Search

SARS-CoV-2 is a novel virus that has been affecting the global population by spreading rapidly and causing severe complications, which require prompt and elaborate emergency treatment. Automatic tools to diagnose COVID-19 could potentially be an important and useful aid. Radiologists and clinicians could potentially rely on interpretable AI technologies to address the diagnosis and monitoring of COVID-19 patients. This paper aims to provide a comprehensive analysis of the state-of-the-art deep learning techniques for COVID-19 classification. The previous studies are methodically evaluated, and a summary of the proposed convolutional neural network (CNN)-based classification approaches is presented. The reviewed papers have presented a variety of CNN models and architectures that were developed to provide an accurate and quick automatic tool to diagnose the COVID-19 virus based on presented CT scan or X-ray images. In this systematic review, we focused on the critical components of the deep learning approach, such as network architecture, model complexity, parameter optimization, explainability, and dataset/code availability. The literature search yielded a large number of studies over the past period of the virus spread, and we summarized their past efforts. State-of-the-art CNN architectures, with their strengths and weaknesses, are discussed with respect to diverse technical and clinical evaluation metrics to safely implement current AI studies in medical practice. [ABSTRACT FROM AUTHOR]

An accurate and efficient automatic brain tumor segmentation algorithm is important for clinical practice. In recent years, there has been much interest in automatic segmentation algorithms that use convolutional neural networks. In this paper, we propose a novel hierarchical multi-scale segmentation network (HMNet), which contains a high-resolution branch and parallel multi-resolution branches. The high-resolution branch can keep track of the brain tumor's spatial details, and the multi-resolution feature exchange and fusion allow the network's receptive fields to adapt to brain tumors of different shapes and sizes. In particular, to overcome the large computational overhead caused by expensive 3D convolution, we propose a lightweight conditional channel weighting block to reduce GPU memory and improve the efficiency of HMNet. We also propose a lightweight multi-resolution feature fusion (LMRF) module to further reduce model complexity and reduce the redundancy of the feature maps. We run tests on the BraTS 2020 dataset to determine how well the proposed network would work. The dice similarity coefficients of HMNet for ET, WT, and TC are 0.781, 0.901, and 0.823, respectively. Many comparative experiments on the BraTS 2020 dataset and other two datasets show that our proposed HMNet has achieved satisfactory performance compared with the SOTA approaches. [ABSTRACT FROM AUTHOR]

Background: This paper presents a novel lightweight approach based on machine learning methods supporting COVID-19 diagnostics based on X-ray images. The presented schema offers effective and quick diagnosis of COVID-19. Methods: Real data (X-ray images) from hospital patients were used in this study. All labels, namely those that were COVID-19 positive and negative, were confirmed by a PCR test. Feature extraction was performed using a convolutional neural network, and the subsequent classification of samples used Random Forest, XGBoost, LightGBM and CatBoost. Results: The LightGBM model was the most effective in classifying patients on the basis of features extracted from X-ray images, with an accuracy of 1.00, a precision of 1.00, a recall of 1.00 and an F1-score of 1.00. Conclusion: The proposed schema can potentially be used as a support for radiologists to improve the diagnostic process. The presented approach is efficient and fast. Moreover, it is not excessively complex computationally. [ABSTRACT FROM AUTHOR]

Automatic and accurate estimation of choroidal thickness plays a very important role in a computer-aided system for eye diseases. One of the most common methods for automatic estimation of choroidal thickness is segmentation-based methods, in which the boundaries of the choroid are first detected from optical coherence tomography (OCT) images. The choroidal thickness is then computed based on the detected boundaries. A shortcoming in the segmentation-based methods is that the estimating precision greatly depends on the segmentation results. To avoid the dependence on the segmentation step, in this paper, we propose a direct method based on convolutional neural networks (CNNs) for estimating choroidal thickness without segmentation. Concretely, a B-scan image is first cropped into several patches. A trained CNN model is then used to estimate the choroidal thickness for each patch. The mean thickness of the choroid in the B-scan is obtained by taking the average of the choroidal thickness on each patch. Then, 150 OCT volumes are collected to evaluate the proposed method. The experiments show that the results obtained by the proposed method are very competitive with those obtained by segmentation-based methods, which indicates that direct estimation of choroidal thickness is very promising. [ABSTRACT FROM AUTHOR]

