Deep Convolutional Neural Networks for Automatic Detection of Orbital Blowout Fractures

CT-based artificial intelligence prediction model for ocular motility score of thyroid eye disease

PURPOSE

Thyroid eye disease (TED) is the most common orbital disease in adults. Ocular motility restriction is the primary complaint of patients, while its evaluation is quite difficult. The present study aimed to introduce an artificial intelligence (AI) model based on orbital computed tomography (CT) images for ocular motility score.

METHODS

A total of 410 sets of CT images and clinical data were obtained from the hospital. To build a triple classification predictive model for ocular motility score, multiple deep learning models were employed to extract features of images and clinical data. Subgroup analyses based on pertinent clinical features were performed to test the efficacy of models.

RESULTS

The ResNet-34 network outperformed Alex-Net and VGG16-Net in prediction of ocular motility score, with the optimal accuracy (ACC) of 0.907, 0.870, and 0.890, respectively. Subgroup analyses indicated no significant difference in ACC between active or inactive phase, functional visual field diplopia or peripheral visual field diplopia (p > 0.05). However, in the gender subgroup, the prediction model performed more accurately in female patients than males (p = 0.02).

CONCLUSION

In conclusion, the AI model based on CT images and clinical data successfully realized automatic scoring of ocular motility in TED patients. This approach potentially enhanced the efficiency and accuracy of ocular motility evaluation, thus facilitating clinical application.

Application of artificial intelligence in oculoplastics

Oculoplastics is a subspecialty of ophthalmology/dermatology concerned with eyelid, orbital, and lacrimal diseases. Artificial intelligence (AI), with its powerful ability to analyze large data sets, has dramatically benefited oculoplastics. The cutting-edge AI technology is widely applied to extract ocular parameters and to use these results for further assessment, such as screening and diagnosis of blepharoptosis and predicting the progression of thyroid eye disease. AI also assists in treatment procedures, such as surgical strategy planning in blepharoptosis. High efficiency and high reliability are the most apparent advantages of AI, with promising prospects. The possibilities of AI in oculoplastics may lie in three-dimensional modeling technology and image generation. We retrospectively summarize AI applications involving eyelid, orbital, and lacrimal diseases in oculoplastics, and we also examine the strengths and weaknesses of AI technology in oculoplastics.

A Deep Learning Model for Screening Computed Tomography Imaging for Thyroid Eye Disease and Compressive Optic Neuropathy

Purpose

Thyroid eye disease (TED) is an autoimmune condition with an array of clinical manifestations, which can be complicated by compressive optic neuropathy. It is important to identify patients with TED early to ensure close monitoring and treatment to prevent potential permanent disability or vision loss. Deep learning artificial intelligence (AI) algorithms have been utilized in ophthalmology and in other fields of medicine to detect disease. This study aims to introduce a deep learning model to evaluate orbital computed tomography (CT) images for the presence of TED and potential compressive optic neuropathy.

Design

Retrospective review and deep learning algorithm modeling.

Subjects

Patients with TED with dedicated orbital CT scans and with an examination by an oculoplastic surgeon over a 10-year period at a single academic institution. Patients with no TED and normal CTs were used as normal controls. Those with other diagnoses, such as tumors or other inflammatory processes, were excluded.

Methods

Orbital CTs were preprocessed and adopted for the Visual Geometry Group-16 network to distinguish patients with no TED, mild TED, and severe TED with compressive optic neuropathy. The primary model included training and testing of all 3 conditions. Binary model performance was also evaluated. An oculoplastic surgeon was also similarly tested with single and serial images for comparison.

Main Outcome Measures

Accuracy of deep learning model discernment of region of interest for CT scans to distinguish TED versus normal control, as well as TED with clinical signs of optic neuropathy.

Results

A total of 1187 photos from 141 patients were used to develop the AI model. The primary model trained on patients with no TED, mild TED, and severe TED had 89.5% accuracy (area under the curve: range, 0.96-0.99) in distinguishing patients with these clinical categories. In comparison, testing of an oculoplastic surgeon in these 3 categories showed decreased accuracy (70.0% accuracy in serial image testing).

Conclusions

The deep learning model developed in the study can accurately detect TED and further detect TED with clinical signs of optic neuropathy based on orbital CT. The model proved superior compared with human expert grading. With further optimization and validation, this TED deep learning model could help guide frontline health care providers in the detection of TED and help stratify the urgency of a referral to an oculoplastic surgeon and endocrinologist.

Financial Disclosures

The authors have no proprietary or commercial interest in any materials discussed in this article.

