Deep learning-based optic disc classification is affected by optic-disc tilt.

We aimed to determine the effect of optic disc tilt on deep learning-based optic disc classification. A total of 2507 fundus photographs were acquired from 2236 eyes of 1809 subjects (mean age of 46 years; 53% men). Among all photographs, 1010 (40.3%) had tilted optic discs. Image annotation was performed to label pathologic changes of the optic disc (normal, glaucomatous optic disc changes, disc swelling, and disc pallor). Deep learning-based classification modeling was implemented to develop optic-disc appearance classification models with the photographs of all subjects and those with and without tilted optic discs. Regardless of deep learning algorithms, the classification models showed better overall performance when developed based on data from subjects with non-tilted discs (AUC, 0.988 ± 0.002, 0.991 ± 0.003, and 0.986 ± 0.003 for VGG16, VGG19, and DenseNet121, respectively) than when developed based on data with tilted discs (AUC, 0.924 ± 0.046, 0.928 ± 0.017, and 0.935 ± 0.008). In classification of each pathologic change, non-tilted disc models had better sensitivity and specificity than the tilted disc models. The optic disc appearance classification models developed based all-subject data demonstrated lower accuracy in patients with the appearance of tilted discs than in those with non-tilted discs. Our findings suggested the need to identify and adjust for the effect of optic disc tilt on the optic disc classification algorithm in future development.

Ganglion cell inner plexiform layer thickness measured by optical coherence tomography to predict visual outcome in chiasmal compression.

We evaluated the prognostic value of the preoperative macular ganglion cell inner plexiform layer (mGCIPL) thickness along with peripapillary retinal nerve fiber layer (pRNFL) thickness measured by optical coherence tomography (OCT) and estimated an optimal cut-off value to predict postoperative visual field (VF) recovery in adult patients with chiasmal compression after decompression surgery. Two hundred forty eyes of 240 patients aged 20 years or older for which preoperative high-definition Cirrus OCT parameters and pre- and postoperative visual function data were available. The prognostic power of pRNFL and mGCIPL thicknesses for complete postoperative VF recovery or significant VF improvement (improvement ≥ 2 dB in the mean deviation) were assessed. The cut-off values for OCT parameters for VF recovery were estimated. The study found that the higher the preoperative pRNFL and mGCIPL thicknesses, the higher the probability of complete postoperative VF recovery (p = 0.0378 and p = 0.0051, respectively) or significant VF improvement (p = 0.0436 and p = 0.0177, respectively). The area under the receiver operating characteristic analysis of preoperative OCT parameters demonstrated that the mGCIPL thickness showed an area under the curve (AUC) of more than 0.7 for complete VF recovery after decompression surgery (AUC = 0.725, 95% CI: 0.655, 0.795), and the optimal mGCIPL thickness cut-off value for complete VF recovery was 77.25 µm (sensitivity 69% and specificity 69%). Preoperative mGCIPL thickness was a powerful predictor of visual functional outcome after decompression surgery for chiasmal compression.

Liver enzymes and risk of ocular motor cranial nerve palsy: a nationwide population-based study.

This study aimed to assess the associations between liver enzymes including γ-glutamyl transferase (GGT) and the development of ocular motor cranial nerve palsy (CNP) using the National Sample Cohort database from Korea's National Health Insurance Service. We analyzed data from 4,233,273 medical screening examinees aged 20 years or more in 2009. Study participants were followed up until December 31, 2018. A Cox proportional hazard regression analysis was performed for quartiles of liver enzymes to determine the linkage between each value and ocular motor CNP using quartile 1 as a reference after adjusting for potential confounders. A total of 5,807 (0.14%) patients developed ocular motor CNP during the follow-up period of 8.22 ± 0.94 years. The incidence of ocular motor CNP gradually increased as the GGT levels increased. The highest quartile of the GGT group had hazard ratio (HR) of 1.245 (95% confidence interval [CI], 1.136-1.365). Regarding alanine aminotransferase (ALT), the highest quartile of the ALT group had HR of 1.141 (95% CI, 1.049-1.241). However, the incidence of ocular motor CNP did not gradually increase as the ALT levels increased. The coexistence of the increased level of GGT, metabolic syndrome, and obesity showed a stronger association with ocular motor CNP development (HR, 1.331; 95% CI, 1.173, 1.511) compared to having a single factor or two factors. In conclusion, our population-based cohort study demonstrated a significant association between serum GGT level and the incidence of ocular motor CNP, suggesting that GGT could be a new clinical marker for predicting the occurrence of ocular motor CNP.

Multimode shunt damping of piezoelectric smart panel for noise reduction.

Multimode shunt damping of piezoelectric smart panel is studied for noise reduction. Piezoelectric smart panel is a plate structure on which a piezoelectric patch is attached with an electrical shunt circuit. When an incidence sound is impinged on the panel structure, the structure vibrates and the attached piezoelectric patch produces an electrical energy, which can be effectively dissipated as heat via the electrical shunt circuit. Since the energy dissipation strongly depends on the vibration mode of the panel structure, many patches are required for multiple vibration modes. Instead of using multiple piezoelectric patches, a single piezoelectric patch is used in conjunction with a blocked shunt circuit for multimode shunt damping. Modeling, shunt parameter tuning, and implementation of the blocked shunt circuit along with an acoustic test of the panel are explained. A remarkable reduction of the transmitted noise was achieved for multiple modes of the panel. Since this technology has many merits in terms of compactness, low cost, robustness, and ease of installation, practical applications in many noise problems can be anticipated.