Application of explainable artificial intelligence for healthcare: A systematic review of the last decade (2011-2022)

Insights into Predicting Tooth Extraction from Panoramic Dental Images: Artificial Intelligence vs. Dentists

OBJECTIVES

Tooth extraction is one of the most frequently performed medical procedures. The indication is based on the combination of clinical and radiological examination and individual patient parameters and should be made with great care. However, determining whether a tooth should be extracted is not always a straightforward decision. Moreover, visual and cognitive pitfalls in the analysis of radiographs may lead to incorrect decisions. Artificial intelligence (AI) could be used as a decision support tool to provide a score of tooth extractability.

MATERIAL AND METHODS

Using 26,956 single teeth images from 1,184 panoramic radiographs (PANs), we trained a ResNet50 network to classify teeth as either extraction-worthy or preservable. For this purpose, teeth were cropped with different margins from PANs and annotated. The usefulness of the AI-based classification as well that of dentists was evaluated on a test dataset. In addition, the explainability of the best AI model was visualized via a class activation mapping using CAMERAS.

RESULTS

The ROC-AUC for the best AI model to discriminate teeth worthy of preservation was 0.901 with 2% margin on dental images. In contrast, the average ROC-AUC for dentists was only 0.797. With a 19.1% tooth extractions prevalence, the AI model's PR-AUC was 0.749, while the dentist evaluation only reached 0.589.

CONCLUSION

AI models outperform dentists/specialists in predicting tooth extraction based solely on X-ray images, while the AI performance improves with increasing contextual information.

CLINICAL RELEVANCE

AI could help monitor at-risk teeth and reduce errors in indications for extractions.

Dynamic and transdiagnostic risk calculator based on Natural Language Processing for the prediction of psychosis in secondary mental health care: development and internal-external validation cohort study

BACKGROUND

Automatic transdiagnostic risk calculators can improve detection of individuals at risk of psychosis. However, they rely on a single point in time assessment and can be refined with dynamic modelling techniques that account for changes in risk over time.

METHODS

We included n=158,139 patients (n=5,007 events) receiving a first index diagnosis of a non-organic and non-psychotic mental disorder within Electronic Health Records from the SLaM NHS Foundation Trust between 01/01/2008 and 10/08/2021. A dynamic Cox landmark model was developed to estimate the 2-year risk of developing psychosis according to TRIPOD statement. The dynamic model included 24 predictors extracted at nine landmark points (baseline, 0, 6, 12, 24, 30, 36, 42, and 48 months): three demographic, one clinical, and 20 Natural Language Processing (NLP) based symptom and substance use predictors. Performance was compared to a static Cox regression model with all predictors assessed at baseline only, indexed via discrimination (C-index), calibration (calibration plots), and potential clinical utility (decision curves) in internal-external validation.

RESULTS

The dynamic model improves discrimination performance compared to the static model at baseline (dynamic: C-index=0.9; static: C-index=0.87) to the final landmark point (dynamic: C-index=0.79; static: C-index=0.76). The dynamic model was also significantly better calibrated (calibration slope=0.97-1.1) than the static model at later landmark points (≥24 months). Net benefit was higher in the dynamic compared to the static model at later landmark points (≥24 months).

CONCLUSION

These findings suggest that dynamic prediction models can improve detection of individuals at risk for psychosis in secondary mental health care.

ML3CNet: Non-local means-assisted automatic framework for lung cancer subtypes classification using histopathological images

BACKGROUND AND OBJECTIVE

Lung cancer (LC) has a high fatality rate that continuously affects human lives all over the world. Early detection of LC prolongs human life and helps to prevent the disease. Histopathological inspection is a common method to diagnose LC. Visual inspection of histopathological diagnosis necessitates more inspection time, and the decision depends on the subjective perception of clinicians. Usually, machine learning techniques mostly depend on traditional feature extraction which is labor-intensive and may not be appropriate for enormous data. In this work, a convolutional neural network (CNN)-based architecture is proposed for the more effective classification of lung tissue subtypes using histopathological images.

METHODS

Authors have utilized the first-time nonlocal mean (NLM) filter to suppress the effect of noise from histopathological images. NLM filter efficiently eliminated noise while preserving the edges of images. Then, the obtained denoised images are given as input to the proposed multi-headed lung cancer classification convolutional neural network (ML3CNet). Furthermore, the model quantization technique is utilized to reduce the size of the proposed model for the storage of the data. Reduction in model size requires less memory and speeds up data processing.

RESULTS

The effectiveness of the proposed model is compared with the other existing state-of-the-art methods. The proposed ML3CNet achieved an average classification accuracy of 99.72%, sensitivity of 99.66%, precision of 99.64%, specificity of 99.84%, F-1 score of 0.9965, and area under the curve of 0.9978. The quantized accuracy of 98.92% is attained by the proposed model. To validate the applicability of the proposed ML3CNet, it has also been tested on the colon cancer dataset.

CONCLUSION

The findings reveal that the proposed approach can be beneficial to automatically classify LC subtypes that might assist healthcare workers in making decisions more precisely. The proposed model can be implemented on the hardware using Raspberry Pi for practical realization.

A deep learning-based model for detecting Leishmania amastigotes in microscopic slides: a new approach to telemedicine

BACKGROUND

Leishmaniasis, an illness caused by protozoa, accounts for a substantial number of human fatalities globally, thereby emerging as one of the most fatal parasitic diseases. The conventional methods employed for detecting the Leishmania parasite through microscopy are not only time-consuming but also susceptible to errors. Therefore, the main objective of this study is to develop a model based on deep learning, a subfield of artificial intelligence, that could facilitate automated diagnosis of leishmaniasis.

METHODS

In this research, we introduce LeishFuNet, a deep learning framework designed for detecting Leishmania parasites in microscopic images. To enhance the performance of our model through same-domain transfer learning, we initially train four distinct models: VGG19, ResNet50, MobileNetV2, and DenseNet 169 on a dataset related to another infectious disease, COVID-19. These trained models are then utilized as new pre-trained models and fine-tuned on a set of 292 self-collected high-resolution microscopic images, consisting of 138 positive cases and 154 negative cases. The final prediction is generated through the fusion of information analyzed by these pre-trained models. Grad-CAM, an explainable artificial intelligence technique, is implemented to demonstrate the model's interpretability.

RESULTS

The final results of utilizing our model for detecting amastigotes in microscopic images are as follows: accuracy of 98.95 1.4%, specificity of 98 2.67%, sensitivity of 100%, precision of 97.91 2.77%, F1-score of 98.92 1.43%, and Area Under Receiver Operating Characteristic Curve of 99 1.33.

CONCLUSION

The newly devised system is precise, swift, user-friendly, and economical, thus indicating the potential of deep learning as a substitute for the prevailing leishmanial diagnostic techniques.

Diagnosis of multiple sclerosis using optical coherence tomography supported by explainable artificial intelligence

BACKGROUND/OBJECTIVES

Study of retinal structure based on optical coherence tomography (OCT) data can facilitate early diagnosis of relapsing-remitting multiple sclerosis (RRMS). Although artificial intelligence can provide highly reliable diagnoses, the results obtained must be explainable.

SUBJECTS/METHODS

The study included 79 recently diagnosed RRMS patients and 69 age matched healthy control subjects. Thickness (Avg) and inter-eye difference (Diff) features are obtained in 4 retinal layers using the posterior pole protocol. Each layer is divided into six analysis zones. The Support Vector Machine plus Recursive Feature Elimination with Leave-One-Out Cross Validation (SVM-RFE-LOOCV) approach is used to find the subset of features that reduces dimensionality and optimises the performance of the classifier.

RESULTS

SVM-RFE-LOOCV was used to identify OCT features with greatest capacity for early diagnosis, determining the area of the papillomacular bundle to be the most influential. A correlation was observed between loss of layer thickness and increase in functional disability. There was also greater functional deterioration in patients with greater asymmetry between left and right eyes. The classifier based on the top-ranked features obtained sensitivity = 0.86 and specificity = 0.90.

CONCLUSIONS

There was consistency between the features identified as relevant by the SVM-RFE-LOOCV approach and the retinotopic distribution of the retinal nerve fibres and the optic nerve head. This simple method contributes to implementation of an assisted diagnosis system and its accuracy exceeds that achieved with magnetic resonance imaging of the central nervous system, the current gold standard. This paper provides novel insights into RRMS affectation of the neuroretina.

Artificial intelligence in retinal screening using OCT images: A review of the last decade (2013-2023)

BACKGROUND AND OBJECTIVES

Optical coherence tomography (OCT) has ushered in a transformative era in the domain of ophthalmology, offering non-invasive imaging with high resolution for ocular disease detection. OCT, which is frequently used in diagnosing fundamental ocular pathologies, such as glaucoma and age-related macular degeneration (AMD), plays an important role in the widespread adoption of this technology. Apart from glaucoma and AMD, we will also investigate pertinent pathologies, such as epiretinal membrane (ERM), macular hole (MH), macular dystrophy (MD), vitreomacular traction (VMT), diabetic maculopathy (DMP), cystoid macular edema (CME), central serous chorioretinopathy (CSC), diabetic macular edema (DME), diabetic retinopathy (DR), drusen, glaucomatous optic neuropathy (GON), neovascular AMD (nAMD), myopia macular degeneration (MMD) and choroidal neovascularization (CNV) diseases. This comprehensive review examines the role that OCT-derived images play in detecting, characterizing, and monitoring eye diseases.

