Automated HEART score determination via ChatGPT: Honing a framework for iterative prompt development

Objectives

This study presents a design framework to enhance the accuracy by which large language models (LLMs), like ChatGPT can extract insights from clinical notes. We highlight this framework via prompt refinement for the automated determination of HEART (History, ECG, Age, Risk factors, Troponin risk algorithm) scores in chest pain evaluation.

Methods

We developed a pipeline for LLM prompt testing, employing stochastic repeat testing and quantifying response errors relative to physician assessment. We evaluated the pipeline for automated HEART score determination across a limited set of 24 synthetic clinical notes representing four simulated patients. To assess whether iterative prompt design could improve the LLMs' ability to extract complex clinical concepts and apply rule-based logic to translate them to HEART subscores, we monitored diagnostic performance during prompt iteration.

Results

Validation included three iterative rounds of prompt improvement for three HEART subscores with 25 repeat trials totaling 1200 queries each for GPT-3.5 and GPT-4. For both LLM models, from initial to final prompt design, there was a decrease in the rate of responses with erroneous, non-numerical subscore answers. Accuracy of numerical responses for HEART subscores (discrete 0-2 point scale) improved for GPT-4 from the initial to final prompt iteration, decreasing from a mean error of 0.16-0.10 (95% confidence interval: 0.07-0.14) points.

Conclusion

We established a framework for iterative prompt design in the clinical space. Although the results indicate potential for integrating LLMs in structured clinical note analysis, translation to real, large-scale clinical data with appropriate data privacy safeguards is needed.

Performance of ChatGPT vs. HuggingChat on OB-GYN Topics

Background While large language models show potential as beneficial tools in medicine, their reliability, especially in the realm of obstetrics and gynecology (OB-GYN), is not fully comprehended. This study seeks to measure and contrast the performance of ChatGPT and HuggingChat in addressing OB-GYN-related medical examination questions, offering insights into their effectiveness in this specialized field. Methods ChatGPT and HuggingChat were subjected to two standardized multiple-choice question banks: Test 1, developed by the National Board of Medical Examiners (NBME), and Test 2, gathered from the Association of Professors of Gynecology & Obstetrics (APGO) Web-Based Interactive Self-Evaluation (uWISE). Responses were analyzed and compared for correctness. Results The two-proportion z-test revealed no statistically significant difference in performance between ChatGPT and HuggingChat on both medical examinations. For Test 1, ChatGPT scored 90%, while HuggingChat scored 85% (p = 0.6). For Test 2, ChatGPT correctly answered 70% of questions, while HuggingChat correctly answered 62% of questions (p = 0.4). Conclusion Awareness of the strengths and weaknesses of artificial intelligence allows for the proper and effective use of its knowledge. Our findings indicate that there is no statistically significant difference in performance between ChatGPT and HuggingChat in addressing medical inquiries. Nonetheless, both platforms demonstrate considerable promise for applications within the medical domain.

Comparing the Performance of Popular Large Language Models on the National Board of Medical Examiners Sample Questions

INTRODUCTION

Large language models (LLMs) have transformed various domains in medicine, aiding in complex tasks and clinical decision-making, with OpenAI's GPT-4, GPT-3.5, Google's Bard, and Anthropic's Claude among the most widely used. While GPT-4 has demonstrated superior performance in some studies, comprehensive comparisons among these models remain limited. Recognizing the significance of the National Board of Medical Examiners (NBME) exams in assessing the clinical knowledge of medical students, this study aims to compare the accuracy of popular LLMs on NBME clinical subject exam sample questions.

METHODS

The questions used in this study were multiple-choice questions obtained from the official NBME website and are publicly available. Questions from the NBME subject exams in medicine, pediatrics, obstetrics and gynecology, clinical neurology, ambulatory care, family medicine, psychiatry, and surgery were used to query each LLM. The responses from GPT-4, GPT-3.5, Claude, and Bard were collected in October 2023. The response by each LLM was compared to the answer provided by the NBME and checked for accuracy. Statistical analysis was performed using one-way analysis of variance (ANOVA).

RESULTS

A total of 163 questions were queried by each LLM. GPT-4 scored 163/163 (100%), GPT-3.5 scored 134/163 (82.2%), Bard scored 123/163 (75.5%), and Claude scored 138/163 (84.7%). The total performance of GPT-4 was statistically superior to that of GPT-3.5, Claude, and Bard by 17.8%, 15.3%, and 24.5%, respectively. The total performance of GPT-3.5, Claude, and Bard was not significantly different. GPT-4 significantly outperformed Bard in specific subjects, including medicine, pediatrics, family medicine, and ambulatory care, and GPT-3.5 in ambulatory care and family medicine. Across all LLMs, the surgery exam had the highest average score (18.25/20), while the family medicine exam had the lowest average score (3.75/5).  Conclusion: GPT-4's superior performance on NBME clinical subject exam sample questions underscores its potential in medical education and practice. While LLMs exhibit promise, discernment in their application is crucial, considering occasional inaccuracies. As technological advancements continue, regular reassessments and refinements are imperative to maintain their reliability and relevance in medicine.

The Role of Artificial Intelligence in Identifying Depression and Anxiety: A Comprehensive Literature Review

This narrative literature review undertakes a comprehensive examination of the burgeoning field, tracing the development of artificial intelligence (AI)-powered tools for depression and anxiety detection from the level of intricate algorithms to practical applications. Delivering essential mental health care services is now a significant public health priority. In recent years, AI has become a game-changer in the early identification and intervention of these pervasive mental health disorders. AI tools can potentially empower behavioral healthcare services by helping psychiatrists collect objective data on patients' progress and tasks. This study emphasizes the current understanding of AI, the different types of AI, its current use in multiple mental health disorders, advantages, disadvantages, and future potentials. As technology develops and the digitalization of the modern era increases, there will be a rise in the application of artificial intelligence in psychiatry; therefore, a comprehensive understanding will be needed. We searched PubMed, Google Scholar, and Science Direct using keywords for this. In a recent review of studies using electronic health records (EHR) with AI and machine learning techniques for diagnosing all clinical conditions, roughly 99 publications have been found. Out of these, 35 studies were identified for mental health disorders in all age groups, and among them, six studies utilized EHR data sources. By critically analyzing prominent scholarly works, we aim to illuminate the current state of this technology, exploring its successes, limitations, and future directions. In doing so, we hope to contribute to a nuanced understanding of AI's potential to revolutionize mental health diagnostics and pave the way for further research and development in this critically important domain.

Enhancing Postoperative Cochlear Implant Care With ChatGPT-4: A Study on Artificial Intelligence (AI)-Assisted Patient Education and Support

BACKGROUND

Cochlear implantation is a critical surgical intervention for patients with severe hearing loss. Postoperative care is essential for successful rehabilitation, yet access to timely medical advice can be challenging, especially in remote or resource-limited settings. Integrating advanced artificial intelligence (AI) tools like Chat Generative Pre-trained Transformer (ChatGPT)-4 in post-surgical care could bridge the patient education and support gap.

AIM

This study aimed to assess the effectiveness of ChatGPT-4 as a supplementary information resource for postoperative cochlear implant patients. The focus was on evaluating the AI chatbot's ability to provide accurate, clear, and relevant information, particularly in scenarios where access to healthcare professionals is limited.

MATERIALS AND METHODS

Five common postoperative questions related to cochlear implant care were posed to ChatGPT-4. The AI chatbot's responses were analyzed for accuracy, response time, clarity, and relevance. The aim was to determine whether ChatGPT-4 could serve as a reliable source of information for patients in need, especially if the patients could not reach out to the hospital or the specialists at that moment.

