An Artificial Intelligence Driven Approach for Classification of Ophthalmic Images using Convolutional Neural Network: An Experimental Study.

BACKGROUND: Early disease detection is emphasized within ophthalmology now more than ever, and as a result, clinicians and innovators turn to deep learning to expedite accurate diagnosis and mitigate treatment delay. Efforts concentrate on the creation of deep learning systems that analyze clinical image data to detect disease-specific features with maximum sensitivity. Moreover, these systems hold promise of early accurate diagnosis and treatment of patients with common progressive diseases. DenseNet, ResNet, and VGG-16 are among a few of the deep learning Convolutional Neural Network (CNN) algorithms that have been introduced and are being investigated for potential application within ophthalmology. METHODS: In this study, the authors sought to create and evaluate a novel ensembled deep learning CNN model that analyzes a dataset of shuffled retinal color fundus images (RCFIs) from eyes with various ocular disease features (cataract, glaucoma, diabetic retinopathy). Our aim was to determine (1) the relative performance of our finalized model in classifying RCFIs according to disease and (2) the diagnostic potential of the finalized model to serve as a screening test for specific diseases (cataract, glaucoma, diabetic retinopathy) upon presentation of RCFIs with diverse disease manifestations. RESULTS: We found adding convolutional layers to an existing VGG-16 model, which was named as a proposed model in this article that, resulted in significantly increased performance with 98% accuracy (p<0.05), including good diagnostic potential for binary disease detection in cataract, glaucoma, diabetic retinopathy. CONCLUSION: The proposed model was found to be suitable and accurate for a decision support system in Ophthalmology Clinical Framework.

The Impact of Surgical and Nonsurgical Interventions on Overall Oral, Palpebral Fissure, and Brow Height Symmetry in a Patient with Facial Synkinesis.

Background: The treatment of facial nerve synkinesis remains complex and variable. Objective: To compare oral, palpebral fissure, and brow symmetry among surgical and nonsurgical interventions in patients with facial synkinesis. Methods: Patients with facial nerve synkinesis at a single tertiary care center between 2008 and 2022 were analyzed before and after interventions using Emotrics software. Symmetry was compared among treatment combination groups (chemodenervation and rehabilitation [CR] vs. chemodenervation and surgery [CS] vs. chemodenervation, surgery, and rehabilitation [CSR]) and among surgical intervention groups (selective neurectomy [SN] vs. selective neurectomy with facelift [SnFa] vs. no surgery). Results: Of the 29 patients meeting inclusion criteria, 72.4% were female and the median age was 60.6 years (interquartile range 49.9-67.6). The median follow-up was 32.6 months; patients who received surgery had a greater follow-up time (57.4 months vs. 26.5 months, p = 0.045). The use of a trimodal approach (CSR) was associated with improved symmetry versus CR for smile angle (p = 0.021). Among surgical interventions, the greatest improvement in palpebral fissure symmetry was in patients who received SN versus no surgery (p = 0.039); the greatest improvement in smile angle was in patients who received SnFa versus no surgery (p = 0.008). Conclusion: We recommend a comprehensive approach to the management of facial synkinesis consisting of chemodenervation, rehabilitation, and surgery tailored to each patient's needs.

Recent Advances in Imaging Macular Atrophy for Late-Stage Age-Related Macular Degeneration.

Age-related macular degeneration (AMD) is a leading cause of blindness worldwide. In late-stage AMD, geographic atrophy (GA) of dry AMD or choroidal neovascularization (CNV) of neovascular AMD eventually results in macular atrophy (MA), leading to significant visual loss. Despite the development of innovative therapies, there are currently no established effective treatments for MA. As a result, early detection of MA is critical in identifying later central macular involvement throughout time. Accurate and early diagnosis is achieved through a combination of clinical examination and imaging techniques. Our review of the literature depicts advances in retinal imaging to identify biomarkers of progression and risk factors for late AMD. Imaging methods like fundus photography; dye-based angiography; fundus autofluorescence (FAF); near-infrared reflectance (NIR); optical coherence tomography (OCT); and optical coherence tomography angiography (OCTA) can be used to detect and monitor the progression of retinal atrophy. These evolving diverse imaging modalities optimize detection of pathologic anatomy and measurement of visual function; they may also contribute to the understanding of underlying mechanistic pathways, particularly the underlying MA changes in late AMD.