METHOD

The 2020 PRISMA guideline was used to structure a systematic review of research on various eye conditions using machine learning (ML) or deep learning (DL) techniques. A thorough search across IEEE, PubMed, Web of Science, and Scopus databases yielded 1787 publications, of which 1136 remained after removing duplicates. Subsequent exclusion of conference papers, review papers, and non-open-access articles reduced the selection to 511 articles. Further scrutiny led to the exclusion of 435 more articles due to lower-quality indexing or irrelevance, resulting in 76 journal articles for the review.

RESULTS

During our investigation, we found that a major challenge for ML-based decision support is the abundance of features and the determination of their significance. In contrast, DL-based decision support is characterized by a plug-and-play nature rather than relying on a trial-and-error approach. Furthermore, we observed that pre-trained networks are practical and especially useful when working on complex images such as OCT. Consequently, pre-trained deep networks were frequently utilized for classification tasks. Currently, medical decision support aims to reduce the workload of ophthalmologists and retina specialists during routine tasks. In the future, it might be possible to create continuous learning systems that can predict ocular pathologies by identifying subtle changes in OCT images.

A machine learning framework for interpretable predictions in patient pathways: The case of predicting ICU admission for patients with symptoms of sepsis

Proactive analysis of patient pathways helps healthcare providers anticipate treatment-related risks, identify outcomes, and allocate resources. Machine learning (ML) can leverage a patient's complete health history to make informed decisions about future events. However, previous work has mostly relied on so-called black-box models, which are unintelligible to humans, making it difficult for clinicians to apply such models. Our work introduces PatWay-Net, an ML framework designed for interpretable predictions of admission to the intensive care unit (ICU) for patients with symptoms of sepsis. We propose a novel type of recurrent neural network and combine it with multi-layer perceptrons to process the patient pathways and produce predictive yet interpretable results. We demonstrate its utility through a comprehensive dashboard that visualizes patient health trajectories, predictive outcomes, and associated risks. Our evaluation includes both predictive performance - where PatWay-Net outperforms standard models such as decision trees, random forests, and gradient-boosted decision trees - and clinical utility, validated through structured interviews with clinicians. By providing improved predictive accuracy along with interpretable and actionable insights, PatWay-Net serves as a valuable tool for healthcare decision support in the critical case of patients with symptoms of sepsis.

An optimized framework for processing multicentric polysomnographic data incorporating expert human oversight

Introduction

Polysomnographic recordings are essential for diagnosing many sleep disorders, yet their detailed analysis presents considerable challenges. With the rise of machine learning methodologies, researchers have created various algorithms to automatically score and extract clinically relevant features from polysomnography, but less research has been devoted to how exactly the algorithms should be incorporated into the workflow of sleep technologists. This paper presents a sophisticated data collection platform developed under the Sleep Revolution project, to harness polysomnographic data from multiple European centers.

Methods

A tripartite platform is presented: a user-friendly web platform for uploading three-night polysomnographic recordings, a dedicated splitter that segments these into individual one-night recordings, and an advanced processor that enhances the one-night polysomnography with contemporary automatic scoring algorithms. The platform is evaluated using real-life data and human scorers, whereby scoring time, accuracy, and trust are quantified. Additionally, the scorers were interviewed about their trust in the platform, along with the impact of its integration into their workflow.

Results

We found that incorporating AI into the workflow of sleep technologists both decreased the time to score by up to 65 min and increased the agreement between technologists by as much as 0.17 κ.

Discussion

We conclude that while the inclusion of AI into the workflow of sleep technologists can have a positive impact in terms of speed and agreement, there is a need for trust in the algorithms.

The State of Artificial Intelligence in Pediatric Surgery: A Systematic Review

BACKGROUND

Artificial intelligence (AI) has been recently shown to improve clinical workflows and outcomes - yet its potential in pediatric surgery remains largely unexplored. This systematic review details the use of AI in pediatric surgery.

METHODS

Nine medical databases were searched from inception until January 2023, identifying articles focused on AI in pediatric surgery. Two authors reviewed full texts of eligible articles. Studies were included if they were original investigations on the development, validation, or clinical application of AI models for pediatric health conditions primarily managed surgically. Studies were excluded if they were not peer-reviewed, were review articles, editorials, commentaries, or case reports, did not focus on pediatric surgical conditions, or did not employ at least one AI model. Extracted data included study characteristics, clinical specialty, AI method and algorithm type, AI model (algorithm) role and performance metrics, key results, interpretability, validation, and risk of bias using PROBAST and QUADAS-2.

RESULTS

Authors screened 8178 articles and included 112. Half of the studies (50%) reported predictive models (for adverse events [25%], surgical outcomes [16%] and survival [9%]), followed by diagnostic (29%) and decision support models (21%). Neural networks (44%) and ensemble learners (36%) were the most commonly used AI methods across application domains. The main pediatric surgical subspecialties represented across all models were general surgery (31%) and neurosurgery (25%). Forty-four percent of models were interpretable, and 6% were both interpretable and externally validated. Forty percent of models had a high risk of bias, and concerns over applicability were identified in 7%.

CONCLUSIONS

While AI has wide potential clinical applications in pediatric surgery, very few published AI algorithms were externally validated, interpretable, and unbiased. Future research needs to focus on developing AI models which are prospectively validated and ultimately integrated into clinical workflows.

LEVEL OF EVIDENCE

2A.

Aleatoric and epistemic uncertainty extraction of patient-specific deep learning-based dose predictions in LDR prostate brachytherapy

Objective.In brachytherapy, deep learning (DL) algorithms have shown the capability of predicting 3D dose volumes. The reliability and accuracy of such methodologies remain under scrutiny for prospective clinical applications. This study aims to establish fast DL-based predictive dose algorithms for low-dose rate (LDR) prostate brachytherapy and to evaluate their uncertainty and stability.Approach.Data from 200 prostate patients, treated with125I sources, was collected. The Monte Carlo (MC) ground truth dose volumes were calculated with TOPAS considering the interseed effects and an organ-based material assignment. Two 3D convolutional neural networks, UNet and ResUNet TSE, were trained using the patient geometry and the seed positions as the input data. The dataset was randomly split into training (150), validation (25) and test (25) sets. The aleatoric (associated with the input data) and epistemic (associated with the model) uncertainties of the DL models were assessed.Main results.For the full test set, with respect to the MC reference, the predicted prostateD90metric had mean differences of -0.64% and 0.08% for the UNet and ResUNet TSE models, respectively. In voxel-by-voxel comparisons, the average global dose difference ratio in the [-1%, 1%] range included 91.0% and 93.0% of voxels for the UNet and the ResUNet TSE, respectively. One forward pass or prediction took 4 ms for a 3D dose volume of 2.56 M voxels (128 × 160 × 128). The ResUNet TSE model closely encoded the well-known physics of the problem as seen in a set of uncertainty maps. The ResUNet TSE rectum D2cchad the largest uncertainty metric of 0.0042.Significance.The proposed DL models serve as rapid dose predictors that consider the patient anatomy and interseed attenuation effects. The derived uncertainty is interpretable, highlighting areas where DL models may struggle to provide accurate estimations. The uncertainty analysis offers a comprehensive evaluation tool for dose predictor model assessment.

Medical long-tailed learning for imbalanced data: Bibliometric analysis

BACKGROUND

In the last decade, long-tail learning has become a popular research focus in deep learning applications in medicine. However, no scientometric reports have provided a systematic overview of this scientific field. We utilized bibliometric techniques to identify and analyze the literature on long-tailed learning in deep learning applications in medicine and investigate research trends, core authors, and core journals. We expanded our understanding of the primary components and principal methodologies of long-tail learning research in the medical field.

METHODS

Web of Science was utilized to collect all articles on long-tailed learning in medicine published until December 2023. The suitability of all retrieved titles and abstracts was evaluated. For bibliometric analysis, all numerical data were extracted. CiteSpace was used to create clustered and visual knowledge graphs based on keywords.

RESULTS

A total of 579 articles met the evaluation criteria. Over the last decade, the annual number of publications and citation frequency both showed significant growth, following a power-law and exponential trend, respectively. Noteworthy contributors to this field include Husanbir Singh Pannu, Fadi Thabtah, and Talha Mahboob Alam, while leading journals such as IEEE ACCESS, COMPUTERS IN BIOLOGY AND MEDICINE, IEEE TRANSACTIONS ON MEDICAL IMAGING, and COMPUTERIZED MEDICAL IMAGING AND GRAPHICS have emerged as pivotal platforms for disseminating research in this area. The core of long-tailed learning research within the medical domain is encapsulated in six principal themes: deep learning for imbalanced data, model optimization, neural networks in image analysis, data imbalance in health records, CNN in diagnostics and risk assessment, and genetic information in disease mechanisms.

CONCLUSION

This study summarizes recent advancements in applying long-tail learning to deep learning in medicine through bibliometric analysis and visual knowledge graphs. It explains new trends, sources, core authors, journals, and research hotspots. Although this field has shown great promise in medical deep learning research, our findings will provide pertinent and valuable insights for future research and clinical practice.