RESULTS

ChatGPT-4 provided responses aligned with current medical guidelines, demonstrating accuracy and relevance. The AI chatbot responded to each query within seconds, indicating its potential as a timely resource. Additionally, the responses were clear and understandable, making complex medical information accessible to non-medical audiences. These findings suggest that ChatGPT-4 could effectively supplement traditional patient education, providing valuable support in postoperative care.

CONCLUSION

The study concluded that ChatGPT-4 has significant potential as a supportive tool for cochlear implant patients post surgery. While it cannot replace professional medical advice, ChatGPT-4 can provide immediate, accessible, and understandable information, which is particularly beneficial in special moments. This underscores the utility of AI in enhancing patient care and supporting cochlear implantation.

Ocular Pathology and Genetics: Transformative Role of Artificial Intelligence (AI) in Anterior Segment Diseases

Artificial intelligence (AI) has become a revolutionary influence in the field of ophthalmology, providing unparalleled capabilities in data analysis and pattern recognition. This narrative review delves into the crucial role that AI plays, particularly in the context of anterior segment diseases with a genetic basis. Corneal dystrophies (CDs) exhibit significant genetic diversity, manifested by irregular substance deposition in the cornea. AI-driven diagnostic tools exhibit promising accuracy in the identification and classification of corneal diseases. Importantly, chat generative pre-trained transformer (ChatGPT)-4.0 shows significant advancement over its predecessor, ChatGPT-3.5. In the realm of glaucoma, AI significantly contributes to precise diagnostics through inventive algorithms and machine learning models, surpassing conventional methods. The incorporation of AI in predicting glaucoma progression and its role in augmenting diagnostic efficiency is readily apparent. Additionally, AI-powered models prove beneficial for early identification and risk assessment in cases of congenital cataracts, characterized by diverse inheritance patterns. Machine learning models achieving exceptional discrimination in identifying congenital cataracts underscore AI's remarkable potential. The review concludes by emphasizing the promising implications of AI in managing anterior segment diseases, spanning from early detection to the tailoring of personalized treatment strategies. These advancements signal a paradigm shift in ophthalmic care, offering optimism for enhanced patient outcomes and more streamlined healthcare delivery.

Generative Artificial Intelligence in Patient Education: ChatGPT Takes on Hypertension Questions

Introduction Uncontrolled hypertension significantly contributes to the development and deterioration of various medical conditions, such as myocardial infarction, chronic kidney disease, and cerebrovascular events. Despite being the most common preventable risk factor for all-cause mortality, only a fraction of affected individuals maintain their blood pressure in the desired range. In recent times, there has been a growing reliance on online platforms for medical information. While providing a convenient source of information, differentiating reliable from unreliable information can be daunting for the layperson, and false information can potentially hinder timely diagnosis and management of medical conditions. The surge in accessibility of generative artificial intelligence (GeAI) technology has led to increased use in obtaining health-related information. This has sparked debates among healthcare providers about the potential for misuse and misinformation while recognizing the role of GeAI in improving health literacy. This study aims to investigate the accuracy of AI-generated information specifically related to hypertension. Additionally, it seeks to explore the reproducibility of information provided by GeAI. Method A nonhuman-subject qualitative study was devised to evaluate the accuracy of information provided by ChatGPT regarding hypertension and its secondary complications. Frequently asked questions on hypertension were compiled by three study staff, internal medicine residents at an ACGME-accredited program, and then reviewed by a physician experienced in treating hypertension, resulting in a final set of 100 questions. Each question was posed to ChatGPT three times, once by each study staff, and the majority response was then assessed against the recommended guidelines. A board-certified internal medicine physician with over eight years of experience further reviewed the responses and categorized them into two classes based on their clinical appropriateness: appropriate (in line with clinical recommendations) and inappropriate (containing errors). Descriptive statistical analysis was employed to assess ChatGPT responses for accuracy and reproducibility. Result Initially, a pool of 130 questions was gathered, of which a final set of 100 questions was selected for the purpose of this study. When assessed against acceptable standard responses, ChatGPT responses were found to be appropriate in 92.5% of cases and inappropriate in 7.5%. Furthermore, ChatGPT had a reproducibility score of 93%, meaning that it could consistently reproduce answers that conveyed similar meanings across multiple runs. Conclusion ChatGPT showcased commendable accuracy in addressing commonly asked questions about hypertension. These results underscore the potential of GeAI in providing valuable information to patients. However, continued research and refinement are essential to evaluate further the reliability and broader applicability of ChatGPT within the medical field.

Future Implications of Artificial Intelligence in Medical Education

Artificial intelligence has experienced explosive growth in the past year that will have implications in all aspects of our lives, including medicine. In order to train a physician workforce that understands these new advancements, medical educators must take steps now to ensure that physicians are adequately trained in medical school, residency, and fellowship programs to become proficient in the usage of artificial intelligence in medical practice. This manuscript discusses the various considerations that leadership within medical training programs should be mindful of when deciding how to best integrate artificial intelligence into their curricula.

Can DALL-E 3 Reliably Generate 12-Lead ECGs and Teaching Illustrations?

The recent integration of the latest image generation model DALL-E 3 into ChatGPT allows text prompts to easily generate the corresponding images, enabling multimodal output from ChatGPT. We explored the feasibility of DALL-E 3 for drawing a 12-lead ECG and found that it can draw rudimentary 12-lead electrocardiograms (ECG) displaying some of the parameters, although the details are not completely accurate. We also explored DALL-E 3's capacity to create vivid illustrations for teaching resuscitation-related medical knowledge. DALL-E 3 produced accurate CPR illustrations emphasizing proper hand placement and technique. For ECG principles, it produced creative heart-shaped waveforms tying ECGs to the heart. With further training, DALL-E 3 shows promise to expand easy-to-understand visual medical teaching materials and ECG simulations for different disease states. In conclusion, DALL-E 3 has the potential to generate realistic 12-lead ECGs and teaching schematics, but expert validation is still needed.

Artificial Intelligence: Knowledge and Attitude Among Lebanese Medical Students

Background Artificial intelligence (AI) has taken on a variety of functions in the medical field, and research has proven that it can address complicated issues in various applications. It is unknown whether Lebanese medical students and residents have a detailed understanding of this concept, and little is known about their attitudes toward AI. Aim This study fills a critical gap by revealing the knowledge and attitude of Lebanese medical students toward AI. Methods A multi-centric survey targeting 365 medical students from seven medical schools across Lebanon was conducted to assess their knowledge of and attitudes toward AI in medicine. The survey consists of five sections: the first part includes socio-demographic variables, while the second comprises the 'Medical Artificial Intelligence Readiness Scale' for medical students. The third part focuses on attitudes toward AI in medicine, the fourth assesses understanding of deep learning, and the fifth targets considerations of radiology as a specialization. Results There is a notable awareness of AI among students who are eager to learn about it. Despite this interest, there exists a gap in knowledge regarding deep learning, albeit alongside a positive attitude towards it. Students who are more open to embracing AI technology tend to have a better understanding of AI concepts (p=0.001). Additionally, a higher percentage of students from Mount Lebanon (71.6%) showed an inclination towards using AI compared to Beirut (63.2%) (p=0.03). Noteworthy are the Lebanese University and Saint Joseph University, where the highest proportions of students are willing to integrate AI into the medical field (79.4% and 76.7%, respectively; p=0.001). Conclusion It was concluded that most Lebanese medical students might not necessarily comprehend the core technological ideas of AI and deep learning. This lack of understanding was evident from the substantial amount of misinformation among the students. Consequently, there appears to be a significant demand for the inclusion of AI technologies in Lebanese medical school courses.

