Nevin Manimala

Neural Netw. 2022 May 6;152:380-393. doi: 10.1016/j.neunet.2022.04.028. Online ahead of print.

ABSTRACT

Traditionally, Convolutional Neural Networks make use of the maximum or arithmetic mean in order to reduce the features extracted by convolutional layers in a downsampling process known as pooling. However, there is no strong argument to settle upon one of the two functions and, in practice, this selection turns to be problem dependent. Further, both of these options ignore possible dependencies among the data. We believe that a combination of both of these functions, as well as of additional ones which may retain different information, can benefit the feature extraction process. In this work, we replace traditional pooling by several alternative functions. In particular, we consider linear combinations of order statistics and generalizations of the Sugeno integral, extending the latter’s domain to the whole real line and setting the theoretical base for their application. We present an alternative pooling layer based on this strategy which we name “CombPool” layer. We replace the pooling layers of three different architectures of increasing complexity by CombPool layers, and empirically prove over multiple datasets that linear combinations outperform traditional pooling functions in most cases. Further, combinations with either the Sugeno integral or one of its generalizations usually yield the best results, proving a strong candidate to apply in most architectures.

PMID:35605303 | DOI:10.1016/j.neunet.2022.04.028

Brief Bioinform. 2022 May 24:bbac185. doi: 10.1093/bib/bbac185. Online ahead of print.

ABSTRACT

Molecular signatures have been excessively reported for diagnosis of many cancers during the last 20 years. However, false-positive signatures are always found using statistical methods or machine learning approaches, and that makes subsequent biological experiments fail. Therefore, signature discovery has gradually become a non-mainstream work in bioinformatics. Actually, there are three critical weaknesses that make the identified signature unreliable. First of all, a signature is wrongly thought to be a gene set, each component of which keeps differential expressions between or among sample groups. Second, there may be many false-positive genes expressed differentially found, even if samples derived from cancer or normal group can be separated in one-dimensional space. Third, cross-platform validation results of a discovered signature are always poor. In order to solve these problems, we propose a new feature selection framework based on ensemble classification to discover signatures for cancer diagnosis. Meanwhile, a procedure for data transform among different expression profiles across different platforms is also designed. Signatures are found on simulation and real data representing different carcinomas across different platforms. Besides, false positives are suppressed. The experimental results demonstrate the effectiveness of our method.

PMID:35605226 | DOI:10.1093/bib/bbac185

Braz J Cardiovasc Surg. 2022 May 23;37(3):356-369. doi: 10.21470/1678-9741-2020-0731.

ABSTRACT

INTRODUCTION: Oral anticoagulants are the treatment of choice for diverse types of coagulation disorders. Warfarin is widely used by the Brazilian population, possibly due to its lower cost than other oral anticoagulants. However, it has a high risk of serious adverse effects if used incorrectly. The Anticoagulation Knowledge Tool (AKT) can assess a patient’s knowledge about her/his oral anticoagulant therapy and can assist health professionals in identifying patients with difficulties in adherence. This study aimed to translate, culturally adapt, and validate the AKT into Brazilian Portuguese.

METHODS: After a standard forward-backward procedure to translate the AKT into Brazilian Portuguese (AKT-Br), a version of the instrument was applied in three groups (patients, pharmacists, and the general population). The reliability of the AKT-Br was tested using an internal consistency measure and test-retest. The validity of the instrument was confirmed with data from the contrasted groups. All statistical analyses were performed with RStudio.

RESULTS: The median scores obtained with the AKT-Br were 29.0, 17.0, and 7.5 for pharmacists, patients, and the general population, respectively (maximum score of 35 points). There was moderate internal consistency for the instrument and test-retest reliability was satisfactory. Analysis of variance for validity of the groups revealed a significant relationship between the total score and the evaluated groups.

CONCLUSION: The ATK-Br is a reliable and valid tool to assess knowledge about oral anticoagulants. AKT-Br can be used in clinical practice as an auxiliary tool to improve patient care through personalised educational interventions.

PMID:35605217 | DOI:10.21470/1678-9741-2020-0731

Braz J Cardiovasc Surg. 2022 May 23;37(3):315-320. doi: 10.21470/1678-9741-2020-0425.