Automatic identification of hypertension and assessment of its secondary effects using artificial intelligence: A systematic review (2013-2023)

Artificial Intelligence (AI) techniques are increasingly used in computer-aided diagnostic tools in medicine. These techniques can also help to identify Hypertension (HTN) in its early stage, as it is a global health issue. Automated HTN detection uses socio-demographic, clinical data, and physiological signals. Additionally, signs of secondary HTN can also be identified using various imaging modalities. This systematic review examines related work on automated HTN detection. We identify datasets, techniques, and classifiers used to develop AI models from clinical data, physiological signals, and fused data (a combination of both). Image-based models for assessing secondary HTN are also reviewed. The majority of the studies have primarily utilized single-modality approaches, such as biological signals (e.g., electrocardiography, photoplethysmography), and medical imaging (e.g., magnetic resonance angiography, ultrasound). Surprisingly, only a small portion of the studies (22 out of 122) utilized a multi-modal fusion approach combining data from different sources. Even fewer investigated integrating clinical data, physiological signals, and medical imaging to understand the intricate relationships between these factors. Future research directions are discussed that could build better healthcare systems for early HTN detection through more integrated modeling of multi-modal data sources.

Deep learning of movement behavior profiles and their association with markers of cardiometabolic health

BACKGROUND

Traditionally, existing studies assessing the health associations of accelerometer-measured movement behaviors have been performed with few averaged values, mainly representing the duration of physical activities and sedentary behaviors. Such averaged values cannot naturally capture the complex interplay between the duration, timing, and patterns of accumulation of movement behaviors, that altogether may be codependently related to health outcomes in adults. In this study, we introduce a novel approach to visually represent recorded movement behaviors as images using original accelerometer outputs. Subsequently, we utilize these images for cluster analysis employing deep convolutional autoencoders.

METHODS

Our method involves converting minute-by-minute accelerometer outputs (activity counts) into a 2D image format, capturing the entire spectrum of movement behaviors performed by each participant. By utilizing convolutional autoencoders, we enable the learning of these image-based representations. Subsequently, we apply the K-means algorithm to cluster these learned representations. We used data from 1812 adult (20-65 years) participants in the National Health and Nutrition Examination Survey (NHANES, 2003-2006 cycles) study who worn a hip-worn accelerometer for 7 seven consecutive days and provided valid accelerometer data.

RESULTS

Deep convolutional autoencoders were able to learn the image representation, encompassing the entire spectrum of movement behaviors. The images were encoded into 32 latent variables, and cluster analysis based on these learned representations for the movement behavior images resulted in the identification of four distinct movement behavior profiles characterized by varying levels, timing, and patterns of accumulation of movement behaviors. After adjusting for potential covariates, the movement behavior profile characterized as "Early-morning movers" and the profile characterized as "Highest activity" both had lower levels of insulin (P < 0.01 for both), triglycerides (P < 0.05 and P < 0.01, respectively), HOMA-IR (P < 0.01 for both), and plasma glucose (P < 0.05 and P < 0.1, respectively) compared to the "Lowest activity" profile. No significant differences were observed for the "Least sedentary movers" profile compared to the "Lowest activity" profile.

CONCLUSIONS

Deep learning of movement behavior profiles revealed that, in addition to duration and patterns of movement behaviors, the timing of physical activity may also be crucial for gaining additional health benefits.

Trustworthy deep learning framework for the detection of abnormalities in X-ray shoulder images

Musculoskeletal conditions affect an estimated 1.7 billion people worldwide, causing intense pain and disability. These conditions lead to 30 million emergency room visits yearly, and the numbers are only increasing. However, diagnosing musculoskeletal issues can be challenging, especially in emergencies where quick decisions are necessary. Deep learning (DL) has shown promise in various medical applications. However, previous methods had poor performance and a lack of transparency in detecting shoulder abnormalities on X-ray images due to a lack of training data and better representation of features. This often resulted in overfitting, poor generalisation, and potential bias in decision-making. To address these issues, a new trustworthy DL framework has been proposed to detect shoulder abnormalities (such as fractures, deformities, and arthritis) using X-ray images. The framework consists of two parts: same-domain transfer learning (TL) to mitigate imageNet mismatch and feature fusion to reduce error rates and improve trust in the final result. Same-domain TL involves training pre-trained models on a large number of labelled X-ray images from various body parts and fine-tuning them on the target dataset of shoulder X-ray images. Feature fusion combines the extracted features with seven DL models to train several ML classifiers. The proposed framework achieved an excellent accuracy rate of 99.2%, F1Score of 99.2%, and Cohen's kappa of 98.5%. Furthermore, the accuracy of the results was validated using three visualisation tools, including gradient-based class activation heat map (Grad CAM), activation visualisation, and locally interpretable model-independent explanations (LIME). The proposed framework outperformed previous DL methods and three orthopaedic surgeons invited to classify the test set, who obtained an average accuracy of 79.1%. The proposed framework has proven effective and robust, improving generalisation and increasing trust in the final results.

A scoping review of applications of artificial intelligence in kinematics and kinetics of ankle sprains - current state-of-the-art and future prospects

BACKGROUND

Despite the existence of evidence-based rehabilitation strategies that address biomechanical deficits, the persistence of recurrent ankle problems in 70% of patients with acute ankle sprains highlights the unresolved nature of this issue. Artificial intelligence (AI) emerges as a promising tool to identify definitive predictors for ankle sprains. This paper aims to summarize the use of AI in investigating the ankle biomechanics of healthy and subjects with ankle sprains.

METHODS

Articles published between 2010 and 2023 were searched from five electronic databases. 59 papers were included for analysis with regards to: i). types of motion tested (functional vs. purposeful ankle movement); ii) types of biomechanical parameters measured (kinetic vs kinematic); iii) types of sensor systems used (lab-based vs field-based); and, iv) AI techniques used.

FINDINGS

Most studies (83.1%) examined biomechanics during functional motion. Single kinematic parameter, specifically ankle range of motion, could obtain accuracy up to 100% in identifying injury status. Wearable sensor exhibited high reliability for use in both laboratory and on-field/clinical settings. AI algorithms primarily utilized electromyography and joint angle information as input data. Support vector machine was the most used supervised learning algorithm (18.64%), while artificial neural network demonstrated the highest accuracy in eight studies.

INTERPRETATIONS

The potential for remote patient monitoring is evident with the adoption of field-based devices. Nevertheless, AI-based sensors are underutilized in detecting ankle motions at risk of sprain. We identify three key challenges: sensor designs, the controllability of AI models, and the integration of AI-sensor models, providing valuable insights for future research.

Explainable artificial intelligence for the automated assessment of the retinal vascular tortuosity

Retinal vascular tortuosity is an excessive bending and twisting of the blood vessels in the retina that is associated with numerous health conditions. We propose a novel methodology for the automated assessment of the retinal vascular tortuosity from color fundus images. Our methodology takes into consideration several anatomical factors to weigh the importance of each individual blood vessel. First, we use deep neural networks to produce a robust extraction of the different anatomical structures. Then, the weighting coefficients that are required for the integration of the different anatomical factors are adjusted using evolutionary computation. Finally, the proposed methodology also provides visual representations that explain the contribution of each individual blood vessel to the predicted tortuosity, hence allowing us to understand the decisions of the model. We validate our proposal in a dataset of color fundus images providing a consensus ground truth as well as the annotations of five clinical experts. Our proposal outperforms previous automated methods and offers a performance that is comparable to that of the clinical experts. Therefore, our methodology demonstrates to be a viable alternative for the assessment of the retinal vascular tortuosity. This could facilitate the use of this biomarker in clinical practice and medical research.

Deep learning-based natural language processing for detecting medical symptoms and histories in emergency patient triage

OBJECTIVE

The manual recording of electronic health records (EHRs) by clinicians in the emergency department (ED) is time-consuming and challenging. In light of recent advancements in large language models (LLMs) such as GPT and BERT, this study aimed to design and validate LLMs for automatic clinical diagnoses. The models were designed to identify 12 medical symptoms and 2 patient histories from simulated clinician-patient conversations within 6 primary symptom scenarios in emergency triage rooms.

MATERIALS AND METHOD

We developed classification models by fine-tuning BERT, a transformer-based pre-trained model. We subsequently analyzed these models using eXplainable artificial intelligence (XAI) and the Shapley additive explanation (SHAP) method. A Turing test was conducted to ascertain the reliability of the XAI results by comparing them to the outcomes of tasks performed and explained by medical workers. An emergency medicine specialist assessed the results of both XAI and the medical workers.

RESULTS

We fine-tuned four pre-trained LLMs and compared their classification performance. The KLUE-RoBERTa-based model demonstrated the highest performance (F1-score: 0.965, AUROC: 0.893) on human-transcribed script data. The XAI results using SHAP showed an average Jaccard similarity of 0.722 when compared with explanations of medical workers for 15 samples. The Turing test results revealed a small 6% gap, with XAI and medical workers receiving the mean scores of 3.327 and 3.52, respectively.

CONCLUSION

This paper highlights the potential of LLMs for automatic EHR recording in Korean EDs. The KLUE-RoBERTa-based model demonstrated superior classification performance. Furthermore, XAI using SHAP provided reliable explanations for model outputs. The reliability of these explanations was confirmed by a Turing test.

