Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

Models for COVID-19 Pandemic: A Comparative Analysis

ArXiv Preprint

COVID-19 pandemic represents an unprecedented global health crisis in the last 100 years. Its economic, social and health impact continues to grow and is likely to end up as one of the worst global disasters since the 1918 pandemic and the World Wars. Mathematical models have played an important role in the ongoing crisis; they have been used to inform public policies and have been instrumental in many of the social distancing measures that were instituted worldwide. In this article we review some of the important mathematical models used to support the ongoing planning and response efforts. These models differ in their use, their mathematical form and their scope.

Aniruddha Adiga, Devdatt Dubhashi, Bryan Lewis, Madhav Marathe, Srinivasan Venkatramanan, Anil Vullikanti


General General

Automated feature detection in dental periapical radiographs by using deep learning.

In Oral surgery, oral medicine, oral pathology and oral radiology ; h5-index 33.0

OBJECTIVE : The aim of this study was to investigate automated feature detection, segmentation, and quantification of common findings in periapical radiographs (PRs) by using deep learning (DL)-based computer vision techniques.

STUDY DESIGN : Caries, alveolar bone recession, and interradicular radiolucencies were labeled on 206 digital PRs by 3 specialists (2 oral pathologists and 1 endodontist). The PRs were divided into "Training and Validation" and "Test" data sets consisting of 176 and 30 PRs, respectively. Multiple transformations of image data were used as input to deep neural networks during training. Outcomes of existing and purpose-built DL architectures were compared to identify the most suitable architecture for automated analysis.

RESULTS : The U-Net architecture and its variant significantly outperformed Xnet and SegNet in all metrics. The overall best performing architecture on the validation data set was "U-Net+Densenet121" (mean intersection over union [mIoU] = 0.501; Dice coefficient = 0.569). Performance of all architectures degraded on the "Test" data set; "U-Net" delivered the best performance (mIoU = 0.402; Dice coefficient = 0.453). Interradicular radiolucencies were the most difficult to segment.

CONCLUSIONS : DL has potential for automated analysis of PRs but warrants further research. Among existing off-the-shelf architectures, U-Net and its variants delivered the best performance. Further performance gains can be obtained via purpose-built architectures and a larger multicentric cohort.

Khan Hassan Aqeel, Haider Muhammad Ali, Ishaq Hamna, Kiyani Amber, Sohail Kanwal, Muhammad Muhammad, Khurram Syed Ali


General General

Impact of lung segmentation on the diagnosis and explanation of COVID-19 in chest X-ray images

ArXiv Preprint

The COVID-19 pandemic is undoubtedly one of the biggest public health crises our society has ever faced. This paper's main objectives are to demonstrate the impact of lung segmentation in COVID-19 automatic identification using CXR images and evaluate which contents of the image decisively contribute to the identification. We have performed lung segmentation using a U-Net CNN architecture, and the classification using three well-known CNN architectures: VGG, ResNet, and Inception. To estimate the impact of lung segmentation, we applied some Explainable Artificial Intelligence (XAI), such as LIME and Grad-CAM. To evaluate our approach, we built a database named RYDLS-20-v2, following our previous publication and the COVIDx database guidelines. We evaluated the impact of creating a COVID-19 CXR image database from different sources, called database bias, and the COVID-19 generalization from one database to another, representing our less biased scenario. The experimental results of the segmentation achieved a Jaccard distance of 0.034 and a Dice coefficient of 0.982. In the best and more realistic scenario, we achieved an F1-Score of 0.74 and an area under the ROC curve of 0.9 for COVID-19 identification using segmented CXR images. Further testing and XAI techniques suggest that segmented CXR images represent a much more realistic and less biased performance. More importantly, the experiments conducted show that even after segmentation, there is a strong bias introduced by underlying factors from the data sources, and more efforts regarding the creation of a more significant and comprehensive database still need to be done.

Lucas O. Teixeira, Rodolfo M. Pereira, Diego Bertolini, Luiz S. Oliveira, Loris Nanni, Yandre M. G. Costa


Surgery Surgery

Reduced postoperative pain using Nociception Level-guided fentanyl dosing during sevoflurane anaesthesia: a randomised controlled trial.

In British journal of anaesthesia ; h5-index 72.0

BACKGROUND : The majority of postoperative patients report moderate to severe pain, possibly related to opioid underdosing or overdosing during surgery. Objective guidance of opioid dosing using the Nociception Level (NOL) index, a multiparameter artificial intelligence-driven index designed to monitor nociception during surgery, may lead to a more appropriate analgesic regimen, with effects beyond surgery. We tested whether NOL-guided opioid dosing during general anaesthesia results in less postoperative pain.

METHODS : In this two-centre RCT, 50 patients undergoing abdominal surgery under fentanyl/sevoflurane anaesthesia were randomised to NOL-guided fentanyl dosing or standard care in which fentanyl dosing was based on haemodynamics. The primary endpoint of the study was postoperative pain assessed in the PACU.