ABSTRACT

INTRODUCTION: In patients undergoing coronary artery bypass grafting (CABG), stroke is a major complication that increases morbidity and mortality. The presence of carotid stenosis (CS) increases risk of stroke, and the optimal treatment remains uncertain due to the lack of randomized clinical trials. The aim of this study is to compare three management approaches to CS in patients submitted to CABG.

METHODS: From 2005 to 2015, 79 consecutive patients with significant CS submitted to CABG were retrospectively evaluated. Patients were divided in three groups, according to CS treatment: 17 underwent staged carotid endarterectomy (CEA)-CABG, 26 underwent synchronous CEA-CABG, and 36 underwent isolated CABG without carotid intervention. The primary outcomes were composed by 30-day postoperative acute myocardial infarction (MI), 30-day postoperative stroke, and death due to all causes during the follow-up.

RESULTS: Patients were evaluated during an average 2.05 years (95% confidence interval = 1.51-2.60) of follow-up. Major adverse cardiac events, including death, postoperative MI, and postoperative stroke, occurred in 76.5% of the staged group, 34.6% of the synchronous group, and 33.3% of the isolated CABG group (P=0.007). As for MI, the rates were 29.4%, 3.85%, and 11.1% (P=0.045), respectively. There was no statistically significant difference in total mortality rates (35.3%, 30.8%, and 25.0%, respectively; P=0,72) and stroke (29.4%, 7.7%, and 8.3%, respectively; P=0,064) between groups.

CONCLUSION: Staged CEA-CABG is associated with higher major adverse cardiac events and MI rate when compared to the strategy of synchronous and isolated CABG, but without statistically difference in total mortality during the entire follow-up.

PMID:35605216 | DOI:10.21470/1678-9741-2020-0425

Braz J Cardiovasc Surg. 2022 May 23;37(1):281-291. doi: 10.21470/1678-9741-2021-0483.

ABSTRACT

INTRODUCTION: Pediatric heart transplantation is the definitive therapy for children with end-stage heart failure. This paper describes our initial experience in pediatric heart transplantation in a tertiary center in Brazil.

METHODS: This is a historical prospective descriptive cohort study based on a review of the medical records of children undergoing heart transplantation at Hospital de Base and Hospital da Criança e Maternidade de São José do Rio Preto. Variables were displayed as frequency, mean, or median. Statistical analysis and Kaplan-Meier actuarial curve were obtained with the aid of Microsoft® Excel® 2019 and STATSDirect version 3.3.5.

RESULTS: Between January 2010 and December 2020, ten children underwent bicaval orthotopic heart transplantation, 30% of which were under one year of age. Nine patients had end-stage heart failure (International Society for Heart and Lung Transplantation-Heart Failure D) and 50% of the recipients were transplanted under conditions of progressive clinical deterioration (Interagency Registry for Mechanically Assisted Circulatory Support ≤ 2). Forty percent of the recipients had a panel-reactive antibody > 20% on virtual crossmatch. In the postoperative period, 80% of patients required high dose of inotropic support (vasoactive-inotropic score > 10) for > 48 hours. The death-free survival rate at 131 months was 77.1±14.4%. Most patients (88.9%) in late follow-up had an episode of active cytomegalovirus infection. Cellular rejection, with or without clinical repercussion, was present in 44.4% of the patients.

CONCLUSION: Pediatric heart transplantation produces acceptable and feasible outcomes as definitive therapy for children with end-stage heart failure.

PMID:35605214 | DOI:10.21470/1678-9741-2021-0483

J Nerv Ment Dis. 2022 May 5. doi: 10.1097/NMD.0000000000001538. Online ahead of print.

ABSTRACT

This study examined the relationship between eight measures of religious involvement and five anxiety disorders among a nationally representative sample of African-Americans (N = 3403). The Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, World Mental Health Composite International Diagnostic Interview was used to assess 12-month and lifetime prevalence for each disorder. Logistic regression indicated weekly service attendance was inversely associated with 12-month and lifetime panic disorder, lifetime agoraphobia, and 12-month and lifetime posttraumatic stress disorder (PTSD). Prayer was inversely associated with 12-month agoraphobia, 12-month social phobia, and lifetime PTSD. Listening to religious radio and looking to God for strength were also inversely related to, respectively, 12-month and lifetime panic disorder. Conversely, reading religious materials was positively associated with 12-month panic disorder, 12-month agoraphobia, lifetime PTSD, and lifetime generalized anxiety disorder. The results are discussed in light of conceptual models that specify multiple and sometimes divergent pathways through which religion impacts health, and suggestions for clinicians addressing anxiety disorders are delineated.

