.3 ± 19.7 µm. These two mean distance values did not differ significantly (Welch t-test, p = 0.27) and was non-inferior (p < 0.0001). The automated ImageJ-based method for determining the MH center was comparable to manual methods. This study showed that automated measurements were non-inferior to manual measurements, and demonstrated a substitutable usefulness, at least for use in clinical practice.
| S-EPMC9181087 | biostudies-literature
Project description:BackgroundThis study aimed to develop deep learning models using macular optical coherence tomography (OCT) images to estimate axial lengths (ALs) in eyes without maculopathy.MethodsA total of 2,664 macular OCT images from 444 patients' eyes without maculopathy, who visited Beijing Hospital between March 2019 and October 2021, were included. The dataset was divided into training, validation, and testing sets with a ratio of 6:2:2. Three pre-trained models (ResNet 18, ResNet 50, and ViT) were developed for binary classification (AL ≥ 26 mm) and regression task. Ten-fold cross-validation was performed, and Grad-CAM analysis was employed to visualize AL-related macular features. Additionally, retinal thickness measurements were used to predict AL by linear and logistic regression models.ResultsResNet 50 achieved an accuracy of 0.872 (95% Confidence Interval [CI], 0.840-0.899), with high sensitivity of 0.804 (95% CI, 0.728-0.867) and specificity of 0.895 (95% CI, 0.861-0.923). The mean absolute error for AL prediction was 0.83 mm (95% CI, 0.72-0.95 mm). The best AUC, and accuracy of AL estimation using macular OCT images (0.929, 87.2%) was superior to using retinal thickness measurements alone (0.747, 77.8%). AL-related macular features were on the fovea and adjacent regions.ConclusionOCT images can be effectively utilized for estimating AL with good performance via deep learning. The AL-related macular features exhibit a localized pattern in the macula, rather than continuous alterations throughout the entire region. These findings can lay the foundation for future research in the pathogenesis of AL-related maculopathy.
| S-EPMC10693454 | biostudies-literature
Project description:Surgical advancements for full-thickness macular hole (FTMH) treatment include vitrectomy, membrane peeling, and the inverted flap technique (IFT). IFT, which involves inverting the internal limiting membrane (ILM) flap over the macular hole (MH) or into the MH, improves success rates and visual recovery. However, issues like mis-aspiration during flap handling have been problematic. We introduce the petaloid technique, to position the ILM flap under air during FTMH surgery to evaluate its outcomes. This retrospective study included 28 eyes, with a mean minimum linear diameter (MLD) 472.04 ± 199.7 μm and basal diameter (BD) of 834.95 ± 593.54 μm. Primary closure of MH was achieved in 96.42% of patients, with 3.57% showing persistent MH during the 6-month follow-up. The VA improved significantly from preoperative levels at each postoperative stage, with notable increases at 3 months (0.86 ± 0.49 logMAR; p = 0.0132) and 6 months (0.77 ± 0.41 logMAR; p = 0.000081). The new closure patterns showed Type A in 28.6%, B in 14.3%, C in 28.6%, and D in 25%. Among different types of closure patterns, although VA improved in all types, significant improvement in VA was noted for type A and type C, with notable improvements at the 6-month follow-up for Type A (0.60 ± 0.23 logMAR; p = 0.02) and at the 1-month follow-up for Type C (0.62 ± 0.28 logMAR; p = 0.02). For macular hole retinal detachment in 5 eyes, the average preoperative VA was 1.86 ± 0.19 logMAR, while the final mean postoperative VA showed a significant improvement to 1.1 ± 0.40 logMAR (p = 0.021, paired t-tests). Retinal reattachment was achieved in all cases (5/5) without recurrent detachment post-silicone oil removal. For the prognostic significance of MH shapes identified by OCT, it can be concluded that the shape of macular holes significantly influences visual acuity outcomes at 6 months post-surgery (p = 0.037). The shape of macular holes, particularly Flask-shaped, significantly impacts visual acuity compared to other shape. The modified petaloid technique for treating FTMH proved safe and effective, with no significant complications noted.
