Transfer learning of medical image segmentation based on optimal transport feature selection

In the unsupervised domain adaptive transfer learning process，domain-independent features lead to the degradation of model segmentation performance，but there is no effective feature selection method for transfer learning segmentation model at present. To solve this problem，a general feature selection module for transfer learning was proposed based on optimal transport，which can be applied to various unsupervised domain adaptive image segmentation models. In this module，the optimal sample subsets of two domains are selected by weighted optimal transport of segmentation accuracy，and then the features of sample subsets are subjected to entropy regularized optimal transport，so as to obtain a descending list of similarity between two domains to remove domain-independent features. The universal feature selection module is applied to three unsupervised domain adaptive models to solve the problem of Covid-19 image segmentation，which improves the model performance to a certain extent. © 2022 Editorial Board of Jilin University. All rights reserved.

Mysterious cause of hypotension in MIS-C

Case Report Multisystem inflammatory syndrome (MIS-C) involves severe multi-organ inflammatory injury 2-6 weeks after COVID-19 infection. Seventy to 85% of patients have cardiovascular involvement, including diminished left ventricular ejection fraction (EF), coronary aneurysm, arrhythmias, valvular dysfunction, and pericardial effusion. Here we present a patient who arrived to the pediatric emergency department (ED) with MIS-C and suspected cardiogenic shock, though without the echocardiogram abnormalities commonly associated with MIS-C. A 7 year old African American male presented for a third time to our ED over the course of 4 days of febrile illness and was found to have MIS-C. During this time, he had no chest pain, palpitations, shortness of breath, or abnormal cardiopulmonary exam. At the first 2 ED visits, he was generally well appearing and after treating fever, had vital signs normal for his age. At his third visit, his vital signs were notable for borderline hypotension 86/48 (threshold 83/39 for his height of 1.25 meters). Troponins, chest X-ray, and EKG were normal. Bedside ultrasound was normal, with EF 55-60% so the hypotension was presumed to be secondary to hypovolemia and sepsis. However, despite 40 mL/kg of fluid boluses and maintenance fluid x1.5, his blood pressure continued to downtrend to a nadir of 79/39. He soon developed an S3 gallop and facial edema indicating fluid overload. His proBNP 4986 pg/mL also resulted at this time, suggesting cardiac injury was present. A formal cardiology echocardiogram confirmed the bedside ultrasound findings, noting normal ventricular size and motion, trivial pericardial effusion, and normal coronary artery size. However, it also detected diastolic dysfunction evident in mildly elevated E/e' of 10.86 of lateral mitral annulus, and 12.7 at medial mitral annulus. Three hours after starting solumedrol for treatment of MIS-C, his blood pressure improved to 110/52. The patient had no further episodes of hypotension, though it is unclear if steroids had resolved this by alleviating the underlying inflammation or as a secondary effect. We present a case of MIS-C that led to diastolic heart failure detected by mild hypotension, elevated proBNP, and subtle findings on formal echocardiogram. Although less common than systolic dysfunction in MIS-C, early recognition of diastolic heart failure is important for effective fluid management and initiation of vasoactive agents in criticallly 'ill patients. Diastolic heart failure with preserved systolic function has been seen on echo of MIS-C patients, and is hypothesized to be the subacute period after recovery of systolic function. However, we did not find clinical symptoms of systolic heart failure prior to the patient's development of diastolic heart failure. It is therefore essential to recognize that a patient with MIS-C may present with diastolic heart failure without preceding symptoms or echo findings of other cardiac anomalies.

