Medical Intelligence Research Articles

Cybersecurity Framework for Synthetic Data in Training Medical AI

Abstract The development of medical artificial intelligence is dependent on the availability of vast quantities of data, a considerable proportion of which is medical data containing sensitive information pertaining to the health and well-being of patients. The use of such data is subject to extensive legal regulation and is further hindered by financial and organisational constraints, which can result in limitations on accessibility. One potential solution to this problem is the use of synthetic data. This article examines the potential for their use in light of cybersecurity requirements derived from horizontal and sectoral EU legislation. The outcome of this analysis is that EU legislation does not contain specific regulations on the use of synthetic data. Consequently, it cannot be concluded that there is any prohibition on their use. Moreover, while the Medical Device Regulation (MDR) contains some general requirements for cybersecurity, these are further specified by the provisions of the AI Act. It is important to note, however, that the AI Act will not apply to Class I medical devices, which are subject only to the MDR. Furthermore, only indirect obligations within the scope under consideration can be derived from the horizontal regulations, which will apply in a limited number of cases.

Patient acceptance of medical service robots in the medical intelligence era: an empirical study based on an extended AI device use acceptance model

In the era of medical intelligence, there are still few studies focusing on medical service robots from a user experience perspective. Guided by the model of artificial intelligence (AI) device use acceptance (AIDUA), this article develops a theoretical model to explain patients’ intention to use medical service robots at hospitals. The proposed model specifically distinguished the dimensions of anthropomorphic attributes of service robots and further introduces two variables, perceived empathy and interaction experience, as a way to construct a three-stage psychological mechanism for patient acceptance of robots. 400 questionnaires from Chinese patients were collected offline and analyzed using structural equation modeling (SEM). The results revealed that the four attributes of anthropomorphism play a differential influence in performance expectations and effort expectations, but all positively contribute to empathy, which in turn positively affects interaction quality. Interaction quality, performance expectations, and effort expectations all influence patients’ emotions, thus having an impact on patients’ intentions to accept service robots. The findings from this study will assist the healthcare sector in upgrading medical service robots, which will improve patient acceptance of these robots.

Medical technologies, telemedicine and artificial intelligence for neurotrauma and neurorehabilitation.

Medical technologies, telemedicine and artificial intelligence for neurotrauma and neurorehabilitation.

Integrating large language models in mental health practice: a qualitative descriptive study based on expert interviews

BackgroundProgress in developing artificial intelligence (AI) products represented by large language models (LLMs) such as OpenAI’s ChatGPT has sparked enthusiasm for their potential use in mental health practice. However, the perspectives on the integration of LLMs within mental health practice remain an underreported topic. Therefore, this study aimed to explore how mental health and AI experts conceptualize LLMs and perceive the use of integrating LLMs into mental health practice.MethodIn February–April 2024, online semi-structured interviews were conducted with 21 experts (12 psychiatrists, 7 mental health nurses, 2 researchers in medical artificial intelligence) from four provinces in China, using snowballing and purposive selection sampling. Respondents’ discussions about their perspectives and expectations of integrating LLMs in mental health were analyzed with conventional content analysis.ResultsFour themes and eleven sub-themes emerged from this study. Firstly, participants discussed the (1) practice and application reform brought by LLMs into mental health (fair access to mental health services, enhancement of patient participation, improvement in work efficiency and quality), and then analyzed the (2) technological-mental health gap (misleading information, lack of professional nuance and depth, user risk). Based on these points, they provided a range of (3) prerequisites for the integration of LLMs in mental health (training and competence, guidelines for use and management, patient engagement and transparency) and expressed their (4) expectations for future developments (reasonable allocation of workload, upgrades and revamps of LLMs).ConclusionThese findings provide valuable insights into integrating LLMs within mental health practice, offering critical guidance for institutions to effectively implement, manage, and optimize these tools, thereby enhancing the quality and accessibility of mental health services.

