Deep learning algorithms for melanoma detection using dermoscopic images: A systematic review and meta-analysis

Background Machine learning (ML) algorithm support vector machine (SVM) performed better than Thrombolysis in Myocardial Infarction (TIMI) score for ASIAN STEMI patients. However, Deep Learning (DL) effectiveness in the multiethnic ASIAN population has yet to be determined. DL has automatic learning of the feature from a given dataset without the need to conduct feature selection. However, the selected features by the algorithm is black box. Identifying features associated with mortality is essential to recognize characteristics of patients with high risk for better patient management. Purpose To develop a DL algorithm for in-hospital mortality in multiethnic STEMI patients using predictors identified from the SVM algorithm. To investigate DL performance constructed using predictors from SVM feature extraction and expert-recommended predictors. Methods We constructed four algorithms; a) DL and SVM algorithms with predictors identified from the SVM variable importance b) DL and SVM using predictors based on expert recommendation. We used registry data from the National Cardiovascular Disease Database of 11397 patient's. Fifty parameters including demographics, cardiovascular risk, medications and clinical variables were considered. The Area under the curve (AUC) is the performance evaluation metric. Algorithms were validated against the TIMI and tested using the same validation data. SVM variable importance with backward elimination was used to select and rank important variables. Results DL algorithms outperform SVM and TIMI on the validation dataset; i) DL with SVM selected predictors (15 predictors, AUC = 0.97), ii) DL with expert-recommended predictors (16 predictors, AUC = 0.96), iii) SVM with selected predictors (15 predictors, AUC = 0.92), iv) SVM with expert-recommended predictors (AUC = 0.89) and TIMI (AUC = 0.82). Common predictors across SVM feature selection, expert-recommendation and TIMI are: age, heart rate, Killip class, fasting blood glucose, systolic blood pressure, comorbid diseases and ST-elevation. SVM feature selection also identified diuretics, PCI and pharmacotherapy drugs as predictors that improve mortality prediction in STEMI patients. Our findings suggest that the TIMI score underestimates patients risk of mortality. DL algorithm using selected predictors classified 35% of nonsurvival patients as high risk (risk probabilities >50%) compared to only 12.7% nonsurvival patients by TIMI (score >5) (Figure below). Conclusions In the ASIAN population, patients with STEMI can be better classified using the DL algorithm compared to the ML and TIMI score. Combining ML feature selection with DL allows the identification of distinct factors in a unique ASIAN population for better mortality prediction than relying solely on an expert recommendation as it is a very subjective approach. Continuous validation on population-specific algorithms using DL and ML is needed before implementing in a real clinical setting. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Technology Development Fund 1 TIMI performance on validation setDL performance on validation set

BackgroundMachine learning (ML) algorithm support vector machine (SVM) performed better than Thrombolysis in Myocardial Infarction (TIMI) score for ASIAN STEMI patients. However, Deep Learning (DL) effectiveness in the multiethnic ASIAN population has yet to be determined. DL has automatic learning of the feature from a given dataset without the need to conduct feature selection. However, the selected features by the algorithm is black box. Identifying features associated with mortality is essential to recognize characteristics of patients with high risk for better patient management.PurposeTo develop a DL algorithm for in-hospital mortality in multiethnic STEMI patients using predictors identified from the SVM algorithm. To investigate DL performance constructed using predictors from SVM feature extraction and expert-recommended predictors.MethodsWe constructed four algorithms; a) DL and SVM algorithms with predictors identified from the SVM variable importance b) DL and SVM using predictors based on expert recommendation. We used registry data from the National Cardiovascular Disease Database of 11397 patient's. Fifty parameters including demographics, cardiovascular risk, medications and clinical variables were considered. The Area under the curve (AUC) is the performance evaluation metric. Algorithms were validated against the TIMI and tested using the same validation data. SVM variable importance with backward elimination was used to select and rank important variables.ResultsDL algorithms outperform SVM and TIMI on the validation dataset; i) DL with SVM selected predictors (15 predictors, AUC = 0.97), ii) DL with expert-recommended predictors (16 predictors, AUC = 0.96), iii) SVM with selected predictors (15 predictors, AUC = 0.92), iv) SVM with expert-recommended predictors (AUC = 0.89) and TIMI (AUC = 0.82). Common predictors across SVM feature selection, expert-recommendation and TIMI are: age, heart rate, Killip class, fasting blood glucose, systolic blood pressure, comorbid diseases and ST-elevation. SVM feature selection also identified diuretics, PCI and pharmacotherapy drugs as predictors that improve mortality prediction in STEMI patients. Our findings suggest that the TIMI score underestimates patients risk of mortality. DL algorithm using selected predictors classified 35% of nonsurvival patients as high risk (risk probabilities >50%) compared to only 12.7% nonsurvival patients by TIMI (score >5) (Figure below).ConclusionsIn the ASIAN population, patients with STEMI can be better classified using the DL algorithm compared to the ML and TIMI score. Combining ML feature selection with DL allows the identification of distinct factors in a unique ASIAN population for better mortality prediction than relying solely on an expert recommendation as it is a very subjective approach. Continuous validation on population-specific algorithms using DL and ML is needed before implementing in a real clinical setting.Funding AcknowledgementType of funding sources: Public grant(s) – National budget only. Main funding source(s): Technology Development Fund 1TIMI performance on validation setDL performance on validation set