Background: To design a novel anomaly detection and localization approach using artificial intelligence methods using optical coherence tomography (OCT) scans for retinal diseases. Methods: High-resolution OCT scans from the publicly available Kaggle dataset and a local dataset were used by four state-of-the-art self-supervised frameworks. The backbone model of all the frameworks was a pre-trained convolutional neural network (CNN), which enabled the extraction of meaningful features from OCT images. Anomalous images included choroidal neovascularization (CNV), diabetic macular edema (DME), and the presence of drusen. Anomaly detectors were evaluated by commonly accepted performance metrics, including area under the receiver operating characteristic curve, F1 score, and accuracy. Results: A total of 25,315 high-resolution retinal OCT slabs were used for training. Test and validation sets consisted of 968 and 4000 slabs, respectively. The best performing across all anomaly detectors had an area under the receiver operating characteristic of 0.99. All frameworks were shown to achieve high performance and generalize well for the different retinal diseases. Heat maps were generated to visualize the quality of the frameworks' ability to localize anomalous areas of the image. Conclusions: This study shows that with the use of pre-trained feature extractors, the frameworks tested can generalize to the domain of retinal OCT scans and achieve high image-level ROC-AUC scores. The localization results of these frameworks are promising and successfully capture areas that indicate the presence of retinal pathology. Moreover, such frameworks have the potential to uncover new biomarkers that are difficult for the human eye to detect. Frameworks for anomaly detection and localization can potentially be integrated into clinical decision support and automatic screening systems that will aid ophthalmologists in patient diagnosis, follow-up, and treatment design. This work establishes a solid basis for further development of automated anomaly detection frameworks for clinical use. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) applied to cardiovascular disease (CVD) is enjoying great success in the field of scientific research. Electrocardiograms (ECGs) are the cornerstone form of examination in cardiology and are the most widely used diagnostic tool because they are widely available, inexpensive, and fast. Applications of AI to ECGs, especially deep learning (DL) methods using convolutional neural networks (CNNs), have been developed in many fields of cardiology in recent years. Deep learning methods provide valuable support for rapid ECG interpretation, demonstrating a diagnostic capability overlapping with specialists in the diagnosis of CVD by a classical analysis of macroscopic changes in the ECG trace. Through photoplethysmography, wearable devices can obtain single-derivative ECGs for the recognition of AI-diagnosed arrhythmias. In addition, CNNs have been developed that recognize no macroscopic electrocardiographic changes and can predict, from a 12-lead ECG, atrial fibrillation, even from sinus rhythm; left and right ventricular function; hypertrophic cardiomyopathy; acute coronary syndromes; or aortic stenosis. The fields of application are many, but numerous are the limitations, mainly associated with the reliability of the acquired data, an inability to verify black box processes, and medico-legal and ethical problems. The challenge of modern medicine is to recognize the limitations of AI and overcome them. [ABSTRACT FROM AUTHOR]

Atrial fibrillation (AF) is the most common sustained arrythmia and one of the strongest risk factors and causal mechanisms of ischemic stroke (IS). Acute IS due to AF tends to be more severe than with other etiology of IS and patients with treated AF have reported to experience worse outcomes after endovascular treatment compared with patients without AF. As cardioembolism accounts for more than a fifth of ISs and the risk of future stroke can be mitigated with effective anticoagulation, which has been shown to be effective and safe in patients with paroxysmal or sustained AF, the screening of patients with cryptogenic IS (CIS) for AF is paramount. Embolic stroke of undetermined source (ESUS) is a subtype of CIS with a high likelihood of cardioembolism. The European Stroke Organization and European Society of Cardiology guidelines recommend at least 72 h of screening when AF is suspected. The longer the screening and the earlier the time point after acute IS, the more likely the AF paroxysm is found. Several methods are available for short-term screening of AF, including in-hospital monitoring and wearable electrocardiogram recorders for home monitoring. Implantable loop monitors provide an effective long-term method to screen patients with high risk of AF after IS and artificial intelligence and convolutional neural networks may enhance the efficacy of AF screening in the future. Direct oral anticoagulants (DOACs) are preferred over vitamin K antagonists in both primary and secondary prevention of IS in AF patients. Recent data from the randomized controlled trials (RCT) also suggest that early initiation of DOAC treatment after acute IS is safe compared to later initiation. Anticoagulation treatment may still predispose for intracranial bleeding, particularly among patients with prior cerebrovascular events. Left atrial appendix closure offers an optional treatment choice for patients with prior intracranial hemorrhage and may offer an alternative to oral anticoagulation even for patients with IS, but these indications await validation in ongoing RCTs. There are still controversies related to the association of found AF paroxysms in CIS patients with prolonged screening, pertaining to the optimal duration of screening and screening strategies with prolonged monitoring techniques in patients with ESUS. In this review, we summarize the current knowledge of epidemiology, screening, and prognosis in AF patients with stroke. [ABSTRACT FROM AUTHOR]

Accurate diagnosis of Left Ventricular Noncompaction Cardiomyopathy (LVNC) is critical for proper patient treatment but remains challenging. This work improves LVNC detection by improving left ventricle segmentation in cardiac MR images. Trabeculated left ventricle indicates LVNC, but automatic segmentation is difficult. We present techniques to improve segmentation and evaluate their impact on LVNC diagnosis. Three main methods are introduced: (1) using full 800 × 800 MR images rather than 512 × 512; (2) a clustering algorithm to eliminate neural network hallucinations; (3) advanced network architectures including Attention U-Net, MSA-UNet, and U-Net++.Experiments utilize cardiac MR datasets from three different hospitals. U-Net++ achieves the best segmentation performance using 800 × 800 images, and it improves the mean segmentation Dice score by 0.02 over the baseline U-Net, the clustering algorithm improves the mean Dice score by 0.06 on the images it affected, and the U-Net++ provides an additional 0.02 mean Dice score over the baseline U-Net. For LVNC diagnosis, U-Net++ achieves 0.896 accuracy, 0.907 precision, and 0.912 F1-score outperforming the baseline U-Net. Proposed techniques enhance LVNC detection, but differences between hospitals reveal problems in improving generalization. This work provides validated methods for precise LVNC diagnosis. [ABSTRACT FROM AUTHOR]

