COVID19-CT-dataset: an open-access chest CT image repository of 1000+ patients with confirmed COVID-19 diagnosis

Development and multicentric external validation of a prognostic COVID-19 severity model based on thoracic CT

BACKGROUND

Risk stratification of COVID-19 patients can support therapeutic decisions, planning and resource allocation in the hospital. In times of high incidence, a prognostic model based on data efficiently retrieved from one source can enable fast decision support.

METHODS

A model was developed to identify patients at risk of developing severe COVID-19 within one month based on their age, sex and imaging features extracted from the thoracic computed tomography (CT). The model was trained on publicly available data from the Study of Thoracic CT in COVID-19 (STOIC) challenge and validated on unseen data from the same study and an external, multicentric dataset. The model, trained on data acquired before any variant of concern dominated, was assessed separately on data collected at later stages of the pandemic when the delta and omicron variants were most prevalent.

RESULTS

A logistic regression based on handcrafted features was found to perform on par with a direct deep learning approach, and the former was selected for simplicity. Volumetric and intensity-based features of lesions and healthy lung parenchyma proved most predictive, in addition to patient age and sex. The model reached an area under the curve of 0.78 on the challenge test set and 0.74 on the external test set. The performance did not drop for the subset acquired at a later stage of the pandemic.

CONCLUSIONS

A logistic regression utilizing features from thoracic CT and its metadata can provide rapid decision support by estimating short-term COVID-19 severity. Its stable performance underscores its potential for real-world clinical integration. By enabling rapid risk stratification using readily available imaging data, this approach can support early clinical decision-making, optimize resource allocation, and improve patient management, particularly during surges in COVID-19 cases. Furthermore, this study provides a foundation for future research on prognostic modelling in respiratory infections.

ScreenDx, an artificial intelligence-based algorithm for the incidental detection of pulmonary fibrosis

BACKGROUND

Nonspecific symptoms and variability in radiographic reporting patterns contribute to a diagnostic delay of the diagnosis of pulmonary fibrosis. An attractive solution is the use of machine-learning algorithms to screen for radiographic features suggestive of pulmonary fibrosis. Thus, we developed and validated a machine learning classifier algorithm (ScreenDx) to screen computed tomography imaging and identify incidental cases of pulmonary fibrosis.

METHODS

ScreenDx is a deep learning convolutional neural network that was developed from a multi-source dataset (cohort A) of 3,658 cases of normal and abnormal CT's, including CT's from patients with COPD, emphysema, and community-acquired pneumonia. Cohort B, a US-based cohort (n = 381) was used for tuning the algorithm, and external validation was performed on cohort C (n = 683), a separate international dataset.

RESULTS

At the optimal threshold, the sensitivity and specificity for detection of pulmonary fibrosis in cohort B was 0.91 (95 % CI 88-94 %) and 0.95 (95 % CI 93-97 %), respectively, with AUC 0.98. In the external validation dataset (cohort C), the sensitivity and specificity were 1.0 (95 % 99.9-100.0) and 0.98 (95 % CI 97.9-99.6), respectively, with AUC 0.997. There were no significant differences in the ability of ScreenDx to identify pulmonary fibrosis based on CT manufacturer (Phillips, Toshiba, GE Healthcare, or Siemens) or slice thickness (2 mm vs 2-4 mm vs 4 mm).

CONCLUSION

Regardless of CT manufacturer or slice thickness, ScreenDx demonstrated high performance across two, multi-site datasets for identifying incidental cases of pulmonary fibrosis. This suggest that the algorithm may be generalizable across patient populations and different healthcare systems.