BackgroundDeep learning (DL) based solutions have been proposed for interpretation of several imaging modalities including radiography, CT, and MR. For chest radiographs, DL algorithms have found success in the evaluation of abnormalities such as lung nodules, pulmonary tuberculosis, cystic fibrosis, pneumoconiosis, and location of peripherally inserted central catheters. Chest radiography represents the most commonly performed radiological test for a multitude of non-emergent and emergent clinical indications. This study aims to assess accuracy of deep learning (DL) algorithm for detection of abnormalities on routine frontal chest radiographs (CXR), and assessment of stability or change in findings over serial radiographs.Methods and findingsWe processed 874 de-identified frontal CXR from 724 adult patients (> 18 years) with DL (Qure AI). Scores and prediction statistics from DL were generated and recorded for the presence of pulmonary opacities, pleural effusions, hilar prominence, and enlarged cardiac silhouette. To establish a standard of reference (SOR), two thoracic radiologists assessed all CXR for these abnormalities. Four other radiologists (test radiologists), unaware of SOR and DL findings, independently assessed the presence of radiographic abnormalities. A total 724 radiographs were assessed for detection of findings. A subset of 150 radiographs with follow up examinations was used to asses change over time. Data were analyzed with receiver operating characteristics analyses and post-hoc power analysis.ResultsAbout 42% (305/ 724) CXR had no findings according to SOR; single and multiple abnormalities were seen in 23% (168/724) and 35% (251/724) of CXR. There was no statistical difference between DL and SOR for all abnormalities (p = 0.2–0.8). The area under the curve (AUC) for DL and test radiologists ranged between 0.837–0.929 and 0.693–0.923, respectively. DL had lowest AUC (0.758) for assessing changes in pulmonary opacities over follow up CXR. Presence of chest wall implanted devices negatively affected the accuracy of DL algorithm for evaluation of pulmonary and hilar abnormalities.ConclusionsDL algorithm can aid in interpretation of CXR findings and their stability over follow up CXR. However, in its present version, it is unlikely to replace radiologists due to its limited specificity for categorizing specific findings.