RESULTS : Median postoperative pain scores were 3.2 (inter-quartile range 1.3-4.3) and 4.8 (3.0-5.3) in NOL-guided and standard care groups, respectively (P=0.006). Postoperative morphine consumption (standard deviation) was 0.06 (0.07) mg kg-1 (NOL-guided group) and 0.09 (0.09) mg kg-1 (control group; P=0.204). During surgery, fentanyl dosing was not different between groups (NOL-guided group: 6.4 [4.2] μg kg-1vs standard care: 6.0 [2.2] μg kg-1, P=0.749), although the variation between patients was greater in the NOL-guided group (% coefficient of variation 66% in the NOL-guided group vs 37% in the standard care group).

CONCLUSIONS : Despite absence of differences in fentanyl and morphine consumption during and after surgery, a 1.6-point improvement in postoperative pain scores was observed in the NOL-guided group. We attribute this to NOL-driven rather than BP- and HR-driven fentanyl dosing during anaesthesia.

CLINICAL TRIAL REGISTRATION : under identifier NL7845.

Meijer Fleur, Honing Maarten, Roor Tessa, Toet Samantha, Calis Paul, Olofsen Erik, Martini Chris, van Velzen Monique, Aarts Leon, Niesters Marieke, Boon Martijn, Dahan Albert


nociception, nociception level-guided anaesthesia, opioid, postoperative pain, stress hormones

General General

When Healthcare Meets Off-the-Shelf WiFi: A Non-Wearable and Low-Costs Approach for In-Home Monitoring

ArXiv Preprint

As elderly population grows, social and health care begin to face validation challenges, in-home monitoring is becoming a focus for professionals in the field. Governments urgently need to improve the quality of healthcare services at lower costs while ensuring the comfort and independence of the elderly. This work presents an in-home monitoring approach based on off-the-shelf WiFi, which is low-costs, non-wearable and makes all-round daily healthcare information available to caregivers. The proposed approach can capture fine-grained human pose figures even through a wall and track detailed respiration status simultaneously by off-the-shelf WiFi devices. Based on them, behavioral data, physiological data and the derived information (e.g., abnormal events and underlying diseases), of the elderly could be seen by caregivers directly. We design a series of signal processing methods and a neural network to capture human pose figures and extract respiration status curves from WiFi Channel State Information (CSI). Extensive experiments are conducted and according to the results, off-the-shelf WiFi devices are capable of capturing fine-grained human pose figures, similar to cameras, even through a wall and track accurate respiration status, thus demonstrating the effectiveness and feasibility of our approach for in-home monitoring.

Lingchao Guo, Zhaoming Lu, Shuang Zhou, Xiangming Wen, Zhihong He


Public Health Public Health

Leading dietary determinants identified using machine learning techniques and a healthy diet score for changes in cardiometabolic risk factors in children: a longitudinal analysis.

In Nutrition journal ; h5-index 45.0

BACKGROUND : Identifying leading dietary determinants for cardiometabolic risk (CMR) factors is urgent for prioritizing interventions in children. We aimed to identify leading dietary determinants for the change in CMR and create a healthy diet score (HDS) to predict CMR in children.

METHODS : We included 5676 children aged 6-13 years in the final analysis with physical examinations, blood tests, and diets assessed at baseline and one year later. CMR score (CMRS) was computed by summing Z-scores of waist circumference, an average of systolic and diastolic blood pressure (SBP and DBP), fasting glucose, high-density lipoprotein cholesterol (HDL-C, multiplying by - 1), and triglycerides. Machine learning was used to identify leading dietary determinants for CMR and an HDS was then computed.

RESULTS : The nine leading predictors for CMRS were refined grains, seafood, fried foods, sugar-sweetened beverages, wheat, red meat other than pork, rice, fungi and algae, and roots and tubers with the contribution ranging from 3.9 to 19.6% of the total variance. Diets high in seafood, rice, and red meat other than pork but low in other six food groups were associated with a favorable change in CMRS. The HDS was computed based on these nine dietary factors. Children with HDS ≥8 had a higher decrease in CMRS (β (95% CI): - 1.02 (- 1.31, - 0.73)), BMI (- 0.08 (- 0.16, - 0.00)), SBP (- 0.46 (- 0.58, - 0.34)), DBP (- 0.46 (- 0.58, - 0.34)), mean arterial pressure (- 0.50 (- 0.62, - 0.38)), fasting glucose (- 0.22 (- 0.32, - 0.11)), insulin (- 0.52 (- 0.71, - 0.32)), and HOMA-IR (- 0.55 (- 0.73, - 0.36)) compared to those with HDS ≦3. Improved HDS during follow-up was associated with favorable changes in CMRS, BMI, percent body fat, SBP, DBP, mean arterial pressure, HDL-C, fasting glucose, insulin, and HOMA-IR.

CONCLUSION : Diets high in seafood, rice, and red meat other than pork and low in refined grains, fried foods, sugar-sweetened beverages, and wheat are leading healthy dietary factors for metabolic health in children. HDS is strongly predictive of CMR factors.

Shang Xianwen, Li Yanping, Xu Haiquan, Zhang Qian, Liu Ailing, Du Songming, Guo Hongwei, Ma Guansheng


Cardiometabolic risk factors, Children, Healthy diet score, Leading dietary determinants, Machine learning