CADS: A Self-Supervised Learner via Cross-Modal Alignment and Deep Self-Distillation for CT Volume Segmentation

Self-supervised learning (SSL) has long had great success in advancing the field of annotation-efficient learning. However, when applied to CT volume segmentation, most SSL methods suffer from two limitations, including rarely using the information acquired by different imaging modalities and providing supervision only to the bottleneck encoder layer. To address both limitations, we design a pretext task to align the information in each 3D CT volume and the corresponding 2D generated X-ray image and extend self-distillation to deep self-distillation. Thus, we propose a self-supervised learner based on Cross-modal Alignment and Deep Self-distillation (CADS) to improve the encoder's ability to characterize CT volumes. The cross-modal alignment is a more challenging pretext task that forces the encoder to learn better image representation ability. Deep self-distillation provides supervision to not only the bottleneck layer but also shallow layers, thus boosting the abilities of both. Comparative experiments show that, during pre-training, our CADS has lower computational complexity and GPU memory cost than competing SSL methods. Based on the pre-trained encoder, we construct PVT-UNet for 3D CT volume segmentation. Our results on seven downstream tasks indicate that PVT-UNet outperforms state-of-the-art SSL methods like MOCOv3 and DiRA, as well as prevalent medical image segmentation methods like nnUNet and CoTr. Code and pre-trained weight will be available at https://github.com/yeerwen/CADS.

Addressing fairness issues in deep learning-based medical image analysis: a systematic review

Deep learning algorithms have demonstrated remarkable efficacy in various medical image analysis (MedIA) applications. However, recent research highlights a performance disparity in these algorithms when applied to specific subgroups, such as exhibiting poorer predictive performance in elderly females. Addressing this fairness issue has become a collaborative effort involving AI scientists and clinicians seeking to understand its origins and develop solutions for mitigation within MedIA. In this survey, we thoroughly examine the current advancements in addressing fairness issues in MedIA, focusing on methodological approaches. We introduce the basics of group fairness and subsequently categorize studies on fair MedIA into fairness evaluation and unfairness mitigation. Detailed methods employed in these studies are presented too. Our survey concludes with a discussion of existing challenges and opportunities in establishing a fair MedIA and healthcare system. By offering this comprehensive review, we aim to foster a shared understanding of fairness among AI researchers and clinicians, enhance the development of unfairness mitigation methods, and contribute to the creation of an equitable MedIA society.

HRCTCov19-a high-resolution chest CT scan image dataset for COVID-19 diagnosis and differentiation

INTRODUCTION

Computed tomography (CT) was a widely used diagnostic technique for COVID-19 during the pandemic. High-Resolution Computed Tomography (HRCT), is a type of computed tomography that enhances image resolution through the utilization of advanced methods. Due to privacy concerns, publicly available COVID-19 CT image datasets are incredibly tough to come by, leading to it being challenging to research and create AI-powered COVID-19 diagnostic algorithms based on CT images.

DATA DESCRIPTION

To address this issue, we created HRCTCov19, a new COVID-19 high-resolution chest CT scan image collection that includes not only COVID-19 cases of Ground Glass Opacity (GGO), Crazy Paving, and Air Space Consolidation but also CT images of cases with negative COVID-19. The HRCTCov19 dataset, which includes slice-level and patient-level labeling, has the potential to assist in COVID-19 research, in particular for diagnosis and a distinction using AI algorithms, machine learning, and deep learning methods. This dataset, which can be accessed through the web at http://databiox.com , includes 181,106 chest HRCT images from 395 patients labeled as GGO, Crazy Paving, Air Space Consolidation, and Negative.

Automatic chest computed tomography image noise quantification using deep learning

PURPOSE

This study aimed to develop a deep learning (DL) method for noise quantification for clinical chest computed tomography (CT) images without the need for repeated scanning or homogeneous tissue regions.

METHODS

A comprehensive phantom CT dataset (three dose levels, six reconstruction methods, amounting to 9240 slices) was acquired and used to train a convolutional neural network (CNN) to output an estimate of local image noise standard deviations (SD) from a single CT scan input. The CNN model consisting of seven convolutional layers was trained on the phantom image dataset representing a range of scan parameters and was tested with phantom images acquired in a variety of different scan conditions, as well as publicly available chest CT images to produce clinical noise SD maps.

RESULTS

Noise SD maps predicted by the CNN agreed well with the ground truth both visually and numerically in the phantom dataset (errors of < 5 HU for most scan parameter combinations). In addition, the noise SD estimates obtained from clinical chest CT images were similar to running-average based reference estimates in areas without prominent tissue interfaces.