The Algorithmic Lung Detective: Artificial Intelligence in the Diagnosis of Pulmonary Embolism

Pulmonary embolism (PE) poses a significant threat as the third leading cause of cardiovascular death, prompting the widespread use of CT pulmonary angiogram for rapid detection. Despite its prevalence, diagnostic accuracy remains variable among radiologists. The emergence of artificial intelligence (AI), notably through convolutional neural networks and deep learning reconstruction, offers a promising avenue to enhance PE detection. AI demonstrates superior sensitivity and negative predictive values, reducing the risk of missed diagnoses. Implementation of AI-based worklist prioritization substantially shortens detection and notification times, streamlining radiological workflows. However, it is crucial to underscore that AI acts as a complement, not a replacement, for radiologists, synergizing with human expertise. As AI integration progresses, it holds the potential to significantly improve diagnostic accuracy and efficiency in pulmonary embolism detection while maintaining the essential role of human judgment in medical decision-making.

Artificial Intelligence in the Early Prediction of Cardiogenic Shock in Acute Heart Failure or Myocardial Infarction Patients: A Systematic Review and Meta-Analysis

Cardiogenic shock (CS) may have a negative impact on mortality in patients with heart failure (HF) or acute myocardial infarction (AMI). Early prediction of CS can result in improved survival. Artificial intelligence (AI) through machine learning (ML) models have shown promise in predictive medicine. Here, we conduct a systematic review and meta-analysis to assess the effectiveness of these models in the early prediction of CS. A thorough search of the PubMed, Web of Science, Cochrane, and Scopus databases was conducted from the time of inception until November 2, 2023, to find relevant studies. Our outcomes were area under the curve (AUC), the sensitivity and specificity of the ML model, the accuracy of the ML model, and the predictor variables that had the most impact in predicting CS. Comprehensive Meta-Analysis (CMA) Version 3.0 was used to conduct the meta-analysis. Six studies were considered in our study. The pooled mean AUC was 0.808 (95% confidence interval: 0.727, 0.890). The AUC in the included studies ranged from 0.77 to 0.91. ML models performed well, with accuracy ranging from 0.88 to 0.93 and sensitivity and specificity of 58%-78% and 88%-93%, respectively. Age, blood pressure, heart rate, oxygen saturation, and blood glucose were the most significant variables required by ML models to acquire their outputs. In conclusion, AI has the potential for early prediction of CS, which may lead to a decrease in the high mortality rate associated with it. Future studies are needed to confirm the results.

Stratified Evaluation of GPT's Question Answering in Surgery Reveals Artificial Intelligence (AI) Knowledge Gaps

Large language models (LLMs) have broad potential applications in medicine, such as aiding with education, providing reassurance to patients, and supporting clinical decision-making. However, there is a notable gap in understanding their applicability and performance in the surgical domain and how their performance varies across specialties. This paper aims to evaluate the performance of LLMs in answering surgical questions relevant to clinical practice and to assess how this performance varies across different surgical specialties. We used the MedMCQA dataset, a large-scale multi-choice question-answer (MCQA) dataset consisting of clinical questions across all areas of medicine. We extracted the relevant 23,035 surgical questions and submitted them to the popular LLMs Generative Pre-trained Transformers (GPT)-3.5 and GPT-4 (OpenAI OpCo, LLC, San Francisco, CA). Generative Pre-trained Transformer is a large language model that can generate human-like text by predicting subsequent words in a sentence based on the context of the words that come before it. It is pre-trained on a diverse range of texts and can perform a variety of tasks, such as answering questions, without needing task-specific training. The question-answering accuracy of GPT was calculated and compared between the two models and across surgical specialties. Both GPT-3.5 and GPT-4 achieved accuracies of 53.3% and 64.4%, respectively, on surgical questions, showing a statistically significant difference in performance. When compared to their performance on the full MedMCQA dataset, the two models performed differently: GPT-4 performed worse on surgical questions than on the dataset as a whole, while GPT-3.5 showed the opposite pattern. Significant variations in accuracy were also observed across different surgical specialties, with strong performances in anatomy, vascular, and paediatric surgery and worse performances in orthopaedics, ENT, and neurosurgery. Large language models exhibit promising capabilities in addressing surgical questions, although the variability in their performance between specialties cannot be ignored. The lower performance of the latest GPT-4 model on surgical questions relative to questions across all medicine highlights the need for targeted improvements and continuous updates to ensure relevance and accuracy in surgical applications. Further research and continuous monitoring of LLM performance in surgical domains are crucial to fully harnessing their potential and mitigating the risks of misinformation.

ChatGPT-4 and Human Researchers Are Equal in Writing Scientific Introduction Sections: A Blinded, Randomized, Non-inferiority Controlled Study

Background Natural language processing models are increasingly used in scientific research, and their ability to perform various tasks in the research process is rapidly advancing. This study aims to investigate whether Generative Pre-trained Transformer 4 (GPT-4) is equal to humans in writing introduction sections for scientific articles. Methods This randomized non-inferiority study was reported according to the Consolidated Standards of Reporting Trials for non-inferiority trials and artificial intelligence (AI) guidelines. GPT-4 was instructed to synthesize 18 introduction sections based on the aim of previously published studies, and these sections were compared to the human-written introductions already published in a medical journal. Eight blinded assessors randomly evaluated the introduction sections using 1-10 Likert scales. Results There was no significant difference between GPT-4 and human introductions regarding publishability and content quality. GPT-4 had one point significantly better scores in readability, which was considered a non-relevant difference. The majority of assessors (59%) preferred GPT-4, while 33% preferred human-written introductions. Based on Lix and Flesch-Kincaid scores, GPT-4 introductions were 10 and two points higher, respectively, indicating that the sentences were longer and had longer words. Conclusion GPT-4 was found to be equal to humans in writing introductions regarding publishability, readability, and content quality. The majority of assessors preferred GPT-4 introductions and less than half could determine which were written by GPT-4 or humans. These findings suggest that GPT-4 can be a useful tool for writing introduction sections, and further studies should evaluate its ability to write other parts of scientific articles.

Artificial Intelligence in the Detection of Barrett's Esophagus: A Systematic Review

Barrett's esophagus (BE) remains a significant precursor to esophageal adenocarcinoma, requiring accurate and efficient diagnosis and management. The increasing application of machine learning (ML) technologies presents a transformative opportunity for diagnosing and treating BE. This systematic review evaluates the effectiveness and accuracy of machine learning technologies in BE diagnosis and management by conducting a comprehensive search across PubMed, Scopus, and Web of Science databases up to the year 2023. The studies were organized into five categories: computer-aided systems, natural language processing and text-based systems, deep learning on histology and biopsy images, real-time and video analysis, and miscellaneous studies. Results indicate high sensitivity and specificity across machine learning applications. Specifically, computer-aided systems showed sensitivities ranging from 84% to 100% and specificities from 64% to 90.7%. Natural language processing and text-based systems achieved an accuracy as high as 98.7%. Deep learning techniques applied to histology and biopsy images displayed sensitivities up to greater than 90% and a specificity of 100%. Furthermore, real-time and video analysis technologies demonstrated high performance with assessment speeds of up to 48 frames per second (fps) and a mean average precision of 75.3%. Overall, the reviewed literature underscores the growing capability and efficiency of machine learning technologies in diagnosing and managing Barrett's esophagus, often outperforming traditional diagnostic methods. These findings highlight the promising future role of machine learning in enhancing clinical practice and improving patient care for Barrett's esophagus.

