Large Language Models and ChatGPT in Medical Sciences: Foundations, Capabilities, and Challenges

Large Language Models and ChatGPT in Medical Sciences: Foundations, Capabilities, and Challenges Authors: Amir Tahavvori Northwell Health, The Feinstein Institutes for Medical Research Reza Judi Chelan YasujUniversity of Medical Sciences Sima Aminoleslami Islamic Azad University, (Ta.C.), Tabriz Omid Fakharzadeh Moghadam Mashhad University of Medical Sciences Leili Haghighi Islamic Azad University Tehran Medical Sciences, Dental Branch Yeganeh Abdian Bahçeşehir University Mohammad Hossein Naderi Faculty of Dentistry, Shahed University Amin Kanani Guilan University of Medical Sciences Nikta Taghipour Jahrom University of Medical Sciences LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 1 Amirali Farshid Ardabil University of Medical Sciences Seyedeh Tabasom Nejati Hormozgan University of Medical Sciences Pouya Kalantari Tarbiat Modares University Pouran Varvani Farahani Cyprus International University Sahar Jafarpour Iran University of Medical Sciences Amirreza Bahari Ardabil University of Medical Sciences Seyyed Erfan Hosseini Asl Ardabil University of Medical Sciences Seyyedeh Baran Hosseini Asl Ardabil University of Medical Sciences Zeinab Mohammadi Aja University of Medical Sciences Houman Bahrami Rad Tarbiat Modares University Sajad Teimoury Shiraz University of Medical Sciences Tara Abdolahyfard Shiraz University of Medical Sciences AMIR TAHAVVORI 2 Edris Habibi HamadanUniversity of Medical Sciences Mohammad Eslami Shahid Beheshti University of Medical Sciences Saman Abdollahpour Shahid Beheshti University of Medical Sciences Sanaz Amiri Marbini University Medical Center Hamburg-Eppendorf Niloofar Taheri Shahroud University of Medical Sciences Dariush Moradi Tehran University of Medical Sciences LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 3 AMIR TAHAVVORI 4 Book Details: Publisher: Kindle Publication Date: November 2025 Language: English Dimensions: 5 x 0.39 x 8 inches © Kindle and PreferPub 2025 ISBN-13: 979-8272445780 This peer-reviewed book is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specially the rights of translation, reprinting, result of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 5 AMIR TAHAVVORI 6 1. Large Language Models and ChatGPT for the Management of Renal Diseases 2. Large Language Models and ChatGPT for the Management of Neurological Diseases 3. Large Language Models and ChatGPT for the Management of Gastrointestinal Diseases 4. Large Language Models and ChatGPT for the Management of Cardiac Diseases 5. Large Language Models and ChatGPT for the Management of Dermatological Diseases 6. Large Language Models and ChatGPT for the Management of Oncological Conditions 7. Large Language Models and ChatGPT for the Management of Oral and Dental Diseases 8. Large Language Models and ChatGPT for the Management of Infectious Diseases 9. Large Language Models and ChatGPT for the Management of Urological and Gynecological Disorders 10. Large Language Models and ChatGPT for the Management of Other Diseases Contents Chapter LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 7 AMIR TAHAVVORI 8 1. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF RENAL DISEASES Background Renal diseases occupy a difficult place in clinical medicine, requiring detailed assessment, longterm monitoring, and constant adjustment to therapy. The complexity of kidney physiology and the vast range of causes behind renal impairment create a heavy demand on medical systems. Physicians must track numerous laboratory values, evaluate imaging, interpret genetic findings, and reconcile these elements into a coherent view of disease progression. This process consumes time and depends on coordination between multiple specialists. In this environment, large language models such as ChatGPT have begun to show potential as tools that can assist clinicians and patients in understanding and managing kidney disorders. Language models are not entirely new to medicine, but their capacity to process and generate natural text at scale has introduced new possibilities. They are trained on immense amounts of written material and learn to understand relationships between words, phrases, and concepts. In practical terms, they can 9 summarize patient notes, organize research findings, and generate explanations that help bridge communication gaps between healthcare professionals and patients. In nephrology, which depends heavily on written documentation and analytical reasoning, these capabilities align well with existing needs. Foundations of LLMs in Medical Context Large language models are neural networks that learn to predict words in a sequence. They operate on patterns rather than rules, which allows them to generate coherent text that resembles human thought. ChatGPT, one of the most recognized examples, has been finetuned using reinforcement learning from human feedback. This process aligns its outputs with acceptable medical and conversational standards. When applied to nephrology, its understanding of structured and unstructured data offers opportunities to improve how information is processed and presented. Traditional data systems in nephrology can be fragmented. A patient’s renal function tests may sit in one database, while imaging, pathology reports, and clinical notes exist elsewhere. LLMs can integrate these streams by transforming raw text into a unified narrative. For instance, a model could summarize the history of a patient with chronic kidney disease, identifying key milestones AMIR TAHAVVORI 10 such as changes in medication or stages of decline in glomerular filtration rate. The generated summary would not replace medical judgment, but it would present essential details concisely, saving clinicians time and reducing oversight. Applications in Diagnosis and Decision Support The early identification of kidney disease is crucial, yet it often goes unnoticed until late stages. Symptoms are nonspecific, and laboratory abnormalities may be subtle. LLMs can aid in the early detection of disease patterns by analyzing longitudinal patient data. A model could flag consistent reports of elevated creatinine or proteinuria in patient records, drawing attention to those requiring follow-up. By doing so, it acts as a supportive system for physicians, ensuring that small but important trends are not overlooked. ChatGPT and similar models can also contribute to differential diagnosis. For a patient presenting with edema, fatigue, and hypertension, the system might outline potential causes, including nephrotic syndrome or diabetic nephropathy, and suggest relevant tests. The goal is not to replace a physician’s judgment but to provide a structured approach that enhances reasoning. By rapidly referencing clinical guidelines and evidence, the model can help doctors refine their diagnostic paths and avoid unnecessary testing. In emergency settings, where quick decisions LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 11 are vital, a language model could assist by offering immediate information about management protocols for acute kidney injury. It might summarize steps for fluid resuscitation, indications for dialysis, or appropriate drug dosing in renal impairment. The model can also clarify