Use of Extended Reality in Medicine During the Covid-19 Pandemic

The ongoing Covid-19 pandemic has generated a strong impetus to digitalize the economy and aspects of our daily life. Clinical medicine traditionally being conservative has seen a limited uptake of these new technologies which has seen widespread adoption in other industries. But due to disruptive nature of the pandemic, clinical medicine has also been forced to adapt and capitalize on these new technologies. Chief amongst them is the utilization of extended reality (XR) technologies, which is an umbrella term that encompasses a spectrum of virtual reality (VR) and Augmented reality (AR) devices that blend the physical world with the digital world. VR technologies immerses users in 3D worlds while AR technologies project 3D objects into the user’s physical environment while permitting full visibility of the user’s surroundings. XR technologies can assist in infection control measures by revolutionizing clinical ward rounds. Patient’s key blood results and vitals can be projected above each patient enhancing the speed of clinical ward rounds for large number of patient’s in community isolation facilities. Examination findings can then be dictated and automatically recorded. XR technologies can also assist clinicians during the planning and execution of highly infective/risky procedures. XR can help proceduralists simulate the procedure, limiting timing spent during the actual procedure. While XR guided robots can actually perform the high risk and delicate procedures, limiting infection risk for the proceduralist. XR technologies can overcome the disruption caused clinical education due to Covid-19 pandemic infection control measures. They can help simulate patient interaction/ clinical scenarios for medical students while keeping both patient and medical students safe from infection. Covid-19 has also generated much psychosocial distress due to the isolation stemming from infection control. XR technologies can be used to help bridge the psychosocial isolation by connecting patients with their family members, hobbies or home towns. This can be especially therapeutic when counselling patients that suffer pandemic related depression/anxiety. Particularly in palliative patients XR technologies can help simulate experiences that would be physically impossible for them. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Epidemiological characteristics of 17 Coronavirus disease 2019 in Puyang, Henan

Objective To explore the epidemiological characteristics of confirmed cases of coronavirus diseases 2019 (COVID-19) in Puyang, Henan Province, so as to provide basis for diseases control and preventive. Methods The epidemiological data of 17 cases of COVID-19 in Puyang were collected, and the time, regional and population distribution characteristics of COVID-19 were described and analyzed by drawing disease sequence diagram and case relationship diagram. Results Among the 17 cases, 7 were male (41.2%);the median age was 36 years old with age ranged from 12 to 66 years. A total of 4 clustered outbreaks occurred, involving 12 cases (70.6%), all of which were family recurrent cases. The median incubation period was 6.5 days, the shortest 3 days and the longest 13 days. The onset time of a second-generation case was 11 days earlier than that of the indicator case. The median time between onset and treatment was 4 days, the shortest was 0 days and the longest was 12 days. Of the 17 cases, 6 had a sojourning history in Hubei Province within 14 days, and 1 had a history of overseas tourism. The other cases were all local infections, 8 of them were close contacts of the confirmed cases, and 2 of them were from unknown sources. Among the close contacts, the cases involved in the family clustering epidemic were transmitted through close contact and respiratory tract. After detailed investigation and inquiry, it was inferred that the transmission route was the staircase droplet transmission. Conclusion In Puyang City, most of the cases were from Hubei Province, and was dominated by family clustering epidemic. There was a possibility of infection in the incubation period. © 2021, Publication Centre of Anhui Medical University. All rights reserved.

COVID⁃19 chest CT image segmentation based on federated learning and blockchain

This paper proposed a COVID-19 chest CT image segmentation method based on Federated Learning (FL) and blockchain to automatically segment the area in lung affected by COVID-19. Firstly, in the situation where the sample data of patients is limited and it is distributed in different institutions, which cannot be easily collected, we applied FL method. Then, we used blockchain network to replace the central server in FL to solve the "single point of failure" problem. Finally, we designed a lightweight separable convolution U-NET to reduce the cost of computation and time. Experimental results show that the method has good performance after training, and its dice metric can achieve 63.26%, which is helpful for diagnosis of COVID-19. © 2021, Jilin University Press. All right reserved.

[Application of remote "Internet+" interactive mode in the management of patients with hypertension during normalized epidemic prevention and control of COVID-19].

