Detecting Hydronephrosis Through Ultrasound Images Using State-of-the-Art Deep Learning Models.

The goal of this study was to assess the feasibility of three models for detecting hydronephrosis through ultrasound images using state-of-the-art deep learning algorithms. The diagnosis of hydronephrosis is challenging because of varying and non-specific presentations. With the characteristics of ready accessibility, no radiation exposure and repeated assessments, point-of-care ultrasound becomes a complementary diagnostic tool for hydronephrosis; however, inter-observer variability still exists after time-consuming training. Artificial intelligence has the potential to overcome the human limitations. A total of 3462 ultrasound frames for 97 patients with hydronephrosis confirmed by the expert nephrologists were included. One thousand six hundred twenty-eight ultrasound frames were also extracted from the 265 controls who had normal renal ultrasonography. We built three deep learning models based on U-Net, Res-UNet and UNet++ and compared their performance. We applied pre-processing techniques including wiping the background to lessen interference by YOLOv4 and standardizing image sizes. Also, post-processing techniques such as adding filter for filtering the small effusion areas were used. The Res-UNet algorithm had the best performance with an accuracy of 94.6% for moderate/severe hydronephrosis with substantial recall rate, specificity, precision, F1 measure and intersection over union. The Res-UNet algorithm has the best performance in detection of moderate/severe hydronephrosis. It would decrease variability among sonographers and improve efficiency under clinical conditions.

Population-based cohort analyses of the bidirectional relationship between type 2 diabetes and depression.

OBJECTIVE: This study addresses the strength of association for the bidirectional relationship between type 2 diabetes and depression. RESEARCH DESIGN AND METHODS: We used two cohort studies with the same source of database to determine the link between depression and type 2 diabetes. The data analyzed included a random sample of 1 million beneficiaries selected from the National Health Insurance claims in 2000. The analysis of diabetes predicting the depression onset consisted of 16,957 diabetic patients and the same number of sex- and age-matched nondiabetic control subjects. The analysis of depression predicting diabetes onset included 5,847 depressive patients and 5,847 sex- and age-matched nondepressive control subjects. The follow-up period was between 2000 and 2006, and onset of end points was identified from ambulatory care claims. The Cox proportional hazards regression model adjusted for potential confounders was used to estimate relative hazards. RESULTS: The first cohort analysis noted an incidence density (ID) of 7.03 per 1,000 person-years (PY) and 5.04 per 1,000 PY for depression in diabetic and nondiabetic subjects, respectively, representing a covariate-adjusted hazard ratio (HR) of 1.43 (95% CI 1.16-1.77). The second cohort analysis noted an ID of 27.59 per 1,000 PY and 9.22 per 1,000 PY for diabetes in depressive and nondepressive subjects, respectively. The covariate-adjusted HR was stronger at 2.02 (1.80-2.27) for incident diabetes associated with baseline depression. CONCLUSIONS: The two cohort studies provided evidence for the bidirectional relationship between diabetes and depression, with a stronger association noted for the depression predicting onset of diabetes.