Exploring the Role of Artificial Intelligence in Mental Healthcare: Progress, Pitfalls, and Promises

The rise of artificial intelligence (AI) heralds a significant revolution in healthcare, particularly in mental health. AI's potential spans diagnostic algorithms, data analysis from diverse sources, and real-time patient monitoring. It is essential for clinicians to remain informed about AI's progress and limitations. The inherent complexity of mental disorders, limited objective data, and retrospective studies pose challenges to the application of AI. Privacy concerns, bias, and the risk of AI replacing human care also loom. Regulatory oversight and physician involvement are needed for equitable AI implementation. AI integration and use in psychotherapy and other services are on the horizon. Patient trust, feasibility, clinical efficacy, and clinician acceptance are prerequisites. In the future, governing bodies must decide on AI ownership, governance, and integration approaches. While AI can enhance clinical decision-making and efficiency, it might also exacerbate moral dilemmas, autonomy loss, and issues regarding the scope of practice. Striking a balance between AI's strengths and limitations involves utilizing AI as a validated clinical supplement under medical supervision, necessitating active clinician involvement in AI research, ethics, and regulation. AI's trajectory must align with optimizing mental health treatment and upholding compassionate care.

Artificial Intelligence and Its Role in the Management of Chronic Medical Conditions: A Systematic Review

Due to the increased burden of chronic medical conditions in recent years, artificial intelligence (AI) is suggested in the medical field to optimize health care. Physicians could implement these automated problem-solving tools for their benefit, reducing their workload, assisting in diagnostics, and supporting clinical decision-making. These tools are being considered for future medical assistance in real life. A literature review was performed to assess the impact of AI on the patient population with chronic medical conditions, using standardized guidelines. A MeSH strategy was created, and the database was searched for appropriate studies using specific inclusion and exclusion criteria. The online database yielded 93 results from various databases, of which 10 moderate to high-quality studies were selected to be included in our systematic review after removing the duplicates, screening titles, and articles. Of the 10 studies, nine recommended using AI after considering the potential limitations such as privacy protection, medicolegal implications, and psychosocial aspects. Due to its non-fatigable nature, AI was found to be of immense help in image recognition. It was also found to be valuable in various disciplines related to administration, physician burden, and patient adherence. The newer technologies of Chatbots and eHealth applications are of great help when used safely and effectively after proper patient education. After a careful review conducted by our team members, it is safe to conclude that implementing AI in daily clinical practice could potentiate the cognitive ability of physicians and decrease the workload through various automated technologies such as image recognition, speech recognition, and voice recognition due to its unmatchable speed and non-fatigable nature when compared to clinicians. Despite its vast benefits to the medical field, a few limitations could hinder its effective implementation into real-life practice, which requires enormous research and strict regulations to support its role as a physician's aid. However, AI should only be used as a medical support system, in order to improve the primary outcomes such as reducing waiting time, healthcare costs, and workload. AI should not be meant to replace physicians.

Bias and Inaccuracy in AI Chatbot Ophthalmologist Recommendations

PURPOSE AND DESIGN

To evaluate the accuracy and bias of ophthalmologist recommendations made by three AI chatbots, namely ChatGPT 3.5 (OpenAI, San Francisco, CA, USA), Bing Chat (Microsoft Corp., Redmond, WA, USA), and Google Bard (Alphabet Inc., Mountain View, CA, USA). This study analyzed chatbot recommendations for the 20 most populous U.S. cities.

METHODS

Each chatbot returned 80 total recommendations when given the prompt "Find me four good ophthalmologists in (city)." Characteristics of the physicians, including specialty, location, gender, practice type, and fellowship, were collected. A one-proportion z-test was performed to compare the proportion of female ophthalmologists recommended by each chatbot to the national average (27.2% per the Association of American Medical Colleges (AAMC)). Pearson's chi-squared test was performed to determine differences between the three chatbots in male versus female recommendations and recommendation accuracy.

RESULTS

Female ophthalmologists recommended by Bing Chat (1.61%) and Bard (8.0%) were significantly less than the national proportion of 27.2% practicing female ophthalmologists (p<0.001, p<0.01, respectively). ChatGPT recommended fewer female (29.5%) than male ophthalmologists (p<0.722). ChatGPT (73.8%), Bing Chat (67.5%), and Bard (62.5%) gave high rates of inaccurate recommendations. Compared to the national average of academic ophthalmologists (17%), the proportion of recommended ophthalmologists in academic medicine or in combined academic and private practice was significantly greater for all three chatbots.

CONCLUSION

This study revealed substantial bias and inaccuracy in the AI chatbots' recommendations. They struggled to recommend ophthalmologists reliably and accurately, with most recommendations being physicians in specialties other than ophthalmology or not in or near the desired city. Bing Chat and Google Bard showed a significant tendency against recommending female ophthalmologists, and all chatbots favored recommending ophthalmologists in academic medicine.

Medical Diagnosis Reimagined as a Process of Bayesian Reasoning and Elimination

This article delves into the interface between the art of medical diagnosis and the mathematical foundations of probability, the Bayes theorem. In a healthcare ecosystem witnessing an artificial intelligence (AI)-driven transformation, understanding the convergence becomes crucial for physicians. Contrary to viewing AI as a mysterious "black box," we demonstrate how every diagnostic decision by a medical practitioner is, in essence, Bayesian reasoning in action. The Bayes theorem is a mathematical translation of systematically updating our belief: it quantifies how an additional piece of information updates our prior belief in something. Using a clinical scenario of Kartagener syndrome, we showcase the parallels between a physician's evolving diagnostic thought process and the mathematical updating of prior beliefs with new evidence. By reimagining medical diagnosis through the lens of Bayes, this paper aims to demystify AI, accentuating its potential role as an enhancer of clinical acumen rather than a replacement. The ultimate vision presented is one of harmony, where AI serves as a symbiotic partner to physicians, with the shared goal of holistic patient care.

Artificial Hallucinations by Google Bard: Think Before You Leap

One of the critical challenges posed by artificial intelligence (AI) tools like Google Bard (Google LLC, Mountain View, California, United States) is the potential for "artificial hallucinations." These refer to instances where an AI chatbot generates fictional, erroneous, or unsubstantiated information in response to queries. In research, such inaccuracies can lead to the propagation of misinformation and undermine the credibility of scientific literature. The experience presented here highlights the importance of cross-checking the information provided by AI tools with reliable sources and maintaining a cautious approach when utilizing these tools in research writing.

Exploring the Potential and Limitations of Chat Generative Pre-trained Transformer (ChatGPT) in Generating Board-Style Dermatology Questions: A Qualitative Analysis

This article investigates the limitations of Chat Generative Pre-trained Transformer (ChatGPT), a language model developed by OpenAI, as a study tool in dermatology. The study utilized ChatPDF, an application that integrates PDF files with ChatGPT, to generate American Board of Dermatology Applied Exam (ABD-AE)-style questions from continuing medical education articles from the Journal of the American Board of Dermatology. A qualitative analysis of the questions was conducted by two board-certified dermatologists, assessing accuracy, complexity, and clarity. Out of 40 questions generated, only 16 (40%) were deemed accurate and appropriate for ABD-AE study preparation. The remaining questions exhibited limitations, including low complexity, lack of clarity, and inaccuracies. The findings highlight the challenges faced by ChatGPT in understanding the domain-specific knowledge required in dermatology. Moreover, the model's inability to comprehend the context and generate high-quality distractor options, as well as the absence of image generation capabilities, further hinders its usefulness. The study emphasizes that while ChatGPT may aid in generating simple questions, it cannot replace the expertise of dermatologists and medical educators in developing high-quality, board-style questions that effectively evaluate candidates' knowledge and reasoning abilities.