Background: Addressing intraoperative bleeding remains a significant challenge in the field of robotic surgery. This research endeavors to pioneer a groundbreaking solution utilizing convolutional neural networks (CNNs). The objective is to establish a system capable of forecasting instances of intraoperative bleeding during robot-assisted radical prostatectomy (RARP) and promptly notify the surgeon about bleeding risks. Methods: To achieve this, a multi-task learning (MTL) CNN was introduced, leveraging a modified version of the U-Net architecture. The aim was to categorize video input as either "absence of blood accumulation" (0) or "presence of blood accumulation" (1). To facilitate seamless interaction with the neural networks, the Bleeding Artificial Intelligence-based Detector (BLAIR) software was created using the Python Keras API and built upon the PyQT framework. A subsequent clinical assessment of BLAIR's efficacy was performed, comparing its bleeding identification performance against that of a urologist. Various perioperative variables were also gathered. For optimal MTL-CNN training parameterization, a multi-task loss function was adopted to enhance the accuracy of event detection by taking advantage of surgical tools' semantic segmentation. Additionally, the Multiple Correspondence Analysis (MCA) approach was employed to assess software performance. Results: The MTL-CNN demonstrated a remarkable event recognition accuracy of 90.63%. When evaluating BLAIR's predictive ability and its capacity to pre-warn surgeons of potential bleeding incidents, the density plot highlighted a striking similarity between BLAIR and human assessments. In fact, BLAIR exhibited a faster response. Notably, the MCA analysis revealed no discernible distinction between the software and human performance in accurately identifying instances of bleeding. Conclusion: The BLAIR software proved its competence by achieving over 90% accuracy in predicting bleeding events during RARP. This accomplishment underscores the potential of AI to assist surgeons during interventions. This study exemplifies the positive impact AI applications can have on surgical procedures. [ABSTRACT FROM AUTHOR]

High myopia is a global ocular disease and one of the most common causes of blindness. Fundus images can be obtained in a noninvasive manner and can be used to monitor and follow up on many fundus diseases, such as high myopia. In this paper, we proposed a computer-aided diagnosis algorithm using deep convolutional neural networks (DCNNs) to grade the risk of high myopia. The input images were automatically classified into three categories: normal fundus images were labeled class 0, low-risk high-myopia images were labeled class 1, and high-risk high-myopia images were labeled class 2. We conducted model training on 758 clinical fundus images collected locally, and the average accuracy reached 98.15% according to the results of fivefold cross-validation. An additional 100 fundus images were used to evaluate the performance of DCNNs, with ophthalmologists performing external validation. The experimental results showed that DCNNs outperformed human experts with an area under the curve (AUC) of 0.9968 for the recognition of low-risk high myopia and 0.9964 for the recognition of high-risk high myopia. In this study, we were able to accurately and automatically perform high myopia classification solely using fundus images. This has great practical significance in terms of improving early diagnosis, prevention, and treatment in clinical practice. [ABSTRACT FROM AUTHOR]

Background: Endoscopic Ultrasound (EUS) is widely used for the diagnosis of bilio-pancreatic and gastrointestinal (GI) tract diseases, for the evaluation of subepithelial lesions, and for sampling of lymph nodes and solid masses located next to the GI tract. The role of Artificial Intelligence in healthcare in growing. This review aimed to provide an overview of the current state of AI in EUS from imaging to pathological diagnosis and training. Methods: AI algorithms can assist in lesion detection and characterization in EUS by analyzing EUS images and identifying suspicious areas that may require further clinical evaluation or biopsy sampling. Deep learning techniques, such as convolutional neural networks (CNNs), have shown great potential for tumor identification and subepithelial lesion (SEL) evaluation by extracting important features from EUS images and using them to classify or segment the images. Results: AI models with new features can increase the accuracy of diagnoses, provide faster diagnoses, identify subtle differences in disease presentation that may be missed by human eyes, and provide more information and insights into disease pathology. Conclusions: The integration of AI in EUS images and biopsies has the potential to improve the diagnostic accuracy, leading to better patient outcomes and to a reduction in repeated procedures in case of non-diagnostic biopsies. [ABSTRACT FROM AUTHOR]

Stroke is an emergency in which delays in treatment can lead to significant loss of neurological function and be fatal. Technologies that increase the speed and accuracy of stroke diagnosis or assist in post-stroke rehabilitation can improve patient outcomes. No resource exists that comprehensively assesses artificial intelligence/machine learning (AI/ML)-enabled technologies indicated for the management of ischemic and hemorrhagic stroke. We queried a United States Food and Drug Administration (FDA) database, along with PubMed and private company websites, to identify the recent literature assessing the clinical performance of FDA-approved AI/ML-enabled technologies. The FDA has approved 22 AI/ML-enabled technologies that triage brain imaging for more immediate diagnosis or promote post-stroke neurological/functional recovery. Technologies that assist with diagnosis predominantly use convolutional neural networks to identify abnormal brain images (e.g., CT perfusion). These technologies perform comparably to neuroradiologists, improve clinical workflows (e.g., time from scan acquisition to reading), and improve patient outcomes (e.g., days spent in the neurological ICU). Two devices are indicated for post-stroke rehabilitation by leveraging neuromodulation techniques. Multiple FDA-approved technologies exist that can help clinicians better diagnose and manage stroke. This review summarizes the most up-to-date literature regarding the functionality, performance, and utility of these technologies so clinicians can make informed decisions when using them in practice. [ABSTRACT FROM AUTHOR]

Prenatal ultrasonography is the most crucial imaging modality during pregnancy. However, problems such as high fetal mobility, excessive maternal abdominal wall thickness, and inter-observer variability limit the development of traditional ultrasound in clinical applications. The combination of artificial intelligence (AI) and obstetric ultrasound may help optimize fetal ultrasound examination by shortening the examination time, reducing the physician's workload, and improving diagnostic accuracy. AI has been successfully applied to automatic fetal ultrasound standard plane detection, biometric parameter measurement, and disease diagnosis to facilitate conventional imaging approaches. In this review, we attempt to thoroughly review the applications and advantages of AI in prenatal fetal ultrasound and discuss the challenges and promises of this new field. [ABSTRACT FROM AUTHOR]