contraindications, preventing medication errors that might worsen renal function. When integrated into electronic health record systems, these functions could support physicians in realtime, improving accuracy and patient safety. Patient Education and Engagement Patient understanding plays a central role in managing chronic kidney disease. Many treatment plans require strict adherence to diet, fluid control, and medication schedules. Patients often find medical explanations confusing or intimidating, which reduces compliance. LLMs like ChatGPT can bridge this gap by translating complex terminology into everyday language. A patient who receives a diagnosis of stage three chronic kidney disease might ask the model to explain what that means. The model could respond with a clear, empathetic explanation of kidney function, emphasizing lifestyle measures such as limiting salt intake and monitoring blood pressure. It could generate customized dietary guidance aligned with established medical advice, helping patients make practical adjustments without feeling overwhelmed. AMIR TAHAVVORI 12 LLMs can also support communication between clinicians and patients outside of appointments. Through secure chat systems, they can answer common questions, remind patients of medication times, and provide reassurance about routine symptoms. When carefully supervised, this form of assistance can extend the reach of care teams, reducing hospital readmissions and promoting continuity of care. Role in Dialysis Management Dialysis is a data-intensive process. Each session produces information on blood flow rates, ultrafiltration volumes, and biochemical results. Reviewing and interpreting these metrics for each patient can be time-consuming. LLMs can analyze and summarize these daily reports, highlighting changes that may need medical attention. For instance, they could flag recurrent hypotension episodes or insufficient clearance rates, prompting a review of the dialysis prescription. For patients undergoing home dialysis, ChatGPT could serve as an accessible support companion. It can guide them through setup procedures, remind them about hygiene protocols, and help them interpret machine readings. These functions can reduce anxiety and improve adherence. The conversational style of a model also makes learning less formal and more interactive, encouraging patients to take active roles in their own care. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 13 LLMs can further assist in resource management for dialysis centers. By reviewing appointment schedules, equipment logs, and staff availability, the system can propose efficient timetables that minimize waiting times and machine downtime. When paired with predictive analytics, it might also anticipate supply shortages based on consumption patterns, allowing smoother operations. Kidney Transplantation and Postoperative Care Renal transplantation demands coordination between multiple disciplines and meticulous monitoring. LLMs can streamline this process by compiling pre-transplant evaluations, summarizing donor compatibility reports, and tracking immunosuppressive therapy adherence. During follow-up, they can identify early signs of graft dysfunction through review of lab data and clinical notes. In patient communication, ChatGPT can explain the significance of post-transplant tests, clarify medication side effects, and provide reminders for lab draws or clinic visits. This interaction is particularly helpful in regions where access to transplant specialists is limited. For example, a patient in a rural area could consult a model to understand fluctuations in creatinine levels or discuss the importance of avoiding certain overthe-counter drugs. AMIR TAHAVVORI 14 For transplant teams, LLMs can serve as research aids by reviewing outcomes data, analyzing rejection rates, and helping in the design of followup protocols. Their ability to summarize multiple studies in natural language allows clinicians to keep pace with evolving literature without dedicating hours to manual reading. Data Analysis and Research Advancement Nephrology research produces a steady flow of studies on topics like glomerular disease mechanisms, dialysis innovations, and population risk modeling. However, the scale of new information makes synthesis difficult. LLMs can process vast numbers of publications and extract key patterns. For example, they can identify recurring biomarkers associated with renal fibrosis or summarize consensus statements from clinical trials. Researchers can use ChatGPT to draft literature reviews, outline grant proposals, or generate preliminary interpretations of data. While these drafts require human revision, they reduce the initial workload and improve efficiency. For largescale genomic or proteomic studies, LLMs can link textual results from research papers to numerical datasets, facilitating hypothesis generation. This connection between unstructured and structured information may accelerate discoveries in precision nephrology. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 15 Integration with Electronic Health Records Electronic health record systems collect a wealth of data but often lack the ability to present it in an intuitive way. Physicians may spend considerable time locating information within multiple pages of reports. LLMs can reorganize this data into clear narratives. When a nephrologist opens a patient file, the model could present a summary: current renal function, medication adherence, blood pressure trends, and recent imaging results. Such systems could also detect inconsistencies or omissions in documentation. For example, if a note mentions new edema but lacks follow-up lab results, the model could prompt the physician to order relevant tests. These functions do not require full autonomy but serve as practical tools to maintain completeness and consistency. When combined with voice recognition technology, ChatGPT could enable physicians to dictate notes during consultations. The model would convert speech into structured text, automatically integrating relevant codes and terms. This feature reduces administrative burden, allowing more time for patient interaction. Addressing Ethical and Technical Challenges Despite their potential, LLMs raise ethical and AMIR TAHAVVORI 16 operational concerns. Medical data must be handled with strict confidentiality, and any use of AI in healthcare must comply with privacy regulations. If ChatGPT or similar models are connected to clinical databases, robust encryption and access control are essential. Institutions must ensure that personal identifiers remain protected, particularly when data is processed through external servers. Another challenge lies in accuracy. While models can produce persuasive text, they sometimes generate incorrect or misleading statements. In medicine, such errors could have serious consequences. Every output must be reviewed by qualified professionals before application in patient care. Maintaining a human in the loop is not only advisable but necessary. Bias within training data also requires attention. If the data used to develop an LLM reflects demographic imbalances, the model may reproduce those biases in its predictions. In nephrology, this could lead to unequal recommendations for certain