Although endoscopic submucosal dissection (ESD) is increasingly utilized to treat early neoplasms of the gastrointestinal tract, its use for duodenal neoplasms is limited by the thin wall and narrow lumen of the duodenum. We have reviewed cases where ESD was used to treat sessile, nonampullary duodenal neoplasms. To do this, we retrospectively reviewed the medical records of patients treated with ESD for adenomas of the duodenum from January 2001 to December 2010, assessing the curative outcomes and complication rates. A total of 14 cases were reviewed. Mean patient age was 56.4 years. The mean size of tumors and mean size of the specimens were 17.1 mm and 26.4 mm, respectively. The en bloc resection rate with ESD was 78.6%, and the complete (R0) resection rate was 85.7%. No patient in the study experienced major bleeding. However, second-look endoscopy revealed minor bleeding requiring endoscopic homeostasis in one case (7.1%). Perforations were observed in five cases (35.7%). Two of the five patients with perforation underwent surgery. The ESD methods yielded acceptable curative resection rates for duodenal adenomas, although ESD was associated with a higher rate of perforation. Therefore, duodenal ESD should be performed with care and only in selected patients to avoid serious complications.

Mammography plays an important role in screening breast cancer among females, and artificial intelligence has enabled the automated detection of diseases on medical images. This study aimed to develop a deep learning model detecting breast cancer in digital mammograms of various densities and to evaluate the model performance compared to previous studies. From 1501 subjects who underwent digital mammography between February 2007 and May 2015, craniocaudal and mediolateral view mammograms were included and concatenated for each breast, ultimately producing 3002 merged images. Two convolutional neural networks were trained to detect any malignant lesion on the merged images. The performances were tested using 301 merged images from 284 subjects and compared to a meta-analysis including 12 previous deep learning studies. The mean area under the receiver-operating characteristic curve (AUC) for detecting breast cancer in each merged mammogram was 0.952 ± 0.005 by DenseNet-169 and 0.954 ± 0.020 by EfficientNet-B5, respectively. The performance for malignancy detection decreased as breast density increased (density A, mean AUC = 0.984 vs. density D, mean AUC = 0.902 by DenseNet-169). When patients’ age was used as a covariate for malignancy detection, the performance showed little change (mean AUC, 0.953 ± 0.005). The mean sensitivity and specificity of the DenseNet-169 (87 and 88%, respectively) surpassed the mean values (81 and 82%, respectively) obtained in a meta-analysis. Deep learning would work efficiently in screening breast cancer in digital mammograms of various densities, which could be maximized in breasts with lower parenchyma density.

Aims The aim of this study was to evaluate the outcome of spinal instrumentation in haemodialyzed patients with native pyogenic spondylodiscitis. Spinal instrumentation in these patients can be dangerous due to rates of complications and mortality, and biofilm formation on the instrumentation. Patients and Methods A total of 134 haemodialyzed patients aged more than 50 years who underwent surgical treatment for pyogenic spondylodiscitis were included in the study. Their mean age was 66.4 years (50 to 83); 66 were male (49.3%) and 68 were female (50.7%). They were divided into two groups according to whether spinal instrumentation was used or not. Propensity score matching was used to attenuate the potential selection bias. The outcome of treatment was compared between these two groups. Results A total of 89 patients (66.4%) underwent non-instrumented surgery and 45 (33.5%) underwent instrumented surgery. There were no significant differences in the rates of postoperative complications, except for an increased rate of wound problems in the instrumented group, which was found in the unmatched cohorts (p = 0.034). There were no significant differences in the rate of recurrent infections (p = 0.328 for the unmatched cohort; p = 0.269 for the matched cohort) and mortality rate, including in-hospital (p = 0.713 for the unmatched cohort; p = 0.738 for the matched cohort) and one-year rates (p = 0.363 for the unmatched cohort; p = 0.787 for the matched cohort), between the groups. However, the interval between the initial diagnosis and the first recurrence was significantly longer in the instrumented group (p = 0.008 for the unmatched cohort; p = 0.032 for the matched cohort). Conclusion Instrumented surgery for haemodialyzed patients with pyogenic spondylodiscitis showed similar outcomes, including recurrence and mortality, to non-instrumented surgery, despite the instrumented group having more severe neurological deficit, a larger number of involved levels, and increased kyphotic angle.