In a previous study, we identified biocular asymmetries in fundus photographs, and macula was discriminative area to distinguish left and right fundus images with > 99.9% accuracy. The purposes of this study were to investigate whether optical coherence tomography (OCT) images of the left and right eyes could be discriminated by convolutional neural networks (CNNs) and to support the previous result. We used a total of 129,546 OCT images. CNNs identified right and left horizontal images with high accuracy (99.50%). Even after flipping the left images, all of the CNNs were capable of discriminating them (DenseNet121: 90.33%, ResNet50: 88.20%, VGG19: 92.68%). The classification accuracy results were similar for the right and left flipped images (90.24% vs. 90.33%, respectively; p = 0.756). The CNNs also differentiated right and left vertical images (86.57%). In all cases, the discriminatory ability of the CNNs yielded a significant p value (< 0.001). However, the CNNs could not well-discriminate right horizontal images (50.82%, p = 0.548). There was a significant difference in identification accuracy between right and left horizontal and vertical OCT images and between flipped and non-flipped images. As this could result in bias in machine learning, care should be taken when flipping images.

Purpose: This study introduces a new machine learning-based auto-merge program (HydraVersion) that automatically combines multiple ocular photographs into single nine-directional ocular photographs. We compared the accuracy and time required to generate ocular photographs between HydraVersion and PowerPoint.Methods: This was a retrospective study of 2,524 sets of 250 nine-directional ocular photographs (134 patients) between March 2016 and June 2022. The test dataset comprised 74 sets of 728 photographs (38 patients). We measured the time taken to generate nine-directional ocular photographs using HydraVersion and PowerPoint, and compared their accuracy.Results: HydraVersion correctly combined 71 (95.95%) of the 74 sets of nine-directional ocular photographs. The average working time for HydraVersion and PowerPoint was 2.40 ± 0.43 and 255.9 ± 26.7 seconds, respectively; HydraVersion was significantly faster than PowerPoint (<i>p</i> < 0.001).Conclusions: Strabismus and neuro-ophthalmology centers are often unable to combine and store photographs, except those of clinically significant cases, because of a lack of time and manpower. This study demonstrated that HydraVersion may facilitate treatment and research because it can quickly and conveniently generate nine-directional ocular photographs.

Purpose: We report the case of a child with idiopathic intracranial hypertension who presented with binocular papillary edema and monocular sixth cranial nerve palsy accompanied by empty sella syndrome evident on brain magnetic resonance imaging.Case summary: A 9-year-old, normal-weight male patient visited the emergency room complaining of headache and diplopia 4 days in duration. The alternative prism cover test revealed esotropia of 16 prism diopters and a -1 right lateral gaze limitation. A fundus examination revealed papilledema and peripapillary hemorrhages in both eyes, and a visual field examination an enlarged, physiological blind spot in the right eye. Brain magnetic resonance imaging revealed elevated cerebrospinal fluid pressure, an empty sella, and posterior scleral flattening. We diagnosed and treated idiopathic intracranial hypertension. After 4 months, the papilledema and peripapillary hemorrhages of both eyes resolved, and the right lateral gaze limitation improved. The empty sella improved on brain magnetic resonance imaging, and we noted no recurrence 8 months after treatment.Conclusions: If a child with suspected idiopathic intracranial hypertension visits a hospital, but it is difficult to perform a lumbar puncture, brain magnetic resonance imaging should be scheduled. If abnormalities are found, these help to determine the course of disease.