populations. Developers and clinicians need to evaluate performance across diverse groups to ensure fairness. Finally, there is the issue of accountability. When AI contributes to medical decision-making, determining responsibility for outcomes becomes complex. Clear frameworks should define how LLM-generated content is used and who bears LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 17 Neurological disorders are the leading cause of disability and the second leading cause of death worldwide. Over the past three decades, mortality and disability rates associated with these conditions have increased substantially. This upward trend is projected to continue globally, driven primarily by demographic changes such as population aging and overall growth. The ongoing expansion in the number of affected individuals highlights a significant gap, as current prevention and management strategies are insufficient to address the effects of these demographic shifts. Tackling this growing public health challenge requires urgent, coordinated, and evidence-based action. Given the existing strain on healthcare systems and limited research capacity, it is essential to establish clear priorities to guide policymakers, governments, and funding organizations in developing and implementing targeted initiatives for prevention, clinical care, and research. Foundations of Large Language Models in Medicine Large language models (LLMs), such as ChatGPT, are popular artificial intelligence (AI) systems designed to generate human-like language. LLMs are capable of processing natural language tasks such as automatic summarization and question answering. They assist clinicians in clinical decision-making by analyzing patient AMIR TAHAVVORI 30 information, medical data, diagnosis, monitoring, and symptom assessment. In addition, LLMs can serve as teaching assistants in patient care. In neurology, AI has the potential to analyze patient records and neuroimaging reports, improve diagnostic accuracy, and provide personalized treatment plans for neurological conditions. Architectures of LLMs LLMs are constructed using neural network architectures that detect complex patterns in natural language and generate coherent text. Although several types of language models have been introduced in the literature, most modern LLMs are based on transformer architectures, which are designed to process long text sequences efficiently. These networks consist of multiple layers, each containing several attention heads and feedforward neural networks. The attention heads determine which parts of the input to prioritize, supporting contextual understanding through matrix operations, while the feedforward layers further analyze these outputs to capture higher-order relationships. Training Methods for LLMs Training LLMs requires extensive natural language datasets to optimize the network’s internal parameters. During training, models predict the next word, fill in masked tokens, or generate entire sentences, minimizing LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 31 the difference between predicted and actual outcomes. Several training approaches are used, including supervised, unsupervised, and selfsupervised learning. Supervised learning is the most common, involving training on labeled datasets where the model learns to predict correct labels. When labeled data are unavailable, unsupervised learning allows models to detect patterns within large text corpora. A more recent method, self-supervised learning, has proven particularly effective in identifying complex relationships between words and phrases. In this approach, models are trained using artificially generated labels to predict missing data, such as answering questions or completing sentences based on contextual cues. Types of LLMs There are several types of LLMs, each with distinct features and applications. ELMo (Embeddings from Language Models) generates contextualized word representations that enhance performance in various NLP tasks. BERT (Bidirectional Encoder Representations from Transformers) is primarily designed for text classification and comprehension through bidirectional encoding. GPT (Generative Pretrained Transformer) focuses on generative NLP applications such as text generation, summarization, and question answering. These models have collectively advanced the field of NLP and expanded its AMIR TAHAVVORI 32 practical applications in medicine and other disciplines. Automated Medical Documentation and Report Generation For models such as GPT-4, which are designed to respond effectively to prompts, performance is improved through human feedback and reinforcement learning. Language modeling, a central NLP task known as autoregression, involves predicting the next word in a sequence based on preceding context. Because LLMs can analyze vast datasets, including medical records and patient interviews, they generate highquality, nuanced text that captures detailed symptoms and experiences. This capability makes them valuable resources for neurological research, clinical reporting, and automated documentation. LLMs in Early Detection and Cognitive Rehabilitation in Neurology LLMs hold significant potential for analyzing language patterns in patients’ speech and writing, which may help detect subtle cognitive changes that human observers often overlook. For example, training LLMs on language data from individuals with Parkinson’s disease or those at high risk for Huntington’s or Alzheimer’s disease could reveal gradual variations in vocabulary, sentence structure, or conceptual complexity. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 33 Early identification of such linguistic shifts may enable timely interventions and personalized rehabilitation strategies. LLMs can also improve clinicians’ ability to identify language deficits following traumatic brain injury or tumor surgery, allowing for more precise cognitive therapy. In support of this, recent initiatives such as ADReSS and ADReSSo have encouraged the development of automated tools for analyzing speech, acoustic, and linguistic features to detect cognitive decline. Researchers have also used GPT-based models to predict dementia from spontaneous speech. Additionally, LLMs can contribute to cognitive rehabilitation by generating word games or storytelling activities tailored to a patient’s language ability. By tracking progress, these models can dynamically adjust task difficulty, offering customized cognitive stimulation. Common neuropsychological assessments for cognitive impairment include tests measuring semantic and phonemic fluency, which can also be adapted for AI-driven applications. Current Trends in NLP Research in Neurology Text classification is one of the most frequent NLP tasks, in which LLMs learn to categorize texts by identifying patterns between examples and corresponding labels. In clinical practice, this involves processing medical records, patient histories, and clinical trial reports. LLMs can AMIR TAHAVVORI 34 extract key information from radiology reports in emergency departments. This capability is particularly important in stroke management, where rapid intervention is crucial and communication may be impaired by neurological deficits. In such situations, LLMs can assist clinicians by emphasizing