Biomedical Engineering and Measurement Technology

Abstract The intersection of biomedical engineering and measurement technology has transformed the medical landscape, offering innovative solutions for diagnostics, treatment, and patient monitoring. Biomedical engineering encompasses a broad spectrum of applications, from molecular biology and tissue engineering to medical robotics and wearable technologies. Measurement technology, meanwhile, enables precise and real-time data collection through advanced sensing techniques, ensuring that these biomedical systems operate at optimal efficiency. In recent years, significant advancements in medical sensors, imaging technologies, and artificial intelligence have paved the way for improved accuracy and performance in medical diagnostics and treatment. This special issue focuses on the latest developments in these fields, showcasing groundbreaking research that enhances our understanding of physiological phenomena and revolutionizes medical measurement applications. Submissions for this topic in Measurement Science and Technology are open for submission from 03 November 2022 to 31 July 2023 and contains 21 outstanding papers in above research fields.

Bias in medical AI: Implications for clinical decision-making.

Biases in medical artificial intelligence (AI) arise and compound throughout the AI lifecycle. These biases can have significant clinical consequences, especially in applications that involve clinical decision-making. Left unaddressed, biased medical AI can lead to substandard clinical decisions and the perpetuation and exacerbation of longstanding healthcare disparities. We discuss potential biases that can arise at different stages in the AI development pipeline and how they can affect AI algorithms and clinical decision-making. Bias can occur in data features and labels, model development and evaluation, deployment, and publication. Insufficient sample sizes for certain patient groups can result in suboptimal performance, algorithm underestimation, and clinically unmeaningful predictions. Missing patient findings can also produce biased model behavior, including capturable but nonrandomly missing data, such as diagnosis codes, and data that is not usually or not easily captured, such as social determinants of health. Expertly annotated labels used to train supervised learning models may reflect implicit cognitive biases or substandard care practices. Overreliance on performance metrics during model development may obscure bias and diminish a model's clinical utility. When applied to data outside the training cohort, model performance can deteriorate from previous validation and can do so differentially across subgroups. How end users interact with deployed solutions can introduce bias. Finally, where models are developed and published, and by whom, impacts the trajectories and priorities of future medical AI development. Solutions to mitigate bias must be implemented with care, which include the collection of large and diverse data sets, statistical debiasing methods, thorough model evaluation, emphasis on model interpretability, and standardized bias reporting and transparency requirements. Prior to real-world implementation in clinical settings, rigorous validation through clinical trials is critical to demonstrate unbiased application. Addressing biases across model development stages is crucial for ensuring all patients benefit equitably from the future of medical AI.

Medical image segmentation network based on feature filtering with low number of parameters

In recent years, the medical image segmentation method based on hybrid convolutional neural network (CNN) and Vision Transformer (ViT) has made great progress, but it still faces the challenge of unbalanced global and local modeling, and excessive parameters. In addition, ViT repeatedly uses the whole feature map to model the global information, thus generating irrelevant and weakly related information, which will weaken the performance of the model when facing small datasets and segmentation targets. Therefore, this paper proposes a feature screening network based on similarity, named Screening Feature (SF)-MixedNet. Specifically, this paper first proposes a new feature extractor, namely Correlation based Similarity Transformer (CSimFormer). On the basis of parameter pruning, it uses the Screening Feature Multi-head Self Attention (SF-MSA) to establish the remote dependency, and calculates the similarity between local elements through the Location-Sensitive Mechanism (LsM) to obtain the weight matrix. Then, the correlation between regional elements is mined by Region Matching and Selection (RMS) mechanism, and the obtained information is filtered according to the corresponding rules to reduce the side effects of redundant information. Extensive experiments on Synapse dataset, ACDC dataset and SegPC-2021 dataset show that the segmentation accuracy reaches 83.51%, 92.20% and 81.27% respectively. Especially in the Synapse dataset, our method is 6.31% higher than the baseline. The method proposed in this paper effectively improves the segmentation accuracy, provides more detailed information for medical diagnosis and promotes the development of medical artificial intelligence technology.

Establishing Medical Intelligence-Leveraging Fast Healthcare Interoperability Resources to Improve Clinical Management: Retrospective Cohort and Clinical Implementation Study.

