Search

Goals/Purpose Patients with adjuvant radiotherapy are significantly more likely to have complications such as infections and wound dehiscence following breast surgery.1 Closed incision negative pressure wound therapy (cINPT) holds incision edges together, prevents external contamination and removes exudate, promoting tensile strength and increased collagen recruitment at the incision site compared to the standard of care dressings.2 cINPT is helpful in surgical closures on high-risk patients.3 We evaluated the complication rates following the use of cINPT in implant-based breast reconstruction. Methods/Technique A cohort was collected retrospectively from a single institution, from January 1, 2018 to October 31 2022. The complication rates were then compared between cINPT dressing and standard-of-care dressings on radiated patients following breast surgery. Results/Complications The data collected consist of a total of 83 radiated breasts that underwent implant based reconstruction. cINPT was used in 41 breasts and standard-of-care dressings were used in 42 breasts. Of the cINPT cohort, 2 patients developed seromas that required bedside aspirations (4.8%) and one patient developed an infection (2.4%) with none of the patients requiring re operations. Of the standard of care dressings cohort, 9 patients developed seromas (21.4%), and 3 patients developed infections (7.1%), of which 3 patients required reoperations (7.1%). Conclusion Patients with prior radiation who received cINPT had lower risk of developing complications compared to patients with prior radiation who received standard-of-care dressings.

Deep learning models have shown their great performances in natural language processing tasks. While much attention has been paid towards improvements in utility, privacy leakage and social bias are two major concerns arising in trained models. In this paper, we protect individual privacy and mitigate gender bias on classification models simultaneously. First, we propose a selective privacy-preserving method that only obscure individuals' sensitive information by adding noise on word embeddings. Then we propose a negative multi-task learning framework to mitigate the gender bias which contains a main task and a gender prediction task. The main task uses a positive loss constraint to ensure utility while the gender prediction task applies a negative loss constraint to remove gender-specific features. We analyze two existing word embeddings and evaluate them on sentiment analysis and medical text classification tasks. Our experimental results show that our negative multi-task learning framework can mitigate the gender bias while keeping models’ utility on both sentiment analysis and medical text classification.

Acute decompensated heart failure (ADHF) is a primary cause of older adults presenting to the emergency department with acute dyspnea. Point-of-care lung ultrasound (LUS) has shown comparable or superior diagnostic accuracy in comparison with a chest x-ray (CXR) in patients presenting with symptoms of ADHF. The systematic review and meta-analysis aimed to elucidate the sensitivity and specificity of LUS in comparison with CXR for diagnosing ADHF and summarize the rapidly growing body of evidence in this domain. A total of 5 databases were searched through February 18, 2021, to identify observational studies that reported on the use of LUS compared with CXR in diagnosing ADHF in patients presenting with shortness of breath. Meta-analysis was conducted on the sensitivities and specificities of each diagnostic method. A total of 8 studies reporting on 2,787 patients were included in this meta-analysis. For patients presenting with signs and symptoms of ADHF, LUS was found to be more sensitive than CXR (91.8% vs 76.5%) and more specific than CXR (92.3% vs 87.0%) for the detection of cardiogenic pulmonary edema. In conclusion, LUS is more sensitive and specific than CXR in detecting pulmonary edema. This highlights the importance of sonographic B-lines, along with the accurate interpretation of clinical data, in the diagnosis of ADHF. In addition to its convenience, reduced costs, and reduced radiation exposure, LUS should be considered an effective alternative to CXR for evaluating patients with dyspnea in the setting of ADHF.

We study conversational group detection in varied social scenes using a message-passing Graph Neural Network (GNN) in combination with the Dominant Sets clustering algorithm. Our approach first describes a scene as an interaction graph, where nodes encode individual features and edges encode pairwise relationship data. Then, it uses a GNN to predict pairwise affinity values that represent the likelihood of two people interacting together, and computes non-overlapping group assignments based on these affinities. We evaluate the proposed approach on the Cocktail Party and MatchNMingle datasets. Our results suggest that using GNNs to leverage both individual and relationship features when computing groups is beneficial, especially when more features are available for each individual.