CONCLUSIONS

Predicting local noise magnitudes without the need for repeated scans is feasible using DL. Our implementation trained with phantom data was successfully applied to open-source clinical data with heterogeneous tissue borders and textures. We suggest that automatic DL noise mapping from clinical patient images could be used as a tool for objective CT image quality estimation and protocol optimization.

Deep learning-based ultrasound transducer induced CT metal artifact reduction using generative adversarial networks for ultrasound-guided cardiac radioablation

In US-guided cardiac radioablation, a possible workflow includes simultaneous US and planning CT acquisitions, which can result in US transducer-induced metal artifacts on the planning CT scans. To reduce the impact of these artifacts, a metal artifact reduction (MAR) algorithm has been developed based on a deep learning Generative Adversarial Network called Cycle-MAR, and compared with iMAR (Siemens), O-MAR (Philips) and MDT (ReVision Radiology), and CCS-MAR (Combined Clustered Scan-based MAR). Cycle-MAR was trained with a supervised learning scheme using sets of paired clinical CT scans with and without simulated artifacts. It was then evaluated on CT scans with real artifacts of an anthropomorphic phantom, and on sets of clinical CT scans with simulated artifacts which were not used for Cycle-MAR training. Image quality metrics and HU value-based analysis were used to evaluate the performance of Cycle-MAR compared to the other algorithms. The proposed Cycle-MAR network effectively reduces the negative impact of the metal artifacts. For example, the calculated HU value improvement percentage for the cardiac structures in the clinical CT scans was 59.58%, 62.22%, and 72.84% after MDT, CCS-MAR, and Cycle-MAR application, respectively. The application of MAR algorithms reduces the impact of US transducer-induced metal artifacts on CT scans. In comparison to iMAR, O-MAR, MDT, and CCS-MAR, the application of developed Cycle-MAR network on CT scans performs better in reducing these metal artifacts.

Radiological Study of Atlas Arch Defects with Meta-Analysis and a Proposed New Classification

This study consists of a retrospective cohort study, a systematic review, and a meta-analysis which were separately conducted. This study aimed to investigate the prevalence of atlas arch defects, generate an evidence-based synthesis, and propose a common classification system for the anterior and combined atlas arch defects. Atlas arch defects are well-corticated gaps in the anterior or posterior arch of the atlas. When both arches are involved, it is known as a combined arch defect. Awareness of these defects is essential for avoiding complications during surgical procedures on the upper spine. The prevalence of arch defects was investigated in an open-access OPC-Radiomics (Radiomic Biomarkers in Oropharyngeal Carcinoma) dataset comprising 606 head and neck computed tomography scans from oropharyngeal cancer patients. A systematic review and meta-analysis were performed to generate prevalence estimates of atlas arch defects and propose a classification system for the anterior and combined atlas arch defects. The posterior arch defect was found in 20 patients (3.3%) out of the 606 patients investigated. The anterior arch defect was not observed in any patient, while a combined arch defect was observed in one patient (0.2%). A meta-analysis of 13,539 participants from 14 studies, including the present study, yielded a pooled-posterior arch defect prevalence of 2.07% (95% confidence interval [CI], 1.22%-2.92%). The prevalences of anterior and combined arch defects were 0.00% (95% CI, 0.00%-0.10%) and 0.14% (95% CI, 0.04%-0.25%), respectively. The anterior and combined arch defects were classified into five subtypes based on their morphology and frequency. The present study showed that atlas arch defects were present in approximately 2% of the general population. For future studies, larger sample sizes should be used for studying arch defects to avoid the small-study effect and to predict the prevalence accurately.