Background: The paper presents a pre-processing method which, based on positron-emission tomography (PET) images of 18F-fluorodeoxyglucose ([18F] FDG) hypermetabolic pulmonary nodules, makes it possible to obtain additional visual characteristics and use them to enhance the specificity of imaging. Material and Methods: A retrospective analysis of 69 FDG-PET/CT scans of solitary hypermetabolic pulmonary nodules (40 cases of lung cancer and 29 benign tumours), where in each case, the standardised uptake value of the hottest voxel within the defined volume of interest was greater than 2.5 (SUVmax > 2.5). No diagnosis could be made based on these SUVmax values. All of the PET DICOM images were transformed by means of the pre-processing method for contouring the uptake levels of [18F] FDG (PCUL-FDG). Next, a multidimensional comparative analysis was conducted using a synthetic variable obtained by calculating the similarities based on the generalised distance measure for non-metric scaling (GDM2) from the pattern object. The calculations were performed with the use of the R language. Results: The PCUL-FDG method revealed 73.9% hypermetabolic nodules definitively diagnosed as either benign or malignant lesions. As for the other 26.1% of the nodules, there was uncertainty regarding their classification (some had features suggesting malignancy, while the characteristics of others made it impossible to confirm malignancy with a high degree of certainty). Conclusions: Application of the PCUL-FDG method enhances the specificity of PET in imaging solitary hypermetabolic pulmonary nodules. Images obtained using the PCUL-FDG method can serve as point of departure for automatic analysis of PET data based on convolutional neural networks. [ABSTRACT FROM AUTHOR]

Preimplantation genetic testing for aneuploidies (PGT-A) is arguably the most effective embryo selection strategy. Nevertheless, it requires greater workload, costs, and expertise. Therefore, a quest towards user-friendly, non-invasive strategies is ongoing. Although insufficient to replace PGT-A, embryo morphological evaluation is significantly associated with embryonic competence, but scarcely reproducible. Recently, artificial intelligence-powered analyses have been proposed to objectify and automate image evaluations. iDAScore v1.0 is a deep-learning model based on a 3D convolutional neural network trained on time-lapse videos from implanted and non-implanted blastocysts. It is a decision support system for ranking blastocysts without manual input. This retrospective, pre-clinical, external validation included 3604 blastocysts and 808 euploid transfers from 1232 cycles. All blastocysts were retrospectively assessed through the iDAScore v1.0; therefore, it did not influence embryologists' decision-making process. iDAScore v1.0 was significantly associated with embryo morphology and competence, although AUCs for euploidy and live-birth prediction were 0.60 and 0.66, respectively, which is rather comparable to embryologists' performance. Nevertheless, iDAScore v1.0 is objective and reproducible, while embryologists' evaluations are not. In a retrospective simulation, iDAScore v1.0 would have ranked euploid blastocysts as top quality in 63% of cases with one or more euploid and aneuploid blastocysts, and it would have questioned embryologists' ranking in 48% of cases with two or more euploid blastocysts and one or more live birth. Therefore, iDAScore v1.0 may objectify embryologists' evaluations, but randomized controlled trials are required to assess its clinical value. [ABSTRACT FROM AUTHOR]

Improving the prediction of blood glucose concentration may improve the quality of life of people living with type 1 diabetes by enabling them to better manage their care. Given the anticipated benefits of such a prediction, numerous methods have been proposed. Rather than attempting to predict glucose concentration, a deep learning framework for prediction is proposed in which prediction is performed using a scale for hypo- and hyper-glycemia risk. Using the blood glucose risk score formula proposed by Kovatchev et al., models with different architectures were trained, including, a recurrent neural network (RNN), a gated recurrent unit (GRU), a long short-term memory (LSTM) network, and an encoder-like convolutional neural network (CNN). The models were trained using the OpenAPS Data Commons data set, comprising 139 individuals, each with tens of thousands of continuous glucose monitor (CGM) data points. The training set was composed of 7% of the data set, while the remaining was used for testing. Performance comparisons between the different architectures are presented and discussed. To evaluate these predictions, performance results are compared with the last measurement (LM) prediction, through a sample-and-hold approach continuing the last known measurement forward. The results obtained are competitive when compared to other deep learning methods. A root mean squared error (RMSE) of 16 mg/dL, 24 mg/dL, and 37 mg/dL were obtained for CNN prediction horizons of 15, 30, and 60 min, respectively. However, no significant improvements were found for the deep learning models compared to LM prediction. Performance was found to be highly dependent on architecture and the prediction horizon. Lastly, a metric to assess model performance by weighing each prediction point error with the corresponding blood glucose risk score is proposed. Two main conclusions are drawn. Firstly, going forward, there is a need to benchmark model performance using LM prediction to enable the comparison between results obtained from different data sets. Secondly, model-agnostic data-driven deep learning models may only be meaningful when combined with mechanistic physiological models; here, it is argued that neural ordinary differential equations may combine the best of both approaches. These findings are based on the OpenAPS Data Commons data set and are to be validated in other independent data sets. [ABSTRACT FROM AUTHOR]