FHIR (Fast Healthcare Interoperability Resources) has been proposed to enable health data interoperability. So far, its applicability has been demonstrated for selected research projects with limited data. This study aimed to design and implement a conceptual medical intelligence framework to leverage real-world care data for clinical decision-making. A Python package for the use of multimodal FHIR data (FHIRPACK [FHIR Python Analysis Conversion Kit]) was developed and pioneered in 5 real-world clinical use cases, that is, myocardial infarction, stroke, diabetes, sepsis, and prostate cancer. Patients were identified based on the ICD-10 (International Classification of Diseases, Tenth Revision) codes, and outcomes were derived from laboratory tests, prescriptions, procedures, and diagnostic reports. Results were provided as browser-based dashboards. For 2022, a total of 1,302,988 patient encounters were analyzed. (1) Myocardial infarction: in 72.7% (261/359) of cases, medication regimens fulfilled guideline recommendations. (2) Stroke: out of 1277 patients, 165 received thrombolysis and 108 thrombectomy. (3) Diabetes: in 443,866 serum glucose and 16,180 glycated hemoglobin A1c measurements from 35,494 unique patients, the prevalence of dysglycemic findings was 39% (13,887/35,494). Among those with dysglycemia, diagnosis was coded in 44.2% (6138/13,887) of the patients. (4) Sepsis: In 1803 patients, Staphylococcus epidermidis was the primarily isolated pathogen (773/2672, 28.9%) and piperacillin and tazobactam was the primarily prescribed antibiotic (593/1593, 37.2%). (5) PC: out of 54, three patients who received radical prostatectomy were identified as cases with prostate-specific antigen persistence or biochemical recurrence. Leveraging FHIR data through large-scale analytics can enhance health care quality and improve patient outcomes across 5 clinical specialties. We identified (1) patients with sepsis requiring less broad antibiotic therapy, (2) patients with myocardial infarction who could benefit from statin and antiplatelet therapy, (3) patients who had a stroke with longer than recommended times to intervention, (4) patients with hyperglycemia who could benefit from specialist referral, and (5) patients with PC with early increases in cancer markers.

Parkinson’s Disease Detection by Using Machine Learning Method based on Local Classification on Class Boundary

Parkinson’s disease (PD) detection has long been an important task in medical intelligence. Recognition methods based on speech signals show great potential in Parkinson’s disease diagnosis. In this paper, based on an efficient machine learning method for Parkinson’s disease detection, we take the use of test data incorporates an efficient Secure Two-Party Computing (S2PC) protocol to protect the privacy of patients. We present two key components, the secure use of data and a local classification methodology, including the description of class boundaries. We conducted experiments on two datasets to validate our proposed method, and the results show well data security protection ability compared to some more sophisticated methods. And the performance of Local Classification on Class Boundary(LCCB) and Hyperplane K-Nearest Neighbor(HKNN) is significantly better than that of both Support Vector Machines(SVM) and Random Forest(RF). When the number of selected features is from 400 to 500, HKNN and LCCB are roughly equal where the accuracy of HKNN is 95.2%, and LCCB has the rate of 94.7%. Then we use Multi-Cluster Feature Selection(MCFS) to analyze and select the important features from D2 dataset. It shows that even if only two features are selected, the boundaries of the two categories are also clear and easy to distinguish.

Fairness in artificial intelligence‐driven multi‐organ image segmentation

AbstractFairness is an emerging consideration when assessing the segmentation performance of machine learning models across various demographic groups. During clinical decision‐making, an unfair segmentation model exhibits risks in that it can pose inappropriate diagnoses and unsuitable treatment plans for underrepresented demographic groups, resulting in severe consequences for patients and society. In medical artificial intelligence (AI), the fairness of multi‐organ segmentation is imperative to augment the integration of models into clinical practice. As the use of multi‐organ segmentation in medical image analysis expands, it is crucial to systematically examine fairness to ensure equitable segmentation performance across diverse patient populations and ensure health equity. However, comprehensive studies assessing the problem of fairness in multi‐organ segmentation remain lacking. This study aimed to provide an overview of the fairness problem in multi‐organ segmentation. We first define fairness and discuss the factors that lead to fairness problems such as individual fairness, group fairness, counterfactual fairness, and max–min fairness in multi‐organ segmentation, focusing mainly on datasets and models. We then present strategies to potentially improve fairness in multi‐organ segmentation. Additionally, we highlight the challenges and limitations of existing approaches and discuss future directions for improving the fairness of AI models for clinically oriented multi‐organ segmentation.