Roadmaps for Guiding Chest Computed Tomography Interpretation involving Pneumonia. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2023;2023:1-4. [PMID: 38083702 DOI: 10.1109/embc40787.2023.10340639] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

CT scanning of the chest is one the most important imaging modalities available for pulmonary disease diagnosis. Lung segmentation plays a crucial step in the pipeline of computer-aided analysis and diagnosis. As deep learning models have achieved human-level accuracy in semantic segmentation of anatomical structures, we propose to use trained deep learning models to predict both healthy and infectious areas in chest CT slices. The semantic segmentation results are summarized and visualized using volume rendering technology in the form of roadmaps. The roadmaps consist of both location and volume information that can be used as a location guidance for inspecting suspected pulmonary lesions of chest CT and can possibly be combined into a rapid triage algorithm for treating acute pulmonary diseases.Clinical Relevance- This research applied trained semantic segmentation models in identifying normal lung and pneumonic infection areas to generate a roadmap for assisting medical doctors in browsing chest CT and prognostication.

Ensemble Deep Learning Derived from Transfer Learning for Classification of COVID-19 Patients on Hybrid Deep-Learning-Based Lung Segmentation: A Data Augmentation and Balancing Framework

BACKGROUND AND MOTIVATION

Lung computed tomography (CT) techniques are high-resolution and are well adopted in the intensive care unit (ICU) for COVID-19 disease control classification. Most artificial intelligence (AI) systems do not undergo generalization and are typically overfitted. Such trained AI systems are not practical for clinical settings and therefore do not give accurate results when executed on unseen data sets. We hypothesize that ensemble deep learning (EDL) is superior to deep transfer learning (TL) in both non-augmented and augmented frameworks.

METHODOLOGY

The system consists of a cascade of quality control, ResNet-UNet-based hybrid deep learning for lung segmentation, and seven models using TL-based classification followed by five types of EDL's. To prove our hypothesis, five different kinds of data combinations (DC) were designed using a combination of two multicenter cohorts-Croatia (80 COVID) and Italy (72 COVID and 30 controls)-leading to 12,000 CT slices. As part of generalization, the system was tested on unseen data and statistically tested for reliability/stability.

RESULTS

Using the K5 (80:20) cross-validation protocol on the balanced and augmented dataset, the five DC datasets improved TL mean accuracy by 3.32%, 6.56%, 12.96%, 47.1%, and 2.78%, respectively. The five EDL systems showed improvements in accuracy of 2.12%, 5.78%, 6.72%, 32.05%, and 2.40%, thus validating our hypothesis. All statistical tests proved positive for reliability and stability.

CONCLUSION

EDL showed superior performance to TL systems for both (a) unbalanced and unaugmented and (b) balanced and augmented datasets for both (i) seen and (ii) unseen paradigms, validating both our hypotheses.

Anatomical Variants Identified on Computed Tomography of Oropharyngeal Carcinoma Patients

Background and Objectives: Anatomical variations in the head, neck and chest are common, and are observed as occasional findings on computed tomography (CT). Although anatomical variations are mostly asymptomatic and do not cause any negative influence on the body function, they may jeopardize diagnosis and may be confused with pathological conditions. The presence of variations may also limit surgical access during tumor removal. The aim of this study was to investigate the prevalence of six anatomical variations—os acromiale, episternal ossicles, cervical rib, Stafne bone cavity, azygos lobe and tracheal bronchus—in an open-access computed tomography dataset obtained from oropharyngeal cancer patients. Materials and Methods: A total of 606 upper-chest and neck computed-tomography scans (79.4% male and 20.6% female) were retrospectively investigated. Sex difference was evaluated using the z-test for two proportions. Results: Os acromiale, episternal ossicles, cervical rib, Stafne bone cavity, azygos lobe, and tracheal bronchus were present in 3.1%, 2.2%, 0.2%, 0%, 0.3% and 0.5%, respectively, of all patients. Os acromiale was identified as meso-acromion in 86.6%, and as pre-acromion in 17.4%, of all acromia. Episternal ossicles were present unilaterally in 58.3%, and bilaterally in 41.7%, of all sterna. Only the cervical rib showed a sex difference in prevalence. Conclusions: awareness of these variations is important for radiologists interpreting head, neck and chest CTs; for example, those of oropharyngeal cancer patients. This study also illustrates the applicability of publicly available datasets in prevalence-based anatomical research. While most of the variations investigated in the present study are well-known, the episternal ossicles are not well explored, and need further investigation.