A scoping review of fair machine learning techniques when using real-world data

OBJECTIVE

The integration of artificial intelligence (AI) and machine learning (ML) in health care to aid clinical decisions is widespread. However, as AI and ML take important roles in health care, there are concerns about AI and ML associated fairness and bias. That is, an AI tool may have a disparate impact, with its benefits and drawbacks unevenly distributed across societal strata and subpopulations, potentially exacerbating existing health inequities. Thus, the objectives of this scoping review were to summarize existing literature and identify gaps in the topic of tackling algorithmic bias and optimizing fairness in AI/ML models using real-world data (RWD) in health care domains.

METHODS

We conducted a thorough review of techniques for assessing and optimizing AI/ML model fairness in health care when using RWD in health care domains. The focus lies on appraising different quantification metrics for accessing fairness, publicly accessible datasets for ML fairness research, and bias mitigation approaches.

RESULTS

We identified 11 papers that are focused on optimizing model fairness in health care applications. The current research on mitigating bias issues in RWD is limited, both in terms of disease variety and health care applications, as well as the accessibility of public datasets for ML fairness research. Existing studies often indicate positive outcomes when using pre-processing techniques to address algorithmic bias. There remain unresolved questions within the field that require further research, which includes pinpointing the root causes of bias in ML models, broadening fairness research in AI/ML with the use of RWD and exploring its implications in healthcare settings, and evaluating and addressing bias in multi-modal data.

CONCLUSION

This paper provides useful reference material and insights to researchers regarding AI/ML fairness in real-world health care data and reveals the gaps in the field. Fair AI/ML in health care is a burgeoning field that requires a heightened research focus to cover diverse applications and different types of RWD.

MRI-Based Radiomics and Deep Learning in Biological Characteristics and Prognosis of Hepatocellular Carcinoma: Opportunities and Challenges

Hepatocellular carcinoma (HCC) is the fifth most common malignancy and the third leading cause of cancer-related death worldwide. HCC exhibits strong inter-tumor heterogeneity, with different biological characteristics closely associated with prognosis. In addition, patients with HCC often distribute at different stages and require diverse treatment options at each stage. Due to the variability in tumor sensitivity to different therapies, determining the optimal treatment approach can be challenging for clinicians prior to treatment. Artificial intelligence (AI) technology, including radiomics and deep learning approaches, has emerged as a unique opportunity to improve the spectrum of HCC clinical care by predicting biological characteristics and prognosis in the medical imaging field. The radiomics approach utilizes handcrafted features derived from specific mathematical formulas to construct various machine-learning models for medical applications. In terms of the deep learning approach, convolutional neural network models are developed to achieve high classification performance based on automatic feature extraction from images. Magnetic resonance imaging offers the advantage of superior tissue resolution and functional information. This comprehensive evaluation plays a vital role in the accurate assessment and effective treatment planning for HCC patients. Recent studies have applied radiomics and deep learning approaches to develop AI-enabled models to improve accuracy in predicting biological characteristics and prognosis, such as microvascular invasion and tumor recurrence. Although AI-enabled models have demonstrated promising potential in HCC with biological characteristics and prognosis prediction with high performance, one of the biggest challenges, interpretability, has hindered their implementation in clinical practice. In the future, continued research is needed to improve the interpretability of AI-enabled models, including aspects such as domain knowledge, novel algorithms, and multi-dimension data sources. Overcoming these challenges would allow AI-enabled models to significantly impact the care provided to HCC patients, ultimately leading to their deployment for clinical use. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY: Stage 2.

Starting the Conversation Around the Ethical Use of Artificial Intelligence in Applied Behavior Analysis

Artificial intelligence (AI) is increasingly a part of our everyday lives. Though much AI work in healthcare has been outside of applied behavior analysis (ABA), researchers within ABA have begun to demonstrate many different ways that AI might improve the delivery of ABA services. Though AI offers many exciting advances, absent from the behavior analytic literature thus far is conversation around ethical considerations when developing, building, and deploying AI technologies. Further, though AI is already in the process of coming to ABA, it is unknown the extent to which behavior analytic practitioners are familiar (and comfortable) with the use of AI in ABA. The purpose of this article is twofold. First, to describe how existing ethical publications (e.g., BACB Code of Ethics) do and do not speak to the unique ethical concerns with deploying AI in everyday, ABA service delivery settings. Second, to raise questions for consideration that might inform future ethical guidelines when developing and using AI in ABA service delivery. In total, we hope this article sparks proactive dialog around the ethical use of AI in ABA before the field is required to have a reactionary conversation.

Global Regulatory Frameworks for the Use of Artificial Intelligence (AI) in the Healthcare Services Sector

The healthcare sector is faced with challenges due to a shrinking healthcare workforce and a rise in chronic diseases that are worsening with demographic and epidemiological shifts. Digital health interventions that include artificial intelligence (AI) are being identified as some of the potential solutions to these challenges. The ultimate aim of these AI systems is to improve the patient's health outcomes and satisfaction, the overall population's health, and the well-being of healthcare professionals. The applications of AI in healthcare services are vast and are expected to assist, automate, and augment several healthcare services. Like any other emerging innovation, AI in healthcare also comes with its own risks and requires regulatory controls. A review of the literature was undertaken to study the existing regulatory landscape for AI in the healthcare services sector in developed nations. In the global regulatory landscape, most of the regulations for AI revolve around Software as a Medical Device (SaMD) and are regulated under digital health products. However, it is necessary to note that the current regulations may not suffice as AI-based technologies are capable of working autonomously, adapting their algorithms, and improving their performance over time based on the new real-world data that they have encountered. Hence, a global regulatory convergence for AI in healthcare, similar to the voluntary AI code of conduct that is being developed by the US-EU Trade and Technology Council, would be beneficial to all nations, be it developing or developed.

Harnessing of Artificial Intelligence for the Diagnosis and Prevention of Hospital-Acquired Infections: A Systematic Review

Healthcare-associated infections (HAIs) are the most common adverse events in healthcare and constitute a major global public health concern. Surveillance represents the foundation for the effective prevention and control of HAIs, yet conventional surveillance is costly and labor intensive. Artificial intelligence (AI) and machine learning (ML) have the potential to support the development of HAI surveillance algorithms for the understanding of HAI risk factors, the improvement of patient risk stratification as well as the prediction and timely detection and prevention of infections. AI-supported systems have so far been explored for clinical laboratory testing and imaging diagnosis, antimicrobial resistance profiling, antibiotic discovery and prediction-based clinical decision support tools in terms of HAIs. This review aims to provide a comprehensive summary of the current literature on AI applications in the field of HAIs and discuss the future potentials of this emerging technology in infection practice. Following the PRISMA guidelines, this study examined the articles in databases including PubMed and Scopus until November 2023, which were screened based on the inclusion and exclusion criteria, resulting in 162 included articles. By elucidating the advancements in the field, we aim to highlight the potential applications of AI in the field, report related issues and shortcomings and discuss the future directions.

Machine learning prediction models for in-hospital postoperative functional outcome after moderate-to-severe traumatic brain injury

AIM

This study aims to utilize machine learning (ML) and logistic regression (LR) models to predict surgical outcomes among patients with traumatic brain injury (TBI) based on admission examination, assisting in making optimal surgical treatment decision for these patients.

METHOD

We conducted a retrospective review of patients hospitalized in our department for moderate-to-severe TBI. Patients admitted between October 2011 and October 2022 were assigned to the training set, while patients admitted between November 2022 and May 2023 were designated as the external validation set. Five ML algorithms and LR model were employed to predict the postoperative Glasgow Outcome Scale (GOS) status at discharge using clinical and routine blood data collected upon admission. The Shapley (SHAP) plot was utilized for interpreting the models.

RESULTS

A total of 416 patients were included in this study, and they were divided into the training set (n = 396) and the external validation set (n = 47). The ML models, using both clinical and routine blood data, were able to predict postoperative GOS outcomes with area under the curve (AUC) values ranging from 0.860 to 0.900 during the internal cross-validation and from 0.801 to 0.890 during the external validation. In contrast, the LR model had the lowest AUC values during the internal and external validation (0.844 and 0.567, respectively). When blood data was not available, the ML models achieved AUCs of 0.849 to 0.870 during the internal cross-validation and 0.714 to 0.861 during the external validation. Similarly, the LR model had the lowest AUC values (0.821 and 0.638, respectively). Through repeated cross-validation analysis, we found that routine blood data had a significant association with higher mean AUC values in all ML and LR models. The SHAP plot was used to visualize the contributions of all predictors and highlighted the significance of blood data in the lightGBM model.

CONCLUSION

The study concluded that ML models could provide rapid and accurate predictions for postoperative GOS outcomes at discharge following moderate-to-severe TBI. The study also highlighted the crucial role of routine blood tests in improving such predictions, and may contribute to the optimization of surgical treatment decision-making for patients with TBI.

Machine learning models for predicting in-hospital outcomes after non-surgical treatment among patients with moderate-to-severe traumatic brain injury

AIM

This study aims to develop prediction models for in-hospital outcomes after non-surgical treatment among patients with moderate-to-severe traumatic brain injury (TBI).

METHOD

We conducted a retrospective review of patients hospitalized for moderate-to-severe TBI in our department from 2011 to 2020. Five machine learning (ML) algorithms and the conventional logistic regression (LR) model were employed to predict in-hospital mortality and the Glasgow Outcome Scale (GOS) functional outcomes. These models utilized clinical and routine blood data collected upon admission.