This study describes the development of a convolutional neural network (CNN) for automated assessment of optic disc photograph quality. Using a code-free deep learning platform, a total of 2377 optic disc photographs were used to develop a deep CNN capable of determining optic disc photograph quality. Of these, 1002 were good-quality images, 609 were acceptable-quality, and 766 were poor-quality images. The dataset was split 80/10/10 into training, validation, and test sets and balanced for quality. A ternary classification model (good, acceptable, and poor quality) and a binary model (usable, unusable) were developed. In the ternary classification system, the model had an overall accuracy of 91% and an AUC of 0.98. The model had higher predictive accuracy for images of good (93%) and poor quality (96%) than for images of acceptable quality (91%). The binary model performed with an overall accuracy of 98% and an AUC of 0.99. When validated on 292 images not included in the original training/validation/test dataset, the model's accuracy was 85% on the three-class classification task and 97% on the binary classification task. The proposed system for automated image-quality assessment for optic disc photographs achieves high accuracy in both ternary and binary classification systems, and highlights the success achievable with a code-free platform. There is wide clinical and research potential for such a model, with potential applications ranging from integration into fundus camera software to provide immediate feedback to ophthalmic photographers, to prescreening large databases before their use in research. [ABSTRACT FROM AUTHOR]

Neovascular age-related macular degeneration (nAMD) and central serous chorioretinopathy (CSC) are two of the most common macular diseases. This study proposes a convolutional neural network (CNN)-based deep learning model for classifying the subtypes of nAMD (polypoidal choroidal vasculopathy, retinal angiomatous proliferation, and typical nAMD) and CSC (chronic CSC and acute CSC) and healthy individuals using single spectral–domain optical coherence tomography (SD–OCT) images. The proposed model was trained and tested using 6063 SD–OCT images from 521 patients and 47 healthy participants. We used three well-known CNN architectures (VGG–16, VGG–19, and ResNet) and two customized classification layers. Additionally, transfer learning and mix–up-based data augmentation were applied to improve robustness and accuracy. Our model demonstrated high accuracies of 99.7% and 91.1% in the nAMD and CSC classification and retinopathy (nAMD and CSC) subtype classification, including normal participants, respectively. Furthermore, we performed an external test to compare the classification accuracy with that of eight ophthalmologists, and our model showed the highest accuracy. The region determined to be important for classification by the model was confirmed using gradient-weighted class activation mapping. The model's clinical criteria were similar to that of the ophthalmologists. [ABSTRACT FROM AUTHOR]

Background: It is well understood that surgical skills largely define patient outcomes both in Minimally Invasive Surgery (MIS) and Robot-Assisted MIS (RAMIS). Non-technical surgical skills, including stress and distraction resilience, decision-making and situation awareness also contribute significantly. Autonomous, technologically supported objective skill assessment can be efficient tools to improve patient outcomes without the need to involve expert surgeon reviewers. However, autonomous non-technical skill assessments are unstandardized and open for more research. Recently, Surgical Data Science (SDS) has become able to improve the quality of interventional healthcare with big data and data processing techniques (capture, organization, analysis and modeling of data). SDS techniques can also help to achieve autonomous non-technical surgical skill assessments. Methods: An MIS training experiment is introduced to autonomously assess non-technical skills and to analyse the workload based on sensory data (video image and force) and a self-rating questionnaire (SURG-TLX). A sensorized surgical skill training phantom and adjacent training workflow were designed to simulate a complicated Laparoscopic Cholecystectomy task; the dissection of the cholecyst's peritonial layer and the safe clip application on the cystic artery in an uncomfortable environment. A total of 20 training sessions were recorded from 7 subjects (3 non-medicals, 2 residents, 1 expert surgeon and 1 expert MIS surgeon). Workload and learning curves were studied via SURG-TLX. For autonomous non-technical skill assessment, video image data with tracked instruments based on Channel and Spatial Reliability Tracker (CSRT) and force data were utilized. An autonomous time series classification was achieved by a Fully Convolutional Neural Network (FCN), where the class labels were provided by SURG-TLX. Results: With unpaired t-tests, significant differences were found between the two groups (medical professionals and control) in certain workload components (mental demands, physical demands, and situational stress, p < 0.0001 , 95% confidence interval, p < 0.05 for task complexity). With paired t-tests, the learning curves of the trials were also studied; the task complexity resulted in a significant difference between the first and the second trials. Autonomous non-technical skill classification was based on the FCN by applying the tool trajectories and force data as input. This resulted in a high accuracy (85%) on temporal demands classification based on the z component of the used forces and 75% accuracy for classifying mental demands/situational stress with the x component of the used forces validated with Leave One Out Cross-Validation. Conclusions: Non-technical skills and workload components can be classified autonomously based on measured training data. SDS can be effective via automated non-technical skill assessment. [ABSTRACT FROM AUTHOR]

(1) Background: Pyoderma gangrenosum (PG) is often situated on the lower legs, and the differentiation from conventional leg ulcers (LU) is a challenging task due to the lack of clear clinical diagnostic criteria. Because of the different therapy concepts, misdiagnosis or delayed diagnosis bears a great risk for patients. (2) Objective: to develop a deep convolutional neural network (CNN) capable of analysing wound photographs to facilitate the PG diagnosis for health professionals. (3) Methods: A CNN was trained with 422 expert-selected pictures of PG and LU. In a man vs. machine contest, 33 pictures of PG and 36 pictures of LU were presented for diagnosis to 18 dermatologists at two maximum care hospitals and to the CNN. The results were statistically evaluated in terms of sensitivity, specificity and accuracy for the CNN and for dermatologists with different experience levels. (4) Results: The CNN achieved a sensitivity of 97% (95% confidence interval (CI) 84.2–99.9%) and outperformed dermatologists, with a sensitivity of 72.7% (CI 54.4–86.7%) significantly (p < 0.03). However, dermatologists achieved a slightly higher specificity (88.9% vs. 83.3%). (5) Conclusions: For the first time, a deep neural network was demonstrated to be capable of diagnosing PG, solely on the basis of photographs, and with a greater sensitivity compared to that of dermatologists. [ABSTRACT FROM AUTHOR]

