Information Bottleneck Attribution for Visual Explanations of Diagnosis and Prognosis

Visual explanation methods have an important role in the prognosis of the patients where the annotated data is limited or unavailable. There have been several attempts to use gradient-based attribution methods to localize pathology from medical scans without using segmentation labels. This research direction has been impeded by the lack of robustness and reliability. These methods are highly sensitive to the network parameters. In this study, we introduce a robust visual explanation method to address this problem for medical applications. We provide an innovative visual explanation algorithm for general purpose and as an example application we demonstrate its effectiveness for quantifying lesions in the lungs caused by the Covid-19 with high accuracy and robustness without using dense segmentation labels. This approach overcomes the drawbacks of commonly used Grad-CAM and its extended versions. The premise behind our proposed strategy is that the information flow is minimized while ensuring the classifier prediction stays similar. Our findings indicate that the bottleneck condition provides a more stable severity estimation than the similar attribution methods. The source code will be publicly available upon publication.

Information Bottleneck Attribution for Visual Explanations of Diagnosis and Prognosis.

Visual explanation methods have an important role in the prognosis of the patients where the annotated data is limited or unavailable. There have been several attempts to use gradient-based attribution methods to localize pathology from medical scans without using segmentation labels. This research direction has been impeded by the lack of robustness and reliability. These methods are highly sensitive to the network parameters. In this study, we introduce a robust visual explanation method to address this problem for medical applications. We provide an innovative visual explanation algorithm for general purpose and as an example application we demonstrate its effectiveness for quantifying lesions in the lungs caused by the Covid-19 with high accuracy and robustness without using dense segmentation labels. This approach overcomes the drawbacks of commonly used Grad-CAM and its extended versions. The premise behind our proposed strategy is that the information flow is minimized while ensuring the classifier prediction stays similar. Our findings indicate that the bottleneck condition provides a more stable severity estimation than the similar attribution methods. The source code will be publicly available upon publication.

The Role of CO-RADS Scoring System in the Diagnosis of COVID-19 Infection and its Correlation with Clinical Signs.

BACKGROUND: Computed tomography (CT) evaluation systematics has become necessary to eliminate the difference of opinion among radiologists in evaluating COVID-19 CT findings. INTRODUCTION: The objectives of this study were to evaluate the efficiency of CO-RADS scoring system in our patients with COVID-19 as well as to examine its correlation with clinical and laboratory findings. METHODS: The CO-RADS category of all patients included in the study was determined by a radiologist who did not know the rtRT-PCR test result of the patients, according to the Covid-19 reporting and data system of Mathias Prokop et al. Results: A total of 1338 patients were included. CT findings were positive in 66.3%, with a mean CO-RADS score of 3,4 ± 1,7. 444 (33.1%) of the patients were in the CO-RADS 1-2, 894 (66.9%) were in the CO-RADS 3-5 group. There were positive correlations between CO-RADS score and age, CMI, hypertension, diabetes mellitus, chronic pulmonary diseases presence of symptoms, symptom duration, presence of cough, shortness of breath, malaise, CRP, and LDH, while CORADS score was negatively correlated with lymphocyte count. The results of the ROC analysis suggested that those with age ≥40 years, symptom duration >2 days, CMI score >1 and/or comorbid conditions were more likely to have a CO-RADS score of 3-5. CONCLUSION: The CO-RADS classification system is a CT findings assessment system that can be used to diagnose COVID-19 in patients with symptoms of cough, shortness of breath, myalgia and fatigue for more than two days.

Prediction of COVID-19 Risk in Public Areas Using IoT and Machine Learning

COVID-19 is a community-acquired infection with symptoms that resemble those of influenza and bacterial pneumonia. Creating an infection control policy involving isolation, disinfection of surfaces, and identification of contagions is crucial in eradicating such pandemics. Incorporating social distancing could also help stop the spread of community-acquired infections like COVID-19. Social distancing entails maintaining certain distances between people and reducing the frequency of contact between people. Meanwhile, a significant increase in the development of different Internet of Things (IoT) devices has been seen together with cyber-physical systems that connect with physical environments. Machine learning is strengthening current technologies by adding new approaches to quickly and correctly solve problems utilizing this surge of available IoT devices. We propose a new approach using machine learning algorithms for monitoring the risk of COVID-19 in public areas. Extracted features from IoT sensors are used as input for several machine learning algorithms such as decision tree, neural network, naïve Bayes classifier, support vector machine, and random forest to predict the risks of the COVID-19 pandemic and calculate the risk probability of public places. This research aims to find vulnerable populations and reduce the impact of the disease on certain groups using machine learning models. We build a model to calculate and predict the risk factors of populated areas. This model generates automated alerts for security authorities in the case of any abnormal detection. Experimental results show that we have high accuracy with random forest of 97.32%, with decision tree of 94.50%, and with the naïve Bayes classifier of 99.37%. These algorithms indicate great potential for crowd risk prediction in public areas.