A novel method of measuring proptosis with computed tomography

BACKGROUND

Computed tomography (CT) can avoid interference factors and has been imported into some software to measure proptosis clinically as the golden standard.

PURPOSE

To establish a new method for semi-automatically measuring the proptosis on CT and evaluate its accuracy and reproducibility.

MATERIAL AND METHODS

A total of 50 orbital CT images were collected of healthy individuals, 25 patients with Graves ophthalmopathy (GO), and 25 patients with orbital fracture (OF). A new image processing software, MedrawHDC, was developed to semi-automatically measure the proptosis (MedrawHDC method). The classic radiological (CR) method (measuring proptosis with the software called Mimics) and MedrawHDC method were applied in all three groups (measured by observer S). Hertel's exophthalmometer (HE) method was also applied in the GO group. Moreover, two other observers were asked to measure the proptosis using MedrawHDC, to evaluate its reproducibility.

RESULTS

The MedrawHDC method was highly consistent with the CR method in measuring proptosis (normal group: intraclass correlation coefficient [ICC] = 0.989; GO group: ICC = 0.979; OF group: ICC = 0.979). In the GO group, the value of proptosis measured by two radiological methods were consistent with that measured by the HE method (CR method: ICC = 0.703; MedrawHDC method: ICC = 0.697). Bland-Altman plots showed similar results. The measurements obtained by three observers were highly reproducible (ICC = 0.995).

CONCLUSION

The newly established MedrawHDC method, with high accessibility, convenience, and repeatability, is reliable in assessing proptosis. It shows high potential for wide application, having clinical value for scientific evaluation of proptosis.

Neural network application for assessing thyroid-associated orbitopathy activity using orbital computed tomography

This study aimed to propose a neural network (NN)-based method to evaluate thyroid-associated orbitopathy (TAO) patient activity using orbital computed tomography (CT). Orbital CT scans were obtained from 144 active and 288 inactive TAO patients. These CT scans were preprocessed by selecting eleven slices from axial, coronal, and sagittal planes and segmenting the region of interest. We devised an NN employing information extracted from 13 pipelines to assess these slices and clinical patient age and sex data for TAO activity evaluation. The proposed NN's performance in evaluating active and inactive TAO patients achieved a 0.871 area under the receiver operating curve (AUROC), 0.786 sensitivity, and 0.779 specificity values. In contrast, the comparison models CSPDenseNet and ConvNeXt were significantly inferior to the proposed model, with 0.819 (p = 0.029) and 0.774 (p = 0.04) AUROC values, respectively. Ablation studies based on the Sequential Forward Selection algorithm identified vital information for optimal performance and evidenced that NNs performed best with three to five active pipelines. This study establishes a promising TAO activity diagnosing tool with further validation.

Is Convolutional Neural Network Accurate for Automatic Detection of Zygomatic Fractures on Computed Tomography?

PURPOSE

Zygomatic fractures involve complex anatomical structures of the mid-face and the diagnosis can be challenging and labor-consuming. This research is aimed to evaluate the performance of an automatic algorithm for the detection of zygomatic fractures based on convolutional neural network (CNN) on spiral computed tomography (CT).

MATERIALS AND METHODS

We designed a cross-sectional retrospective diagnostic trial study. Clinical records and CT scans of patients with zygomatic fractures were reviewed. The sample consisted of two types of patients with different zygomatic fractures statuses (positive or negative) in Peking University School of Stomatology from 2013 to 2019. All CT samples were randomly divided into three groups at a ratio of 6:2:2 as training set, validation set, and test set, respectively. All CT scans were viewed and annotated by three experienced maxillofacial surgeons, serving as the gold standard. The algorithm consisted of two modules as follows: (1) segmentation of the zygomatic region of CT based on U-Net, a type of CNN model; (2) detection of fractures based on Deep Residual Network 34(ResNet34). The region segmentation model was used first to detect and extract the zygomatic region, then the detection model was used to detect the fracture status. The Dice coefficient was used to evaluate the performance of the segmentation algorithm. The sensitivity and specificity were used to assess the performance of the detection model. The covariates included age, gender, duration of injury, and the etiology of fractures.

RESULTS

A total of 379 patients with an average age of 35.43 ± 12.74 years were included in the study. There were 203 nonfracture patients and 176 fracture patients with 220 sites of zygomatic fractures (44 patients underwent bilateral fractures). The Dice coefficientof zygomatic region detection model and gold standard verified by manual labeling were 0.9337 (coronal plane) and 0.9269 (sagittal plane), respectively. The sensitivity and specificity of the fracture detection model were 100% (p＞.05).