Faricimab Effectively Resolves Intraretinal Fluid and Preserves Vision in Refractory, Recalcitrant, and Nonresponsive Neovascular Age-Related Macular Degeneration.

PURPOSE: To evaluate the functional and anatomic outcomes of faricimab treatment in patients with neovascular age-related macular degeneration (nAMD) who are unresponsive to other anti-vascular endothelial growth factor (VEGF) therapies. METHODS: A retrospective interventional study was conducted on patients with refractory nAMD who were initially treated with intravitreal bevacizumab, ranibizumab, or aflibercept. These patients were switched to monthly faricimab injections. The central subfield thickness (CST), intraretinal fluid (IRF) or subretinal fluid (SRF) height, and visual acuities were compared before and after faricimab treatment. RESULTS: A total of 13 eyes (eight right eyes and five left eyes) from 11 patients were followed for 10.4 ± 6.9 months after bevacizumab treatment and 40.3 ± 28.7 months after aflibercept treatment before switching to faricimab. The follow-up time for patients receiving a mean number of 3.7 ± 1.3 faricimab injections was 3.4 ± 1.2 months. The overall median CST was reduced by 18µm (p=0.001) from 342µm to 318µm, along with a reduction of 89µm (p=0.03) in IRF/SRF height from 97µm to 40µm. Following three consecutive injections, the CST showed a significant reduction of 21.5µm (p=0.004) from 344µm to 322.5µm, and IRF/SRF height was reduced by 89µm (p=0.03) from 104µm to 18.5µm. The intraretinal fluid size decreased and leakage stopped, as seen on fluorescein angiography. Visual acuity remained stable after switching to faricimab treatment (0.59 ± 0.45 logMAR vs 0.58 ± 0.45 logMAR, p=1). CONCLUSIONS: Faricimab has proven to be an effective treatment for nAMD patients resistant to other anti-VEGF agents. It demonstrates significant anatomical improvement and vision preservation in this challenging patient population.

Variance in 3D anatomic localization of surgical margins based on conventional margin labeling in head and neck squamous cell carcinoma.

OBJECTIVE: In head and neck cancer (HNC), positive margins are strongly predictive of treatment failure. We sought to measure the accuracy of localization of margin sampling sites based on conventional anatomic labels using a digital 3D-model. METHODS: Preoperative CT scans for 9 patients with HNC treated operatively at our institution were imported into a multiplanar radiology software, which was used to render a digital 3D model of each tumor intended to represent the resection specimen. Surgical margin labels recorded during the operative case were collected from pathology records. Margin labels (N = 64) were presented to participating physicians.Participants were asked to mark the anatomic location of each surgical margin using the 3D-model and corresponding radiographic planes for reference.For each individual margin, the 3D coordinates of each participant's marker were used to calculate a mean localization point called the geometric centroid. Mean distance from individual markers to the centroid was compared between participantsand margin types. RESULTS: Amongst 7 surgeons, markers were placed a mean distance of 12.6 mm ([SD] = 7.5) from the centroid.Deep margins were marked with a greater mean distance than mucosal/skin margins (19.6 [24.8] mm vs. 15.3 [14.9] mm, p = 0.034). When asked to relocate a margin following re-resection, surgeons marked a point an average of 20.6 [12.4] mm from their first marker with a range of 3.9- 45.1 mm. CONCLUSIONS: Retrospective localization of conventionally labeled margins is an imprecise process with variability across the care team. Future interventions targeting margin documentation and communication may improve sampling precision.

Recurrence of premalignant oral cavity and oropharynx lesions after pulsed diode laser treatment.