Introduction/objectives: Diabetic Kidney Disease (DKD) is the leading cause of end-stage kidney disease. Despite optimal glycemic control and blood pressure management, progression to DKD cannot be halted in some patients. We aimed to find the association of modifiable and non-modifiable risk factors and comorbid conditions in patients with DKD. Methods: Retrospective medical record review of adult patients with diabetes mellitus (DM) was performed who visited our internal medicine office between January 1, 2020 and December 31, 2020. Results: Among 728 patients with DM, 471 (64.7%) patients had DKD, and 257 (35.3%) patients were without DKD. Among the group of patients with DKD, the majority were in CKD stage G1A2 (34.6%), followed equally by G2A2 and G3aA1 (16.8% each). Mean age of the patients with DKD was significantly greater than the patients without DKD (69.4 years vs 62.2 years; P Conclusions: In patients with diabetes, increasing age, female sex, and lack of moderate alcohol consumption were associated with increased odds of DKD. Higher frequencies of association of hypertension, hyperlipidemia, coronary artery disease, cerebrovascular accidents, congestive heart failure, carotid artery stenosis, aortic aneurysm, peripheral artery disease, gout, and osteoarthritis were also seen in patients with DKD.

Introduction: The purpose of this study is to produce a bibliometric review of the 30 most cited articles related to 6 major domains of body contouring—abdominoplasty, thighplasty, brachioplasty, gluteoplasty, body lift, and liposuction—for resident and fellow education. Methods: The authors utilized the Web of Science Citation Index to identify the 30 most cited articles related to surgery for body contouring published from 1975 to 2020. Articles were classified according to their level of evidence, type of study, and country of publication. Results: A total of 336 articles were reviewed to compile our list. The mean number of citations across the articles was 114.7 ± SD 86.1. The highest prevalence of the papers was published between 2000 and 2009 (n = 15, 50%). The country with the highest number of contributions was the United States (n = 22, 73%). Plastic and Reconstructive Surgery served as the main journal of publication for these papers (n = 22, 73.3%). The majority of articles were designated for clinical-type studies (n = 26, 86.7%). No basic science or prevalence study design papers were listed. In terms of level of evidence (LoE), most papers were assigned IV (n = 11, 36.7%) and III (n = 7, 23.3%). Conclusions: Our study reveals that the most cited papers in body contouring are of LoE III and IV. Although the LoE of plastic surgery research, in general, has improved, in the past decade, a call for higher quality papers remains. Overall, this analysis provides an easy, electronic starting point for residents and fellows interested in understanding the field's evolution.

Introduction Famotidine is a competitive histamine H2-receptor antagonist most commonly used for gastric acid suppression but thought to have potential efficacy in treating patients with Coronavirus disease 2019 (COVID-19). The aims of this systematic review and meta-analysis are to summarize the current literature and report clinical outcomes on the use of famotidine for treatment of hospitalized patients with COVID-19. Methods Five databases were searched through February 12, 2021 to identify observational studies that reported on associations of famotidine use with outcomes in COVID-19. Meta-analysis was conducted for composite primary clinical outcome (e.g. rate of death, intubation, or intensive care unit admissions) and death separately, where either aggregate odds ratio (OR) or hazard ratio (HR) was calculated. Results Four studies, reporting on 46,435 total patients and 3,110 patients treated with famotidine, were included in this meta-analysis. There was no significant association between famotidine use and composite outcomes in patients with COVID-19: HR 0.63 (95% CI: 0.35, 1.16). Across the three studies that reported mortality separated from other endpoints, there was no association between famotidine use during hospitalization and risk of death—HR 0.67 (95% CI: 0.26, 1.73) and OR 0.79 (95% CI: 0.19, 3.34). Heterogeneity ranged from 83.69% to 88.07%. Conclusion Based on the existing observational studies, famotidine use is not associated with a reduced risk of mortality or combined outcome of mortality, intubation, and/or intensive care services in hospitalized individuals with COVID-19, though heterogeneity was high, and point estimates suggested a possible protective effect for the composite outcome that may not have been observed due to lack of power. Further randomized controlled trials (RCTs) may help determine the efficacy and safety of famotidine as a treatment for COVID-19 patients in various care settings of the disease.