Deep learning model for extensive smartphone-based diagnosis and triage of cataracts and multiple corneal diseases.

AIM: To develop an artificial intelligence (AI) algorithm that diagnoses cataracts/corneal diseases from multiple conditions using smartphone images. METHODS: This study included 6442 images that were captured using a slit-lamp microscope (6106 images) and smartphone (336 images). An AI algorithm was developed based on slit-lamp images to differentiate 36 major diseases (cataracts and corneal diseases) into 9 categories. To validate the AI model, smartphone images were used for the testing dataset. We evaluated AI performance that included sensitivity, specificity and receiver operating characteristic (ROC) curve for the diagnosis and triage of the diseases. RESULTS: The AI algorithm achieved an area under the ROC curve of 0.998 (95% CI, 0.992 to 0.999) for normal eyes, 0.986 (95% CI, 0.978 to 0.997) for infectious keratitis, 0.960 (95% CI, 0.925 to 0.994) for immunological keratitis, 0.987 (95% CI, 0.978 to 0.996) for cornea scars, 0.997 (95% CI, 0.992 to 1.000) for ocular surface tumours, 0.993 (95% CI, 0.984 to 1.000) for corneal deposits, 1.000 (95% CI, 1.000 to 1.000) for acute angle-closure glaucoma, 0.992 (95% CI, 0.985 to 0.999) for cataracts and 0.993 (95% CI, 0.985 to 1.000) for bullous keratopathy. The triage of referral suggestion using the smartphone images exhibited high performance, in which the sensitivity and specificity were 1.00 (95% CI, 0.478 to 1.00) and 1.00 (95% CI, 0.976 to 1.000) for 'urgent', 0.867 (95% CI, 0.683 to 0.962) and 1.00 (95% CI, 0.971 to 1.000) for 'semi-urgent', 0.853 (95% CI, 0.689 to 0.950) and 0.983 (95% CI, 0.942 to 0.998) for 'routine' and 1.00 (95% CI, 0.958 to 1.00) and 0.896 (95% CI, 0.797 to 0.957) for 'observation', respectively. CONCLUSIONS: The AI system achieved promising performance in the diagnosis of cataracts and corneal diseases.

Deep-learning approach to detect childhood glaucoma based on periocular photograph.

Childhood glaucoma is one of the major causes of blindness in children, however, its diagnosis is of great challenge. The study aimed to demonstrate and evaluate the performance of a deep-learning (DL) model for detecting childhood glaucoma based on periocular photographs. Primary gaze photographs of children diagnosed with glaucoma with appearance features (corneal opacity, corneal enlargement, and/or globe enlargement) were retrospectively collected from the database of a single referral center. DL framework with the RepVGG architecture was used to automatically recognize childhood glaucoma from photographs. The average receiver operating characteristic curve (AUC) of fivefold cross-validation was 0.91. When the fivefold result was assembled, the DL model achieved an AUC of 0.95 with a sensitivity of 0.85 and specificity of 0.94. The DL model showed comparable accuracy to the pediatric ophthalmologists and glaucoma specialists in diagnosing childhood glaucoma (0.90 vs 0.81, p = 0.22, chi-square test), outperforming the average of human examiners in the detection rate of childhood glaucoma in cases without corneal opacity (72% vs. 34%, p = 0.038, chi-square test), with a bilateral corneal enlargement (100% vs. 67%, p = 0.03), and without skin lesions (87% vs. 64%, p = 0.02). Hence, this DL model is a promising tool for diagnosing missed childhood glaucoma cases.

Assessment of the corneal biomechanical features of granular corneal dystrophy type 2 using dynamic ultra-high-speed Scheimpflug imaging.

PURPOSE: To evaluate the corneal biomechanical features of eyes with granular corneal dystrophy type 2 (GCD2) by analyzing corneal biomechanical indices obtained using a Corvis ST (CST) dynamic ultra-high-speed Scheimpflug imaging device. METHODS: In this retrospective case-control study, 35 CST parameters were compared in normal eyes (control) and eyes of patients with GCD2 treated at Osaka University Hospital, Osaka, Japan. The parameters included the Corvis Biomechanical Index (CBI), which is important in differentiating eyes with keratoconus from normal eyes. We measured the deposition rates of lesions in the central 7-mm region of the eye and assessed the correlation between the deposition rate and the CBI. RESULTS: Twenty-one eyes with GCD2 and 23 control eyes were analyzed. Eyes with GCD2 showed significantly less corneal stiffness in 15 CST parameters than did control eyes. In particular, the CBI was remarkably higher in eyes with GCD2 than in control eyes (P = 0.000006). Additionally, the deposition rate and the CBI were positively correlated. CONCLUSIONS: GCD2 eyes had softer corneas than did control eyes in most biomechanical CST parameters, and one of the parameters (the CBI) was linked to the rate of deposited lesions. Since IOP may be underestimated in GCD2 eyes, management should be especially careful in GCD2 cases complicated by glaucoma.