Deep Convolutional Neural Networks for Detecting COVID-19 Using Medical Images: A Survey

Coronavirus Disease 2019 (COVID-19), which is caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-COV-2), surprised the world in December 2019 and has threatened the lives of millions of people. Countries all over the world closed worship places and shops, prevented gatherings, and implemented curfews to stand against the spread of COVID-19. Deep Learning (DL) and Artificial Intelligence (AI) can have a great role in detecting and fighting this disease. Deep learning can be used to detect COVID-19 symptoms and signs from different imaging modalities, such as X-Ray, Computed Tomography (CT), and Ultrasound Images (US). This could help in identifying COVID-19 cases as a first step to curing them. In this paper, we reviewed the research studies conducted from January 2020 to September 2022 about deep learning models that were used in COVID-19 detection. This paper clarified the three most common imaging modalities (X-Ray, CT, and US) in addition to the DL approaches that are used in this detection and compared these approaches. This paper also provided the future directions of this field to fight COVID-19 disease.

Robust Classification and Detection of Big Medical Data Using Advanced Parallel K-Means Clustering, YOLOv4, and Logistic Regression

Big-medical-data classification and image detection are crucial tasks in the field of healthcare, as they can assist with diagnosis, treatment planning, and disease monitoring. Logistic regression and YOLOv4 are popular algorithms that can be used for these tasks. However, these techniques have limitations and performance issue with big medical data. In this study, we presented a robust approach for big-medical-data classification and image detection using logistic regression and YOLOv4, respectively. To improve the performance of these algorithms, we proposed the use of advanced parallel k-means pre-processing, a clustering technique that identified patterns and structures in the data. Additionally, we leveraged the acceleration capabilities of a neural engine processor to further enhance the speed and efficiency of our approach. We evaluated our approach on several large medical datasets and showed that it could accurately classify large amounts of medical data and detect medical images. Our results demonstrated that the combination of advanced parallel k-means pre-processing, and the neural engine processor resulted in a significant improvement in the performance of logistic regression and YOLOv4, making them more reliable for use in medical applications. This new approach offers a promising solution for medical data classification and image detection and may have significant implications for the field of healthcare.

The choice of an autocorrelation length in dark-field lung imaging

Respiratory diseases are one of the most common causes of death, and their early detection is crucial for prompt treatment. X-ray dark-field radiography (XDFR) is a promising tool to image objects with unresolved micro-structures such as lungs. Using Talbot-Lau XDFR, we imaged inflated porcine lungs together with Polymethylmethacrylat (PMMA) microspheres (in air) of diameter sizes between 20 and 500 [Formula: see text] over an autocorrelation range of 0.8-5.2 [Formula: see text]. The results indicate that the dark-field extinction coefficient of porcine lungs is similar to that of densely-packed PMMA spheres with diameter of [Formula: see text], which is approximately the mean alveolar structure size. We evaluated that, in our case, the autocorrelation length would have to be limited to [Formula: see text] in order to image [Formula: see text]-thick lung tissue without critical visibility reduction (signal saturation). We identify the autocorrelation length to be the critical parameter of an interferometer that allows to avoid signal saturation in clinical lung dark-field imaging.

Grayscale Image Statistical Attributes Effectively Distinguish the Severity of Lung Abnormalities in CT Scan Slices of COVID-19 Patients