RESULTS

This study included a total of 196 patients who received only non-surgical treatment after moderate-to-severe TBI. When predicting mortality, ML models achieved area under the curve (AUC) values of 0.921 to 0.994 using clinical and routine blood data, and 0.877 to 0.982 using only clinical data. In comparison, LR models yielded AUCs of 0.762 and 0.730 respectively. When predicting the GOS outcome, ML models achieved AUCs of 0.870 to 0.915 using clinical and routine blood data, and 0.858 to 0.927 using only clinical data. In comparison, the LR model yielded AUCs of 0.798 and 0.787 respectively. Repeated internal validation showed that the contributions of routine blood data for prediction models may depend on different prediction algorithms and different outcome measurements.

CONCLUSION

The study reported ML-based prediction models that provided rapid and accurate predictions on short-term outcomes after non-surgical treatment among patients with moderate-to-severe TBI. The study also highlighted the superiority of ML models over conventional LR models and proposed the complex contributions of routine blood data in such predictions.

Research on symptoms composition, time series evolution, and network visualisation of interstitial cystitis based on complex network community discovery algorithm

We analyzed the symptoms composition of Interstitial Cystitis (IC), the regularity of the evolution of symptoms before and after treatment, and the visualization of the community network, to provide a reference for clinical diagnosis and treatment of Interstitial Cystitis. Based on the outpatient electronic case data of 552 patients with Interstitial Cystitis, we used a complex network community discovery algorithm, directed weighted complex network, and Sankey map to mine the data of the symptoms composition of Interstitial Cystitis, the evolution of symptoms before and after treatment and the visualization of the community network, to analyze the epidemiological characteristics of interstitial cystitis symptoms in the real world. By the community division of the complex network of interstitial cystitis symptoms, We finally obtained three core symptom communities. Among them, symptom community A (bladder-related symptoms) is the symptom community with the highest proportion of nodes (60.00%) in the complex network of Interstitial Cystitis, symptom community B (non-bladder-related symptoms 1) ranks second (32.00%) in a complex network of Interstitial Cystitis, and symptom community C (non-bladder-related symptoms 2) has the lowest proportion (8.00%). There is a complex evolutionary relationship between the symptoms of Interstitial Cystitis before and after treatment. Among the single symptoms before and after treatment, the decreased rate of Day frequency is 93.22%, and the reduced urgency rate is 93.07%. The decline rate of Nocturia was 82.33%. From the perspective of different communities, the overall symptoms of symptom community A decreased by 34.39% after treatment, the general symptoms of symptom community B decreased by 35.37%, and the prevalent symptoms of symptom community C decreased by 71.43%. In the case of using diet regulation treatment to treat bladder pain, the cure rate of bladder pain can reach 22.67%. The cure rate of burning in bladders can get 15.38% with Percutaneous Sacral neuromodulation, 96.95% with diet regulation treatment, and 100% with Percutaneous Sacral neuromodulation. When using behavioral physiotherapy to treat bladder pain, 3.57% of the patient's symptoms change to bladder discomfort; 4% of the patient's symptoms change to bladder discomfort when using oral medicine to treat bladder pain.Symptom research methods based on community findings can effectively explore the rule of symptom outcome of Interstitial Cystitis before and after treatment, and the results are highly interpretable by professionals. The cover image is based on the Original Article Research on symptoms composition, time series evolution, and network visualisation of interstitial cystitis based on complex network community discovery algorithm by Lei Pang et al., https://doi.org/10.1049/syb2.12083.

Application of artificial intelligence techniques for non-alcoholic fatty liver disease diagnosis: A systematic review (2005-2023)

BACKGROUND AND OBJECTIVES

Non-alcoholic fatty liver disease (NAFLD) is a common liver disease with a rapidly growing incidence worldwide. For prognostication and therapeutic decisions, it is important to distinguish the pathological stages of NAFLD: steatosis, steatohepatitis, and liver fibrosis, which are definitively diagnosed on invasive biopsy. Non-invasive ultrasound (US) imaging, including US elastography technique, and clinical parameters can be used to diagnose and grade NAFLD and its complications. Artificial intelligence (AI) is increasingly being harnessed for developing NAFLD diagnostic models based on clinical, biomarker, or imaging data. In this work, we systemically reviewed the literature for AI-enabled NAFLD diagnostic models based on US (including elastography) and clinical (including serological) data.

METHODS

We performed a comprehensive search on Google Scholar, Scopus, and PubMed search engines for articles published between January 2005 and June 2023 related to AI models for NAFLD diagnosis based on US and/or clinical parameters using the following search terms: "non-alcoholic fatty liver disease", "non-alcoholic steatohepatitis", "deep learning", "machine learning", "artificial intelligence", "ultrasound imaging", "sonography", "clinical information".

RESULTS

We reviewed 64 published models that used either US (including elastography) or clinical data input to detect the presence of NAFLD, non-alcoholic steatohepatitis, and/or fibrosis, and in some cases, the severity of steatosis, inflammation, and/or fibrosis as well. The performances of the published models were summarized, and stratified by data input and algorithms used, which could be broadly divided into machine and deep learning approaches.

CONCLUSION

AI models based on US imaging and clinical data can reliably detect NAFLD and its complications, thereby reducing diagnostic costs and the need for invasive liver biopsy. The models offer advantages of efficiency, accuracy, and accessibility, and serve as virtual assistants for specialists to accelerate disease diagnosis and reduce treatment costs for patients and healthcare systems.

AI-Enhanced Detection of Clinically Relevant Structural and Functional Anomalies in MRI: Traversing the Landscape of Conventional to Explainable Approaches

Anomaly detection in medical imaging, particularly within the realm of magnetic resonance imaging (MRI), stands as a vital area of research with far-reaching implications across various medical fields. This review meticulously examines the integration of artificial intelligence (AI) in anomaly detection for MR images, spotlighting its transformative impact on medical diagnostics. We delve into the forefront of AI applications in MRI, exploring advanced machine learning (ML) and deep learning (DL) methodologies that are pivotal in enhancing the precision of diagnostic processes. The review provides a detailed analysis of preprocessing, feature extraction, classification, and segmentation techniques, alongside a comprehensive evaluation of commonly used metrics. Further, this paper explores the latest developments in ensemble methods and explainable AI, offering insights into future directions and potential breakthroughs. This review synthesizes current insights, offering a valuable guide for researchers, clinicians, and medical imaging experts. It highlights AI's crucial role in improving the precision and speed of detecting key structural and functional irregularities in MRI. Our exploration of innovative techniques and trends furthers MRI technology development, aiming to refine diagnostics, tailor treatments, and elevate patient care outcomes. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY: Stage 1.

Predicting disease onset from electronic health records for population health management: a scalable and explainable Deep Learning approach

Introduction

The move from a reactive model of care which treats conditions when they arise to a proactive model which intervenes early to prevent adverse healthcare events will benefit from advances in the predictive capabilities of Artificial Intelligence and Machine Learning. This paper investigates the ability of a Deep Learning (DL) approach to predict future disease diagnosis from Electronic Health Records (EHR) for the purposes of Population Health Management.

Methods

In this study, embeddings were created using a Word2Vec algorithm from structured vocabulary commonly used in EHRs e.g., Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) codes. This study is based on longitudinal medical data from ~50 m patients in the USA. We introduced a novel method of including binned observation values into an embeddings model. We also included novel features associated with wider determinants of health. Patient records comprising these embeddings were then fed to a Bidirectional Gated Recurrent Unit (GRU) model to predict the likelihood of patients developing Type 2 Diabetes Mellitus, Chronic Obstructive Pulmonary Disorder (COPD), Hypertension or experiencing an Acute Myocardial Infarction (MI) in the next 3 years. SHapley Additive exPlanations (SHAP) values were calculated to achieve model explainability.

Results

Increasing the data scope to include binned observations and wider determinants of health was found to improve predictive performance. We achieved an area under the Receiver Operating Characteristic curve value of 0.92 for Diabetes prediction, 0.94 for COPD, 0.92 for Hypertension and 0.94 for MI. The SHAP values showed that the models had learned features known to be associated with these outcomes.

Discussion

The DL approach outlined in this study can identify clinically-relevant features from large-scale EHR data and use these to predict future disease outcomes. This study highlights the promise of DL solutions for identifying patients at future risk of disease and providing clinicians with the means to understand and evaluate the drivers of those predictions.