A Case of Endonasal Endoscopic Surgery for Intraorbital Metastasis of Gastric Ring Cell Carcinoma.

Gastric signet ring cell carcinoma has well-known metastatic features, including peritoneal dissemination and carcinomatous lymphangitis of the lung, but no intraorbital metastases were reported previously. A woman in her 60s developed left eye pain, sudden vision loss, and headache 12 years after gastric cancer treatment. Symptoms did not improve despite steroid pulses. Craniotomy showed no malignant findings. The patient was referred to our department for symptomatic relief and biopsy due to the lack of a definitive diagnosis and no improvement in her ocular pain. Endonasal endoscopic surgery was performed for diagnostic purposes and to relieve symptoms through orbital decompression. Preoperative computed tomography examination revealed a tumor at the left medial orbit, extending to the orbital apex. Orbital decompression through the open left medial orbital wall was performed with biopsy of the intraorbital tumor. Pathological findings were consistent with metastatic signet ring cell carcinoma. Pain and subjective improvement of visual acuity were noted the day after surgery. Twelve months postoperatively, diplopia remains, but there has been no worsening of symptoms.

Skull Base Metastasis of Breast Cancer With Oculomotor and Trochlear Nerve Palsy.

Skull base metastatic tumors are rare. Breast cancer in particular can cause bone metastases after a long period of time. A 70-year-old woman presented with multiple cranial nerve palsy. Magnetic resonance imaging revealed a lesion that extended from the orbit to the base of the skull, and the patient was referred to our department. Ophthalmological evaluation showed left visual acuity impairment, left oculomotor nerve palsy, and left trochlear nerve palsy. Endoscopic biopsy performed 5 years after the completion of breast cancer treatment revealed skull base metastases. In unilateral multiple cranial nerve palsy, the possibility of skull base metastases should be considered.

[Two Cases of Venous Malformation of the Orbit Treated by Endoscopic Endonasal Surgery].

Venous malformation of the orbit(VMO), previously called orbital cavernous hemangioma, has been classified as a vascular malformation according to the International Society for the Study of Vascular Anomalies. Among various surgical approaches for VMO, endoscopic endonasal surgery(EES)has recently been developed, especially for those in the inferomedial quadrant of the orbit. Two 67-year-old and 69-year-old women presented with decreased visual acuity and visual field deficit, respectively. Their CT and MRI scans revealed retrobulbar masses, suggestive of the inferomedial type of VMO. The first case was diagnosed as an intraconal VMO, and subtotal removal was achieved through binostril EES using a two-surgeon four-handed technique after palliative partial resection through a prior frontal craniotomy. In the second case, diagnosed as an extraconal VMO, total en bloc removal was achieved using the same surgical technique as above. In both cases, the visual functions improved after the procedures, with uneventful postoperative courses. Although the inferomedial VMO is an uncommon type, EES is well indicated for this condition. The international consensus of surgical techniques and staging from a surgical point of view should be established in the near future.

Spontaneous orbital decompression in thyroid eye disease: new measurement methods and its influential factors.

PURPOSE: To evaluate spontaneous decompression of the medial orbital wall and orbital floor in thyroid eye disease using new measurement methods and to analyze the influential factors. METHODS: This retrospective study included 86 patients (172 sides). Regarding evaluation of spontaneous medial orbital decompression, an anteroposterior line was drawn between the posterior lacrimal crest and the junction between the ethmoid bone and corpus ossis sphenoidalis. The bulged and/or dented areas from that line were measured. Regarding evaluation of spontaneous orbital floor decompression, the length of the perpendicular distance from a line that was drawn between the inferior orbital rim and the orbital process of palatal bone to the tip of the superior bulge of the orbital floor was measured. RESULTS: Multivariate linear regression analysis revealed that the maximum cross-sectional areas of the superior rectus/levator palpebrae superioris complex (P = 0.020) and medial rectus muscle (P = 0.028) were influential factors for spontaneous decompression of medial orbital wall (adjusted r2 = 0.090; P < 0.001), whereas the number of cycles of steroid pulse therapy (P = 0.002) and the maximum cross-sectional area of the inferior rectus muscle (P = 0.007) were the ones for that of the orbital floor (adjusted r2 = 0.096; P < 0.001). CONCLUSION: We believe that the identification of multiple influential factors of spontaneous decompression of the medial orbital wall and orbital floor will be helpful for better understanding and planned management of thyroid eye disease patients undergoing orbital decompression surgery.