Grayscale statistical attributes analysed for 513 extract images taken from pulmonary computed tomography (CT) scan slices of 57 individuals (49 confirmed COVID-19 positive; eight confirmed COVID-19 negative) are able to accurately predict a visual score (VS from 0 to 4) used by a clinician to assess the severity of lung abnormalities in the patients. Some of these attributes can be used graphically to distinguish useful but overlapping distributions for the VS classes. Using machine and deep learning (ML/DL) algorithms with twelve grayscale image attributes as inputs enables the VS classes to be accurately distinguished. A convolutional neural network achieves this with better than 96% accuracy (only 18 images misclassified out of 513) on a supervised learning basis. Analysis of confusion matrices enables the VS prediction performance of ML/DL algorithms to be explored in detail. Those matrices demonstrate that the best performing ML/DL algorithms successfully distinguish between VS classes 0 and 1, which clinicians cannot readily do with the naked eye. Just five image grayscale attributes can also be used to generate an algorithmically defined scoring system (AS) that can also graphically distinguish the degree of pulmonary impacts in the dataset evaluated. The AS classification illustrated involves less overlap between its classes than the VS system and could be exploited as an automated expert system. The best-performing ML/DL models are able to predict the AS classes with better than 99% accuracy using twelve grayscale attributes as inputs. The decision tree and random forest algorithms accomplish that distinction with just one classification error in the 513 images tested.

A Survey on AI Techniques for Thoracic Diseases Diagnosis Using Medical Images

Thoracic diseases refer to disorders that affect the lungs, heart, and other parts of the rib cage, such as pneumonia, novel coronavirus disease (COVID-19), tuberculosis, cardiomegaly, and fracture. Millions of people die every year from thoracic diseases. Therefore, early detection of these diseases is essential and can save many lives. Earlier, only highly experienced radiologists examined thoracic diseases, but recent developments in image processing and deep learning techniques are opening the door for the automated detection of these diseases. In this paper, we present a comprehensive review including: types of thoracic diseases; examination types of thoracic images; image pre-processing; models of deep learning applied to the detection of thoracic diseases (e.g., pneumonia, COVID-19, edema, fibrosis, tuberculosis, chronic obstructive pulmonary disease (COPD), and lung cancer); transfer learning background knowledge; ensemble learning; and future initiatives for improving the efficacy of deep learning models in applications that detect thoracic diseases. Through this survey paper, researchers may be able to gain an overall and systematic knowledge of deep learning applications in medical thoracic images. The review investigates a performance comparison of various models and a comparison of various datasets.

Explainable multi-instance and multi-task learning for COVID-19 diagnosis and lesion segmentation in CT images

Coronavirus Disease 2019 (COVID-19) still presents a pandemic trend globally. Detecting infected individuals and analyzing their status can provide patients with proper healthcare while protecting the normal population. Chest CT (computed tomography) is an effective tool for screening of COVID-19. It displays detailed pathology-related information. To achieve automated COVID-19 diagnosis and lung CT image segmentation, convolutional neural networks (CNNs) have become mainstream methods. However, most of the previous works consider automated diagnosis and image segmentation as two independent tasks, in which some focus on lung fields segmentation and the others focus on single-lesion segmentation. Moreover, lack of clinical explainability is a common problem for CNN-based methods. In such context, we develop a multi-task learning framework in which the diagnosis of COVID-19 and multi-lesion recognition (segmentation of CT images) are achieved simultaneously. The core of the proposed framework is an explainable multi-instance multi-task network. The network learns task-related features adaptively with learnable weights, and gives explicable diagnosis results by suggesting local CT images with lesions as additional evidence. Then, severity assessment of COVID-19 and lesion quantification are performed to analyze patient status. Extensive experimental results on real-world datasets show that the proposed framework outperforms all the compared approaches for COVID-19 diagnosis and multi-lesion segmentation.

Unsupervised Domain Adaptation for Vertebrae Detection and Identification in 3D CT Volumes Using a Domain Sanity Loss