Deep learning techniques in PET/CT imaging: A comprehensive review from sinogram to image space

Positron emission tomography/computed tomography (PET/CT) is increasingly used in oncology, neurology, cardiology, and emerging medical fields. The success stems from the cohesive information that hybrid PET/CT imaging offers, surpassing the capabilities of individual modalities when used in isolation for different malignancies. However, manual image interpretation requires extensive disease-specific knowledge, and it is a time-consuming aspect of physicians' daily routines. Deep learning algorithms, akin to a practitioner during training, extract knowledge from images to facilitate the diagnosis process by detecting symptoms and enhancing images. This acquired knowledge aids in supporting the diagnosis process through symptom detection and image enhancement. The available review papers on PET/CT imaging have a drawback as they either included additional modalities or examined various types of AI applications. However, there has been a lack of comprehensive investigation specifically focused on the highly specific use of AI, and deep learning, on PET/CT images. This review aims to fill that gap by investigating the characteristics of approaches used in papers that employed deep learning for PET/CT imaging. Within the review, we identified 99 studies published between 2017 and 2022 that applied deep learning to PET/CT images. We also identified the best pre-processing algorithms and the most effective deep learning models reported for PET/CT while highlighting the current limitations. Our review underscores the potential of deep learning (DL) in PET/CT imaging, with successful applications in lesion detection, tumor segmentation, and disease classification in both sinogram and image spaces. Common and specific pre-processing techniques are also discussed. DL algorithms excel at extracting meaningful features, and enhancing accuracy and efficiency in diagnosis. However, limitations arise from the scarcity of annotated datasets and challenges in explainability and uncertainty. Recent DL models, such as attention-based models, generative models, multi-modal models, graph convolutional networks, and transformers, are promising for improving PET/CT studies. Additionally, radiomics has garnered attention for tumor classification and predicting patient outcomes. Ongoing research is crucial to explore new applications and improve the accuracy of DL models in this rapidly evolving field.

Antimicrobial learning systems: an implementation blueprint for artificial intelligence to tackle antimicrobial resistance

The proliferation of various forms of artificial intelligence (AI) brings many opportunities to improve health care. AI models can harness complex evolving data, inform and augment human actions, and learn from health outcomes such as morbidity and mortality. The global public health challenge of antimicrobial resistance (AMR) needs large-scale optimisation of antimicrobial use and wider infection care, which could be enabled by carefully constructed AI models. As AI models become increasingly useful and robust, health-care systems remain challenging places for their deployment. An implementation gap exists between the promise of AI models and their use in patient and population care. Here, we outline an adaptive implementation and maintenance framework for AI models to improve antimicrobial use and infection care as a learning system. The roles of AMR problem identification, law and regulation, organisational support, data processing, and AI development, assessment, maintenance, and scalability in the implementation of AMR-targeted AI models are considered.

An Interventional Radiologist's Primer of Critical Appraisal of Artificial Intelligence Research

Recent advances in artificial intelligence (AI) are expected to cause a significant paradigm shift in all digital data-driven aspects of information gain, processing, and decision making in both clinical healthcare and medical research. The field of interventional radiology (IR) will be enmeshed in this innovation, yet the collective IR expertise in the field of AI remains rudimentary because of lack of training. This primer provides the clinical interventional radiologist with a simple guide for critically appraising AI research and products by identifying 12 fundamental items that should be considered: (a) need for AI technology to address the clinical problem, (b) type of applied Al algorithm, (c) data quality and degree of annotation, (d) reporting of accuracy, (e) applicability of standardized reporting, (f) reproducibility of methodology and data transparency, (g) algorithm validation, (h) interpretability, (i) concrete impact on IR, (j) pathway toward translation to clinical practice, (k) clinical benefit and cost-effectiveness, and (l) regulatory framework.

Enlightening the path to NSCLC biomarkers: Utilizing the power of XAI-guided deep learning

BACKGROUND AND OBJECTIVE

The early diagnosis of Non-small cell lung cancer (NSCLC) is of prime importance to improve the patient's survivability and quality of life. Being a heterogeneous disease at the molecular and cellular level, the biomarkers responsible for the heterogeneity aid in distinguishing NSCLC into its prominent subtypes-adenocarcinoma and squamous cell carcinoma. Moreover, if identified, these biomarkers could pave the path to targeted therapy. Through this work, a novel explainable AI (XAI)-guided deep learning framework is proposed that assists in discovering a set of significant NSCLC-relevant biomarkers using methylation data.

METHODS

The proposed framework is divided into two blocks- the first block combines an autoencoder and a neural network to classify NSCLC instances. The second block utilizes various eXplainable AI (XAI) methods, namely IntegratedGradients, GradientSHAP, and DeepLIFT, to discover a set of seven significant biomarkers.

RESULTS

The classification performance of the biomarkers discovered using the proposed framework is evaluated by employing multiple machine learning algorithms, among which the Multilayer Perceptron (MLP) algorithm-based model outperforms others, yielding a 10-fold cross-validation accuracy of 91.53%. An improved accuracy of 96.37% is achieved by integrating RNA-Seq, CNV, and methylation data. On performing statistical analysis using the Friedman and Nemenyi tests, the MLP model is found to be significantly better than other machine learning-based models. Further, the clinical efficacy of the resultant biomarkers is established based on their potential druggability, the likelihood of predicting NSCLC patients' survival, gene-disease association, and biological pathways targeted by them. While the biomarkers C18orf18, CCNT2, THOP1, and TNPO2, are found potentially druggable, the biomarkers CCDC15, SNORA9, THOP1, and TNPO2 are found prognostically relevant. On further analysis, some of the discovered biomarkers are found to be associated with around 104 diseases. Moreover, five KEGG, ten Reactome, and three Wiki pathways are found to be triggered by the biomarkers discovered.

CONCLUSION

In summary, the proposed framework uncovers a set of clinically effective biomarkers that accurately classify NSCLC. As a future course of work, efforts would be made to combine a variety of omics data with histopathological data to unveil more precise biomarkers for devising personalized therapy.

Shedding light on the black box of a neural network used to detect prostate cancer in whole slide images by occlusion-based explainability

Diagnostic histopathology faces increasing demands due to aging populations and expanding healthcare programs. Semi-automated diagnostic systems employing deep learning methods are one approach to alleviate this pressure. The learning models for histopathology are inherently complex and opaque from the user's perspective. Hence different methods have been developed to interpret their behavior. However, relatively limited attention has been devoted to the connection between interpretation methods and the knowledge of experienced pathologists. The main contribution of this paper is a method for comparing morphological patterns used by expert pathologists to detect cancer with the patterns identified as important for inference of learning models. Given the patch-based nature of processing large-scale histopathological imaging, we have been able to show statistically that the VGG16 model could utilize all the structures that are observable by the pathologist, given the patch size and scan resolution. The results show that the neural network approach to recognizing prostatic cancer is similar to that of a pathologist at medium optical resolution. The saliency maps identified several prevailing histomorphological features characterizing carcinoma, e.g., single-layered epithelium, small lumina, and hyperchromatic nuclei with halo. A convincing finding was the recognition of their mimickers in non-neoplastic tissue. The method can also identify differences, i.e., standard patterns not used by the learning models and new patterns not yet used by pathologists. Saliency maps provide added value for automated digital pathology to analyze and fine-tune deep learning systems and improve trust in computer-based decisions.

Reimagining Core Entrustable Professional Activities for Undergraduate Medical Education in the Era of Artificial Intelligence

The proliferation of generative artificial intelligence (AI) and its extensive potential for integration into many aspects of health care signal a transformational shift within the health care environment. In this context, medical education must evolve to ensure that medical trainees are adequately prepared to navigate the rapidly changing health care landscape. Medical education has moved toward a competency-based education paradigm, leading the Association of American Medical Colleges (AAMC) to define a set of Entrustable Professional Activities (EPAs) as its practical operational framework in undergraduate medical education. The AAMC's 13 core EPAs for entering residencies have been implemented with varying levels of success across medical schools. In this paper, we critically assess the existing core EPAs in the context of rapid AI integration in medicine. We identify EPAs that require refinement, redefinition, or comprehensive change to align with the emerging trends in health care. Moreover, this perspective proposes a set of "emerging" EPAs, informed by the changing landscape and capabilities presented by generative AI technologies. We provide a practical evaluation of the EPAs, alongside actionable recommendations on how medical education, viewed through the lens of the AAMC EPAs, can adapt and remain relevant amid rapid technological advancements. By leveraging the transformative potential of AI, we can reshape medical education to align with an AI-integrated future of medicine. This approach will help equip future health care professionals with technological competence and adaptive skills to meet the dynamic and evolving demands in health care.

Existing Barriers Faced by and Future Design Recommendations for Direct-to-Consumer Health Care Artificial Intelligence Apps: Scoping Review

BACKGROUND

Direct-to-consumer (DTC) health care artificial intelligence (AI) apps hold the potential to bridge the spatial and temporal disparities in health care resources, but they also come with individual and societal risks due to AI errors. Furthermore, the manner in which consumers interact directly with health care AI is reshaping traditional physician-patient relationships. However, the academic community lacks a systematic comprehension of the research overview for such apps.

OBJECTIVE

This paper systematically delineated and analyzed the characteristics of included studies, identified existing barriers and design recommendations for DTC health care AI apps mentioned in the literature and also provided a reference for future design and development.

METHODS

This scoping review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews guidelines and was conducted according to Arksey and O'Malley's 5-stage framework. Peer-reviewed papers on DTC health care AI apps published until March 27, 2023, in Web of Science, Scopus, the ACM Digital Library, IEEE Xplore, PubMed, and Google Scholar were included. The papers were analyzed using Braun and Clarke's reflective thematic analysis approach.