Objective: To explore the effect of remote "Internet+" interactive management strategy on blood pressure control in patients with hypertension during normalized epidemic prevention and control of COVID-19. Methods: This is a randomized controlled study. A total of 394 patients with hypertension who were treated in Chinese People's Liberation Army General Hospital from October 2019 to December 2020 were randomly divided into experimental group (197 cases) and control group (197 cases). The experimental group adopted remote "Internet+" interaction mode to carry out remote blood pressure intervention, and the control group received traditional blood pressure control mode, and the intervention time was 6 months. Evaluation indicators included blood pressure level, blood pressure lowering speed, time to target blood pressure, blood pressure measurement times, communication times with doctors, medication compliance, blood pressure measurement compliance and disease awareness after 6 months of intervention. The evaluation indexes of the two groups were compared, and the bivariate Pearson correlation analysis was used to explore the relationship between the speed of blood pressure reduction and the times of blood pressure measurement and doctor communication in all patients. Results: A total of 394 patients with hypertension were included in this study, including 209 males, aged (67.6±2.8) years old. After 6 months of intervention, the systolic and diastolic blood pressure of the two groups were both lower than the baseline blood pressure before intervention (both P<0.05), the systolic blood pressure ((125.7±11.7) mmHg (1 mmHg=0.133 kPa) vs. (132.6±12.9) mmHg, P<0.001) and diastolic blood pressure ((72.4±10.7) mmHg vs. (79.8±11.6) mmHg, P<0.001) in the experimental group were lower than those in the control group. The blood pressure reduction speed of the experimental group was faster than that of the control group ((18.63±1.59) mmHg/d vs. (13.26±2.85) mmHg/d, P<0.001), and the time to reach the target blood pressure in the experimental group was shorter than that in the control group ((23.69±2.93) d vs. (47.12±5.81) d, P<0.001). Compared with the control group, the blood pressure measurement times ((0.98±0.13) times/d vs. (0.20±0.40) times/d, P<0.05) and the number of communications with doctors ((0.97±0.16) times/week vs. (0.12±0.32) times/week, P<0.05) were significantly higher in the experimental group. Correlation analysis showed that the speed of blood pressure reduction was positively correlated with the number of blood pressure measurements (r=0.419, P<0.01) and the number of communications with doctors (r=0.857, P<0.01). The proportion of standardized medication (93.91% (185/197) vs. 51.78% (102/197), P<0.001), timely measurement (97.46% (192/197) vs. 47.21% (93/197), P<0.001) and high-degree disease awareness (94.42% (186/197) vs. 49.24% (97/197), P<0.001) were significantly higher in the experimental group than those in the control group. Conclusions: The remote "Internet+" interactive management strategy can effectively improve patients' blood pressure control. The doctor-patient interaction can improve medication compliance and measurement compliance of patients, and help shorten the time to reach the target blood pressure.

A weakly supervised learning method based on attention fusion for COVID-19 segmentation in CT images

Since the end of 2019, the COVID-19, which has swept across the world, has caused serious impacts on public health and economy. Although Reverse Transcription-Polymerase Chain Reaction (RT-PCR) is the gold standard for clinical diagnosis, it is very time-consuming and labor-intensive. At the same time, more and more people have doubted the sensitivity of RT-PCR. Therefore, Computed Tomography (CT) images are used as a substitute for RT-PCR. Powered by the research of the field of artificial intelligence, deep learning, which is a branch of machine learning, has made a great success on medical image segmentation. However, general full supervision methods require pixel-level point-by-point annotations, which is very costly. In this paper, we put forward an image segmentation method based on weakly supervised learning for CT images of COVID-19, which can effectively segment the lung infection area and doesn't require pixel-level labels. Our method is contrasted with another four weakly supervised learning methods in recent years, and the results have been significantly improved.

Discussion on the Management Method of COVID-19 Infection Prevention in Medical Residents

Background: To explore the homogenized management method of infection prevention and control in medical resident under COVID-19 epidemic situation. Methods: Overall, 268 members in Ningbo Medical Group which was participating in the management of novel coronavirus pneumonia in Wuhan, China were managed homogeneously in terms of personal cleaning and disinfection of articles, prevention and control of infection in life, training and assessment of relevant infection knowledge in hospitals. Results: In the epidemic situation, the homogenization management of infection prevention and control in the resident of medical team is an important measure to block the transmission link. Conclusion: 1. Conduct homogenization management on the training of hospital infection knowledge for medical staff. 2. Conduct homogeneous management of effective assessment of training knowledge. 3. Conduct homogenization management from personal cleaning and disinfection, as well as infection prevention and control in daily life.