Corticosteroid Withdrawal after Complete Resection of Recurrent IgG4-Related Ophthalmic Disease.

Corticosteroids are the first-line treatment for immunoglobulin G4-related ophthalmic disease (IgG4-ROD). However, the therapeutic effect of corticosteroids is unpredictable, and recurrences occur frequently. We present a case of a 60-year-old woman with IgG4-ROD of the lacrimal fossa. She was first treated with oral prednisolone, which was tapered within three months. However, proptosis recurred two months after prednisolone withdrawal. En bloc excision of the lesion was performed via the eyelid crease approach. The patient is lesion-free without corticosteroids at 18 months after surgery. Complete resection can be a powerful strategy for IgG4-ROD if a well-defined lesion is located in accessible areas.

The Role of Overriding Preseptal Orbicularis Oculi Muscle in Development of Involutional Lower Eyelid Entropion: Microscopic Viewpoints.

The aim of this study was to demonstrate the extent of preseptal orbicularis oculi muscle (OOM) override onto the pretarsal OOM in the lower eyelids. In this experimental microscopic study, 22 exenterated specimens from 22 Japanese cadavers were prepared as full-thickness sagittal sections. All exenterated specimens were devoid of lower eyelid entropion. The tarsal dimension and distance from the lower tarsal edge to the tip of overriding OOM were microscopically measured. The rotation axis of the tarsus was presumed to pass through the tarsal centroid and the distance from an estimated tarsal centroid to the lower tarsal edge was calculated. Consequently, 2 eyelids did not exhibit any overriding of the OOM. In the remaining 20 eyelids with OOM override, the average distance of the overriding OOM was found at a level covering 48.0% of the whole tarsal height. In 15 out of the 20 eyelids, the tip of the overriding OOM was located lower than the centroid. These microscopic findings suggest that overriding of the preseptal OOM does not always produce an inward rotational force on the tarsus.

Impact of Canthal Height on Horizontal Upper Eyelid Peak Position: A Guide for Blepharoptosis Surgery.

This study was aimed at examining the relationship between the horizontal upper eyelid peak position (Δpeak-pupil) and the difference between the medial and lateral canthal heights (Δcanthi) in Japanese individuals. Images of the left eyelid of 71 participants without eyelid disease were acquired with a camera with a slit-lamp microscope. Multivariate linear regression analysis was performed to determine the influence of age, sex, distance from the pupil center to the upper eyelid margin (margin reflex distance-1), Δcanthi, and palpebral fissure width on Δpeak-pupil. In addition, Δpeak-pupil distribution in participants with Δcanthi <3 (Group A), 3 to 6 (Group B), and >6 mm (Group C) was calculated as the average±2x the standard deviation. The mean Δpeak-pupil was 1.12 ±â€Š1.29 mm (range: -1.44 to 6.12) and the mean Δcanthi was 4.50 ±â€Š1.71 mm (1.02-7.96). Δcanthi was a significant determinant factor of Δpeak-pupil (P < 0.0001), whereas age (P = 0.62), sex (P = 0.88), margin reflex distance-1 (P = 0.29), and the horizontal palpebral fissure width (P = 0.65) were not related to Δpeak-pupil. The normal Δpeak-pupil ranges were -1.47 to 1.52, -0.95 to 3.22, and -0.84 to 4.73 mm in Groups A, B, and C, respectively. There was a statistically significant difference between the Group A and B ranges and between the Group A and C ranges (Kruskal-Wallis test, P < 0.01). It was concluded that the normal eyelid peak position differs according to medial-to-lateral canthal positional difference. This finding may help intraoperative determination of the upper eyelid peak position in blepharoptosis surgery.

Computed Tomography-Based Prediction of Exophthalmos Reduction After Deep Lateral Orbital Wall Decompression for Graves' Orbitopathy.