RESULTS

Of the 2898 papers retrieved, 32 (1.1%) covering this emerging field were included. The included papers were recently published (2018-2023), and most (23/32, 72%) were from developed countries. The medical field was mostly general practice (8/32, 25%). In terms of users and functionalities, some apps were designed solely for single-consumer groups (24/32, 75%), offering disease diagnosis (14/32, 44%), health self-management (8/32, 25%), and health care information inquiry (4/32, 13%). Other apps connected to physicians (5/32, 16%), family members (1/32, 3%), nursing staff (1/32, 3%), and health care departments (2/32, 6%), generally to alert these groups to abnormal conditions of consumer users. In addition, 8 barriers and 6 design recommendations related to DTC health care AI apps were identified. Some more subtle obstacles that are particularly worth noting and corresponding design recommendations in consumer-facing health care AI systems, including enhancing human-centered explainability, establishing calibrated trust and addressing overtrust, demonstrating empathy in AI, improving the specialization of consumer-grade products, and expanding the diversity of the test population, were further discussed.

CONCLUSIONS

The booming DTC health care AI apps present both risks and opportunities, which highlights the need to explore their current status. This paper systematically summarized and sorted the characteristics of the included studies, identified existing barriers faced by, and made future design recommendations for such apps. To the best of our knowledge, this is the first study to systematically summarize and categorize academic research on these apps. Future studies conducting the design and development of such systems could refer to the results of this study, which is crucial to improve the health care services provided by DTC health care AI apps.

New Horizons in artificial intelligence in the healthcare of older people

Artificial intelligence (AI) in healthcare describes algorithm-based computational techniques which manage and analyse large datasets to make inferences and predictions. There are many potential applications of AI in the care of older people, from clinical decision support systems that can support identification of delirium from clinical records to wearable devices that can predict the risk of a fall. We held four meetings of older people, clinicians and AI researchers. Three priority areas were identified for AI application in the care of older people. These included: monitoring and early diagnosis of disease, stratified care and care coordination between healthcare providers. However, the meetings also highlighted concerns that AI may exacerbate health inequity for older people through bias within AI models, lack of external validation amongst older people, infringements on privacy and autonomy, insufficient transparency of AI models and lack of safeguarding for errors. Creating effective interventions for older people requires a person-centred approach to account for the needs of older people, as well as sufficient clinical and technological governance to meet standards of generalisability, transparency and effectiveness. Education of clinicians and patients is also needed to ensure appropriate use of AI technologies, with investment in technological infrastructure required to ensure equity of access.

PS-VTS: particle swarm with visit table strategy for automated emotion recognition with EEG signals

Recognizing emotions accurately in real life is crucial in human-computer interaction (HCI) systems. Electroencephalogram (EEG) signals have been extensively employed to identify emotions. The researchers have used several EEG-based emotion identification datasets to validate their proposed models. In this paper, we have employed a novel metaheuristic optimization approach for accurate emotion classification by applying it to select both channel and rhythm of EEG data. In this work, we have proposed the particle swarm with visit table strategy (PS-VTS) metaheuristic technique to improve the effectiveness of EEG-based human emotion identification. First, the EEG signals are denoised using a low pass filter, and then rhythm extraction is done using discrete wavelet transform (DWT). The continuous wavelet transform (CWT) approach transforms each rhythm signal into a rhythm image. The pre-trained MobilNetv2 model has been pre-trained for deep feature extraction, and a support vector machine (SVM) is used to classify the emotions. Two models are developed for optimal channels and rhythm sets. In Model 1, optimal channels are selected separately for each rhythm, and global optima are determined in the optimization process according to the best channel sets of the rhythms. The best rhythms are first determined for each channel, and then the optimal channel-rhythm set is selected in Model 2. Our proposed model obtained an accuracy of 99.2871% and 97.8571% for the classification of HA (high arousal)-LA (low arousal) and HV (high valence)-LV (low valence), respectively with the DEAP dataset. Our generated model obtained the highest classification accuracy compared to the previously reported methods.

Explainable machine learning for profiling the immunological synapse and functional characterization of therapeutic antibodies

Therapeutic antibodies are widely used to treat severe diseases. Most of them alter immune cells and act within the immunological synapse; an essential cell-to-cell interaction to direct the humoral immune response. Although many antibody designs are generated and evaluated, a high-throughput tool for systematic antibody characterization and prediction of function is lacking. Here, we introduce the first comprehensive open-source framework, scifAI (single-cell imaging flow cytometry AI), for preprocessing, feature engineering, and explainable, predictive machine learning on imaging flow cytometry (IFC) data. Additionally, we generate the largest publicly available IFC dataset of the human immunological synapse containing over 2.8 million images. Using scifAI, we analyze class frequency and morphological changes under different immune stimulation. T cell cytokine production across multiple donors and therapeutic antibodies is quantitatively predicted in vitro, linking morphological features with function and demonstrating the potential to significantly impact antibody design. scifAI is universally applicable to IFC data. Given its modular architecture, it is straightforward to incorporate into existing workflows and analysis pipelines, e.g., for rapid antibody screening and functional characterization.

The promise of explainable deep learning for omics data analysis: Adding new discovery tools to AI

Deep learning has already revolutionised the way a wide range of data is processed in many areas of daily life. The ability to learn abstractions and relationships from heterogeneous data has provided impressively accurate prediction and classification tools to handle increasingly big datasets. This has a significant impact on the growing wealth of omics datasets, with the unprecedented opportunity for a better understanding of the complexity of living organisms. While this revolution is transforming the way these data are analyzed, explainable deep learning is emerging as an additional tool with the potential to change the way biological data is interpreted. Explainability addresses critical issues such as transparency, so important when computational tools are introduced especially in clinical environments. Moreover, it empowers artificial intelligence with the capability to provide new insights into the input data, thus adding an element of discovery to these already powerful resources. In this review, we provide an overview of the transformative effects explainable deep learning is having on multiple sectors, ranging from genome engineering and genomics, from radiomics to drug design and clinical trials. We offer a perspective to life scientists, to better understand the potential of these tools, and a motivation to implement them in their research, by suggesting learning resources they can use to move their first steps in this field.

An explainability framework for deep learning on chemical reactions exemplified by enzyme-catalysed reaction classification

Assigning or proposing a catalysing enzyme given a chemical or biochemical reaction is of great interest to life sciences and chemistry alike. The exploration and design of metabolic pathways and the challenge of finding more sustainable enzyme-catalysed alternatives to traditional organic reactions are just two examples of tasks that require an association between reaction and enzyme. However, given the lack of large and balanced annotated data sets of enzyme-catalysed reactions, assigning an enzyme to a reaction still relies on expert-curated rules and databases. Here, we present a data-driven explainable human-in-the-loop machine learning approach to support and ultimately automate the association of a catalysing enzyme with a given biochemical reaction. In addition, the proposed method is capable of predicting enzymes as candidate catalysts for organic reactions amendable to biocatalysis. Finally, the introduced explainability and visualisation methods can easily be generalised to support other machine-learning approaches involving chemical and biochemical reactions.

Artificial Cognition for Detection of Mental Disability: A Vision Transformer Approach for Alzheimer's Disease

Alzheimer's disease is a common neurological disorder and mental disability that causes memory loss and cognitive decline, presenting a major challenge to public health due to its impact on millions of individuals worldwide. It is crucial to diagnose and treat Alzheimer's in a timely manner to improve the quality of life of both patients and caregivers. In the recent past, machine learning techniques have showed potential in detecting Alzheimer's disease by examining neuroimaging data, especially Magnetic Resonance Imaging (MRI). This research proposes an attention-based mechanism that employs the vision transformer approach to detect Alzheimer's using MRI images. The presented technique applies preprocessing to the MRI images and forwards them to a vision transformer network for classification. This network is trained on the publicly available Kaggle dataset, and it illustrated impressive results with an accuracy of 99.06%, precision of 99.06%, recall of 99.14%, and F1-score of 99.1%. Furthermore, a comparative study is also conducted to evaluate the performance of the proposed method against various state-of-the-art techniques on diverse datasets. The proposed method demonstrated superior performance, outperforming other published methods when applied to the Kaggle dataset.

Explainable artificial intelligence to predict and identify prostate cancer tissue by gene expression

BACKGROUND AND OBJECTIVE

Prostate cancer is one of the most prevalent forms of cancer in men worldwide. Traditional screening strategies such as serum PSA levels, which are not necessarily cancer-specific, or digital rectal exams, which are often inconclusive, are still the screening methods used for the disease. Some studies have focused on identifying biomarkers of the disease but none have been reported for diagnosis in routine clinical practice and few studies have provided tools to assist the pathologist in the decision-making process when analyzing prostate tissue. Therefore, a classifier is proposed to predict the occurrence of PCa that provides physicians with accurate predictions and understandable explanations.

METHODS

A selection of 47 genes was made based on differential expression between PCa and normal tissue, GO gene ontology as well as the literature to be used as input predictors for different machine learning methods based on eXplainable Artificial Intelligence. These methods were trained using different class-balancing strategies to build accurate classifiers using gene expression data from 550 samples from 'The Cancer Genome Atlas'. Our model was validated in four external cohorts with different ancestries, totaling 463 samples. In addition, a set of SHapley Additive exPlanations was provided to help clinicians understand the underlying reasons for each decision.

RESULTS

An in-depth analysis showed that the Random Forest algorithm combined with majority class downsampling was the best performing approach with robust statistical significance. Our method achieved an average sensitivity and specificity of 0.90 and 0.8 with an AUC of 0.84 across all databases. The relevance of DLX1, MYL9 and FGFR genes for PCa screening was demonstrated in addition to the important role of novel genes such as CAV2 and MYLK.