| S-EPMC11456521 | biostudies-literature
Project description:BackgroundTo describe the features of foveal microvasculature using optical coherence tomography angiography (OCTA) and to determine the related clinical factors in eyes with surgically closed macular hole (MH).MethodsA retrospective case series of 18 patients with unilateral MH was reviewed. The patients maintained complete hole closure after vitrectomy with inner limiting membrane (ILM) peeling for at least 12 months. The healthy fellow eyes were studied as controls. The foveal microvasculature of both eyes was examined by OCTA. The area of the foveal avascular zone (FAZ) and the vascular density (VD) ratio in the superficial and deep capillary plexuses (SCP and DCP) were determined after surgery. Several clinical factors including age, stage and dimensions of MH, papillofoveal distance, the extent of nasal displacement of the fovea after surgery, postoperative central foveal thickness, and outer-retina integrity were evaluated to determine any relationships with the OCTA parameters.ResultsThe mean FAZ area in both the SCP and DCP (0.29 ± 0.11 mm2 and 0.39 ± 0.14 mm2) was significantly smaller than those of the controls (0.45 ± 0.14 mm2 and 0.62 ± 0.22 mm2) (p = 0.001 and <0.001, respectively). The mean VD ratio in the SCP (0.270 ± 0.349) was similar to that of the controls (0.321 ± 0.189) (p = 0.231); however, that in the DCP (0.321 ± 0.189) was significantly lower than that of the controls (0.331 ± 0.119) (p = 0.025). Only the extent of nasal displacement of the fovea was correlated with the DCP FAZ-area difference values between the study group and the controls (correlation coefficient = 0.577; p = 0.012).ConclusionsAfter successful MH surgery, the FAZ area in both the SCP and DCP was smaller and the VD ratio of the DCP was lower, suggesting a possible DCP vulnerability to tractional stress. As the FAZ area of the DCP in closed-MH eyes became smaller than that in the controls, the fovea was less displaced toward the optic disc, possibly reflecting a lack of retinal redundancy caused by horizontal stretching accompanied by foveal displacement.
| S-EPMC5704531 | biostudies-literature
Project description:Macular holes, one of the most common macular diseases, require timely treatment. The morphological changes on optical coherence tomography (OCT) images provided an opportunity for direct observation of the disease, and accurate segmentation was needed to identify and quantify the lesions. Developments of such algorithms had been obstructed by a lack of high-quality datasets (the OCT images and the corresponding gold standard macular hole segmentation labels), especially for supervised learning-based segmentation algorithms. In such context, we established a large OCT image macular hole segmentation (OIMHS) dataset with 3859 B-scan images of 119 patients, and each image provided four segmentation labels: retina, macular hole, intraretinal cysts, and choroid. This dataset offered an excellent opportunity for investigating the accuracy and reliability of different segmentation algorithms for macular holes and a new research insight into the further development of clinical research for macular diseases, which included the retina, lesions, and choroid in quantitative analyses.
| S-EPMC10628143 | biostudies-literature
Project description:PurposeTo assess the feasibility and interest of measuring macular hole (MH) size using en face optical coherence tomography (OCT) compared with manual diameter measurements on B-scans.MethodsAmong our previously published series of 100 patients operated for primary MH, patients whose images were acquired with Cirrus 5000 HD-OCT (Carl Zeiss Meditec) with a quality signal strength >5/10 were included. Three segmentations (internal limiting membrane, horizontal, and retinal pigment epithelium (RPE)) were compared for obtaining the most appropriate en face image. MH surface areas were measured using ImageJ software. Mean diameters calculated from surface areas (diameter=2 × √(surface area/π)) were compared with those measured on B-scans.ResultsNineteen patients were included with a mean age of 72±8 years (56-86) and a female predominance (3/16). The mean absolute difference between horizontal and vertical diameters measured on B-scans was of 54±47 μm (0-180) without reaching significance (P=0.874). RPE segmentation provided the best en face image and was feasible without and with adjustment, respectively, in 79% and 100% (cases with vitreomacular traction) of cases. No significant difference in mean diameters was observed between those calculated from en face images (435±143 μm (195-693)) and those measured on horizontal B-scans (426±139 μm (214-705), P=0.482).ConclusionMeasuring MH size on en face OCT images is feasible, reliable, and eliminates the potential bias related to manual measurements on B-scans. Its integration into OCT devices would offer an automated and easy-to-use option for clinical practice.
| S-EPMC5848286 | biostudies-literature
Project description:BackgroundWhile deep learning has delivered promising results in the field of ophthalmology, the hurdle to complete a deep learning study is high. In this study, we aim to facilitate small scale model trainings by exploring the role of preprocessing to reduce computational burden and accelerate learning.MethodsA small subset of a previously published dataset containing optical coherence tomography images of choroidal neovascularization, drusen, diabetic macula edema, and normal macula was modified using Fourier transformation and bandpass filter, producing high frequency images, original images, and low frequency images. Each set of images was trained with the same model, and their performances were compared.ResultsCompared to that with the original image dataset, the model trained with the high frequency image dataset achieved an improved final performance and reached maximum performance much earlier (in fewer epochs). The model trained with low frequency images did not achieve a meaningful performance.ConclusionAppropriate preprocessing of training images can accelerate the training process and can potentially facilitate modeling using artificial intelligence when limited by sample size or computational power.