(1) Background: The Alberta Stroke Program Early CT Score (ASPECTS) is a standardized scoring tool used to evaluate the severity of acute ischemic stroke (AIS) on non-contrast CT (NCCT). Our aim in this study was to automate ASPECTS. (2) Methods: We utilized a total of 258 patient images with suspected AIS symptoms. Expert ASPECTS readings on NCCT were used as ground truths. A deep learning-based automatic detection (DLAD) algorithm was developed for automated ASPECTS scoring based on 168 training patient images using a convolutional neural network (CNN) architecture. An additional 90 testing patient images were used to evaluate the performance of the DLAD algorithm, which was then compared with ASPECTS readings on NCCT as performed by physicians. (3) Results: The sensitivity, specificity, and accuracy of DLAD for the prediction of ASPECTS were 65%, 82%, and 80%, respectively. These results demonstrate that the DLAD algorithm was not inferior to radiologist-read ASPECTS on NCCT. With the assistance of DLAD, the individual sensitivity of the ER physician, neurologist, and radiologist improved. (4) Conclusion: The proposed DLAD algorithm exhibits a reasonable ability for ASPECTS scoring on NCCT images in patients presenting with AIS symptoms. The DLAD algorithm could be a valuable tool to improve and accelerate the decision-making process of front-line physicians. [ABSTRACT FROM AUTHOR]

(1) Background: Single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) is a long-established estimation methodology for medical diagnosis using image classification illustrating conditions in coronary artery disease. For these procedures, convolutional neural networks have proven to be very beneficial in achieving near-optimal accuracy for the automatic classification of SPECT images. (2) Methods: This research addresses the supervised learning-based ideal observer image classification utilizing an RGB-CNN model in heart images to diagnose CAD. For comparison purposes, we employ VGG-16 and DenseNet-121 pre-trained networks that are indulged in an image dataset representing stress and rest mode heart states acquired by SPECT. In experimentally evaluating the method, we explore a wide repertoire of deep learning network setups in conjunction with various robust evaluation and exploitation metrics. Additionally, to overcome the image dataset cardinality restrictions, we take advantage of the data augmentation technique expanding the set into an adequate number. Further evaluation of the model was performed via 10-fold cross-validation to ensure our model's reliability. (3) Results: The proposed RGB-CNN model achieved an accuracy of 91.86%, while VGG-16 and DenseNet-121 reached 88.54% and 86.11%, respectively. (4) Conclusions: The abovementioned experiments verify that the newly developed deep learning models may be of great assistance in nuclear medicine and clinical decision-making. [ABSTRACT FROM AUTHOR]

The rapid spread of COVID-19 across the globe since its emergence has pushed many countries' healthcare systems to the verge of collapse. To restrict the spread of the disease and lessen the ongoing cost on the healthcare system, it is critical to appropriately identify COVID-19-positive individuals and isolate them as soon as possible. The primary COVID-19 screening test, RT-PCR, although accurate and reliable, has a long turn-around time. More recently, various researchers have demonstrated the use of deep learning approaches on chest X-ray (CXR) for COVID-19 detection. However, existing Deep Convolutional Neural Network (CNN) methods fail to capture the global context due to their inherent image-specific inductive bias. In this article, we investigated the use of vision transformers (ViT) for detecting COVID-19 in Chest X-ray (CXR) images. Several ViT models were fine-tuned for the multiclass classification problem (COVID-19, Pneumonia and Normal cases). A dataset consisting of 7598 COVID-19 CXR images, 8552 CXR for healthy patients and 5674 for Pneumonia CXR were used. The obtained results achieved high performance with an Area Under Curve (AUC) of 0.99 for multi-class classification (COVID-19 vs. Other Pneumonia vs. normal). The sensitivity of the COVID-19 class achieved 0.99. We demonstrated that the obtained results outperformed comparable state-of-the-art models for detecting COVID-19 on CXR images using CNN architectures. The attention map for the proposed model showed that our model is able to efficiently identify the signs of COVID-19. [ABSTRACT FROM AUTHOR]

Non-melanoma skin cancer, and basal cell carcinoma in particular, is one of the most common types of cancer. Although this type of malignancy has lower metastatic rates than other types of skin cancer, its locally destructive nature and the advantages of its timely treatment make early detection vital. The combination of multispectral imaging and artificial intelligence has arisen as a powerful tool for the detection and classification of skin cancer in a non-invasive manner. The present study uses hyperspectral images to discern between healthy and basal cell carcinoma hyperspectral signatures. Upon the combined use of convolutional neural networks, with a final support vector machine activation layer, the present study reaches up to 90% accuracy, with an area under the receiver operating characteristic curve being calculated at 0.9 as well. While the results are promising, future research should build upon a dataset with a larger number of patients. [ABSTRACT FROM AUTHOR]

Background: The aim of this study was to evaluate the performance of an automated COVID-19 detection method based on a transfer learning technique that makes use of chest computed tomography (CT) images. Method: In this study, we used a publicly available multiclass CT scan dataset containing 4171 CT scans of 210 different patients. In particular, we extracted features from the CT images using a set of convolutional neural networks (CNNs) that had been pretrained on the ImageNet dataset as feature extractors, and we then selected a subset of these features using the Information Gain filter. The resulting feature vectors were then used to train a set of k Nearest Neighbors classifiers with 10-fold cross validation to assess the classification performance of the features that had been extracted by each CNN. Finally, a majority voting approach was used to classify each image into two different classes: COVID-19 and NO COVID-19. Results: A total of 414 images of the test set (10% of the complete dataset) were correctly classified, and only 4 were misclassified, yielding a final classification accuracy of 99.04%. Conclusions: The high performance that was achieved by the method could make it feasible option that could be used to assist radiologists in COVID-19 diagnosis through the use of CT images. [ABSTRACT FROM AUTHOR]