A variety of medical computer vision applications analyze 2D slices of computed tomography (CT) scans, whereas axial slices from the body trunk region are usually identified based on their relative position to the spine. A limitation of such systems is that either the correct slices must be extracted manually or labels of the vertebrae are required for each CT scan to develop an automated extraction system. In this paper, we propose an unsupervised domain adaptation (UDA) approach for vertebrae detection and identification based on a novel Domain Sanity Loss (DSL) function. With UDA the model’s knowledge learned on a publicly available (source) data set can be transferred to the target domain without using target labels, where the target domain is defined by the specific setup (CT modality, study protocols, applied pre- and processing) at the point of use (e.g., a specific clinic with its specific CT study protocols). With our approach, a model is trained on the source and target data set in parallel. The model optimizes a supervised loss for labeled samples from the source domain and the DSL loss function based on domain-specific “sanity checks” for samples from the unlabeled target domain. Without using labels from the target domain, we are able to identify vertebra centroids with an accuracy of 72.8%. By adding only ten target labels during training the accuracy increases to 89.2%, which is on par with the current state-of-the-art for full supervised learning, while using about 20 times less labels. Thus, our model can be used to extract 2D slices from 3D CT scans on arbitrary data sets fully automatically without requiring an extensive labeling effort, contributing to the clinical adoption of medical imaging by hospitals.

A Soft Labeling Approach to Develop Automated Algorithms that Incorporate Uncertainty in Pulmonary Opacification on Chest CT using COVID-19 Pneumonia

RATIONALE AND OBJECTIVES

Hard data labels for automated algorithm training are binary and cannot incorporate uncertainty between labels. We proposed and evaluated a soft labeling methodology to quantify opacification and percent well-aerated lung (%WAL) on chest CT, that considers uncertainty in segmenting pulmonary opacifications and reduces labeling burden.

MATERIALS AND METHODS

We retrospectively sourced 760 COVID-19 chest CT scans from five international centers between January and June 2020. We created pixel-wise labels for >27,000 axial slices that classify three pulmonary opacification patterns: pure ground-glass, crazy-paving, consolidation. We also quantified %WAL as the total area of lung without opacifications. Inter-user hard label variability was quantified using Shannon entropy (range=0-1.39, low-high entropy/variability). We incorporated a soft labeling and modeling cycle following an initial model with hard labels and compared performance using point-wise accuracy and intersection-over-union of opacity labels with ground-truth, and correlation with ground-truth %WAL.

RESULTS

Hard labels annotated by 12 radiologists demonstrated large inter-user variability (3.37% of pixels achieved complete agreement). Our soft labeling approach increased point-wise accuracy from 60.0% to 84.3% (p=0.01) compared to hard labeling at predicting opacification type and area involvement. The soft label model accurately predicted %WAL (R=0.900) compared to the hard label model (R=0.856), but the improvement was not statistically significant (p=0.349).

CONCLUSION

Our soft labeling approach increased accuracy for automated quantification and classification of pulmonary opacification on chest CT. Although we developed the model on COVID-19, our intent is broad application for pulmonary opacification contexts and to provide a foundation for future development using soft labeling methods.

Federated Learning Approach with Pre-Trained Deep Learning Models for COVID-19 Detection from Unsegmented CT images

(1) Background: Coronavirus disease 2019 (COVID-19) is an infectious disease caused by SARS-CoV-2. Reverse transcription polymerase chain reaction (RT-PCR) remains the current gold standard for detecting SARS-CoV-2 infections in nasopharyngeal swabs. In Romania, the first reported patient to have contracted COVID-19 was officially declared on 26 February 2020. (2) Methods: This study proposes a federated learning approach with pre-trained deep learning models for COVID-19 detection. Three clients were locally deployed with their own dataset. The goal of the clients was to collaborate in order to obtain a global model without sharing samples from the dataset. The algorithm we developed was connected to our internal picture archiving and communication system and, after running backwards, it encountered chest CT changes suggestive for COVID-19 in a patient investigated in our medical imaging department on the 28 January 2020. (4) Conclusions: Based on our results, we recommend using an automated AI-assisted software in order to detect COVID-19 based on the lung imaging changes as an adjuvant diagnostic method to the current gold standard (RT-PCR) in order to greatly enhance the management of these patients and also limit the spread of the disease, not only to the general population but also to healthcare professionals.

Artificial intelligence-based CT metrics used in predicting clinical outcome of COVID-19 in young and middle-aged adults

Background

Currently, most researchers mainly analyzed coronavirus disease 2019 (COVID‐19) pneumonia visually or qualitatively, probably somewhat time‐consuming and not precise enough.