Resistance to medical artificial intelligence: Integrating AI awareness, AI risks, and displacement of responsibility

Resistance to medical artificial intelligence: Integrating AI awareness, AI risks, and displacement of responsibility

Assessment of communication skills in health professions education; Ottawa 2024 consensus statement

Despite the increasing inclusion of communication skills in accreditation standards and an increase in time dedicated to teaching these skills, communication is often regarded as a separate skill and is therefore, not consistently represented in overall systems of assessment in Health Professions Education (HPE). The ascendence of competency-based medical education, programmatic assessment, artificial intelligence, and widespread use of telehealth, alongside changing patient expectations warrant an update in thinking about the assessment of communication skills in health professions education. This consensus statement draws on existing literature, expert pinion, and emerging challenges to situate the assessment of communication skills in the contemporary health professions education context. The statement builds on previous work to offer an update on the topic and include new developments related to assessment, particularly: the challenges and opportunities associated with systems of assessment; patient and peer perspectives in assessment; assessment of interprofessional communication, cross-cultural communication, digital communication; and assessment using digital technologies. Consensus was reached through extensive discussion among the authors and other experts in HPE, exploration of the literature, and discussion during an Ottawa 2024 conference workshop. The statement puts forward a summary of available evidence with suggestions for what educators and curriculum developers should consider in their planning and design of the assessment of communication.

PrescDRL: deep reinforcement learning for herbal prescription planning in treatment of chronic diseases

Treatment planning for chronic diseases is a critical task in medical artificial intelligence, particularly in traditional Chinese medicine (TCM). However, generating optimized sequential treatment strategies for patients with chronic diseases in different clinical encounters remains a challenging issue that requires further exploration. In this study, we proposed a TCM herbal prescription planning framework based on deep reinforcement learning for chronic disease treatment (PrescDRL). PrescDRL is a sequential herbal prescription optimization model that focuses on long-term effectiveness rather than achieving maximum reward at every step, thereby ensuring better patient outcomes. We constructed a high-quality benchmark dataset for sequential diagnosis and treatment of diabetes and evaluated PrescDRL against this benchmark. Our results showed that PrescDRL achieved a higher curative effect, with the single-step reward improving by 117% and 153% compared to doctors. Furthermore, PrescDRL outperformed the benchmark in prescription prediction, with precision improving by 40.5% and recall improving by 63%. Overall, our study demonstrates the potential of using artificial intelligence to improve clinical intelligent diagnosis and treatment in TCM.

Problematic aspects of medical artificial intelligence. Part 2

The capabilities of artificial intelligence (AI) and machine learning are growing at an unprecedented pace. These technologies have many useful applications, from machine translation to medical image analysis. A large number of such applications are currently being developed, and an increasing number of such applications is expected in the long term. Unfortunately, weaknesses and other unpleasant aspects of AI have received insufficient attention. In this review, we consider a whole range of already known problems and possible risks associated with the use of innovative neural network technologies, paying special attention to the ways of preventing real dangers and potential threats in order to expand the range of stakeholders and profile experts involved in the discussion of current issues of medical AI cybersecurity, formation of responsible approach to the vulnerabilities of neural network platforms, increasing the reliability of equipment protection for its safe use, as well as the importance of legal and ethical aspects of regulating the use of AI. Despite certain challenges described in our review, it is clear that AI will be an important element of the healthcare future. As the population continues to age and the demand for healthcare services grows, neural networks are expected to drive healthcare very soon, especially in the areas of medical image analysis, virtual assistants, drug development, treatment recommendations and patients' data processing. We would like to emphasize that while we recognize the innovative role that digital technologies and AI can and should play in strengthening the domestic healthcare system, we should not overlook the importance of timely and accurate assessment of their beneficial or negative impact on the industry to ensure such management decisions do not unnecessarily divert our attention and resources from non-digital approaches and research. This article is a continuation of the article by Berdutin VA, Romanova TE, Romanov SV, Abaeva OP. Problematic aspects of medical artificial intelligence. Part 1. Sociology of Medicine. 2023;22(2):202–211. DOI: https://doi.org/10.17816/socm619132

Innovations in ultrasound training in obstetrics.