critical neuroimaging findings. Applications of LLMs in the Management of Neurodegenerative Disorders Neurodegenerative diseases (ND) are irreversible, age-related neurological disorders that lead to the progressive loss of neurons. These include Alzheimer’s disease, Parkinson’s disease, Amyotrophic Lateral Sclerosis (ALS), Multiple Sclerosis (MS), and Huntington’s disease (HD), with Alzheimer’s and Parkinson’s being the most common types. LLMs such as ChatGPT offer opportunities for improving disease management by assisting clinicians, supporting early diagnosis, and enabling personalized treatment plans. Alzheimer’s Disease (AD) Alzheimer’s disease (AD) is the most common form of dementia. It is characterized by cognitive decline, memory loss, and abnormal behavior. LLMs offer greater potential for earlier and more accurate AD diagnosis by deeply analyzing and integrating multiple data types, such as LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 35 clinical information, neuroimaging, and genetic data, capturing subtle cognitive features that traditional methods might miss. Additionally, LLMs provide fine-grained linguistic analysis that can detect language-based cognitive decline, leading to higher diagnostic accuracy. They can also design novel drug molecules for AD therapeutics. LLMs can act as assistants to significantly support patients with AD by providing emotional support to reduce loneliness associated with dementia or by helping with daily tasks. Moreover, information retrieval is one of the key capabilities of LLMs, allowing them to compile and resynthesize knowledge and present it in an understandable form. LLMs can not only aid clinicians in developing therapy strategies but can also deliver reminiscence therapy, which may enhance emotional well-being and improve cognitive function. Parkinson’s Disease (PD) Parkinson’s disease (PD) is the second most common neurodegenerative disorder. It is characterized by the progressive loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc) and the accumulation of misfolded α-synuclein, leading to motor symptoms such as bradykinesia, tremor, rigidity, and later postural instability. LLMs have the capability to process extensive data and generate valuable insights through near real-time analysis, AMIR TAHAVVORI 36 which enables the development of a consistent ontology for PD monitoring. Furthermore, LLMs provide effective personalized PD management by tailoring care to patient-specific needs and adjusting medication dosage and frequency. They can also analyze linguistic features in depth, offering early detection since language impairments often precede physical symptoms. Moreover, regression analyses have demonstrated that large language model–derived linguistic feature spaces can predict Parkinson’s disease severity, as quantified by UPDRS scores. Multiple Sclerosis (MS) Multiple sclerosis (MS) is an acquired autoimmune disease of the central nervous system (CNS), particularly affecting young adults. MS is a neurodegenerative disorder that leads to demyelination and axonal loss within the CNS. Similar to other neurological disorders, LLMs have potential for accurate diagnosis and effective management of MS, thereby assisting specialists in clinical examination. In addition, LLMs can act as rehabilitation therapists by creating personalized rehabilitation plans for MS patients, considering disease progression and individual patient conditions. Moreover, LLMs possess the ability to translate complex medical terminology used by clinicians into more understandable language for patients, helping them gain a clearer understanding of their condition. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 37 Amyotrophic Lateral Sclerosis (ALS) Amyotrophic Lateral Sclerosis (ALS) is a progressive, heterogeneous neurodegenerative disorder characterized by the loss of upper and lower motor neurons. LLMs can promptly transform condensed user input into fluent dialogue, significantly reducing the motor communication burden for ALS patients who rely on eye-typing systems. Additionally, LLMs, particularly ChatGPT, which utilizes Natural Language Processing (NLP) techniques, can enhance ALS patient care by analyzing and accelerating emotional responses, thereby assisting clinicians in both the management and diagnosis of ALS. Applications of LLMs in the Management of Epilepsy Epilepsy is a chronic brain disorder that can develop at any age. It has various causes and is characterized by seizures, which may present in multiple forms. LLMs can extract clinical information from diverse text sources, such as health records and patient reports, leading to the early and accurate diagnosis of epilepsy. Additionally, LLMs assist specialists in developing personalized treatment plans based on the individual patient’s condition and the stage of the disorder by analyzing the vast amount of available clinical data. One study suggests that AMIR TAHAVVORI 38 fine-tuning an LLM offers a novel method for retrieving seizure frequency data from electronic health records, which helps clinicians obtain more accurate information about how often patients experience seizures. This precise data extraction facilitates the analysis of the effectiveness of anti-seizure medications, ultimately improving treatment strategies for epilepsy. Moreover, LLMs can analyze linguistic features, providing additional diagnostic insights by examining how patients express their experiences. Strengths and Opportunities LLMs, including specialized systems and ChatGPT, have demonstrated more accurate and efficient diagnostic performance compared to active neurologists in complex clinical settings. These models have achieved higher scores in differential diagnosis and provided credible and relevant resources at a significantly faster rate. It is noteworthy that these models have demonstrated performance exceeding the human average in functional neurology board exams, excelled in both low and high-level cognitive tasks, and shown significant potential for supporting clinical decision-making and integrating neurological knowledge. The strengths of LLMs in the field of neurology include the ability to quickly combine vast amounts of data, accurately process natural language in medical documentation, support automated screenings, and help simplify LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 39 complex, data-driven workflows. By providing accurate and empathetic explanations, they enhance the patient education process while simultaneously helping physicians stay up to date with rapidly changing medical knowledge. Significant opportunities exist in the field of customizing LLMs for specialized branches of neurology, including developing their applications in neuropsychology, neuropsychiatry, and various research areas, as well as utilizing these models for rapid and widespread extraction of scientific literature, generating innovative hypotheses, and conducting remote assessments. This will be particularly efficient in resource-limited settings and telemedicine services. As efforts continue to overcome current limitations in reasoning depth and expertise in specific domains, LLMs are predicted to play a key role in significantly improving quality, increasing productivity, and expanding access to