PMID:35605210 | DOI:10.1097/NMD.0000000000001538

JMIR Public Health Surveill. 2022 May 17. doi: 10.2196/37668. Online ahead of print.

ABSTRACT

BACKGROUND: Most studies of long covid (symptoms of COVID-19 beyond 4 weeks) have focused on people hospitalised in their initial illness. Long covid is thought to be under-recorded in UK primary care electronic records.

OBJECTIVE: We sought to determine which symptoms people present to primary care following COVID-19, and whether presentation differs in people who were not hospitalised, and post-long covid mortality.

METHODS: We used routine data from the nationally representative Primary Care Sentinel Cohort of the Oxford-Royal College of General Practitioners Research and Surveillance Centre (N=7.4million), applying a pre-defined long covid phenotype and grouped by whether the illness index was in hospital or community. We included COVID-19 cases between 1st-March-2020 and 1st-April-2021. We conducted a before and after analysis of pre-specified long covid symptoms identified by the Office of National Statistics, comparing symptoms presented between one and six months after their index infection matched with the same months one year previously. We conducted logistic regression analysis, quoting odds ratios with 95% confidence intervals, reporting differences between those with an index community infection compared to those who had been hospitalised, and separately associations with all-cause mortality.

RESULTS: 5.6% (416,505/7,396,702) and 1.8% (7,623/416,505) of patients respectively had a coded diagnosis of COVID-19 and diagnosis or referral for long covid. People coded as having long covid were significantly more likely to have presented the pre-specified symptoms after vs before COVID-19 infection (odds ratios 2.66 [2.46-2.88] for those with index community infection and 2.42 [2.03-2.89] for those hospitalised). Following an index community infection, patients were more likely to present with non-specific symptoms (odds ratio 3.44 [3.00-3.95], P<.001) than following a hospital admission (odds ratio 2.09 [1.56-2.80], P<.001). Mental health sequelae were more commonly associated with hospital admission index infections (odds ratio 2.21 [1.64-2.96]) compared to community (odds ratio 1.36 [1.21-1.53], P<.001). People presenting to primary care following hospital infection were more likely to be male (odds ratio 1.43 [1.25-1.64], P<.001), more socioeconomically deprived (odds ratio 1.42 [1.24-1.63], P<.001); and to have multi-morbidity (odds ratio 1.41 [1.26-1.57], P<.001) than those presenting after an index community infection. All-cause mortality in people with long covid was associated with increasing age; male gender (odds ratio 3.32 [1.34-9.24], P<.01) and higher multi-morbidity score (odds ratio 2.11 [1.34-3.29], P<.001). One or more vaccine doses was associated with reduced odds of mortality (odds ratio 0.10 [0.03-0.35], P<.001).

CONCLUSIONS: The low percentage of people recorded as having long covid following COVID-19 reflects either low prevalence or under-recording. The characteristics and comorbidities of those presenting with long covid following a community infection are different from those who were hospitalised with their index infection. This study provides insights into the presentation of long covid in primary care and implications for workload.

CLINICALTRIAL: Not applicable.

PMID:35605170 | DOI:10.2196/37668

Aust J Rural Health. 2022 May 23. doi: 10.1111/ajr.12875. Online ahead of print.

ABSTRACT

OBJECTIVE: To explore reasons for survival disparities for patients with prostate cancer in the Barwon South West area of Victoria.

DESIGN, SETTING AND PARTICIPANTS: We have described incidence, diagnostics, treatment pathways, and survival for four regions of the Barwon South Western Victoria. Analysis included all newly diagnosed prostate cancer patients from 2009 to 2015 in the Evaluation of Cancer Outcomes Barwon South West Registry. Regions included 1: Queenscliffe 2: Geelong, Colac Otway and Corangamite 3: Moyne, Warrnambool and Southern Grampians and 4: Glenelg. Across the four regions, variables were compared using a chi square statistic or analysis of variance and survival data was assessed with the Kaplan-Meier curves.

MAIN OUTCOME MEASURES: Incidence, treatment pathways and survival for prostate cancer patients.