Ultrasound technology is critical in obstetrics, enabling detailed examination of the fetus and maternal anatomy. However, increasing complexity demands specialised training to maximise its potential. This study explores innovative approaches to ultrasound training in obstetrics, focussing on enhancing diagnostic skills and patient safety. This review examines recent innovations in ultrasound training, including competency-based medical education (CBME), simulation technologies, technology-based resources, artificial intelligence (AI), and online-learning platforms. Traditional training methods such as theoretical learning, practical experience, and peer learning are also discussed to provide a comprehensive view of current practises. Innovations in ultrasound training include the use of high-fidelity simulators, virtual reality (VR), augmented reality (AR), and hybrid-learning platforms. Simulation technologies offer reproducibility, risk-free learning, diverse scenarios, and immediate feedback. AI and machine learning facilitate personalised-learning paths, real-time feedback, and automated-image analysis. Online-learning platforms and e-learning methods provide flexible, accessible, and cost-effective education. Gamification enhances learning motivation and engagement through educational games and virtual competitions. The integration of innovative technologies in ultrasound training significantly improves diagnostic skills, learner confidence, and patient safety. However, challenges such as high costs, the need for comprehensive instructor training, and integration into existing programs must be addressed. Standardisation and certification ensure high-quality and consistent training. Future developments in AI, VR, and 3D printing promise further advancements in ultrasound education. Innovations in ultrasound training in obstetrics offer significant improvements in medical education and patient care. The successful implementation and continuous development of these technologies are crucial to meet the growing demands of modern obstetrics.

The role of artificial intelligence in optimizing management of atrial fibrillation in acute ischemic stroke.

Atrial fibrillation (AF) is a severe condition associated with high morbidity and mortality, including an increased risk of stroke and poor outcomes poststroke. Our understanding of the prognosis in AF remains poor. Machine learning (ML) has been applied to the diagnosis, management, and prognosis of AF in the context of stroke but remains suboptimal for clinical use. This article endeavors to provide a comprehensive overview of current ML applications to AF patients at risk of stroke, as well as poststroke patients without AF. Strategies to develop effective ML involve the validation of a variety of ML algorithms across internal and external datasets as well as exploring their predictive powers in hypothetical and realistic settings. Recent literature of this rapidly evolving field has displayed much promise. However, further testing and innovation of medical artificial intelligence are required before its imminent introduction to ensure complete patient trust within the community. Prioritizing this research is imperative for advancing the optimization of ongoing care for AF patients, as well as the management of stroke patients with AF.

Accuracy of ChatGPT responses on tracheotomy for patient education.

To investigate the accuracy of information provided by ChatGPT-4o to patients about tracheotomy. Twenty common questions of patients about tracheotomy were presented to ChatGPT-4o twice (7-day intervals). The accuracy, clarity, relevance, completeness, referencing, and usefulness of responses were assessed by a board-certified otolaryngologist and a board-certified intensive care unit practitioner with the Quality Analysis of Medical Artificial Intelligence (QAMAI) tool. The interrater reliability and the stability of the ChatGPT-4o responses were evaluated with intraclass correlation coefficient (ICC) and Pearson correlation analysis. The total scores of QAMAI were 22.85 ± 4.75 for the intensive care practitioner and 21.45 ± 3.95 for the otolaryngologist, which consists of moderate-to-high accuracy. The otolaryngologist and the ICU practitioner reported high ICC (0.807; 95%CI: 0.655-0.911). The highest QAMAI scores have been found for clarity and completeness of explanations. The QAMAI scores for the accuracy of the information and the referencing were the lowest. The information related to the post-laryngectomy tracheostomy remains incomplete or erroneous. ChatGPT-4o did not provide references for their responses. The stability analysis reported high stability in regenerated questions. The accuracy of ChatGPT-4o is moderate-to-high in providing information related to the tracheotomy. However, patients using ChatGPT-4o need to be cautious about the information related to tracheotomy care, steps, and the differences between temporary and permanent tracheotomies.

Advances and prospects of multi-modal ophthalmic artificial intelligence based on deep learning: a review.