Echocardiographic strain imaging reflects myocardial deformation and is a sensitive measure of cardiac function and wall-motion abnormalities. Deep learning (DL) algorithms could automate the interpretation of echocardiographic strain imaging. We developed and trained an automated DL-based algorithm for left ventricular (LV) strain measurements in an internal dataset. Global longitudinal strain (GLS) was validated externally in (i) a real-world Taiwanese cohort of participants with and without heart failure (HF), (ii) a core-lab measured dataset from the multinational prevalence of microvascular dysfunction-HF and preserved ejection fraction (PROMIS-HFpEF) study, and regional strain in (iii) the HMC-QU-MI study of patients with suspected myocardial infarction. Outcomes included measures of agreement [bias, mean absolute difference (MAD), root-mean-squared-error (RMSE), and Pearson's correlation (R)] and area under the curve (AUC) to identify HF and regional wall-motion abnormalities. The DL workflow successfully analysed 3741 (89%) studies in the Taiwanese cohort, 176 (96%) in PROMIS-HFpEF, and 158 (98%) in HMC-QU-MI. Automated GLS showed good agreement with manual measurements (mean ± SD): -18.9 ± 4.5% vs. -18.2 ± 4.4%, respectively, bias 0.68 ± 2.52%, MAD 2.0 ± 1.67, RMSE = 2.61, R = 0.84 in the Taiwanese cohort; and -15.4 ± 4.1% vs. -15.9 ± 3.6%, respectively, bias -0.65 ± 2.71%, MAD 2.19 ± 1.71, RMSE = 2.78, R = 0.76 in PROMIS-HFpEF. In the Taiwanese cohort, automated GLS accurately identified patients with HF (AUC = 0.89 for total HF and AUC = 0.98 for HF with reduced ejection fraction). In HMC-QU-MI, automated regional strain identified regional wall-motion abnormalities with an average AUC = 0.80. DL algorithms can interpret echocardiographic strain images with similar accuracy as conventional measurements. These results highlight the potential of DL algorithms to democratize the use of cardiac strain measurements and reduce time-spent and costs for echo labs globally.

Objectives: To create a novel preoperative prediction model based on a deep learning algorithm to predict neoplasm T staging and grading in patients with upper tract urothelial carcinoma (UTUC). Methods: We performed a retrospective cohort study of patients diagnosed with UTUC between 2001 and 2012 at our institution. Five deep learning algorithms (CGRU, BiGRU, CNN-BiGRU, CBiLSTM, and CNN-BiLSTM) were used to develop a preoperative prediction model for neoplasm T staging and grading. The Matthews correlation coefficient (MMC) and the receiver-operating characteristic curve with the area under the curve (AUC) were used to evaluate the performance of each prediction model. Results: The clinical data of a total of 884 patients with pathologically confirmed UTUC were collected. The T-staging prediction model based on CNN-BiGRU achieved the best performance, and the MMC and AUC were 0.598 (0.592–0.604) and 0.760 (0.755–0.765), respectively. The grading prediction model [1973 World Health Organization (WHO) grading system] based on CNN-BiGRU achieved the best performance, and the MMC and AUC were 0.612 (0.609–0.615) and 0.804 (0.801–0.807), respectively. The grading prediction model [2004 WHO grading system] based on BiGRU achieved the best performance, and the MMC and AUC were 0.621 (0.616–0.626) and 0.824 (0.819–0.829), respectively. Conclusions: We developed an accurate UTUC preoperative prediction model to predict neoplasm T staging and grading based on deep learning algorithms, which will help urologists to make appropriate treatment decisions in the early stage.

Detection of metastasis of mediastinal lymph nodes in lung cancer patients with an artificial intelligence model.

Detection of Lisfranc malalignment leading to the instability of the joint, particularly in subtle cases, has been a concern for foot and ankle care providers. X-ray radiographs are the mainstay in the diagnosis of these injuries; thus, improving the performance of clinicians in interpreting radiographs can noticeably affect the quality of health care in these patients. Here we assessed the performance of deep learning algorithms on weightbearing radiographs for detection of Lisfranc joint malalignment in patients with Lisfranc instability. In a retrospective study, 640 patients with Lisfranc malalignment leading to instability were recruited plus 640 individuals with uninjured feet and healthy Lisfranc joint as the control group. All radiographs were screened by orthopaedic surgeons. Two deep learning models were trained, validated, and tested (in a ratio 80:10:10) using a single-view (anteroposterior) and 3-view (anteroposterior, lateral, oblique) radiographs. The performances of the models were reported as sensitivity, specificity, positive and negative predictive values, accuracy, F score, and area under the curve (AUC). No significant differences were observed between the patients and the controls regarding age, gender, race, and body mass index. The best deep learning algorithm outperformed our human interpreters (<1% vs ~10% misdiagnosis), 94.8% sensitivity, 96.9% specificity, 98.6% accuracy, 95.8% F score, and 99.4% AUC. Deep learning methods have shown promising potential in acting as an assistant interpreter of radiographic images in patients with Lisfranc malalignment. Developing these algorithms can hasten and improve the accuracy of diagnosis and reduce further costs and burdens on the patients and health care system. Level III, case-control Machine Learning study.

