Deep Learning Model for Static Ocular Torsion Detection Using Synthetically Generated Fundus Images.

Purpose: The objective of the study is to develop deep learning models using synthetic fundus images to assess the direction (intorsion versus extorsion) and amount (physiologic versus pathologic) of static ocular torsion. Static ocular torsion assessment is an important clinical tool for classifying vertical ocular misalignment; however, current methods are time-intensive with steep learning curves for frontline providers. Methods: We used a dataset (n = 276) of right eye fundus images. The disc-foveal angle was calculated using ImageJ to generate synthetic images via image rotation. Using synthetic datasets (n = 12,740 images per model) and transfer learning (the reuse of a pretrained deep learning model on a new task), we developed a binary classifier (intorsion versus extorsion) and a multiclass classifier (physiologic versus pathologic intorsion and extorsion). Model performance was evaluated on unseen synthetic and nonsynthetic data. Results: On the synthetic dataset, the binary classifier had an accuracy and area under the receiver operating characteristic curve (AUROC) of 0.92 and 0.98, respectively, whereas the multiclass classifier had an accuracy and AUROC of 0.77 and 0.94, respectively. The binary classifier generalized well on the nonsynthetic data (accuracy = 0.94; AUROC = 1.00). Conclusions: The direction of static ocular torsion can be detected from synthetic fundus images using deep learning methods, which is key to differentiate between vestibular misalignment (skew deviation) and ocular muscle misalignment (superior oblique palsies). Translational Relevance: Given the robust performance of our models on real fundus images, similar strategies can be adopted for deep learning research in rare neuro-ophthalmologic diseases with limited datasets.

Pearls & Oy-sters: Deep Phenotyping of Abnormal Eye Movements Advances the Detection of Gerstmann-Sträussler-Scheinker Syndrome.

A 58-year-old previously healthy woman presents with 3 years of rapidly progressive ataxia, parkinsonism, dysautonomia, peripheral neuropathy, leg weakness, spasticity, hyperreflexia, and mild vertical-gaze palsy. She has a matrilineal family history of neurodegenerative diseases. She was initially postulated to have spinocerebellar ataxia or atypical parkinsonism with cerebellar features. However, on closer inspection, her abnormal extraocular eye movements suggested rare mimicking disorders such as prion disease as part of the differential diagnosis, requiring further evaluation. This case highlights how deep phenotyping can open new diagnostic considerations, inform additional workup, and yield the precise diagnosis of Gerstmann-Sträussler-Scheinker syndrome (GSS).

aEYE: A deep learning system for video nystagmus detection.

Background: Nystagmus identification and interpretation is challenging for non-experts who lack specific training in neuro-ophthalmology or neuro-otology. This challenge is magnified when the task is performed via telemedicine. Deep learning models have not been heavily studied in video-based eye movement detection. Methods: We developed, trained, and validated a deep-learning system (aEYE) to classify video recordings as normal or bearing at least two consecutive beats of nystagmus. The videos were retrospectively collected from a subset of the monocular (right eye) video-oculography (VOG) recording used in the Acute Video-oculography for Vertigo in Emergency Rooms for Rapid Triage (AVERT) clinical trial (#NCT02483429). Our model was derived from a preliminary dataset representing about 10% of the total AVERT videos (n = 435). The videos were trimmed into 10-sec clips sampled at 60 Hz with a resolution of 240 × 320 pixels. We then created 8 variations of the videos by altering the sampling rates (i.e., 30 Hz and 15 Hz) and image resolution (i.e., 60 × 80 pixels and 15 × 20 pixels). The dataset was labeled as "nystagmus" or "no nystagmus" by one expert provider. We then used a filtered image-based motion classification approach to develop aEYE. The model's performance at detecting nystagmus was calculated by using the area under the receiver-operating characteristic curve (AUROC), sensitivity, specificity, and accuracy. Results: An ensemble between the ResNet-soft voting and the VGG-hard voting models had the best performing metrics. The AUROC, sensitivity, specificity, and accuracy were 0.86, 88.4, 74.2, and 82.7%, respectively. Our validated folds had an average AUROC, sensitivity, specificity, and accuracy of 0.86, 80.3, 80.9, and 80.4%, respectively. Models created from the compressed videos decreased in accuracy as image sampling rate decreased from 60 Hz to 15 Hz. There was only minimal change in the accuracy of nystagmus detection when decreasing image resolution and keeping sampling rate constant. Conclusion: Deep learning is useful in detecting nystagmus in 60 Hz video recordings as well as videos with lower image resolutions and sampling rates, making it a potentially useful tool to aid future automated eye-movement enabled neurologic diagnosis.

Clinical Reasoning: A 32-Year-Old Woman With Tunnel Vision and Back Pain.

The incidence of new onset visual disturbances in emergency departments across the country is frequent. A detailed history of events and thoughtful physical examination may produce a diagnosis; however, atypical cases may require further diagnostic testing to explain symptoms. We present a case of presumed increased intracranial pressure with atypical findings on diagnostic testing, which allowed our team to explore a broader differential diagnosis. This clinical reasoning article will benefit students, residents, and attendings alike to continue to uncover etiologies for symptoms of increased intracranial pressure and review differential diagnoses in similar presentations.

Pearls & Oy-sters: Vertical Diplopia and Ocular Torsion: Peripheral vs Central Localization.