Interpretable Machine Learning for COVID-19: An Empirical Study on Severity Prediction Task

The black-box nature of machine learning models hinders the deployment of some high-accuracy medical diagnosis algorithms. It is risky to put one's life in the hands of models that medical researchers do not fully understand or trust. However, through model interpretation, black-box models can promptly reveal significant biomarkers that medical practitioners may have overlooked due to the surge of infected patients in the COVID-19 pandemic. This research leverages a database of 92 patients with confirmed SARS-CoV-2 laboratory tests between 18th January 2020 and 5th March 2020, in Zhuhai, China, to identify biomarkers indicative of infection severity prediction. Through the interpretation of four machine learning models, decision tree, random forests, gradient boosted trees, and neural networks using permutation feature importance, partial dependence plot, individual conditional expectation, accumulated local effects, local interpretable model-agnostic explanations, and Shapley additive explanation, we identify an increase in N-terminal pro-brain natriuretic peptide, C-reaction protein, and lactic dehydrogenase, a decrease in lymphocyte is associated with severe infection and an increased risk of death, which is consistent with recent medical research on COVID-19 and other research using dedicated models. We further validate our methods on a large open dataset with 5644 confirmed patients from the Hospital Israelita Albert Einstein, at São Paulo, Brazil from Kaggle, and unveil leukocytes, eosinophils, and platelets as three indicative biomarkers for COVID-19.

ChatGPT's Ability to Assess Quality and Readability of Online Medical Information: Evidence From a Cross-Sectional Study

Introduction Artificial Intelligence (AI) platforms have gained widespread attention for their distinct ability to generate automated responses to various prompts. However, its role in assessing the quality and readability of a provided text remains unclear. Thus, the purpose of this study is to evaluate the proficiency of the conversational generative pre-trained transformer (ChatGPT) in utilizing the DISCERN tool to evaluate the quality of online content regarding shock wave therapy for erectile dysfunction. Methods Websites were generated using a Google search of "shock wave therapy for erectile dysfunction" with location filters disabled. Readability was analyzed using Readable software (Readable.com, Horsham, United Kingdom). Quality was assessed independently by three reviewers using the DISCERN tool. The same plain text files collected were inputted into ChatGPT to determine whether they produced comparable metrics for readability and quality. Results The study results revealed a notable disparity between ChatGPT's readability assessment and that obtained from a reliable tool, Readable.com (p<0.05). This indicates a lack of alignment between ChatGPT's algorithm and that of established tools, such as Readable.com. Similarly, the DISCERN score generated by ChatGPT differed significantly from the scores generated manually by human evaluators (p<0.05), suggesting that ChatGPT may not be capable of accurately identifying poor-quality information sources regarding shock wave therapy as a treatment for erectile dysfunction. Conclusion ChatGPT's evaluation of the quality and readability of online text regarding shockwave therapy for erectile dysfunction differs from that of human raters and trusted tools. Therefore, ChatGPT's current capabilities were not sufficient for reliably assessing the quality and readability of textual content. Further research is needed to elucidate the role of AI in the objective evaluation of online medical content in other fields. Continued development in AI and incorporation of tools such as DISCERN into AI software may enhance the way patients navigate the web in search of high-quality medical content in the future.

Advancing Artificial Intelligence for Clinical Knowledge Retrieval: A Case Study Using ChatGPT-4 and Link Retrieval Plug-In to Analyze Diabetic Ketoacidosis Guidelines

Introduction This case study aimed to enhance the traceability and retrieval accuracy of ChatGPT-4 in medical text by employing a step-by-step systematic approach. The focus was on retrieving clinical answers from three international guidelines on diabetic ketoacidosis (DKA). Methods A systematic methodology was developed to guide the retrieval process. One question was asked per guideline to ensure accuracy and maintain referencing. ChatGPT-4 was utilized to retrieve answers, and the 'Link Reader' plug-in was integrated to facilitate direct access to webpages containing the guidelines. Subsequently, ChatGPT-4 was employed to compile answers while providing citations to the sources. This process was iterated 30 times per question to ensure consistency. In this report, we present our observations regarding the retrieval accuracy, consistency of responses, and the challenges encountered during the process. Results Integrating ChatGPT-4 with the 'Link Reader' plug-in demonstrated notable traceability and retrieval accuracy benefits. The AI model successfully provided relevant and accurate clinical answers based on the analyzed guidelines. Despite occasional challenges with webpage access and minor memory drift, the overall performance of the integrated system was promising. The compilation of the answers was also impressive and held significant promise for further trials. Conclusion The findings of this case study contribute to the utilization of AI text-generation models as valuable tools for medical professionals and researchers. The systematic approach employed in this case study and the integration of the 'Link Reader' plug-in offer a framework for automating medical text synthesis, asking one question at a time before compilation from different sources, which has led to improving AI models' traceability and retrieval accuracy. Further advancements and refinement of AI models and integration with other software utilities hold promise for enhancing the utility and applicability of AI-generated recommendations in medicine and scientific academia. These advancements have the potential to drive significant improvements in everyday medical practice.

Eyes on AI: ChatGPT's Transformative Potential Impact on Ophthalmology

ChatGPT, a large language model by OpenAI, has been adopted in various domains since its release in November 2022, but its application in ophthalmology remains less explored. This editorial assesses ChatGPT's potential applications and limitations in ophthalmology across clinical, educational, and research settings. In clinical settings, ChatGPT can serve as an assistant, offering diagnostic and therapeutic suggestions based on patient data and assisting in patient triage. However, its tendencies to generate inaccurate results and its inability to keep up with recent medical guidelines render it unsuitable for standalone clinical decision-making. Data security and compliance with the Health Insurance Portability and Accountability Act (HIPAA) also pose concerns, given ChatGPT's potential to inadvertently expose sensitive patient information. In education, ChatGPT can generate practice questions, provide explanations, and create patient education materials. However, its performance in answering domain-specific questions is suboptimal. In research, ChatGPT can facilitate literature reviews, data analysis, manuscript development, and peer review, but issues of accuracy, bias, and ethics need careful consideration. Ultimately, ensuring accuracy, ethical integrity, and data privacy is essential when integrating ChatGPT into ophthalmology.

Experimental analysis of machine learning methods to detect Covid-19 from x-rays

To automate the detection of covid-19 patients most have proposed deep learning neural networks to classify patients using large databases of chest x-rays. Very few used classical machine learning methods. Machine learning methods may require less computational power and perform well if the data set is small. We experiment with classical machine learning methods on three different data sources varying in size from 55 to almost 4000 samples. We experiment with four feature extraction methods of Gabor, SURF, LBP, and HOG. Backpropagation neural networks and k-nearest neighbor classifiers are combined using one of the four combining methods of bagging, RSM, ARCx4 boosting and Ada-boosting. Results show that using the proper feature extraction and feature selection methods very high performance can be reached using simple backpropagation neural network classifiers. Regardless of combiner method used, the best classification rate achieved was 99.06% for the largest data set, and 100% for the smallest data set.

Overview of Early ChatGPT's Presence in Medical Literature: Insights From a Hybrid Literature Review by ChatGPT and Human Experts

ChatGPT, an artificial intelligence chatbot, has rapidly gained prominence in various domains, including medical education and healthcare literature. This hybrid narrative review, conducted collaboratively by human authors and ChatGPT, aims to summarize and synthesize the current knowledge of ChatGPT in the indexed medical literature during its initial four months. A search strategy was employed in PubMed and EuropePMC databases, yielding 65 and 110 papers, respectively. These papers focused on ChatGPT's impact on medical education, scientific research, medical writing, ethical considerations, diagnostic decision-making, automation potential, and criticisms. The findings indicate a growing body of literature on ChatGPT's applications and implications in healthcare, highlighting the need for further research to assess its effectiveness and ethical concerns.