Early detection and treatment of sarcopenia are crucial for improving patient outcomes, yet current diagnostic methods often lack the accuracy, accessibility, and efficiency needed for widespread clinical use. The aim of this study was to develop an accurate, secure, and evidence-based multimodal AI model using a point-of-care ultrasound (POCUS) framework combining muscle imaging properties with physical performance for sarcopenia diagnosis. The model uses clinical data and POCUS images. Clinical data consisted of age, gender, height, weight, body mass index (BMI) and data on physical performance by Short Physical Performance Battery (SPPB) scores. SPPB scores were chosen since it is recommended by both the European Working Group of Sarcopenia in Older People 2 and the Asian Working Group for Sarcopenia. POCUS data consisted of images from the dominant thigh, focusing on the rectus femoris muscle, using longitudinal and transverse projections. Various Machine Learning (ML) and Deep Learning (DL) algorithms and multimodal architectures were tested. Explainable AI (XAI) methods, including Grad-CAM for ultrasound images and feature-attribution analysis for clinical variables, were integrated to provide transparent interpretation of the multimodal model’s diagnostic decisions. The final model was implemented as part of the Sarcopenia Artificial Intelligence Diagnostic Decision Support System (SAID DSS). Participants (24) were mostly women (63%) with a mean age of 81 years (± 5.2), (age range: 71–91 years) a mean body mass index of 26 kg/m2 (± 4.1), and mean SPPB scores of 5 (± 1.6) and 9 (± 1.6) for sarcopenic and controls. 1060 and 2414 longitudinal and transverse ultrasound events for sarcopenic and control participants, respectively, were used, demonstrating a robust dataset despite the small number of participants. Comprehensive experimental results showed that a feature-level fusion technique using a multilayer perceptron network as classifier and Xception architectures for image feature extraction demonstrated the best performance. The final model yielded a diagnostic accuracy of 85%, an F1-score of 0.85 and an area under the curve (AUC) of 0.84, higher than previous models. This study is the first to introduce a clinically oriented, AI-based multimodal model for sarcopenia detection, demonstrating improved performance over existing approaches. In addition, we provided an explanation of the decisions generated by the best-performing detection model. By integrating this model into SAID DSS, we provide a practical and scalable tool with potential for direct application in clinical workflows, supporting early and accurate identification of sarcopenia. Not applicable.

Deep learning has gained immense popularity in scientific computing, and its algorithms are widely used in complex problem-solving industries. Every deep learning algorithm use different types of neural networks to perform indented tasks. Deep learning (DL) algorithms have emerged from different machine learning and soft computing methodologies. Since then, a number of deep learning (DL) algorithms have been recently introduced in the scientific community and applied in various application fields. Today, the use of DLs has become indispensable due to their intelligence, effective learning, accuracy and reliability in model creation. However, a comprehensive list of DL algorithms has not yet been presented in the scientific literature. This article lists the most popular DL algorithms and their application areas. Deep learning uses ANN artificial neural networks to perform convoluted calculations on huge amounts of data. It is a type of machine learning based on the structure and function of the human brain. Deep learning algorithms train machines by learning from examples. Industries such as healthcare, e-commerce, entertainment and advertising often use deep learning.

To develop a multitask deep learning (DL) algorithm to automatically classify mammography imaging findings and predict the existence of extensive intraductal component (EIC) in invasive breast cancer. Mammograms with invasive breast cancers from 2010 to 2019 were downloaded for two radiologists performing image segmentation and imaging findings annotation. Images were randomly split into training, validation, and test datasets. A multitask approach was performed on the EfficientNet-B0 neural network mainly to predict EIC and classify imaging findings. Three more models were trained for comparison, including a single-task model (predicting EIC), a two-task model (predicting EIC and cell receptor status), and a three-task model (combining the abovementioned tasks). Additionally, these models were trained in a subgroup of invasive ductal carcinoma. The DeLong test was used to examine the difference in model performance. This study enrolled 1459 breast cancers on 3076 images. The EIC-positive rate was 29.0%. The three-task model was the best DL model with an area under the curve (AUC) of EIC prediction of 0.758 and 0.775 at the image and breast (patient) levels, respectively. Mass was the most accurately classified imaging finding (AUC = 0.915), followed by calcifications and mass with calcifications (AUC = 0.878 and 0.824, respectively). Cell receptor status prediction was less accurate (AUC = 0.625-0.653). The multitask approach improves the model training compared to the single-task model, but without significant effects. A mammography-based multitask DL model can perform simultaneous imaging finding classification and EIC prediction. The study results demonstrated the potential of deep learning to extract more information from mammography for clinical decision-making. • Extensive intraductal component (EIC) is an independent risk factor of local tumor recurrence after breast-conserving surgery. • A mammography-based deep learning model was trained to predict extensive intraductal component close to radiologists' reading. • The developed multitask deep learning model could perform simultaneous imaging finding classification and extensive intraductal component prediction.