CONCLUSION

The performance of the algorithm based on CNNs was not statistically different from the gold standard (manual diagnosis) for zygomatic fracture detection in order for the algorithm to be applied clinically.

Prognostic factors of traumatic optic neuropathy based on multimodal analysis-Especially the influence of postoperative dressing change and optic nerve blood supply on prognosis

Objective

To investigate the critical prognostic factors of patients with traumatic optic neuropathy (TON) treated with endoscopic transnasal optic canal decompression (ETOCD) and to perform multimodal analysis based on imaging examinations of optical coherence tomography angiography (OCTA) and CT scan. Subsequently, a new prediction model was established.

Methods

The clinical data of 76 patients with TON who underwent decompression surgery with the endoscope-navigation system in the Department of Ophthalmology, Shanghai Ninth People's Hospital from January 2018 to December 2021 were retrospectively analyzed. The clinical data included demographic characteristics, reasons for injury, interval between injury and surgery, multimode imaging information of CT scan and OCTA, including orbital fracture, optical canal fractures, vessel density of optic disc and macula, and the times of postoperative dressing change. Binary logistic regression was used to establish a model for best corrected visual acuity (BCVA) after treatment as a predictor of TON outcome.

Results

Postoperative BCVA improved in 60.5% (46/76) patients and did not improve in 39.5% (30/76) patients. The times of postoperative dressing change had a significant impact on the prognosis. Other factors affecting the prognosis were microvessel density of the central optic disc, the cause of injury, and the microvessel density above the macula. The area under the raw current curves of the predictive model was 0.7596.

Conclusions

The times of dressing changes after the operation, i.e., continuous treatment, is the key factor affecting prognosis. The microvessel density in the center of the optic disc and superior macula, quantitatively analyzed by OCTA, is the prognostic factor of TON and may be used as a prognostic marker of TON.

A Minority Class Balanced Approach Using the DCNN-LSTM Method to Detect Human Wrist Fracture

The emergency department of hospitals receives a massive number of patients with wrist fracture. For the clinical diagnosis of a suspected fracture, X-ray imaging is the major screening tool. A wrist fracture is a significant global health concern for children, adolescents, and the elderly. A missed diagnosis of wrist fracture on medical imaging can have significant consequences for patients, resulting in delayed treatment and poor functional recovery. Therefore, an intelligent method is needed in the medical department to precisely diagnose wrist fracture via an automated diagnosing tool by considering it a second option for doctors. In this research, a fused model of the deep learning method, a convolutional neural network (CNN), and long short-term memory (LSTM) is proposed to detect wrist fractures from X-ray images. It gives a second option to doctors to diagnose wrist facture using the computer vision method to lessen the number of missed fractures. The dataset acquired from Mendeley comprises 192 wrist X-ray images. In this framework, image pre-processing is applied, then the data augmentation approach is used to solve the class imbalance problem by generating rotated oversamples of images for minority classes during the training process, and pre-processed images and augmented normalized images are fed into a 28-layer dilated CNN (DCNN) to extract deep valuable features. Deep features are then fed to the proposed LSTM network to distinguish wrist fractures from normal ones. The experimental results of the DCNN-LSTM with and without augmentation is compared with other deep learning models. The proposed work is also compared to existing algorithms in terms of accuracy, sensitivity, specificity, precision, the F1-score, and kappa. The results show that the DCNN-LSTM fusion achieves higher accuracy and has high potential for medical applications to use as a second option.

A quantitative study of the effect of ICL orientation selection on post-operative vault and model-assisted vault prediction

Background

Appropriate vault height of implantable collamer lens (ICL) implantation matters for it has risks of corneal endothelial cell loss, cataract formation and intraocular pressure elevation, which could lead to irreversible damage to optic nerve. Therefore, pre-operative prediction for an ideal vault height is a hotspot. However, few data exist regarding quantitative effect of ICL orientation on vault height. This study is aimed to quantitatively investigate the effect of ICL implantation orientation on vault height, and built a machine-learning (ML)-based vault prediction model taking implantation orientation into account.

Methods

473 consecutive case series treated with ICL implantation were retrospectively analyzed (408 were horizontally implanted, and 65 were vertically implanted). Multivariable logistic regression analysis was performed to determine the association between ICL orientation and achieved vault. ML was performed to develop a new vault height prediction model taking ICL orientation into account. Receiver operating characteristic curve (ROC) and net reclassification index (NRI) were obtained to assess the prediction ability.