RESULTS: A total of 1776 patients were diagnosed with prostate cancer from 2009 to 2015 in the Barwon South West area. In regions 1-4, there were 298 (1.04%), 1085 (0.92%), 273 (0.97%) and 120 (1.2%) cases, respectively. There was no significant difference in Gleason score and treatment. The 5-year survival rate was 85%, 76%, 71% and 80%, respectively, as compared with the national average of 95%. PSA scores >20 ng/ml at diagnosis, as a surrogate for high-risk disease, occurred in 23%, 29%, 22% and 21%, respectively (p < 0.01). The proportions presenting with stage IV disease were 17%, 26%, 21% and 6%, respectively (p = 0.10).

CONCLUSION: Men diagnosed with prostate cancer in South West Victoria have a considerably lower 5-year survival compared with the national average with later disease at presentation in some areas.

PMID:35605157 | DOI:10.1111/ajr.12875

Dent Traumatol. 2022 May 23. doi: 10.1111/edt.12759. Online ahead of print.

ABSTRACT

BACKGROUND/AIMS: Pediatric oral and maxillofacial surgeons have faced severe challenges in ward management due to their high risk of exposure during the COVID-19 epidemic. The aim of this study was to analyze and summarize the treatment methods and infection prevention and control measures applied in emergency cases in the Department of Pediatric Oral and Maxillofacial Surgery, Children’s Hospital of Chongqing Medical University, during the COVID-19 epidemic.

METHODS: In this retrospective study, information was collected from 256 pediatric emergency patients who were treated from January 23, 2020 to August 9, 2021. The patients’ data were statistically analyzed according to age, gender, disease and pathogenesis, operation time, and the main treatment applied in pediatric oral and maxillofacial emergency cases during the COVID-19 epidemic.

RESULTS: During the epidemic period, 256 pediatric emergency patients were successfully treated. Among them, there were 170 boys and 86 girls. In all, 182 patients were diagnosed with oral or facial lacerations; 43 had jaw fractures; 26 had maxillofacial infections; and five had dento-alveolar fractures. A total of 246 patients underwent surgery under negative pressure with level 3 protection standards. No doctors or patients infected with COVID-19 were found throughout the stury period.

CONCLUSIONS: Pediatric oral and maxillofacial emergency in-patients mainly experienced maxillofacial trauma during the COVID-19 epidemic, followed by infection. Effective diagnosis and treatment, and avoidance of COVID-19 infection can be achieved by strictly following epidemic prevention and treatment procedures.

PMID:35605156 | DOI:10.1111/edt.12759

Scand J Gastroenterol. 2022 May 23:1-6. doi: 10.1080/00365521.2022.2070436. Online ahead of print.

ABSTRACT

BACKGROUND: Artificial intelligence (AI) for polyp detection is being introduced to colonoscopy, but there is uncertainty how this affects endoscopists’ ability to detect polyps and neoplasms. We performed a video-based study to address whether AI improved the endoscopists’ performance to detect polyps.

METHODS: We established a dataset of 200 colonoscopy videos (length 5 s; 100 without polyps and 100 with one polyp). About 33 early-career endoscopists (50-400 colonoscopies performed) from 10 European countries classified each video as either ‘polyp present’ or ‘polyp not present’. The video assessment was performed twice with a four-week interval. The first assessment was performed without any AI tool, whereas the second was performed with an AI tool for polyp detection. The primary endpoint was early-career endoscopists’ sensitivity to detect polyps. Gold standard for presence and histology of polyps were confirmed by two expert endoscopists and pathologists, respectively. McNemar’s test was used for statistical significance.

RESULTS: There were 86 neoplastic and 14 non-neoplastic polyps (mean size 5.6 mm) in the 100 videos with polyps. Early-career endoscopists’ sensitivity to detect polyps increased from 86.3% (95% confidence interval [CI]: 85.1-87.5%) to 91.7% (95%CI: 90.7-92.6%) with the AI aid (p < .0001). Their sensitivity to detect neoplastic polyps increased from 85.4% (95% CI: 84.0-86.7%) to 92.1% (95%CI: 91.1-93.1%) with the AI aid (p < .0001).

CONCLUSION: The polyp detection AI tool helped early-career endoscopists to increase their sensitivity to identify all polyps and neoplastic polyps during colonoscopy.

PMID:35605150 | DOI:10.1080/00365521.2022.2070436