neurological care services. Ethical and Legal Considerations Ethical, legal, and regulatory challenges are among the issues that LLMs like ChatGPT face in the field of neurology. The main ethical concerns include the potential disclosure of sensitive patient data, the presence of potential information biases, and the generation of AI hallucinations, which can negatively impact clinical decision-making. Additionally, it is essential that models be rigorously validated AMIR TAHAVVORI 40 using representative datasets and that their performance limitations and boundaries across different populations be transparently labeled. Legally, responsibilities related to errors, data misuse, or the incorrect use of outputs from LLMs are still not precisely defined. Therefore, we need to develop clearer guidelines to specify responsibilities among developers, physicians, and institutions. Additionally, adhering to privacy regulations such as the Health Insurance Portability and Accountability Act (HIPAA) and the General Data Protection Regulation (GDPR) is crucial for maintaining the security of individuals' identifiable data when training and deploying LLMs in the field of neurology. For responsible adoption, it is essential to establish unified ethical frameworks that include principles such as transparency, interpretability, traceability, privacy preservation, and fair treatment. Additionally, quality control and continuous monitoring, scientific and ethical integrity, intellectual property rights, and active stakeholder participation are considered important components of these frameworks. Therefore, it is expected that regulatory oversight bodies will establish flexible and proportionate regulations for large language models in the field of clinical neurology to keep pace with advancements. Limitations and Challenges LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 41 Despite the advantages of LLMs in the management of neurological diseases, their application in neurology presents several limitations and challenges. As mentioned earlier, LLMs can analyze linguistic features to offer accurate diagnoses, but the limited size of speech data from patients makes this approach less reliable. Additionally, LLMs sometimes generate hallucinations, producing outputs that seem reasonable but are entirely incorrect. Moreover, patients with ALS often experience impairments in facial expressions and body language, making emotion recognition more difficult. This limitation can result in misinterpretation of the patient’s emotional and mental state, thereby reducing the reliability of AI-assisted assessments. The use of patient data also increases the risk of private patient information being leaked. Anonymizing health data does not guarantee complete security, as some algorithms have been capable of re-identifying patients. Furthermore, LLMs cannot fully replace physicians because their outputs require validation, and humans remain responsible for oversight and ethical decisionmaking. Conclusion Large language models (LLMs) are transforming neurological care by improving diagnostic precision, aiding in personalized treatment approaches, and strengthening communication AMIR TAHAVVORI 42 between patients and clinicians. Their ability to analyze extensive and varied data sources, spanning clinical records, imaging reports, and patient narratives, provides unique opportunities to identify disease characteristics that traditional methods might overlook. Nonetheless, the integration of LLMs into clinical practice is still in its early stages and is hindered by issues such as data privacy concerns, hallucinations, and the need for stringent validation. Ultimately, LLMs should be seen as supplementary tools rather than replacements for physicians. They enhance medical decision-making through advanced computational insights. Future research that integrates LLMs with multimodal technologies and establishes clear ethical and regulatory frameworks could transform these models into powerful allies for improving outcomes in patients with neurological disorders. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 43 AMIR TAHAVVORI 44 Background Gastrointestinal diseases affect millions worldwide, encompassing a spectrum that ranges from mild functional disorders to severe chronic illnesses like inflammatory bowel disease, cirrhosis, and gastrointestinal cancers. The complexity of the digestive system and its constant exposure to external and internal influences make diagnosis and treatment challenging. Many conditions share overlapping symptoms, and diagnostic work often relies on combining endoscopic findings, imaging, histopathology, and biochemical data. These processes can be fragmented across healthcare systems and time-consuming for physicians. In recent years, artificial intelligence has begun to address these inefficiencies. Among the emerging technologies, large language models such as ChatGPT have shown potential to reshape how clinicians and patients interact with medical information related to gastrointestinal health. Language models are designed to interpret and produce human-like text. When applied to healthcare, they can summarize medical 3. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF GASTROINTESTINAL DISEASES 45 notes, assist in decision-making, and translate technical language into terms patients can understand. The digestive system’s disorders, which require ongoing communication between gastroenterologists, radiologists, nutritionists, and patients, present a fertile ground for such technology. ChatGPT, a product of OpenAI, has gained attention for its ability to process complex input and provide structured, readable output that supports education, documentation, and patient care. Understanding LLMs in the Context of Medicine Large language models are trained on immense datasets that include books, articles, and, increasingly, medical literature. Their architecture allows them to recognize subtle associations between concepts. When exposed to medical information, they can reproduce patterns of reasoning similar to expert analysis. In the field of gastroenterology, these systems can assist in integrating clinical data, laboratory findings, and imaging reports to generate summaries that are understandable and relevant to the user’s question. Unlike conventional data processing tools, which depend on strict coding and structured formats, LLMs excel at handling unstructured text. A single hospital encounter may generate several reports, including discharge summaries, AMIR TAHAVVORI 46 pathology interpretations, and patient messages. A language model can merge these sources into one coherent report, identifying the main findings and highlighting follow-up needs. This function reduces the time physicians spend sorting through records and allows them to focus on interpretation and patient counseling. ChatGPT’s conversational interface also provides a natural way to access information. When a clinician asks for a summary of the latest guidelines on Barrett’s esophagus or colon cancer screening, the model can produce a concise answer referencing up-to-date recommendations. This efficiency gives doctors more opportunities to focus on individualized decision-making rather than manual searches. Diagnostic Assistance and Decision Support The diagnostic process in gastroenterology often involves analyzing symptoms that can be vague or overlapping. A patient presenting with abdominal pain, bloating, or changes in bowel habits could have a range of possible conditions, from irritable bowel syndrome to celiac disease or colon cancer. LLMs can help narrow possibilities by aligning patient symptoms with established diagnostic criteria. For instance, they can compare reported features to the Rome IV criteria for functional bowel disorders and suggest appropriate next steps. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 47 When paired with clinical databases, LLMs can assist in recognizing patterns that point to specific pathologies. For example, if a patient’s records show intermittent rectal bleeding, iron-deficiency anemia, and a history of polyps, the system could prompt the physician to consider colorectal malignancy and verify whether screening colonoscopy is due. This function acts as a supportive safety net, reducing the likelihood of missed diagnoses. In acute settings, where time is limited, LLMs can help by summarizing standard protocols. A physician managing a patient with gastrointestinal bleeding could request guidance on fluid resuscitation, transfusion thresholds, or medication adjustments. The system might produce a summary consistent with clinical guidelines, helping reinforce best practices. Although the physician must confirm every recommendation, such tools can reduce cognitive load and improve consistency in care delivery. Patient Communication and Education Gastrointestinal disorders often require detailed explanation and patient cooperation. Diet, lifestyle, and adherence to medication all affect outcomes. Yet many patients find the medical terminology confusing and feel anxious about their condition. ChatGPT can help by rephrasing complex medical language into accessible AMIR TAHAVVORI 48 explanations. A patient newly diagnosed with ulcerative colitis may ask, “What does this mean for me?” The model could generate a plain-language description of how inflammation affects the colon and what treatments aim to achieve. It might also list common triggers and provide general dietary guidance drawn from reliable sources. This interaction allows patients to learn at their own pace, reducing dependence on rushed clinical encounters. In chronic conditions like irritable bowel syndrome, ChatGPT could assist in ongoing selfmanagement. Patients could ask daily questions about meal choices or symptoms, receiving responses aligned with established medical advice. When integrated into healthcare systems, such AI assistants could remind patients to take medications, schedule follow-ups, or record symptoms. These functions strengthen engagement and continuity of care, two factors strongly associated with better outcomes in gastrointestinal health. Documentation and Administrative Support The administrative burden in gastroenterology can be substantial. Clinicians often spend more time documenting than speaking to patients. Endoscopy reports, discharge summaries, and multidisciplinary meeting notes all require LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 49 cardiovascular medicine by supporting clinical decision-making, formulating personalized therapies, and enhancing patient–clinician communication. However, ethical challenges, including data privacy and algorithmic fairness, require rigorous scrutiny to realize AI’s potential. Computational systems, such as ChatGPT, are designed to synthesize medical knowledge and patient-specific insights. Integrating ChatGPT into cardiac surgery improves outcome prediction and risk assessment by incorporating patientspecific factors, including mental health and social support. Nevertheless, cognitive biases, information constraints, and systemic barriers, such as technical infrastructure and policy frameworks, demand careful evaluation to realize its benefits. The rapid evolution of machine learning (ML) technologies has profoundly transformed medical research and clinical practice. Large language models (LLMs) excel in drug target discovery, accelerating the identification of novel therapeutic pathways. These models show strong performance in complex clinical scenarios, passing USMLE-level examinations and offering insightful explanations. ML approaches consistently outperform traditional risk stratification models, improving predictive accuracy for cardiovascular outcomes, renal insufficiency, and other chronic conditions. AMIR TAHAVVORI 62 They generate rapid, reliable responses based on guidelines, such as for cardiopulmonary resuscitation (CPR). However, their use in complex clinical scenarios requiring advanced expertise remains underexplored. By adapting protocols to include open-ended assessments, ML tools can enhance diagnostic and prognostic capabilities, potentially improve patient outcomes, and reduce healthcare costs. Cardiovascular diseases, particularly multivessel coronary artery disease (CAD) and severe aortic stenosis, present complex clinical challenges due to their intricate interplay of anatomical, procedural, and clinical factors. Few studies have explored the potential of LLM methods to deliver rapid, evidence-based recommendations in these intricate scenarios. Previous research has demonstrated promising concordance between LLM-derived insights and heart team (HT) decisions, especially in managing valvular heart diseases. However, applying these models to multivessel CAD, with its distinct complexities, remains underexamined. As cardiovascular technology advances, it generates vast datasets, significantly increasing the workload of medical professionals. This surge in data renders accurate and timely detection of cardiovascular disease increasingly demanding. LLM, a subfield dedicated to enabling computer programs to learn and interpret data features, has emerged as a LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 63 valuable tool for assisting in the diagnosis, prediction, and classification of cardiovascular diseases. Nevertheless, LLM approaches rely on manual feature engineering, including extraction, selection, and reduction, which often struggle to identify the most relevant features from patient data. In cardiology, LLMs and ChatGPT demonstrate wide-ranging potential across education, diagnostics, and research. They can provide patients with informative material on cardiovascular pathology, behavioral modifications, and emergency response steps, helping them better understand their health conditions. Studies indicate that these systems generally produce accurate, detailed, and safe responses to typical questions regarding risks and prevention strategies. For clinicians, LLMs can propose differential diagnoses, suggest diagnostic tests, and recommend treatment pathways consistent with established guidelines. In managing heart failure, potential applications include risk assessment, symptom interpretation, and the generation of personalized medication recommendations, which may reduce readmission rates by improving self-care support. These tools also enhance operational efficiency by managing records, generating summaries, and assisting in identifying appropriate candidates for clinical research. AMIR TAHAVVORI 64 Their strengths are