Mediating role of positivity in the relationship between state anxiety and problematic social media use during the COVID-19 pandemic.

This study investigated the mediating role of positivity in the relationship between state anxiety and problematic social media use during the COVID-19 pandemic. The Positivity Scale, State-Trait Anxiety Inventory, and Bergen Social Media Addiction Scale were used to collect data from 834 social media users. SEM-based mediation analysis was used to test hypothesized relationships. Results indicated that positivity has a direct negative effect on problematic social media use. Furthermore, the results indicated that state anxiety has a direct negative effect on positivity and state anxiety has a direct positive effect on problematic social media use. Positivity significantly mediates the relationship between state anxiety and problematic social media use.

Acute Limb Ischemia in Hospitalized COVID-19 Patients.

BACKGROUND: COVID-19 is a multisystemic disorder. Hematologic and cardiovascular involvement of COVID-19 causes thromboembolic events across multiple organs which mainly manifest as venous thromboembolism, and rarely, peripheral arterial thromboembolic events. In-situ thrombosis of a healthy, non-atherosclerotic native artery is rare, and COVID-19 has been reported to be a cause of this phenomenon. We aimed to report our institutional experience with COVID-19 patients who developed acute limb ischemia (ALI) during hospitalization or after discharge. METHODS: This was a single-center cross-sectional study. Records of all patients ≥18 years of age admitted to a tertiary center with a confirmed diagnosis of COVID-19 infection between September 1 and December 31, 2020 were retrospectively examined. Data regarding patient demographics, co-morbidities and outcomes were collected. Patients were followed-up during index hospitalization and for 30 days postdischarge. Acute limb ischemia was diagnosed by means of duplex ultrasound and computed tomography angiography in the presence of a clinical suspicion. RESULTS: A total of 681 consecutive patients (38.5% women) were hospitalized with a confirmed diagnosis of COVID-19 during the study period. Median age was 63 years (IQR, 52-74). In-hospital mortality occurred in 94 (13.8%) patients. Ninety (13.2%) patients required intensive care unit admission at some point of their hospital stay. Six (0.9%) patients (one woman) with a median age of 62 years experienced ALI (IQR, 59-64.3). All patients were receiving low molecular weight heparin when they developed ALI. The median of duration between COVID-19 diagnosis and ALI symptom onset was 13 days (IQR, 11.3-14). Three patients underwent emergent surgical thrombectomy combined with systemic anticoagulation, and 3 received systemic anticoagulation alone. Two patients with ALI did not survive to hospital discharge. Among survivors, 1 patient underwent bilateral major amputations, and another underwent a minor amputation within 1 month of hospital discharge. Symptoms of ALI completely resolved in 2 patients without sequelae. CONCLUSIONS: COVID-19 is a multisystemic disorder with involvement of hematologic and cardiovascular systems. Despite widespread use of thromboprophylaxis, hospitalized patients with COVID-19 are at increased risk of ALI, and subsequent limb loss or even death.

Acute limb ischemia in a COVID-19 patient

Introduction Acute limb ischemia (ALI) is defined as an abrupt decrease in arterial perfusion of a limb with a threat to viability of the limb. Coronavirus disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2, and has been declared as a global pandemic by the World Health Organization. Patients with COVID-19 have deranged blood coagulation parameters and are prone to thromboembolic events. This hypercoagulable state caused by COVID-19 mainly manifests as venous thromboembolism. Peripheral arterial involvement is less frequent. We present a case of a spontaneous ALI in a COVID-19 patient. Case A 62-year-old man with an insignificant past medical history presented with ALI 12 days after an initial diagnosis of COVID-19. He was on therapeutic doses of low molecular weight heparin when ischemic symptoms developed. A surgical thrombectomy was unsuccessful. He partially benefited from intravenous unfractionated heparin and iloprost infusions. He was discharged home on postoperative day 14, and is scheduled to have an amputaion of the 1st toe. Conclusions COVID-19 infection is associated with an increase incidence of thromboembolic events, including ALI. Even young and otherwise healthy patients may develop ALI despite the use of prophylactic anticoagulation. Management of ALI in COVID-19 patients might be harder than expected, due to the hypercoagulable state. Patients may benefit from prolonged postoperative unfractionated heparin administration.