Improving the performance of machine learning algorithms to overcome the curse of dimensionality while maintaining interpretability is still a challenging issue for researchers in artificial intelligence. Patch learning (PL), based on the improved adaptive network-based fuzzy inference system (ANFIS) and continuous local optimization for the input domain, is characterized by high accuracy. However, PL can only handle low-dimensional data set regression. Based on the parallel and serial ensembles, two deep patch learning algorithms with embedded adaptive fuzzy systems (DPLFSs) are proposed in this paper. First, using the maximum information coefficient (MIC) and Pearson's correlation coefficients for feature selection, the variables with the least relationship (linear or nonlinear) are excluded. Second, principal component analysis is used to reduce the complexity further of DPLFSs. Meanwhile, fuzzy C-means clustering is used to enhance the interpretability of DPLFSs. Then, an improved PL method is put forward for the training of each sub-fuzzy system in a fashion of bottom-up layer-by-layer, and finally, the structure optimization is performed to significantly improve the interpretability of DPLFSs. Experiments on several benchmark data sets show the advantages of a DPLFS: (1) it can handle medium-scale data sets; (2) it can overcome the curse of dimensionality faced by PL; (3) its precision and generalization are greatly improved; and (4) it can overcome the poor interpretability of deep learning networks. Compared with shallow and deep learning algorithms, DPLFSs have the advantages of interpretability, self-learning, and high precision. DPLFS1 is superior for medium-scale data; DPLFS2 is more efficient and effective for high-dimensional problems, has a faster convergence, and is more interpretable.

Wearable electroencephalography (EEG) devices for seizure detection accuracy and reliability are deep learning (DL) applications in the field of epilepsy diagnosis. In this study, we sought to increase the accuracy of seizure detection using advanced DL algorithms on the Children’s Hospital Boston - Massachusetts Institute of Technology (CHB-MIT) EEG database. First, a fully convolutional network (FCN) was trained and assessed using accuracy and recall/precision metrics, and the early stopping technique was used to avoid overfitting. To assess the performance, the FCN was evaluated in terms of various metrics, including accuracy, precision, recall, F1-score, and receiver operating characteristic (ROC)-area under the curve (AUC). In addition, two-dimensional (2D) convolutional neural networks (CNNs) and long short-term memory (LSTM) models were used to model the database, and their performance was thoroughly measured using different metrics, graphs, and confusion matrices. Using LSTM variants, such as standard LSTM, bidirectional LSTM, stacked LSTM, and LSTM attention mechanisms, hybrid convolutional LSTM (ConvLSTM) models were trained and compared. The comparison was conducted based on the training and validation accuracy and loss, as well as the graphs resulting from the precision–recall curves. Apart from DL approaches, EEG signal analysis using time–frequency techniques, such as wavelet transform and short-time Fourier transform, has also been investigated. These methods assisted in the analysis of the time–frequency features of EEG signals in combination with DL models. This study demonstrates that the performance of wearable EEG devices can be augmented using a combination of DL and seizure signal processing techniques. The FCN achieved an accuracy of 92%, with a recall for seizures of 33%, an F1-score of 0.03, and strong ROC-AUC results. The 2D CNN achieved 96% accuracy, a seizure recall of 70%, an F1-score of 0.12, and an ROC-AUC score of 78%. The baseline LSTM struggled with effectiveness at 53% accuracy with a seizure recall of 18%. In contrast, the LSTM model, which incorporated synthetic minority oversampling technique (SMOTE) balancing, was able to reach improvements of up to 89% accuracy, with a precision of 91%, a recall of 86%, an F1-score of 0.89, and a strong ROC curve performance. Among the models, the LSTM with SMOTE was the best performer, with 89% accuracy, 91% precision, 86% recall, and an F1-score of 0.89. These results provide evidence that applying techniques for data balancing in combination with certain DL network architectures significantly improves the detection of seizures using wearable EEG devices worn on the body. We believe that real-time monitoring and high-performance systems are feasible using optimized DL frameworks. The analysis of the performance of different models allows for the understanding of the possibilities of optimizing the architectures of DL algorithms for the modern diagnosis of epilepsy in real time. The source code used to carry out the experiments is publicly available at CHB-MIT EEG Dataset Python Scripts (https://www.kaggle.com/code/adnankust/adnaneeg1).