most evident in controlled settings, where evaluations have shown strong agreement with expert opinions in developing care strategies for complex clinical scenarios, maintaining stable performance across varying levels of difficulty. These models have successfully completed portions of professional examinations, demonstrating proficiency in analyzing guidelines and explaining concepts with clarity. Integrating LLMs with diverse inputs, such as health records and sensor data, allows for more accurate forecasting of complications and identification of at-risk populations for timely intervention. In academic research, they support the compilation of literature reviews, hypothesis generation, and manuscript preparation, thereby accelerating the dissemination of scientific knowledge. However, several obstacles limit their complete integration into cardiac care. A primary concern is their tendency to produce inaccurate information or fabricate references, which can mislead users and pose safety risks. Biases within training data may exacerbate existing healthcare inequalities, while ethical concerns persist regarding data privacy, informed consent, and accountability for AI-generated recommendations. Current versions also face difficulties in visual interpretation, limiting their usefulness in analyzing electrocardiograms or ultrasound images, both of which are critical for cardiovascular assessment. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 65 Foundations of LLMs in Cardiovascular Medicine LLMs are built upon deep learning architectures that predict word sequences based on context. They rely on layers of interconnected nodes, each refining patterns and relationships between words and ideas. ChatGPT, one of the most advanced LLMs, is based on a transformer architecture Furthermore, gaps in contextual understanding and expert-level reasoning can result in errors during high-stakes decision-making situations that demand nuanced evaluation. Financial barriers, such as high implementation costs, along with the potential reduction of human interaction in patient care, also warrant careful consideration. Addressing these challenges requires comprehensive validation, regulatory oversight, and the integration of human expertise with computational systems. Future improvements may include advanced versions equipped with image analysis capabilities and real-time data integration to broaden their scope of application. This section examines the foundational principles of LLMs and ChatGPT, evaluates their specific roles in managing cardiovascular conditions, and discusses the challenges associated with their adoption, aiming to guide both researchers and clinicians in their responsible and effective implementation. AMIR TAHAVVORI 66 that enables it to process large amounts of data efficiently. By training on medical literature and clinical data, such models can interpret terminology and apply reasoning to real-world situations. Cardiology produces enormous amounts of text and numerical information, from imaging reports to electrocardiograms, lab results, and discharge summaries. LLMs can process these diverse data types by focusing on the text-based components and summarizing them into clinically meaningful narratives. For instance, they can synthesize the history of a patient with coronary artery disease, noting when angina symptoms first appeared, how medications have changed, and how recent test results align with previous patterns. This level of summarization allows healthcare teams to work more efficiently, with less time spent searching for key facts across multiple reports. LLMs also provide the foundation for predictive and supportive functions when paired with other data models. They can be linked with electronic health record systems to help interpret evolving patient data and alert clinicians to potential issues such as drug interactions or deteriorating cardiac function. Although the final judgment must always rest with the physician, these tools can help identify early signs of risk before symptoms become severe. Diagnostic Assistance and LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 67 Decision Support Diagnosing cardiac conditions often requires synthesizing information from multiple sources. Patients may present with chest pain, shortness of breath, or fatigue—symptoms that could signify anything from benign anxiety to acute myocardial infarction. LLMs can assist in these complex evaluations by providing structured overviews of likely causes based on symptom descriptions, medical history, and known risk factors. When paired with electronic records, ChatGPT can highlight patterns that suggest disease progression. For example, if a patient’s records show increasing shortness of breath, declining ejection fraction, and elevated BNP levels, the model could flag possible worsening heart failure. It could also remind clinicians of recommended next steps according to established guidelines. Such applications do not replace human expertise but function as intelligent assistants, offering prompts that ensure thorough consideration of relevant data. In emergency departments, where time is critical, ChatGPT could support physicians by summarizing key elements from a patient’s file, including prior cardiac interventions, medication lists, and known allergies. This rapid review reduces time spent navigating digital systems during life-threatening situations. The model could also generate concise discharge summaries or patient instructions after stabilization, AMIR TAHAVVORI 68 improving clarity and communication between healthcare teams. For diagnostic imaging, while LLMs cannot interpret raw echocardiograms or angiograms on their own, they can provide structured descriptions when paired with visual AI systems. The combination of text and image analysis can help radiologists and cardiologists produce consistent, accurate reports and reduce variability between readers. Patient Education and Self-Management Heart disease management extends far beyond the clinic. Patients must understand their condition, adhere to medication schedules, and make significant lifestyle changes. However, many struggle to interpret medical language or remember complex instructions. ChatGPT and other language models can help bridge this communication gap. A patient recently diagnosed with hypertension, for instance, could use ChatGPT to understand how blood pressure affects the heart and why adherence to medication is crucial. The model could explain the difference between systolic and diastolic values or describe how sodium intake influences vascular resistance. Such conversations make medical advice more approachable, encouraging long-term compliance. For those living with chronic conditions like LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 69 heart failure, ChatGPT can serve as a daily support tool. Patients could ask questions about diet restrictions, physical activity, or symptom monitoring. With proper safeguards and accurate training data, the model could provide consistent guidance, reminding patients to weigh themselves, report swelling, or avoid certain foods. Language models can also assist caregivers, offering clear explanations about medication regimens or early signs of deterioration. In lowresource