| S-EPMC10108538 | biostudies-literature
Project description:PurposeTo develop a deep learning (DL) model to detect morphologic patterns of diabetic macular edema (DME) based on optical coherence tomography (OCT) images.MethodsIn the training set, 12,365 OCT images were extracted from a public data set and an ophthalmic center. A total of 656 OCT images were extracted from another ophthalmic center for external validation. The presence or absence of three OCT patterns of DME, including diffused retinal thickening, cystoid macular edema, and serous retinal detachment, was labeled with 1 or 0, respectively. A DL model was trained to detect three OCT patterns of DME. The occlusion test was applied for the visualization of the DL model.ResultsApplying 5-fold cross-validation method in internal validation, the area under the receiver operating characteristic curve for the detection of three OCT patterns (i.e., diffused retinal thickening, cystoid macular edema, and serous retinal detachment) was 0.971, 0.974, and 0.994, respectively, with an accuracy of 93.0%, 95.1%, and 98.8%, respectively, a sensitivity of 93.5%, 94.5%, and 96.7%, respectively, and a specificity of 92.3%, 95.6%, and 99.3%, respectively. In external validation, the area under the receiver operating characteristic curve was 0.970, 0.997, and 0.997, respectively, with an accuracy of 90.2%, 95.4%, and 95.9%, respectively, a sensitivity of 80.1%, 93.4%, and 94.9%, respectively, and a specificity of 97.6%, 97.2%, and 96.5%, respectively. The occlusion test showed that the DL model could successfully identify the pathologic regions most critical for detection.ConclusionOur DL model demonstrated high accuracy and transparency in the detection of OCT patterns of DME. These results emphasized the potential of artificial intelligence in assisting clinical decision-making processes in patients with DME.
| S-EPMC8078116 | biostudies-literature
Project description:AimTo develop a fully automatic algorithm to segment retinal cavitations on optical coherence tomography (OCT) images of macular telangiectasia type 2 (MacTel2).MethodsThe dataset consisted of 99 eyes from 67 participants enrolled in an international, multicentre, phase 2 MacTel2 clinical trial (NCT01949324). Each eye was imaged with spectral-domain OCT at three time points over 2 years. Retinal cavitations were manually segmented by a trained Reader and the retinal cavitation volume was calculated. Two convolutional neural networks (CNNs) were developed that operated in sequential stages. In the first stage, CNN1 classified whether a B-scan contained any retinal cavitations. In the second stage, CNN2 segmented the retinal cavitations in a B-scan. We evaluated the performance of the proposed method against alternative methods using several performance metrics and manual segmentations as the gold standard.ResultsThe proposed method was computationally efficient and accurately classified and segmented retinal cavitations on OCT images, with a sensitivity of 0.94, specificity of 0.80 and average Dice similarity coefficient of 0.94±0.07 across all time points. The proposed method produced measurements that were highly correlated with the manual measurements of retinal cavitation volume and change in retinal cavitation volume over time.ConclusionThe proposed method will be useful to help clinicians quantify retinal cavitations, assess changes over time and further investigate the clinical significance of these early structural changes observed in MacTel2.
| S-EPMC8365558 | biostudies-literature
Project description:The choroid layer is a vascular layer in human retina and its main function is to provide oxygen and support to the retina. Various studies have shown that the thickness of the choroid layer is correlated with the diagnosis of several ophthalmic diseases. For example, diabetic macular edema (DME) is a leading cause of vision loss in patients with diabetes. Despite contemporary advances, automatic segmentation of the choroid layer remains a challenging task due to low contrast, inhomogeneous intensity, inconsistent texture and ambiguous boundaries between the choroid and sclera in Optical Coherence Tomography (OCT) images. The majority of currently implemented methods manually or semi-automatically segment out the region of interest. While many fully automatic methods exist in the context of choroid layer segmentation, more effective and accurate automatic methods are required in order to employ these methods in the clinical sector. This paper proposed and implemented an automatic method for choroid layer segmentation in OCT images using deep learning and a series of morphological operations. The aim of this research was to segment out Bruch's Membrane (BM) and choroid layer to calculate the thickness map. BM was segmented using a series of morphological operations, whereas the choroid layer was segmented using a deep learning approach as more image statistics were required to segment accurately. Several evaluation metrics were used to test and compare the proposed method against other existing methodologies. Experimental results showed that the proposed method greatly reduced the error rate when compared with the other state-of-the-art methods.
| S-EPMC6395677 | biostudies-literature