A Case Study Demonstrating Applications of ChatGPT in the Clinical Management of Treatment-Resistant Schizophrenia

Chat Generative Pre-trained Transformer, also known as ChatGPT, is a new artificial intelligence (AI) program that responds to user inquiry with discourse resembling human language. The range of ChatGPT capabilities caught the interest of the medical world after it demonstrated its ability to pass medical boards examinations. In this case report, we present the clinical treatment of a 22-year-old male diagnosed with treatment-resistant schizophrenia (TRS) and compare the medical management suggested by ChatGPT to current standards of care in order to assess the program's ability to identify the disorder, evaluate potential medical and psychiatric work-up, and develop a treatment plan addressing the distinct nuances of our patient. In our inquiry with ChatGPT, we found that it can accurately identify our patient as having TRS and order appropriate tests to methodically rule out alternative causes of acute psychosis. Furthermore, the AI program suggests pharmacologic treatment options including clozapine with adjuvant medications, and nonpharmacologic treatment options including electroconvulsive therapy (ECT), repetitive transcranial magnetic stimulation (rTMS), and psychotherapy which align with current standards of care. Lastly, ChatGPT provides a comprehensive list of side effects associated with antipsychotics and mood stabilizers used to treat TRS. We found both potential for and limitations in the clinical application of ChatGPT to assist in the assessment and management of complex medical conditions. Overall, ChatGPT may serve as a powerful tool to organize medical data in a meaningful and palatable format for medical professionals to reference during patient care.

Colorectal Cancer Diagnostic Methods: The Present and Future

To meet the needs of the colorectal cancer (CRC) patient population, colorectal cancer screening is continuously updated. The most significant advice is to start CRC screening exams at age 45 for people at average risk for CRC. CRC testing is divided into two categories: stool-based tests and visual inspections. High-sensitivity guaiac-based fecal occult blood testing, fecal immunochemical testing, and multitarget stool DNA testing are stool-based assays. Colon capsule endoscopy and flexible sigmoidoscopy are visualization examinations. There have been arguments about the importance of these tests in detecting and managing precursor lesions because of the lack of validation of screening results. Recent advancements in artificial intelligence and genetics have prompted the creation of newer diagnostic tests, which require validation in diverse populations and cohorts. In this article, we have discussed the present and emerging diagnostic tests.

Saudi Arabia Health Systems: Challenging and Future Transformations With Artificial Intelligence

The healthcare system in Saudi Arabia is facing several challenges, including an aging population, an increase in chronic diseases, and a shortage of healthcare professionals. To address these challenges, the government is taking proactive steps, including expanding healthcare infrastructure, promoting the use of technology, improving the quality of healthcare services, and emphasizing the importance of preventive healthcare. In addition, the adoption of artificial intelligence (AI) solutions can play a crucial role in transforming the healthcare system by improving efficiency, reducing costs, and enhancing the quality of care. However, the adoption of AI solutions comes with challenges such as the need for high-quality data and the development of regulations and guidelines. The government needs to continue to invest in healthcare and AI solutions to build a more efficient and effective healthcare system that benefits all citizens.

The Integration of Artificial Intelligence Into Patient Care: A Case of Atrial Fibrillation Caught by a Smartwatch

Artificial intelligence (AI) offers a wide range of applications in clinical practice, and new technologies are rapidly evolving the healthcare industry and enhancing outcomes. Smartwatches represent the most popular type of wearable AI device that can assist people in detecting cardiac arrhythmias via constant monitoring of heart activity. Numerous advantages result from integrating AI into healthcare systems, including improved patient care, lower rates of medical errors, better treatment recommendations, and more accurate diagnosis of diseases. However, doubts still remain regarding the adoption of AI into patient care due to the challenges it poses. In this paper, we report a case of atrial fibrillation (AF) in a young patient that was detected by his smartwatch. We also highlight some of the benefits and challenges of AI applications in healthcare.

Artificial Hallucinations in ChatGPT: Implications in Scientific Writing

While still in its infancy, ChatGPT (Generative Pretrained Transformer), introduced in November 2022, is bound to hugely impact many industries, including healthcare, medical education, biomedical research, and scientific writing. Implications of ChatGPT, that new chatbot introduced by OpenAI on academic writing, is largely unknown. In response to the Journal of Medical Science (Cureus) Turing Test - call for case reports written with the assistance of ChatGPT, we present two cases one of homocystinuria-associated osteoporosis, and the other is on late-onset Pompe disease (LOPD), a rare metabolic disorder. We tested ChatGPT to write about the pathogenesis of these conditions. We documented the positive, negative, and rather troubling aspects of our newly introduced chatbot's performance.

A Comprehensive Review of the Role of Artificial Intelligence in Obstetrics and Gynecology

The exponential growth of artificial intelligence (AI) has fascinated its application in various fields and so in the field of healthcare. Technological advancements in theories and learning algorithms and the availability of processing through huge datasets have created a breakthrough in the medical field with computing systems. AI can potentially drive clinicians and practitioners with appropriate decisions in managing cases and reaching a diagnosis, so its application is extensively spread in the medical field. Thus, computerized algorithms have made predictions so simple and accurate. This is because AI can proffer information accurately even to many patients. Furthermore, the subsets of AI, namely, machine learning (ML) and deep learning (DL) methods, have aided in detecting complex patterns from huge datasets and using such patterns in making predictions. Despite numerous challenges, AI implementation in obstetrics and gynecology is found to have a spellbound development. Therefore, this review propounds exploring the implementation of AI in obstetrics and gynecology to improve the outcomes and clinical experience. In that context, the evolution and progress of AI, the role of AI in ultrasound diagnosis in distinct phases of pregnancy, clinical benefits, preterm birth postpartum period, and applications of AI in gynecology are elucidated in this review with future recommendations.

ChatGPT Output Regarding Compulsory Vaccination and COVID-19 Vaccine Conspiracy: A Descriptive Study at the Outset of a Paradigm Shift in Online Search for Information

BACKGROUND

Being on the verge of a revolutionary approach to gathering information, ChatGPT (an artificial intelligence (AI)-based language model developed by OpenAI, and capable of producing human-like text) could be the prime motive of a paradigm shift on how humans will acquire information. Despite the concerns related to the use of such a promising tool in relation to the future of the quality of education, this technology will soon be incorporated into web search engines mandating the need to evaluate the output of such a tool. Previous studies showed that dependence on some sources of online information (e.g., social media platforms) was associated with higher rates of vaccination hesitancy. Therefore, the aim of the current study was to describe the output of ChatGPT regarding coronavirus disease 2019 (COVID-19) vaccine conspiracy beliefs. and compulsory vaccination.

METHODS

The current descriptive study was conducted on January 14, 2023 using the ChatGPT from OpenAI (OpenAI, L.L.C., San Francisco, CA, USA). The output was evaluated by two authors and the degree of agreement regarding the correctness, clarity, conciseness, and bias was evaluated using Cohen's kappa.

RESULTS

The ChatGPT responses were dismissive of conspiratorial ideas about severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) origins labeling it as non-credible and lacking scientific evidence. Additionally, ChatGPT responses were totally against COVID-19 vaccine conspiracy statements. Regarding compulsory vaccination, ChatGPT responses were neutral citing the following as advantages of this strategy: protecting public health, maintaining herd immunity, reducing the spread of disease, cost-effectiveness, and legal obligation, and on the other hand, it cited the following as disadvantages of compulsory vaccination: ethical and legal concerns, mistrust and resistance, logistical challenges, and limited resources and knowledge.