CONCLUSIONS

This model has shown good performance in 4 independent external cohorts of different ancestries and the explanations provided are consistent with each other and with the literature, opening a horizon for its application in clinical practice. In the near future, these genes, in combination with our model, could be applied to liquid biopsy to improve PCa screening.

Machine Learning Algorithm Predicts Mortality Risk in Intensive Care Unit for Patients with Traumatic Brain Injury

BACKGROUND

Numerous mortality prediction tools are currently available to assist patients with moderate to severe traumatic brain injury (TBI). However, an algorithm that utilizes various machine learning methods and employs diverse combinations of features to identify the most suitable predicting outcomes of brain injury patients in the intensive care unit (ICU) has not yet been well-established.

METHOD

Between January 2016 and December 2021, we retrospectively collected data from the electronic medical records of Chi Mei Medical Center, comprising 2260 TBI patients admitted to the ICU. A total of 42 features were incorporated into the analysis using four different machine learning models, which were then segmented into various feature combinations. The predictive performance was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) curve and validated using the Delong test.

RESULT

The AUC for each model under different feature combinations ranged from 0.877 (logistic regression with 14 features) to 0.921 (random forest with 22 features). The Delong test indicated that the predictive performance of the machine learning models is better than that of traditional tools such as APACHE II and SOFA scores.

CONCLUSION

Our machine learning training demonstrated that the predictive accuracy of the LightGBM is better than that of APACHE II and SOFA scores. These features are readily available on the first day of patient admission to the ICU. By integrating this model into the clinical platform, we can offer clinicians an immediate prognosis for the patient, thereby establishing a bridge for educating and communicating with family members.

A historical perspective of biomedical explainable AI research

The black-box nature of most artificial intelligence (AI) models encourages the development of explainability methods to engender trust into the AI decision-making process. Such methods can be broadly categorized into two main types: post hoc explanations and inherently interpretable algorithms. We aimed at analyzing the possible associations between COVID-19 and the push of explainable AI (XAI) to the forefront of biomedical research. We automatically extracted from the PubMed database biomedical XAI studies related to concepts of causality or explainability and manually labeled 1,603 papers with respect to XAI categories. To compare the trends pre- and post-COVID-19, we fit a change point detection model and evaluated significant changes in publication rates. We show that the advent of COVID-19 in the beginning of 2020 could be the driving factor behind an increased focus concerning XAI, playing a crucial role in accelerating an already evolving trend. Finally, we present a discussion with future societal use and impact of XAI technologies and potential future directions for those who pursue fostering clinical trust with interpretable machine learning models.

INSOMNet: Automated insomnia detection using scalogram and deep neural networks with ECG signals

Sleep is a natural state of rest for the body and mind. It is essential for a human's physical and mental health because it helps the body restore itself. Insomnia is a sleep disorder that causes difficulty falling asleep or staying asleep and can lead to several health problems. Conventional sleep monitoring and insomnia detection systems are expensive, laborious, and time-consuming. This is the first study that integrates an electrocardiogram (ECG) scalogram with a convolutional neural network (CNN) to develop a model for the accurate measurement of the quality of sleep in identifying insomnia. Continuous wavelet transform has been employed to convert 1-D time-domain ECG signals into 2-D scalograms. Obtained scalograms are fed to AlexNet, MobileNetV2, VGG16, and newly developed CNN for automated detection of insomnia. The proposed INSOMNet system is validated on the cyclic alternating pattern (CAP) and sleep disorder research center (SDRC) datasets. Six performance measures, accuracy (ACC), false omission rate (FOR), sensitivity (SEN), false discovery rate (FDR), specificity (SPE), and threat score (TS), have been calculated to evaluate the developed model. Our developed system attained the classifications ACC of 98.91%, 98.68%, FOR of 1.5, 0.66, SEN of 98.94%, 99.31%, FDR of 0.80, 2.00, SPE of 98.87%, 98.08%, and TS 0.98, 0.97 on CAP and SDRC datasets, respectively. The developed model is less complex and more accurate than transfer-learning networks. The prototype is ready to be tested with a huge dataset from diverse centers.

An eXplainability Artificial Intelligence approach to brain connectivity in Alzheimer's disease

The advent of eXplainable Artificial Intelligence (XAI) has revolutionized the way human experts, especially from non-computational domains, approach artificial intelligence; this is particularly true for clinical applications where the transparency of the results is often compromised by the algorithmic complexity. Here, we investigate how Alzheimer's disease (AD) affects brain connectivity within a cohort of 432 subjects whose T1 brain Magnetic Resonance Imaging data (MRI) were acquired within the Alzheimer's Disease Neuroimaging Initiative (ADNI). In particular, the cohort included 92 patients with AD, 126 normal controls (NC) and 214 subjects with mild cognitive impairment (MCI). We show how graph theory-based models can accurately distinguish these clinical conditions and how Shapley values, borrowed from game theory, can be adopted to make these models intelligible and easy to interpret. Explainability analyses outline the role played by regions like putamen, middle and superior temporal gyrus; from a class-related perspective, it is possible to outline specific regions, such as hippocampus and amygdala for AD and posterior cingulate and precuneus for MCI. The approach is general and could be adopted to outline how brain connectivity affects specific brain regions.

Artificial intelligence and machine learning in prehospital emergency care: A scoping review

Our scoping review provides a comprehensive analysis of the landscape of artificial intelligence (AI) applications in prehospital emergency care (PEC). It contributes to the field by highlighting the most studied AI applications and identifying the most common methodological approaches across 106 included studies. The findings indicate a promising future for AI in PEC, with many unique use cases, such as prognostication, demand prediction, resource optimization, and the Internet of Things continuous monitoring systems. Comparisons with other approaches showed AI outperforming clinicians and non-AI algorithms in most cases. However, most studies were internally validated and retrospective, highlighting the need for rigorous prospective validation of AI applications before implementation in clinical settings. We identified knowledge and methodological gaps using an evidence map, offering a roadmap for future investigators. We also discussed the significance of explainable AI for establishing trust in AI systems among clinicians and facilitating real-world validation of AI models.

LSGP-USFNet: Automated Attention Deficit Hyperactivity Disorder Detection Using Locations of Sophie Germain's Primes on Ulam's Spiral-Based Features with Electroencephalogram Signals

Anxiety, learning disabilities, and depression are the symptoms of attention deficit hyperactivity disorder (ADHD), an isogenous pattern of hyperactivity, impulsivity, and inattention. For the early diagnosis of ADHD, electroencephalogram (EEG) signals are widely used. However, the direct analysis of an EEG is highly challenging as it is time-consuming, nonlinear, and nonstationary in nature. Thus, in this paper, a novel approach (LSGP-USFNet) is developed based on the patterns obtained from Ulam's spiral and Sophia Germain's prime numbers. The EEG signals are initially filtered to remove the noise and segmented with a non-overlapping sliding window of a length of 512 samples. Then, a time-frequency analysis approach, namely continuous wavelet transform, is applied to each channel of the segmented EEG signal to interpret it in the time and frequency domain. The obtained time-frequency representation is saved as a time-frequency image, and a non-overlapping n × n sliding window is applied to this image for patch extraction. An n × n Ulam's spiral is localized on each patch, and the gray levels are acquired from this patch as features where Sophie Germain's primes are located in Ulam's spiral. All gray tones from all patches are concatenated to construct the features for ADHD and normal classes. A gray tone selection algorithm, namely ReliefF, is employed on the representative features to acquire the final most important gray tones. The support vector machine classifier is used with a 10-fold cross-validation criteria. Our proposed approach, LSGP-USFNet, was developed using a publicly available dataset and obtained an accuracy of 97.46% in detecting ADHD automatically. Our generated model is ready to be validated using a bigger database and it can also be used to detect other children's neurological disorders.

Automated Urine Cell Image Classification Model Using Chaotic Mixer Deep Feature Extraction

Microscopic examination of urinary sediments is a common laboratory procedure. Automated image-based classification of urinary sediments can reduce analysis time and costs. Inspired by cryptographic mixing protocols and computer vision, we developed an image classification model that combines a novel Arnold Cat Map (ACM)- and fixed-size patch-based mixer algorithm with transfer learning for deep feature extraction. Our study dataset comprised 6,687 urinary sediment images belonging to seven classes: Cast, Crystal, Epithelia, Epithelial nuclei, Erythrocyte, Leukocyte, and Mycete. The developed model consists of four layers: (1) an ACM-based mixer to generate mixed images from resized 224 × 224 input images using fixed-size 16 × 16 patches; (2) DenseNet201 pre-trained on ImageNet1K to extract 1,920 features from each raw input image, and its six corresponding mixed images were concatenated to form a final feature vector of length 13,440; (3) iterative neighborhood component analysis to select the most discriminative feature vector of optimal length 342, determined using a k-nearest neighbor (kNN)-based loss function calculator; and (4) shallow kNN-based classification with ten-fold cross-validation. Our model achieved 98.52% overall accuracy for seven-class classification, outperforming published models for urinary cell and sediment analysis. We demonstrated the feasibility and accuracy of deep feature engineering using an ACM-based mixer algorithm for image preprocessing combined with pre-trained DenseNet201 for feature extraction. The classification model was both demonstrably accurate and computationally lightweight, making it ready for implementation in real-world image-based urine sediment analysis applications.