Purpose

This study aimed to excavate more information, such as differences in distribution, density, and severity of pneumonia lesions between males and females in a specific age group using artificial intelligence (AI)‐based computed tomography (CT) metrics. Besides, these metrics were incorporated into a clinical regression model to predict the short‐term outcome.

Materials and methods

The clinical, laboratory information and a series of HRCT images from 49 patients, aged from 20 to 50 years and confirmed with COVID‐19, were collected. The volumes and percentages of infection (POIs) among bilateral lungs and each bronchopulmonary segment were extracted using uAI‐Discover‐NCP software (version R001). The POI in three HU ranges (i.e., <−300, −300–49, and ≥50 HU representing ground‐glass opacity [GGO], mixed opacity, and consolidation) were also extracted. Hospital stay was predicted with several POI after adjusting days from illness onset to admission, leucocytes, lymphocytes, C‐reactive protein, age, and gender using a multiple linear regression model. A total of 91 patients aged 20–50 from public database were selected.

Results

Right lower lobes had the highest POI, followed by left lower lobes, right upper lobes, middle lobes, and left upper lobes. The distributions in lung lobes and segments were different between the sexes. Men had a higher total POI and GGO of the lungs, but less consolidation than women in initial CT (all p < 0.05). The total POI, percentage of consolidation on initial CT, and changed POI were positively correlated with hospital stay in the model. A total of 91 patients aged 20–50 years in the public database were selected, and AI segmentation was performed. The POI of the lower lobes was obviously higher than that in the upper lobes; the POI of each segment of the right upper lobe in the males was higher than that in the females, which was consistent with the result of the 49 patients previously.

Conclusion

Both men and women had characteristic distributions in lung lobes and bronchopulmonary segments. AI‐based CT quantitative metrics can provide more precise information regarding lesion distribution and severity to predict clinical outcome.

A comparative study of medical image enhancement algorithms and quality assessment metrics on COVID-19 CT images

Medical imaging can help doctors in better diagnosis of several conditions. During the present COVID-19 pandemic, timely detection of novel coronavirus is crucial, which can help in curing the disease at an early stage. Image enhancement techniques can improve the visual appearance of COVID-19 CT scans and speed-up the process of diagnosis. In this study, we analyze some state-of-the-art image enhancement techniques for their suitability in enhancing the CT scans of COVID-19 patients. Six quantitative metrics, Entropy, SSIM, AMBE, PSNR, EME, and EMEE, are used to evaluate the enhanced images. Two experienced radiologists were involved in the study to evaluate the performance of the enhancement techniques and the quantitative metrics used to assess them.

Anatomical variants identified on chest computed tomography of 1000+ COVID-19 patients from an open-access dataset

Chest computed tomography (CT) has been the preferred imaging modality during the pandemic owing to its sensitivity in detecting COVID‐19 infections. Recently, a large number of COVID‐19 imaging datasets have been deposited in public databases, leading to rapid advances in COVID‐19 research. However, the application of these datasets beyond COVID‐19‐related research has been little explored. The authors believe that they could be used in anatomical research to elucidate the link between anatomy and disease and to study disease‐related alterations to normal anatomy. Therefore, the present study was designed to investigate the prevalence of six well‐known anatomical variants in the thorax using open‐access CT images obtained from over 1000 Iranian COVID‐19 patients aged between 6 and 89 years (60.9% male and 39.1% female). In brief, we found that the azygos lobe, tracheal bronchus, and cardiac bronchus were present in 0.8%, 0.2%, and 0% of the patients, respectively. Variations of the sternum, including sternal foramen, episternal ossicles, and sternalis muscle, were observed in 9.6%, 2.9%, and 1.5%, respectively. We believe anatomists could benefit from using open‐access datasets as raw materials for research because these datasets are freely accessible and are abundant, though further research is needed to evaluate the uses of other datasets from different body regions and imaging modalities. Radiologists should also be aware of these common anatomical variants when examining lung CTs, especially since the use of this imaging modality has increased during the pandemic.