CONCLUSIONS

The current study showed that ChatGPT could be a source of information to challenge COVID-19 vaccine conspiracies. For compulsory vaccination, ChatGPT resonated with the divided opinion in the scientific community toward such a strategy; nevertheless, it detailed the pros and cons of this approach. As it currently stands, the judicious use of ChatGPT could be utilized as a user-friendly source of COVID-19 vaccine information that could challenge conspiracy ideas with clear, concise, and non-biased content. However, ChatGPT content cannot be used as an alternative to the original reliable sources of vaccine information (e.g., the World Health Organization [WHO] and the Centers for Disease Control and Prevention [CDC]).

Artificial Intelligence in Medicine: Text Mining of Health Care Workers' Opinions

BACKGROUND

Artificial intelligence (AI) is being increasingly adopted in the health care industry for administrative tasks, patient care operations, and medical research.

OBJECTIVE

We aimed to examine health care workers' opinions about the adoption and implementation of AI-powered technology in the health care industry.

METHODS

Data were comments about AI posted on a web-based forum by 905 health care professionals from at least 77 countries, from May 2013 to October 2021. Structural topic modeling was used to identify the topics of discussion, and hierarchical clustering was performed to determine how these topics cluster into different groups.

RESULTS

Overall, 12 topics were identified from the collected comments. These comments clustered into 2 groups: impact of AI on health care system and practice and AI as a tool for disease screening, diagnosis, and treatment. Topics associated with negative sentiments included concerns about AI replacing human workers, impact of AI on traditional medical diagnostic procedures (ie, patient history and physical examination), accuracy of the algorithm, and entry of IT companies into the health care industry. Concerns about the legal liability for using AI in treating patients were also discussed. Positive topics about AI included the opportunity offered by the technology for improving the accuracy of image-based diagnosis and for enhancing personalized medicine.

CONCLUSIONS

The adoption and implementation of AI applications in the health care industry are eliciting both enthusiasm and concerns about patient care quality and the future of health care professions. The successful implementation of AI-powered technologies requires the involvement of all stakeholders, including patients, health care organization workers, health insurance companies, and government regulatory agencies.

Comprehensive Review on the Use of Artificial Intelligence in Ophthalmology and Future Research Directions

BACKGROUND

Having several applications in medicine, and in ophthalmology in particular, artificial intelligence (AI) tools have been used to detect visual function deficits, thus playing a key role in diagnosing eye diseases and in predicting the evolution of these common and disabling diseases. AI tools, i.e., artificial neural networks (ANNs), are progressively involved in detecting and customized control of ophthalmic diseases. The studies that refer to the efficiency of AI in medicine and especially in ophthalmology were analyzed in this review.

MATERIALS AND METHODS

We conducted a comprehensive review in order to collect all accounts published between 2015 and 2022 that refer to these applications of AI in medicine and especially in ophthalmology. Neural networks have a major role in establishing the demand to initiate preliminary anti-glaucoma therapy to stop the advance of the disease.

RESULTS

Different surveys in the literature review show the remarkable benefit of these AI tools in ophthalmology in evaluating the visual field, optic nerve, and retinal nerve fiber layer, thus ensuring a higher precision in detecting advances in glaucoma and retinal shifts in diabetes. We thus identified 1762 applications of artificial intelligence in ophthalmology: review articles and research articles (301 pub med, 144 scopus, 445 web of science, 872 science direct). Of these, we analyzed 70 articles and review papers (diabetic retinopathy (N = 24), glaucoma (N = 24), DMLV (N = 15), other pathologies (N = 7)) after applying the inclusion and exclusion criteria.

CONCLUSION

In medicine, AI tools are used in surgery, radiology, gynecology, oncology, etc., in making a diagnosis, predicting the evolution of a disease, and assessing the prognosis in patients with oncological pathologies. In ophthalmology, AI potentially increases the patient's access to screening/clinical diagnosis and decreases healthcare costs, mainly when there is a high risk of disease or communities face financial shortages. AI/DL (deep learning) algorithms using both OCT and FO images will change image analysis techniques and methodologies. Optimizing these (combined) technologies will accelerate progress in this area.

Artificial Intelligence in Cancer Research: Trends, Challenges and Future Directions

The World Health Organization (WHO), in their 2022 report, identified cancer as one of the leading causes of death, accounting for about 16% of deaths worldwide. The Cancer-Moonshot community aims to reduce the cancer death rate by half in the next 25 years and wants to improve the lives of cancer-affected people. Cancer mortality can be reduced if detected early and treated appropriately. Cancers like breast cancer and cervical cancer have high cure probabilities when treated early in accordance with best practices. Integration of artificial intelligence (AI) into cancer research is currently addressing many of the challenges where medical experts fail to bring cancer to control and cure, and the outcomes are quite encouraging. AI offers many tools and platforms to facilitate more understanding and tackling of this life-threatening disease. AI-based systems can help pathologists in diagnosing cancer more accurately and consistently, reducing the case error rates. Predictive-AI models can estimate the likelihood for a person to get cancer by identifying the risk factors. Big data, together with AI, can enable medical experts to develop customized treatments for cancer patients. The side effects from this kind of customized therapy will be less severe in comparison with the generalized therapies. However, many of these AI tools will remain ineffective in fighting against cancer and saving the lives of millions of patients unless they are accessible and understandable to biologists, oncologists, and other medical cancer researchers. This paper presents the trends, challenges, and future directions of AI in cancer research. We hope that this paper will be of help to both medical experts and technical experts in getting a better understanding of the challenges and research opportunities in cancer diagnosis and treatment.

Readiness to Embrace Artificial Intelligence Among Medical Doctors and Students: Questionnaire-Based Study

BACKGROUND

Similar to understanding how blood pressure is measured by a sphygmomanometer, physicians will soon have to understand how an artificial intelligence-based application has come to the conclusion that a patient has hypertension, diabetes, or cancer. Although there are an increasing number of use cases where artificial intelligence is or can be applied to improve medical outcomes, the extent to which medical doctors and students are ready to work and leverage this paradigm is unclear.

OBJECTIVE

This research aims to capture medical students' and doctors' level of familiarity toward artificial intelligence in medicine as well as their challenges, barriers, and potential risks linked to the democratization of this new paradigm.

METHODS

A web-based questionnaire comprising five dimensions-demographics, concepts and definitions, training and education, implementation, and risks-was systematically designed from a literature search. It was completed by 207 participants in total, of which 105 (50.7%) medical doctors and 102 (49.3%) medical students trained in all continents, with most of them in Europe, the Middle East, Asia, and North America.

RESULTS

The results revealed no significant difference in the familiarity of artificial intelligence between medical doctors and students (P=.91), except that medical students perceived artificial intelligence in medicine to lead to higher risks for patients and the field of medicine in general (P<.001). We also identified a rather low level of familiarity with artificial intelligence (medical students=2.11/5; medical doctors=2.06/5) as well as a low attendance to education or training. Only 2.9% (3/105) of medical doctors attended a course on artificial intelligence within the previous year, compared with 9.8% (10/102) of medical students. The complexity of the field of medicine was considered one of the biggest challenges (medical doctors=3.5/5; medical students=3.8/5), whereas the reduction of physicians' skills was the most important risk (medical doctors=3.3; medical students=3.6; P=.03).

