A real-world study for timely assessing the diabetic macular edema refractory to intravitreal anti-VEGF treatment.

Background: Early Identifying and characterizing patients with diabetic macular edema (DME) is essential for individualized treatment and outcome optimization. This study aimed to timely investigate optical coherence tomography (OCT) biomarkers of DME refractory to intravitreal anti-vascular endothelial growth factor (VEGF) therapy. Methods: We retrospective reviewed 72 eyes from 44 treatment-naïve patients who were treated with intravitreal anti-VEGF for DME. OCT scans prior to anti-VEGF were evaluated for serous retinal detachment (SRD), size of outer nuclear layer cystoid changes, diffuse retinal thickening, integrity of the inner segment-outer segment (IS-OS) junction, quantity and location of hyperreflective foci, vitreomacular interface abnormalities, and epiretinal membrane (ERM). The Baseline best-corrected visual acuity (BCVA) and central macular thickness was recorded at baseline and 4 months after treatment with anti-VEGF. The main outcome measure was the correlation between spectral-domain OCT measurements and BCVA response at baseline and after anti-VEGF treatment (mean change from baseline; ≥ 10 Early Treatment Diabetic Retinopathy Study letters in BCVA). Results: Partially continuous IS-OS layers (partially vs. completely continuous: ß, -0.138; Wald chi-square, 16.392; P<0.001) was predictor of better response to anti-VEGF treatment. In contrast, ERM (present vs. absent ERM: ß, 0.215; Wald chi-square, 5.921; P=0.015) and vitreomacular traction (vitreomacular traction vs. posterior vitreous detachment: ß=0.259; Wald chi-square=5.938; P=0.015) were the predictors of poor response. The improvement of BCVA trended toward the OCT predictive value of central macular thickness reduction; however, this was not significant. Conclusion: Partially continuous IS-OS layers is predictive of better response to anti-VEGF therapy in DME. Meanwhile, ERM is a significant predictor of poor response.

Association between vaginal or cesarean delivery and later development of stress urinary incontinence or pelvic organ prolapse: A retrospective population-based cohort study.

INTRODUCTION AND HYPOTHESIS: Child delivery mode may be associated with pelvic floor disorders. We explored the association between different delivery modes and later development of stress urinary incontinence (SUI) and pelvic organ prolapse (POP) in Taiwanese women. METHODS: This was a retrospective population-based cohort study. Women who delivered babies between January 1, 2000, and December 31, 2018, were selected for this study. The study used Taiwan's National Health Insurance (NHI) Research Database. After propensity score matching, 51,587 women who underwent cesarean section (C/S) and 51,587 women who underwent vaginal delivery (VD) were recruited. Primary outcomes were the presence of SUI and POP after delivery. RESULTS: The incidence of SUI (1.6/1000 person-years) and POP (1.5/1000 person-years) was higher in the VD group than in the C/S group (0.8 and 0.6 in 1000 person-years). VD was associated with an increased risk of SUI [hazard ratio (HR): 2.79, 95% confidence interval (CI): 2.45-3.17] and POP (HR: 1.96, 95% CI: 1.75-2.19) compared to C/S. We also found that age (HR: 1.06, 95% CI: 1.05-1.08 in SUI, HR: 1.08, 95% CI: 1.07-1.09 in POP) and Charlson Comorbidity Index (CCI) (HR: 1.28, 95% CI: 1.12-1.46 in SUI, HR: 1.27, 95% CI: 1.13-1.43 in POP) were associated with an increased risk of SUI and POP. The cumulative incidence of SUI and POP was higher in the VD group than in the C/S group (log-rank test, P < 0.05). CONCLUSIONS: The current study was the largest retrospective cohort study regarding the influence of delivery mode on SUI and POP so far. VD was found to be associated with an increased risk of SUI and POP compared with C/S. Postpartum care for pelvic physical therapy should be provided particularly to women undergoing VD.

Automatic Meter Reading from UAV Inspection Photos in the Substation by Combining YOLOv5s and DeeplabV3.

The combination of unmanned aerial vehicles (UAVs) and artificial intelligence is significant and is a key topic in recent substation inspection applications; and meter reading is one of the challenging tasks. This paper proposes a method based on the combination of YOLOv5s object detection and Deeplabv3+ image segmentation to obtain meter readings through the post-processing of segmented images. Firstly, YOLOv5s was introduced to detect the meter dial area and the meter was classified. Following this, the detected and classified images were passed to the image segmentation algorithm. The backbone network of the Deeplabv3+ algorithm was improved by using the MobileNetv2 network, and the model size was reduced on the premise that the effective extraction of tick marks and pointers was ensured. To account for the inaccurate reading of the meter, the divided pointer and scale area were corroded first, and then the concentric circle sampling method was used to flatten the circular dial area into a rectangular area. Several analog meter readings were calculated by flattening the area scale distance. The experimental results show that the mean average precision of 50 (mAP50) of the YOLOv5s model with this method in this data set reached 99.58%, that the single detection speed reached 22.2 ms, and that the mean intersection over union (mIoU) of the image segmentation model reached 78.92%, 76.15%, 79.12%, 81.17%, and 75.73%, respectively. The single segmentation speed reached 35.1 ms. At the same time, the effects of various commonly used detection and segmentation algorithms on the recognition of meter readings were compared. The results show that the method in this paper significantly improved the accuracy and practicability of substation meter reading detection in complex situations.