With recent advancements in machine learning, especially in deep learning, the prediction of eye diseases based on fundus photography using deep convolutional neural networks (DCNNs) has attracted great attention. However, studies focusing on identifying the right disease among several candidates, which is a better approximation of clinical diagnosis in practice comparing with the case that aims to distinguish one particular eye disease from normal controls, are limited. The performance of existing algorithms for multi-class classification of fundus images is at most mediocre. Moreover, in many studies consisting of different eye diseases, labeled images are quite limited mainly due to privacy concern of patients. In this case, it is infeasible to train huge DCNNs, which usually have millions of parameters. To address these challenges, we propose to utilize a lightweight deep learning architecture called MobileNetV2 and transfer learning to distinguish four common eye diseases, including Glaucoma, Maculopathy, Pathological Myopia, and Retinitis Pigmentosa, from normal controls using a small training data. We also apply a visualization approach to highlight the loci that are most related to the disease labels to make the model more explainable. The highlighted area chosen by the algorithm itself may give some hints for further fundus image studies. Our experimental results show that our system achieves an average accuracy of 96.2%, sensitivity of 90.4%, and specificity of 97.6% on the test data via five independent runs, and outperforms two other deep learning-based algorithms both in terms of accuracy and efficiency. [ABSTRACT FROM AUTHOR]

Recently, the estimation of bone maturation using deep learning has been actively conducted. However, many studies have considered hand–wrist radiographs, while a few studies have focused on estimating cervical vertebral maturation (CVM) using lateral cephalograms. This study proposes the use of deep learning models for estimating CVM from lateral cephalograms. As the second, third, and fourth cervical vertebral regions (denoted as C2, C3, and C4, respectively) are considerably smaller than the whole image, we propose a stepwise segmentation-based model that focuses on the C2–C4 regions. We propose three convolutional neural network-based classification models: a one-step model with only CVM classification, a two-step model with region of interest (ROI) detection and CVM classification, and a three-step model with ROI detection, cervical segmentation, and CVM classification. Our dataset contains 600 lateral cephalogram images, comprising six classes with 100 images each. The three-step segmentation-based model produced the best accuracy (62.5%) compared to the models that were not segmentation-based. [ABSTRACT FROM AUTHOR]

The purpose of this study is to evaluate and compare the performance of six state-of-the-art convolutional neural network (CNN)-based deep learning models for cervical vertebral maturation (CVM) on lateral cephalometric radiographs, and implement visualization of CVM classification for each model using gradient-weighted class activation map (Grad-CAM) technology. A total of 600 lateral cephalometric radiographs obtained from patients aged 6–19 years between 2013 and 2020 in Pusan National University Dental Hospital were used in this study. ResNet-18, MobileNet-v2, ResNet-50, ResNet-101, Inception-v3, and Inception-ResNet-v2 were tested to determine the optimal pre-trained network architecture. Multi-class classification metrics, accuracy, recall, precision, F1-score, and area under the curve (AUC) values from the receiver operating characteristic (ROC) curve were used to evaluate the performance of the models. All deep learning models demonstrated more than 90% accuracy, with Inception-ResNet-v2 performing the best, relatively. In addition, visualizing each deep learning model using Grad-CAM led to a primary focus on the cervical vertebrae and surrounding structures. The use of these deep learning models in clinical practice will facilitate dental practitioners in making accurate diagnoses and treatment plans. [ABSTRACT FROM AUTHOR]

Background: Computed tomography angiography (CTA) is one of the most commonly used imaging technique for the management of vascular diseases. Here, we aimed to develop a hybrid method combining a feature-based expert system with a supervised deep learning (DL) algorithm to enable a fully automatic segmentation of the abdominal vascular tree. Methods: We proposed an algorithm based on the hybridization of a data-driven convolutional neural network and a knowledge-based model dedicated to vascular system segmentation. By using two distinct datasets of CTA from patients to evaluate independence to training dataset, the accuracy of the hybrid method for lumen and thrombus segmentation was evaluated compared to the feature-based expert system alone and to the ground truth provided by a human expert. Results: The hybrid approach demonstrated a better accuracy for lumen segmentation compared to the expert system alone (volume similarity: 0.8128 vs. 0.7912, p = 0.0006 and Dice similarity coefficient: 0.8266 vs. 0.7942, p < 0.0001). The accuracy for thrombus segmentation was also enhanced using the hybrid approach (volume similarity: 0.9404 vs. 0.9185, p = 0.0027 and Dice similarity coefficient: 0.8918 vs. 0.8654, p < 0.0001). Conclusions: By enabling a robust and fully automatic segmentation, the method could be used to develop real-time decision support to help in the management of vascular diseases. [ABSTRACT FROM AUTHOR]