Government support for SMEs in response to COVID-19: theoretical model using Wang transform

Purpose Many governments are taking measures in support of small and medium-sized enterprises (SMEs) to mitigate the economic impact of the COVID-19 outbreak. This paper presents a theoretical model for evaluating various government measures, including insurance for bank loans, interest rate subsidy, bridge loans and relief of tax burdens. Design/methodology/approach This paper distinguishes a firm's intrinsic value and book value, where a firm can lose its intrinsic value when it encounters cash-flow crunch. Wang transform is applied to (1) calculating the appropriate level of interest rate subsidy payable to incentivize banks to issue more loans to SMEs and to extend the loan maturity of current debt to the SMEs, (2) describing the frailty distribution for SMEs and (3) defining banks' underwriting capability and overlap index in risk selection. Findings Government support for SMEs can be in the form of an appropriate level of interest rate subsidy payable to incentivize banks to issue more loans to SMEs and to extend the loan maturity of current debt to the SMEs. Research limitations/implications More available data on bank loans would have helped strengthen the empirical studies. Practical implications This paper makes policy recommendations of establishing policy-oriented banks or investment funds dedicated to supporting SMEs, developing risk indices for SMEs to facilitate refined risk underwriting, providing SMEs with long-term tax relief and early-stage equity-type investments. Social implications The model highlights the importance of providing bridge loans to SMEs during the COVID-19 disruption to prevent massive business closures. Originality/value This paper provides an analytical framework using Wang transform for analyzing the most effective form of government support for SMEs.

[Epidemic characteristics and trend analysis of the COVID-19 in Hubei province].

Objective: By describing and analyzing the epidemic characteristics and trends of the attack rate, the crude mortality and relevant indexes in Hubei province during the pandemic of COVID-19 to provide comprehensive evaluations of the epidemic trends and the effects of intervention measures. Methods: Based on the case data reported in Hubei province during the COVID-19 epidemic, combined with the important time of major interventions and event, the cumulative attack rate, the sequential increase rate of new cases, baseline increase rate of new cases, the observation- confirmed case conversion rate, the cumulative crude mortality, the daily severe case rate, and the ratio of death to severe were used to describe and analyze the epidemic characteristics in different phases of the COVID-19 epidemic. Results: The epidemic experienced an outbreak phase from January 10 to February 3 with large amount of case reported, a peak phase from February 4 to February 19 with continuous increasing number of new cases and deaths, a platform phase from February 20 to March 3 with balanced diagnosis and treatment number, and a descending phase from March 4 to March 18 with decreased diagnosis and increased treatment number. Up to March 18, the cumulative attack rate of the COVID-19 epidemic in Hubei province increased from 0.03/10 000 on January 19 to 11.46/10 000, from 0.04/10 000 on January 10 to 45.13/10 000 in Wuhan city, and from 0.002/ 10 000 on January 20 to 3.70/ 10 000 in other areas of Hubei province other than Wuhan city. The increase rate of new cases fluctuated during the epidemic period and reached the highest at February 12 in Hubei province. The cumulative crude mortality in Hubei Province increased rapidly from 1.01% on January 19 to 5.13% on January 26, then decreased to 2.54% on February 13, and then slowly increased to 4.62% on March 18, and similar trend was also observed in Wuhan city. The daily severe rate in Hubei Province increased from 26.88% on January 27 to 34.27% on March 18. The ratio of death to severe decreased from 7.37% on January 23 to 0.35% on March 18. Conclusions: The epidemic cycle of COVID-19 in Hubei province proposed to be 60 days, which was about 1.76 times of the combination of the longest incubation period or isolation period (14 d) and the average hospitalization time of confirmed patients in Hubei province (20 d). It suggested that the major anti-epidemic decisions made in China were effective.

[Epidemiological investigation of a family clustering of COVID-19].

Objective: To investigate the epidemiological characteristics of a family clustering of COVID-19. Methods: Field epidemiological survey was conducted. Results: Case 1 of the long-term residents from Hubei province was the source of infection of this family clustering. There were 6 cases (from case 2 to case 7) infected in the whole incubation period. The incubation period was more than 14 days for 3 of the second-generation cases. Routes of transmission included respiratory droplets (from case 1 transmitted to case 6, from case 1 to her family members) and close contact (from case 1 to other cases in her family). All the age groups were generally susceptible, while elderly were easier to progress to critically ill. Besides respiratory symptoms, there were also gastrointestinal symptoms, of which diarrhea was the most common one. Conclusions: Family clustering had been an important part for COVID-19 cases.