PURPOSE: To examine the predictability of exophthalmos reduction using preoperative computed tomography (CT) in deep lateral orbital wall decompression for Graves' orbitopathy. METHODS: This was a retrospective, observational, case-control study conducted at a single institution. Forty-three patients (43 orbits) who were treated with deep lateral decompression with (27 patients) and without (16 patients) fat removal. Multivariate linear regression analyses were used to identify factors influencing exophthalmos reduction 3 months postoperatively. Variables investigated included age; smoking history; history of corticosteroid therapy and/or radiotherapy; preoperative diplopia; width, depth, and height of the greater wing of the sphenoid bone (trigone); and volume of fat removed. The influence of these parameters on the area of the orbital opening was also analyzed. RESULTS: Trigone width and amount of fat removed were positively correlated with exophthalmos reduction (both, P < .050); none of the other variables exhibited significant correlations. The predictive equation for postoperative reduction in Hertel exophthalmometric values was 0.57 + 0.15 × trigone width (mm) + 1.1 × fat removed (cm3). Depth and height were positively correlated (both, P < .050) and width was negatively correlated (P = .0045) with orbital opening area. CONCLUSIONS: Trigone width and amount of fat removed were positive predictors of exophthalmos reduction after deep lateral decompression. Trigone width was inversely associated with orbital width, and results showed that larger exophthalmos reductions could be achieved in narrower orbits. This CT-based prediction method will assist preoperative decision-making regarding additional fat removal and/or removal of another orbital wall.

A Case of Compressive Optic Neuropathy Secondary to Lymphoma of the Extraocular Muscles.

Compressive optic neuropathy is an urgent condition that may lead to blindness. We report a rare case of compressive optic neuropathy secondary to lymphoma of the extraocular muscles in which the vision was successfully rescued. A 67-year-old man presented with visual acuity of hand motion in the right eye for 1 month. Magnetic resonance imaging demonstrated enlargement of the medial and inferior rectus muscles with their tendons. Tendon enlargement, which is uncommon as Graves' orbitopathy, motivated us to perform a biopsy. Intravenous steroids following biopsy and chemotherapy after definitive diagnosis allowed for recovery of vision. Clinicians should be aware that enlarged extraocular muscles due to lymphoma can cause compressive optic neuropathy.

Involutional lower eyelid entropion: causative factors and therapeutic management.

PURPOSE: To summarize proposed causative factors and the outcomes of surgical practices for involutional lower eyelid entropion. METHODS: We reviewed the literature on proposed causative factors and the outcomes of surgical practices for involutional lower eyelid entropion, searched on PubMed. RESULTS: Vertical and horizontal laxities of the lower eyelid, and overriding of the preseptal orbicularis oculi muscle onto the pretarsal orbicularis oculi muscle have been proposed as the major causes of involutional lower eyelid entropion. Treatment procedures have been developed over the years to address one or more of these causative factors. CONCLUSIONS: Various causative factors and treatment procedures have been advocated to explain and correct involutional lower eyelid entropion. The appropriate procedure is chosen according to the patient's condition, such as the presence of vertical laxity, horizontal laxity, and orbicularis oculi muscle overriding. A combination of these procedures to correct multiple factors further decreases the recurrence rate.

Bony Orbital Decompression Following Lateral Canthotomy and Cantholysis for Traumatic Orbital Compartment Syndrome.

PURPOSE: To describe the clinical course of patients with traumatic orbital compartment syndrome who underwent bony orbital decompression due to persistently abnormal pupillary light reflex after lateral canthotomy and cantholysis. METHODS: Four consecutive patients were retrospectively reviewed. The authors performed bony orbital decompression as there was no improvement in the pupillary light reflex and a tight orbit persisted even after lateral canthotomy and cantholysis. RESULTS: The median interval between injury and bony decompression was 8.8 hours (range, 7-12 hours). All patients showed a preoperative intraocular pressure of 40 mm Hg or greater, which decreased to 20 mm Hg or less the next day. Two patients showed globe tenting with a posterior globe angle of 110° or less, which was resolved on the following day. Two patients with initial visual acuity of counting fingers or better showed complete visual recovery. By contrast, only 1 of the 2 patients with no light perception slightly improved to light perception while the other showed no improvement after surgery. CONCLUSIONS: Bony orbital decompression is effective for the treatment of traumatic orbital compartment syndrome in patients whose preoperative visual acuity is counting fingers or better.

Rescue technique for complete removal of an accidentally ruptured orbital dumbbell deep dermoid cyst: A case report.