Objectives: To retrospectively assess radiographic data and to prospectively classify radiographs (namely, panoramic, bitewing, periapical, and cephalometric images), we compared three deep learning architectures for their classification performance. Methods: Our dataset consisted of 31,288 panoramic, 43,598 periapical, 14,326 bitewing, and 1176 cephalometric radiographs from two centers (Berlin/Germany; Lucknow/India). For a subset of images L (32,381 images), image classifications were available and manually validated by an expert. The remaining subset of images U was iteratively annotated using active learning, with ResNet-34 being trained on L, least confidence informative sampling being performed on U, and the most uncertain image classifications from U being reviewed by a human expert and iteratively used for re-training. We then employed a baseline convolutional neural networks (CNN), a residual network (another ResNet-34, pretrained on ImageNet), and a capsule network (CapsNet) for classification. Early stopping was used to prevent overfitting. Evaluation of the model performances followed stratified k-fold cross-validation. Gradient-weighted Class Activation Mapping (Grad-CAM) was used to provide visualizations of the weighted activations maps. Results: All three models showed high accuracy (>98%) with significantly higher accuracy, F1-score, precision, and sensitivity of ResNet than baseline CNN and CapsNet (p < 0.05). Specificity was not significantly different. ResNet achieved the best performance at small variance and fastest convergence. Misclassification was most common between bitewings and periapicals. For bitewings, model activation was most notable in the inter-arch space for periapicals interdentally, for panoramics on bony structures of maxilla and mandible, and for cephalometrics on the viscerocranium. Conclusions: Regardless of the models, high classification accuracies were achieved. Image features considered for classification were consistent with expert reasoning. [ABSTRACT FROM AUTHOR]

Objectives: The present study aimed to train deep convolutional neural networks (CNNs) to detect caries lesions on Near-Infrared Light Transillumination (NILT) imagery obtained either in vitro or in vivo and to assess the models' generalizability. Methods: In vitro, 226 extracted posterior permanent human teeth were mounted in a diagnostic model in a dummy head. Then, NILT images were generated (DIAGNOcam, KaVo, Biberach), and images were segmented tooth-wise. In vivo, 1319 teeth from 56 patients were obtained and segmented similarly. Proximal caries lesions were annotated pixel-wise by three experienced dentists, reviewed by a fourth dentist, and then transformed into binary labels. We trained ResNet classification models on both in vivo and in vitro datasets and used 10-fold cross-validation for estimating the performance and generalizability of the models. We used GradCAM to increase explainability. Results: The tooth-level prevalence of caries lesions was 41% in vitro and 49% in vivo, respectively. Models trained and tested on in vivo data performed significantly better (mean ± SD accuracy: 0.78 ± 0.04) than those trained and tested on in vitro data (accuracy: 0.64 ± 0.15; p < 0.05). When tested in vitro, the models trained in vivo showed significantly lower accuracy (0.70 ± 0.01; p < 0.01). Similarly, when tested in vivo, models trained in vitro showed significantly lower accuracy (0.61 ± 0.04; p < 0.05). In both cases, this was due to decreases in sensitivity (by −27% for models trained in vivo and −10% for models trained in vitro). Conclusions: Using in vitro setups for generating NILT imagery and training CNNs comes with low accuracy and generalizability. Clinical significance: Studies employing in vitro imagery for developing deep learning models should be critically appraised for their generalizability. Applicable deep learning models for assessing NILT imagery should be trained on in vivo data. [ABSTRACT FROM AUTHOR]

Hollow organ perforation can precipitate a life-threatening emergency due to peritonitis followed by fulminant sepsis and fatal circulatory collapse. Pneumoperitoneum is typically detected as subphrenic free air on frontal chest X-ray images; however, treatment is reliant on accurate interpretation of radiographs in a timely manner. Unfortunately, it is not uncommon to have misdiagnoses made by emergency physicians who have insufficient experience or who are too busy and overloaded by multitasking. It is essential to develop an automated method for reviewing frontal chest X-ray images to alert emergency physicians in a timely manner about the life-threatening condition of hollow organ perforation that mandates an immediate second look. In this study, a deep learning-based approach making use of convolutional neural networks for the detection of subphrenic free air is proposed. A total of 667 chest X-ray images were collected at a local hospital, where 587 images (positive/negative: 267/400) were used for training and 80 images (40/40) for testing. This method achieved 0.875, 0.825, and 0.889 in sensitivity, specificity, and AUC score, respectively. It may provide a sensitive adjunctive screening tool to detect pneumoperitoneum on images read by emergency physicians who have insufficient clinical experience or who are too busy and overloaded by multitasking. [ABSTRACT FROM AUTHOR]

Artificial Intelligence in Gastroenterology - Walking into the Room of Little Miracles The surge of artificial intelligence (AI) in medicine stands on a lengthy, and frequently reticent, buildout. Undoubtedly, AI has come to "democratize" GI endoscopy by improving patients' outcomes irrespective of place/time of care, enhancing the professional experience for clinicians, and assisting healthcare systems with capacity issues. 31593701 2 Nakase H., Hirano T., Wagatsuma K., Ichimiya T., Yamakawa T., Yokoyama Y., Hayashi Y., Hirayama D., Kazama T., Yoshii S. Artificial intelligence-assisted endoscopy changes the definition of mucosal healing in ulcerative colitis. [Extracted from the article]

In fields using functional near-infrared spectroscopy (fNIRS), there is a need for an easy-to-understand method that allows visual presentation and rapid analysis of data and test results. This preliminary study examined whether deep learning (DL) could be applied to the analysis of fNIRS-derived brain activity data. To create a visual presentation of the data, an imaging program was developed for the analysis of hemoglobin (Hb) data from the prefrontal cortex in healthy volunteers, obtained by fNIRS before and after tooth clenching. Three types of imaging data were prepared: oxygenated hemoglobin (oxy-Hb) data, deoxygenated hemoglobin (deoxy-Hb) data, and mixed data (using both oxy-Hb and deoxy-Hb data). To differentiate between rest and tooth clenching, a cross-validation test using the image data for DL and a convolutional neural network was performed. The network identification rate using Hb imaging data was relatively high (80‒90%). These results demonstrated that a method using DL for the assessment of fNIRS imaging data may provide a useful analysis system. [ABSTRACT FROM AUTHOR]