PURPOSE: Leukoplakia is common with a 1 % incidence in the population and may harbor preneoplastic changes. Diode lasers provide both precision and coagulation for excision of superficial lesions in clinical and operative settings with little damage to deeper tissue. We aim to determine the rate of oral and oropharyngeal hyperkeratosis and dysplasia recurrence after treatment with diode laser. MATERIALS AND METHODS: Patients with oral or oropharyngeal hyperkeratotic or mild dysplastic lesions treated with pulsed diode laser between 2013 and 2020 at a tertiary academic institution were analyzed. The main outcome measure was recurrence of hyperkeratosis and dysplasia after treatment. RESULTS: Fourteen patients received diode laser treatment for hyperkeratotic or mild dysplastic lesions of the oral cavity or oropharynx. Demographic features included 9 (64.3 %) females and mean age of 66.6 years. In these 14 patients, 18 distinct lesions were identified. Eleven (61.1 %) lesions were located on the oral tongue, 4 (22.2 %) on the buccal mucosa, 2 (11.1 %) on the hard palate, and 1 (5.6 %) on the soft palate. Average time from lesion clinical diagnosis to the first diode laser treatment was 8.3 months with an average number of 1.4 treatments per lesion. Three (16.7 %) lesions experienced recurrence after the most recent treatment. None of the lesions underwent malignant transformation. None of the patients experienced bleeding, tethering, or dysarthria after treatment. One patient developed pyogenic granuloma and reported chronic tongue pain. CONCLUSIONS: Pulsed diode laser treatment of leukoplakia was well tolerated with low complication rates and reasonable control of precancerous lesions.

Rate of Fat Graft Volume Loss After Parotidectomy.

BACKGROUND: Reconstruction after parotidectomy can include fat grafting, which allows for symmetry, but grafts have demonstrated volume loss over time. OBJECTIVES: To provide quantitative evidence for the rate of volume loss of fat grafts. METHODS: Patients who received parotidectomy with fat graft reconstruction at a single institution from August 2016 to October 2020 were identified. Relationships between clinical factors and the logarithmic rate of fat graft volume loss were analyzed. RESULTS: Twelve patients received parotidectomy, fat graft reconstruction, and underwent a postoperative magnetic resonance imaging (MRI) scan. Rate of fat graft volume loss was a mean of 1.8% per month (standard deviation [SD]: 2.1% per month). Total parotid fat graft volume loss was a mean of 57.4% (SD: 67.5%). The mean follow-up time was 35.5 months (range: 9-89.8 months). Correlations between body mass index (BMI), history of smoking, and history of alcohol consumption and logarithmic rates of fat graft volume loss were increased but not significantly. CONCLUSIONS: Fat grafts have the potential of 60% volume loss at approximately 1 year. If there is clinical suspicion that patients will require adjuvant radiation or have clinical factors such as a smoking or alcohol-use history, volume requirements may be even greater to maintain adequate parotid volume for aesthetic purposes.

Otologic opioid usage and pain control in the postoperative period: An observational prospective study.

PURPOSE: Evaluate opioid prescribing patterns, opioid consumption, and patient pain patterns following otologic surgery. MATERIALS AND METHODS: Patients were included if they were ≥18 years old and received otologic surgery between November 2019 and August 2020. Patients were provided a survey which included a visual analog scale for recording their pain postoperatively and the amount of opioid they had remaining. Patients who did not complete all portions of the survey were excluded. RESULTS: Ninety-one patients completed the post-operative questionnaire. Collectively, patients were prescribed 5797 morphine milligram equivalents and used 3092: approximately 47% went unused. Of patients receiving a transcanal incision (n = 28/91, 31%), 70% went unused, whereas patients receiving a postauricular incision (n = 57/91, 63%), 38% went unutilized. The utilization difference between transcanal and postauricular cohorts was significant (p = 0.002). On multivariate analysis, patients who received a postauricular incision had 60% more opioid usage (p < 0.001), whereas those with a transcanal incision had an average reduction of 40% in opioid usage (p < 0.001). CONCLUSIONS: A significant amount of opioid medication went unused in this study. Patients with postauricular incisions had significantly increased opioid utilization as compared to those with transcanal incisions. Otologists may be able to successfully manage pain in the postoperative period with a reduced opioid prescription multimodal analgesia and increased patient education. Further study is needed to support this suggestion.