Deep Learning for Contrast Enhanced Mammography - A Systematic Review

Background Combining the artificial intelligence algorithm with the known clinical risk factors may provide enhanced accuracy for prediction of the hidden atrial fibrillation (AF) in patients with embolic stroke with undetermined source (ESUS). Purpose We aimed to develop enhanced prediction models for AF with deep learning algorithm (DLA) and clinical predictors in patients with ESUS. The DLA was created to identify the patients with paroxysmal AF based on their electrocardiograms (ECG) during sinus rhythm. Methods We analyzed the 221 patients who underwent insertable cardiac monitor (ICM) for AF detection following ESUS. The DLA was previously developed with sinus rhythm ECGs of 10,605 paroxysmal AF patients and 50,522 non-AF patients. The convolutional neural network was used for the DLA. The primary endpoint was defined as any AF episode lasting over 5 min by ICM. The atrial ectopic burden (AEB) was calculated as the percentage of the number of conducted QRS by atrial ectopy on Holter monitoring. Results AF (≥5 min) was detected by ICM in 32 patients (14.5%) during follow-up period of 15.1±8.6 months. AF patients had higher AEB (0.199% vs 0.023%, p&amp;lt;0.001), larger left atrial diameters (LAD, 41.2 mm vs 35.7 mm, p&amp;lt;0.001), and larger left atrial volume index (LAVI, 46.4 ml/m2 vs 32.3 ml/m2, p&amp;lt;0.001) than those without AF. The means of calculated probabilities of AF by DLA were higher in patients with AF than those without AF (63.8% vs 40.2%, p&amp;lt;0.001). In the receiver operating characteristic curve analysis, the areas under the curve (AUC) were the highest in DLA (0.824) followed by AEB (0.784), LAVI (0.780), and LAD (0.768). The multivariable model with AEB, LAVI, and DLA demonstrated excellent prediction accuracy for paroxysmal AF (AUC: 0.902, Figure 1) Conclusions In patients with ESUS, the DLA outperformed other clinical risk factors for prediction of AF. Combining DLA with AEB, LAD and LAVI could is a potential useful tool to predict AF in ESUS patients. Funding Acknowledgement Type of funding sources: None.

To automatically detect and classify the lesions of diabetic retinopathy (DR) in fundus fluorescein angiography (FFA) images using deep learning algorithm through comparing 3 convolutional neural networks (CNNs). A total of 4067 FFA images from Eye Center at the Second Affiliated Hospital of Zhejiang University School of Medicine were annotated with 4 kinds of lesions of DR, including non-perfusion regions (NP), microaneurysms, leakages, and laser scars. Three CNNs including DenseNet, ResNet50, and VGG16 were trained to achieve multi-label classification, which means the algorithms could identify 4 retinal lesions above at the same time. The area under the curve (AUC) of DenseNet reached 0.8703, 0.9435, 0.9647, and 0.9653 for detecting NP, microaneurysms, leakages, and laser scars, respectively. For ResNet50, AUC was 0.8140 for NP, 0.9097 for microaneurysms, 0.9585 for leakages, and 0.9115 for laser scars. And for VGG16, AUC was 0.7125 for NP, 0.5569 for microaneurysms, 0.9177 for leakages, and 0.8537 for laser scars. Experimental results demonstrate that DenseNet is a suitable model to automatically detect and distinguish retinal lesions in the FFA images with multi-label classification, which lies the foundation of automatic analysis for FFA images and comprehensive diagnosis and treatment decision-making for DR.