PURPOSE: To report a rescue technique for complete removal of an accidentally ruptured orbital dumbbell deep dermoid cyst. OBSERVATIONS: A 33-year-old female presented with left proptosis with retrobulbar discomfort for 3 months. Computed tomography images showed an orbital dumbbell deep dermoid cyst. A lateral orbitotomy was performed under general anesthesia. The cyst was ruptured during osteotomy of the lateral orbital rim. The cyst was opened vertically from the ruptured site using a Stephen's tenotomy scissors to visually confirm the internal wall of the cyst and to keep the epithelial lining intact while separating the external wall of the cyst from the bone. There was a small defect of the epithelial lining at the inferoposterior margin of the cyst. Granulomatous inflammation of the lacrimal gland was found adjacent to the defect site. The cyst was completely removed and the lacrimal inflammation subsided after the operation. CONCLUSIONS AND IMPORTANCE: Opening of the cyst is a useful technique to remove the entire epithelial lining of an accidentally ruptured dumbbell dermoid cyst.

Eyelid Myokymia with Concomitant Cerebral Tumour: A Case Report.

Eyelid myokymia is a localised movement disorder of the orbicularis oculi muscle with involuntary, fine, continuous, and undulating contractions. Although this entity is thought to be peripheral nerve origin, it rarely occurs with an intracranial lesion. The authors report a case of eyelid myokymia with concomitant cerebral tumour. A 52-year-old woman had a 6-month history of left eyelid myokymia accompanied by upper eyelid ptosis and lower eyelid reverse ptosis. Magnetic resonance imaging showed a solid mass measuring 20 × 25 × 20 mm in the temporal lobe of the cerebral cortex, showing isointense on T1-weighted and hyperintense on T2-weighted images. The clinical diagnosis was cerebral astrocytoma.

Differences in Common Orbital Blowout Fracture Sites by Age.

BACKGROUND: The purpose of this study was to compare orbital blowout fracture sites among different age groups. METHODS: This retrospective, observational, comparative study included 671 sides of 663 patients with an orbital blowout fracture. The authors classified patients into five groups based on age (childhood, 0 to 9 years; adolescence, 10 to 18 years; early adulthood, 19 to 44 years; middle adulthood, 45 to 64 years; and late adulthood, 65 years or older); causes of injury into five categories (i.e., sports, traffic accident, assault, fall, and others); and orbital blowout fracture sites into four regions [i.e., the orbital floor lateral (lateral floor) or medial to the infraorbital groove/canal (medial floor); the inferomedial orbital strut (strut); and the medial orbital wall (medial wall)]. RESULTS: A single medial floor fracture was observed in 70.3 percent of patients in the childhood age group. However, the proportion of patients with a single medial wall fracture or multiple fracture sites increased with age. A medial wall fracture was more common than a medial floor fracture in late adulthood. The distribution of orbital blowout fracture sites, the ratio of single or multiple fracture sites, and the causes of injury were statistically different among the age groups (p = 0.005, p = 0.003, and p < 0.001, respectively, Pearson chi-square test). CONCLUSIONS: Common orbital blowout fracture sites varied among age groups. These differences may be attributed to multiple factors, including differences in the morphology of the orbit and paranasal sinuses, composition of the facial bone, and causes of injury among age groups.

Lagophthalmos caused by cicatricial adhesion of orbital adipose tissue to orbital roof: A case report.

PURPOSE: To report a case of lagophthalmos caused by cicatricial adhesion of orbital adipose tissue to the orbital roof. OBSERVATIONS: A 23-year-old female was presented with right lagophthalmos. Five months prior to consult at our clinic, she suffered from a penetrating trauma to the frontal lobe of the brain through the right orbital roof with cerebrospinal fluid leakage. Decompressive craniectomy was performed immediately after the injury using a coronal incision, which was followed by reconstruction with an artificial bone 1 month later. On examination at our clinic, she showed right exposure keratopathy with best corrected visual acuity of 20/100 due to corneal opacity. The palpebral contracted scar was first elongated using Z-plasty technique but excursion of the upper eyelid under a finger force assistance was insufficient to eliminate lagophthalmos. However, complete eyelid closure under a finger force assistance was achieved after sharp dissection of the cicatrized adipose tissue from the orbital roof. An autogenous dermis-fat was grafted on the orbital roof and superior orbital rim to avoid adhesion of orbital adipose tissue onto the bone again. CONCLUSION AND IMPORTANCE: Cicatricial adhesion of an orbital adipose tissue to the orbital roof is one of the possible causes of posttraumatic lagophthalmos in patients with an orbital roof fracture. Surgeons need to be aware of this condition in planning of surgical repair when such a fracture is encountered.