Results

95% of all the patients achieved 20/20 uncorrected distance visual acuity (UDVA) or better. No complications including cataract formation, dispersion or optic nerve injury were observed in any cases. Sex, sphere power, cylinder power, axis, ICL size and ICL orientation were all significant risk factors associated to vault height, and age was positively co-related. Of note, ICL size and ICL orientation were the top-ranking risk factors. Comparing to conventional horizontal implantation, vertical implantation could reduce the achieved vault by 81.187 μm (p < 0.001). In regarding to different ICL sizes, vertical implantation had no good to vault reduction when using ICL of 12.1 mm. However, it could reduce the vault by 59.351 μm and 160.992 μm respectively when ICL of 12.6mm and 13.2 mm were implanted (p = 0.0097 and p = 0.0124). For prediction of vault height, ML based model significantly outperformed traditional multivariable regression model.

Conclusion

We provide quantitative evidence that vertical implantation of ICL could effectively reduce the achieved vault height, especially when large size ICL was implanted, comparing to traditional horizontal implantation. ML is extremely applicable in development of vault prediction model.

Orbital and eyelid diseases: The next breakthrough in artificial intelligence?

Orbital and eyelid disorders affect normal visual functions and facial appearance, and precise oculoplastic and reconstructive surgeries are crucial. Artificial intelligence (AI) network models exhibit a remarkable ability to analyze large sets of medical images to locate lesions. Currently, AI-based technology can automatically diagnose and grade orbital and eyelid diseases, such as thyroid-associated ophthalmopathy (TAO), as well as measure eyelid morphological parameters based on external ocular photographs to assist surgical strategies. The various types of imaging data for orbital and eyelid diseases provide a large amount of training data for network models, which might be the next breakthrough in AI-related research. This paper retrospectively summarizes different imaging data aspects addressed in AI-related research on orbital and eyelid diseases, and discusses the advantages and limitations of this research field.

Approach to Orbital Fractures After Athletic Injuries

The bony orbit is commonly involved in athletic injuries. Evaluation should include a comprehensive history and ocular examination. Computed tomography imaging is the gold standard for diagnostic testing. Urgent surgical intervention for orbital floor fractures should occur after "white eye" trapdoor fractures or if oculocardiac response occurs. Most orbital fractures do not require urgent intervention and repair can be completed within 2 weeks of injury. There are many approaches to repair orbital fractures, and consideration of techniques depends on the unique fracture pattern. Intraoperative computed tomography has become frequently used and can lead to increased identification of plate malpositioning intraoperatively.

Deep Learning-Based CT Radiomics for Feature Representation and Analysis of Aging Characteristics of Asian Bony Orbit

This paper puts forward a new method for automatic segmentation of bony orbit as well as automatic extraction and classification of aging features of segmented orbit contour based on depth learning, with which the aging mode of bony orbit contour is preliminarily validated.

Improving rib fracture detection accuracy and reading efficiency with deep learning-based detection software: a clinical evaluation

OBJECTIVES

To investigate the impact of deep learning (DL) on radiologists' detection accuracy and reading efficiency of rib fractures on CT.

METHODS

Blunt chest trauma patients (n = 198) undergoing thin-slice CT were enrolled. Images were read by two radiologists (R1, R2) in three sessions: S1, unassisted reading; S2, assisted by DL as the concurrent reader; S3, DL as the second reader. The fractures detected by the readers and total reading time were documented. The reference standard for rib fractures was established by an expert panel. The sensitivity and false-positives per scan were calculated and compared among S1, S2, and S3.

RESULTS

The reference standard identified 865 fractures on 713 ribs (102 patients) The sensitivity of S1, S2, and S3 was 82.8, 88.9, and 88.7% for R1, and 83.9, 88.7, and 88.8% for R2, respectively. The sensitivity of S2 and S3 was significantly higher compared to S1 for both readers (all p < 0.05). The sensitivity between S2 and S3 did not differ significantly (both p > 0.9). The false-positive per scan had no difference between sessions for R1 (p = 0.24) but was lower for S2 and S3 than S1 for R2 (both p < 0.05). Reading time decreased by 36% (R1) and 34% (R2) in S2 compared to S1.

CONCLUSIONS

Using DL as a concurrent reader can improve the detection accuracy and reading efficiency for rib fracture.

ADVANCES IN KNOWLEDGE

DL can be integrated into the radiology workflow to improve the accuracy and reading efficiency of CT rib fracture detection.