Depressed Mood Modulates Impact of Chronic Rhinosinusitis Symptoms on Quality of Life.

BACKGROUND: It is unknown how severity of depressed mood affects the well-recognized relationship between chronic rhinosinusitis (CRS) symptom burden and decreased general health-related quality of life (QOL). OBJECTIVE: The objective of this study was to determine whether depressed mood would affect the relationship between CRS symptom burden and decreased general health-related QOL. METHODS: For this cross-sectional study, 610 participants with CRS were prospectively recruited. CRS symptom burden was measured with the 22-item Sino-Nasal Outcome Test (SNOT-22). General health-related QOL was measured with the EuroQol 5-dimensional health utility value (EQ-5D HUV) and visual analog scale (EQ-5D VAS). Depressed mood was measured using the Patient Health Questionnaire-2 (PHQ-2). Participants were stratified as having well-controlled CRS symptoms (SNOT-22 < 35) and poorly controlled CRS symptoms (SNOT-22 ≥ 35). Good general health-related QOL was determined as EQ-5D HUV ≥ 0.9 or EQ-5D VAS ≥ 80, in contrast to low general health-related QOL. RESULTS: In participants with well-controlled CRS symptoms, both SNOT-22 and PHQ-2 were significantly and negatively associated with good general health-related QOL (P < .020 in all cases using multivariable regression). In participants with poorly controlled CRS symptoms, the PHQ-2 was significantly and negatively associated with good general health-related QOL using the EQ-5D HUV (adjusted odds ratio [OR] = 0.37, 95% confidence interval [CI]: 0.22-0.62, P < .001) or the EQ-5D VAS (adjusted OR = 0.74, 95% CI: 0.57-0.96, P = .024). In contrast, SNOT-22 was not associated with general health-related QOL in these patients. CONCLUSIONS: Depressed mood modulates the association between CRS symptom burden and general health-related QOL. Our results suggest a threshold of CRS symptom burden or control, beyond which depressed mood-not CRS symptom burden-drives the association with general health-related QOL.

Relationship between chronic rhinosinusitis exacerbation frequency and asthma control.

OBJECTIVES/HYPOTHESIS: To determine the association between the frequency of acute chronic rhinosinusitis (CRS) exacerbations (AECRS) and the degree of asthma control in asthmatic CRS patients. STUDY DESIGN: Cross-sectional study. METHODS: We prospectively recruited 108 asthmatic CRS patients as participants. Asthma control was assessed using the Asthma Control Test (ACT). The frequency of AECRS was assessed using three previously described indirect metrics for AECRS: the frequency of patient-reported sinus infections, CRS-related antibiotics use, and CRS-related oral corticosteroids use in the last 3 months. CRS symptom severity was measured using the 22-item Sinonasal Outcome Test (SNOT-22). Associations between ACT score and metrics for AECRS were performed using linear regression while controlling for clinical and demographic characteristics, including SNOT-22 score. RESULTS: ACT score was significantly and negatively associated with the frequency of patient-reported sinus infections (adjusted linear regression coefficient [ß] = -1.2, 95% confidence interval [CI]: -2.3 to -0.1, P = .033), CRS-related antibiotics courses (adjusted ß = -1.4, 95% CI: -2.3 to -0.5, P = .004), and CRS-related oral corticosteroid courses (adjusted ß = -1.5, 95% CI: -2.5 to -0.5, P = .004) in the last 3 months, independent of characteristics including SNOT-22 score. Poor asthma control could be detected using one or more sinus infections (70.6% sensitivity, 47.3% specificity), CRS-related antibiotics (50.0% sensitivity, 73.0% specificity), or CRS-related oral corticosteroids (58.8% sensitivity, 71.6% specificity) in the last 3 months. CONCLUSIONS: AECRS are negatively associated with the level of asthma control in asthmatic CRS patients, independent of CRS symptom severity. These results highlight AECRS as a distinct clinical manifestation of CRS that should be routinely assessed in CRS patients. LEVEL OF EVIDENCE: 2c. Laryngoscope, 128:1033-1038, 2018.