CONCLUSIONS

The question is not whether artificial intelligence will be used in medicine, but when it will become a standard practice for optimizing health care. The low level of familiarity with artificial intelligence identified in this study calls for the implementation of specific education and training in medical schools and hospitals to ensure that medical professionals can leverage this new paradigm and improve health outcomes.

A Review of Applications of Artificial Intelligence in Gastroenterology

Artificial intelligence (AI) is the science that deals with creating 'intelligent machines'. AI has revolutionized medicine because of its application in several fields across medicine like radiology, neurology, ophthalmology, orthopedics and gastroenterology. In this review, we intend to summarize the basics of AI, the application of AI in various gastrointestinal pathologies till date as well as challenges/ problems related to the application of AI in medicine. Literature search using keywords like artificial intelligence, gastroenterology, applications, etc. were used. The literature search was done using Google Scholar, PubMed and ScienceDirect. All the relevant articles were gathered and relevant data were extracted from them. We concluded AI has achieved major feats in the past few decades. It has helped clinicians in diagnosing complex diseases, managing treatments as well as in predicting outcomes, all in all, which helps doctors from all over the globe in dispensing better healthcare services.

Clinical Wide-Field Retinal Image Deep Learning Classification of Exudative and Non-Exudative Age-Related Macular Degeneration

Background: Age-related macular degeneration (AMD) is a disease that currently affects approximately 196 million individuals and is projected to affect 288 million in 2040. As a result, better and earlier detection methods for this disease are needed in an effort to provide a higher quality of care. One way to achieve this is through the utilization of machine learning. A deep neural network, specifically a convoluted neural network (CNN) can be trained to differentiate between different types of AMD images given the proper training data.

Methods: In this study, a CNN was trained on 420 Optos wide-field retinal images for 70 epochs in order to classify between exudative and non-exudative AMD. These images were obtained and labeled by ophthalmologists from the Martel Eye Clinic in Rancho Cordova, CA. 

Results: After completing the study, a model was created with 88% accuracy. Both the training and validation loss started above 1 and ended below 0.2. Despite only analyzing a single image at a time, the model was still able to accurately identify if the individual had AMD in both eyes or one eye only. The model had the most trouble with bilateral non-exudative AMD. Overall the model was fairly accurate in the other categories. It was noted that the neural network was able to further differentiate from a single image if the disease is present in left, right, or both eyes. This is a point of contention for further investigation as it is impossible for the artificial intelligence (AI) to extrapolate the condition of both eyes from only one image.

Conclusion: This research fostered the development of a CNN that was able to differentiate between exudative and non-exudative AMD. As well as determine if the disease is present in the right, left, or both eyes with a relatively high degree of accuracy. The model was trained on clinical data and can theoretically be used to classify other clinical images it has never encountered before. 

Artificial Intelligence-Based Application Provides Accurate Medical Triage Advice When Compared to Consensus Decisions of Healthcare Providers

Accurate medical triage is essential for improving patient outcomes and efficient healthcare delivery. Patients increasingly rely on artificial intelligence (AI)-based applications to access healthcare information, including medical triage advice. We assessed the accuracy of triage decisions provided by an AI-based application. We presented 50 clinical vignettes to the AI-based application, seven emergency medicine providers, and five internal medicine physicians. We compared the triage decisions of the AI-based application to those of the individual providers as well as their consensus decisions. When compared to the human clinicians’ consensus triage decisions, the AI-based application performed equal or better than individual human clinicians.

Utilizing Artificial Intelligence in Critical Care: Adding A Handy Tool to Our Armamentarium

We have witnessed rapid advancement in technology over the last few decades. With the advent of artificial intelligence (AI), newer avenues have opened for researchers. AI has added an entirely new dimension to this technological boom. Researchers in medical science have been excited about the tantalizing prospect of utilizing AI for the benefit of patient care. Lately, we have come across studies trying to test and validate various models based on AI to improve patient care strategies in critical care medicine as well. Thus, in this review, we will attempt to succinctly review current literature discussing AI in critical care medicine and analyze its future utility based on prevailing evidence.

Can Big Data and Machine Learning Improve Our Understanding of Acute Respiratory Distress Syndrome?

Acute respiratory distress syndrome (ARDS) accounts for 10% of all diagnoses in the Intensive Care Unit, and about 40% of the patients succumb to the disease. Clinical methods alone can result in the under-recognition of this heterogeneous syndrome. The purpose of this study is to evaluate the role that big data and machine learning (ML) have played in understanding the heterogeneity of the disease and the development of various prediction algorithms. Most of the work in the field of ML in ARDS has been in the development of prediction models that have comparable efficacies to that of traditional models. Prediction algorithms have been useful in identifying new variables that may be important to consider in the future, supplementing the unknown information with the help of available noninvasive parameters, as well as predicting mortality. Phenotype identification using an unsupervised ML algorithm has been pivotal in classifying the heterogeneous population into more homogenous classes. Big data generated from ventilators in the form of ventilator waveform analysis and images in the form of radiomics have also been leveraged for the identification of the syndrome and can be incorporated into a clinical decision support system. Although the results are promising, lack of generalizability, "black box" nature of algorithms and concerns about "alarm fatigue" should be addressed for more mainstream adoption of these models.

On classifying sepsis heterogeneity in the ICU: insight using machine learning

OBJECTIVES

Current machine learning models aiming to predict sepsis from electronic health records (EHR) do not account 20 for the heterogeneity of the condition despite its emerging importance in prognosis and treatment. This work demonstrates the added value of stratifying the types of organ dysfunction observed in patients who develop sepsis in the intensive care unit (ICU) in improving the ability to recognize patients at risk of sepsis from their EHR data.

MATERIALS AND METHODS

Using an ICU dataset of 13 728 records, we identify clinically significant sepsis subpopulations with distinct organ dysfunction patterns. We perform classification experiments with random forest, gradient boost trees, and support vector machines, using the identified subpopulations to distinguish patients who develop sepsis in the ICU from those who do not.

RESULTS

The classification results show that features selected using sepsis subpopulations as background knowledge yield a superior performance in distinguishing septic from non-septic patients regardless of the classification model used. The improved performance is especially pronounced in specificity, which is a current bottleneck in sepsis prediction machine learning models.

CONCLUSION

Our findings can steer machine learning efforts toward more personalized models for complex conditions including sepsis.

A novel approach for quantitative assessment of mucosal damage in inflammatory bowel disease

AIMS

One of the main reliable histological features to suggest the diagnosis of inflammatory bowel disease is the presence of significant distortion of the crypt architecture indicating the chronic nature of the disease resulting in mucosal damage. This feature has a considerable intra-observer and inter-observer variability leading to significant subjectivity in colonic biopsy assessment. In this paper, we present a novel automated system to assess mucosal damage and architectural distortion in inflammatory bowel disease (IBD).

METHODS

The proposed system relies on advanced image understating and processing techniques to segment digitally acquired images of microscopic biopsies, then, to extract key features to quantify the crypts irregularities in shape and distribution. These features were used as inputs to an artificial intelligent classifier that, after a training phase, can carry out the assessment automatically.

RESULTS

The developed system was evaluated using 118 IBD biopsies. 116 out of 118 biopsies were correctly classified as compared to the consensus of three expert pathologists, achieving an overall precision of 98.31%.

CONCLUSIONS

An automated intelligent system to quantitatively assess inflammatory bowel disease was developed. The proposed system utilized advanced image understanding techniques together with an intelligent classifier to conduct the assessment. The developed system proved to be reliable, robust, and minimizes subjectivity and inter- and intra-observer variability.

VIRTUAL SLIDES

The virtual slide(s) for this article can be found here: http://www.diagnosticpathology.diagnomx.eu/vs/1797721309305023.