In recent years, ophthalmology has emerged as a new frontier in medical artificial intelligence (AI) with multi-modal AI in ophthalmology garnering significant attention across interdisciplinary research. This integration of various types and data models holds paramount importance as it enables the provision of detailed and precise information for diagnosing eye and vision diseases. By leveraging multi-modal ophthalmology AI techniques, clinicians can enhance the accuracy and efficiency of diagnoses, and thus reduce the risks associated with misdiagnosis and oversight while also enabling more precise management of eye and vision health. However, the widespread adoption of multi-modal ophthalmology poses significant challenges. In this review, we first summarize comprehensively the concept of modalities in the field of ophthalmology, the forms of fusion between modalities, and the progress of multi-modal ophthalmic AI technology. Finally, we discuss the challenges of current multi-modal AI technology applications in ophthalmology and future feasible research directions. In the field of ophthalmic AI, evidence suggests that when utilizing multi-modal data, deep learning-based multi-modal AI technology exhibits excellent diagnostic efficacy in assisting the diagnosis of various ophthalmic diseases. Particularly, in the current era marked by the proliferation of large-scale models, multi-modal techniques represent the most promising and advantageous solution for addressing the diagnosis of various ophthalmic diseases from a comprehensive perspective. However, it must be acknowledged that there are still numerous challenges associated with the application of multi-modal techniques in ophthalmic AI before they can be effectively employed in the clinical setting.

Brain Tumor Detection and Classification using YOLOv10 and AI Chatbot Using LLMs

The integration of advanced medical imaging techniques and artificial intelligence (AI) has greatly improved the early detection and diagnosis of brain tumors. This paper introduces a novel system for brain tumor detection and classification using the YOLOv10 (You Only Look Once) model, enhanced by an AI chatbot powered by Large Language Models (LLMs). Leveraging the real-time object detection capabilities of YOLOv10, the system accurately classifies brain tumors from MRI images into four categories: Glioma, Meningioma, Pituitary, and No Tumor. This deep learning-based approach ensures swift and precise analysis of complex medical images. In addition, an AI chatbot is integrated to provide seamless interaction and information retrieval for both patients and healthcare professionals. Utilizing LLMs, the chatbot offers advanced conversational abilities, delivering detailed explanations about tumor types, potential treatments, and further medical advice based on the detected tumor characteristics. This dual-system approach aims to enhance diagnostic accuracy while providing real-time, accessible support, ultimately improving decision-making for medical practitioners and enhancing patient outcomes. The proposed system exemplifies the synergy between cutting-edge AI technologies and medical diagnostics, showcasing the potential to revolutionize patient care.

Readiness, knowledge, and perception towards artificial intelligence of medical students at faculty of medicine, Pelita Harapan University, Indonesia: a cross sectional study

IntroductionArtificial intelligence (AI) enables machines to perform many complicated human skills which require various levels of human intelligence. In the field of medicine, AI helps physicians in making diagnoses and treatments for patients with more efficiency, accuracy, and precision. In order to prepare medical students who are the future healthcare workforce, it is important to enhance their readiness, knowledge and perception toward AI. This study aims to assess Pelita Harapan University (PHU) medical students’ readiness, knowledge, and perception toward AI.MethodsA quantitative cross-sectional study was conducted to assess respondents’ readiness, knowledge and perception toward AI. An online questionnaire was distributed via Google Forms to all batch of medical students. Medical Artificial Intelligence Readiness Scale for Medical Students (MAIRS-MS) questionnaire was used to evaluate AI readiness, while an adapted questionnaires was used to evaluate knowledge and perception toward AI. Data were then analyzed using IBM Statistical Package for Social Sciences (SPSS) version 23.0.ResultsA total of 650 respondents were included in this study. Most respondents were in pre-clinical phase (88%) while the remaining were in clinical phase (12%). Overall, the total mean score for AI readiness was 73.34 of 100. Respondents had a mean score 24.52 ± 5.26 of 40, 27.78 ± 4.65 of 40, 10.57 ± 2.07 of 15, and 10.47 ± 2.00 of 15 in the cognitive, ability, vision, and ethics domain respectively. Generally, respondents had sufficient knowledge and positive perception toward AI. There were also significant correlation between readiness and knowledge with gender, having studied coding previously in high school, and having family or close friends working in AI field. Social media also had a good influence on enchancing readiness in the domain of ability and ethics, and perception towards AI.ConclusionMedical students of PHU mostly showed neutral to favorable response on readiness, knowledge, and perception towards AI. Incorporating AI into high school and medical curriculum is an important step to prepare medical students’ encounter and partnership with AI as the future workforce in medicine.