We describe a case with torsional deviation of the eyes from a brainstem lesion. Torsional eye movement refers to changes in the position of the eyes in the roll plane around the visual axis. When the head is tilted laterally (that is, rolled towards the shoulder), the eyes roll in the opposite direction as part of the torsional vestibulo-ocular reflex known as the ocular counter-roll (OCR). Pathologies that affect the otolith-ocular pathway can lead to a torsional deviation of the eyes as part of the ocular tilt reaction (OTR) that also cause vertical deviation of the eyes (skew deviation) and head tilt. Lesions caudal to the pontomedullary junction (such as the labyrinth, eighth cranial nerve, or vestibular nucleus) result in an OTR with ipsiversive torsional deviation, whereas lesions rostral to the junction result in an OTR with contraversive torsional deviation. Furthermore, torsional deviation of the eyes in OTR is conjugate (incyclotorsion in the higher eye and excyclotorsion in the lower eye), whereas torsional deviation from ocular palsy in peripheral lesions is disconjugate (e.g., exyclotorsion only in the higher eye with superior oblique palsy). Therefore, the pattern of torsional eye deviation can be helpful in localizing the lesion. Several techniques including fundus photography, double Maddox rod testing, optical coherence tomography (OCT), and video-oculography are used to measure torsional eye position.

Opinion and Special Articles: Remote Evaluation of Acute Vertigo: Strategies and Technological Considerations.

Patients with acute vestibular disorders are often a diagnostic challenge for neurologists, especially when the evaluation must be conducted remotely. The clinical dilemma remains: Does the patient have a benign peripheral inner ear problem or a worrisome central vestibular disorder, such as a stroke? The use of a focused history and the virtual HINTS (head impulse test, nystagmus evaluation, and test of skew) examination are key steps towards correctly diagnosing and triaging the acute vertiginous patient. When looking for signs of vestibulo-ocular dysfunction, there are important technological and practical considerations for an effective clinical interpretation.

HINTS Examination in Acute Vestibular Neuritis: Do Not Look Too Hard for the Skew.

BACKGROUND: An ocular tilt reaction (OTR) is a triad of a skew deviation, head tilt, and ocular counter-roll that can be partial or complete. An OTR can occur anywhere along the utriculo-ocular motor pathways from the labyrinth to the interstitial nucleus of Cajal but is almost always central in origin. In acute vestibular neuritis (AVN), case reports have described patients with an OTR due to AVN, although it is unclear whether this examination finding is common or rare. METHODS: The vestibular and ocular motor features of 7 patients presenting with AVN are described. RESULTS: Each of the 7 patients presented with typical features of AVN, including contralesional unidirectional spontaneous nystagmus and an ipsilesional abnormal head impulse test, although each patient also had a complete OTR. None of the patients had vertical diplopia or a skew deviation that was measurable with alternate cover testing (i.e., abnormal "test of skew" according to the Head Impulse, Nystagmus, Test of Skew examination); however, all had a subtle 1 prism diopter hyperphoria that was only measurable with a Maddox rod test. CONCLUSION: Seven cases of typical AVN with an OTR are presented, and in the authors' experience, the presence of a subtle OTR is a common feature of AVN in these patients.

Consensus on Virtual Management of Vestibular Disorders: Urgent Versus Expedited Care.

The virtual practice has made major advances in the way that we care for patients in the modern era. The culture of virtual practice, consulting, and telemedicine, which had started several years ago, took an accelerated leap as humankind was challenged by the novel coronavirus pandemic (COVID19). The social distancing measures and lockdowns imposed in many countries left medical care providers with limited options in evaluating ambulatory patients, pushing the rapid transition to assessments via virtual platforms. In this novel arena of medical practice, which may form new norms beyond the current pandemic crisis, we found it critical to define guidelines on the recommended practice in neurotology, including remote methods in examining the vestibular and eye movement function. The proposed remote examination methods aim to reliably diagnose acute and subacute diseases of the inner-ear, brainstem, and the cerebellum. A key aim was to triage patients into those requiring urgent emergency room assessment versus non-urgent but expedited outpatient management. Physicians who had expertise in managing patients with vestibular disorders were invited to participate in the taskforce. The focus was on two topics: (1) an adequate eye movement and vestibular examination strategy using virtual platforms and (2) a decision pathway providing guidance about which patient should seek urgent medical care and which patient should have non-urgent but expedited outpatient management.

Eight Syndrome: Horizontal Gaze Palsy Plus Ipsilateral Seventh Nerve Palsy.

A 62-year-old woman developed a right horizontal gaze palsy and ipsilateral facial nerve palsy due to a right pontine tegmentum infarct. This constitutes a forme fruste of the eight-and-a-half syndrome that we have termed the eight syndrome.

Treatment of Blepharospasm/Hemifacial Spasm.

OPINION STATEMENT: The treatment of both hemifacial spasm (HFS) and blepharospasm (BEB) requires making the appropriate clinical diagnosis. Advance imaging and electrophysiologic studies are useful; however, one's clinical suspicion is paramount. The purpose of this review is to summarize current and emerging therapies for both entities. Botulinum toxin (BTX) remains the first-line therapy to treat both conditions. If chemodenervation has failed, surgery may be considered. Due to the risks associated with surgery, the benefits of this option must be carefully weighed. Better surgical outcomes are possible when procedures are performed at tertiary centers with experienced surgeons and advanced imaging techniques. Microvascular decompression is an efficacious method to treat HFS, and myectomy is an option for medication-refractory BEB; the risks of the latter may outweigh any meaningful clinical benefits. Oral agents only provide short-term relief and can cause several unwanted effects; they are reserved for patients who cannot receive BTX and/or surgery. Transcranial magnetic stimulation has gained some traction in the treatment of BEB and may provide safer non-invasive options for refractory patients in the future.