settings, where access to cardiologists is limited, these systems could deliver crucial information that helps patients seek timely medical attention. Role in Rehabilitation and Lifestyle Modification Cardiac rehabilitation is an essential part of recovery after myocardial infarction or heart surgery. Unfortunately, adherence to rehabilitation programs remains low due to logistical barriers and limited understanding of their importance. LLMs can help patients stay engaged by providing ongoing motivation, reminders, and education. ChatGPT could help design personalized plans that include diet, exercise, and stress reduction techniques tailored to individual needs. When linked with wearable devices, the model could summarize daily activity and provide AMIR TAHAVVORI 70 encouragement or feedback. For example, it might say, “You walked 5,000 steps today, which is an improvement from yesterday. Maintaining this pace can help strengthen your heart.” Such reinforcement, while simple, can improve patient morale and commitment to rehabilitation. In addition to supporting individuals, language models can aid rehabilitation teams by summarizing patient progress and identifying those at risk of dropping out. This combination of automation and human oversight enhances efficiency and personalizes the rehabilitation experience. Clinical Documentation and Workflow Support Documentation is one of the most timeconsuming aspects of cardiology practice. From progress notes to procedural reports, clinicians must record details that are accurate and comprehensive. LLMs can help reduce this burden by generating drafts or structured summaries. During consultations, physicians can dictate findings, and ChatGPT could transcribe and organize the information, ensuring all necessary details are included. The model could format the report to align with institutional standards, leaving the physician to review and confirm its accuracy. In complex cardiac procedures, such as catheterization or bypass surgery, ChatGPT could LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 71 help prepare operative reports by summarizing preoperative evaluations, procedural steps, and postoperative recommendations. The result is a standardized and clear document that facilitates communication among multidisciplinary teams. When applied to inpatient care, LLMs can track daily changes in lab results, medications, and vital signs. They can automatically highlight significant variations, such as rising creatinine levels in patients receiving diuretics, helping physicians identify complications early. Research and Knowledge Synthesis Cardiovascular research generates a continuous flow of publications, spanning basic science, clinical trials, and epidemiology. LLMs can analyze and summarize this expanding body of knowledge. A researcher might request summaries of recent trials on antiplatelet therapy or outcomes of new heart failure drugs, and the model could provide concise overviews including sample sizes, endpoints, and conclusions. By processing vast amounts of literature, ChatGPT can identify emerging trends and help researchers design new studies. For instance, it might detect that most recent investigations on atrial fibrillation focus on non-vitamin K anticoagulants and suggest underexplored areas like patient adherence or device-based prevention. In clinical practice, LLMs can serve as quick references for evidence-based decisions. A AMIR TAHAVVORI 72 physician uncertain about the appropriate dosing of a novel medication can consult the model for guideline-aligned information, reducing dependence on time-consuming manual searches. LLMs can also help draft academic manuscripts, abstracts, or grant proposals. They can outline arguments, ensure clarity, and summarize results, allowing researchers to focus on interpretation rather than formatting. While human oversight remains essential, this assistance accelerates scientific communication. Predictive and Preventive Applications Prevention is the cornerstone of cardiology. Identifying risk factors early can save lives. LLMs can contribute to prevention by analyzing health records and identifying patients at high risk for developing cardiovascular disease. For instance, ChatGPT could be trained to recognize combinations of factors—such as hypertension, diabetes, obesity, and smoking— that suggest elevated risk for coronary artery disease. It could then prompt physicians to initiate preventive interventions, such as statin therapy or lifestyle counseling. When linked to wearable devices or mobile health applications, language models can provide personalized feedback on activity levels, sleep quality, and heart rate trends. A patient’s smartwatch might send data that ChatGPT LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 73 translates into a message like, “Your resting heart rate has been increasing this week; it might be useful to review your stress and caffeine intake.” While such feedback requires validation, it offers a glimpse into how AI could make prevention more interactive. In hospital settings, predictive systems can use LLMs to monitor for signs of impending cardiac arrest or acute decompensation. By analyzing nursing notes, lab data, and vital signs, the model could alert staff to subtle deteriorations, supporting rapid response teams. Cardiac Imaging and Data Interpretation Imaging plays a central role in cardiology. Echocardiograms, CT scans, and MRIs produce enormous amounts of data that require interpretation. While LLMs do not analyze images directly, they can support radiologists and cardiologists by generating or refining textual reports. A system combining visual AI with ChatGPT could automatically describe findings in structured language. After detecting reduced wall motion on echocardiography, the model could generate a draft report stating, “Left ventricular ejection fraction is moderately reduced, with regional hypokinesis of the anterior wall.” The clinician would then verify the findings before finalizing the report. AMIR TAHAVVORI 74 This integration reduces errors in documentation and ensures uniformity in terminology. It can also assist in training by providing consistent examples for junior doctors learning to interpret imaging results. Ethical and Legal Considerations The introduction of LLMs into cardiac care raises ethical and legal concerns. Medical information is sensitive, and any system handling patient data must adhere to privacy regulations. Institutions must ensure that AI systems operate within secure frameworks that protect personal identifiers. Transparency is another priority. Clinicians and patients should understand how AI-generated recommendations are derived and be aware of their limitations. When ChatGPT produces text based on its training data, it cannot always provide exact sources or explain reasoning. This opacity can create challenges in clinical accountability. Regulatory bodies will need to define standards for AI-assisted decision-making. Determining responsibility when errors occur is complex. A balance must be found between harnessing innovation and safeguarding ethical integrity. Bias is also a significant issue. If the model’s training data underrepresents certain populations, its outputs may be less accurate for those groups. This could perpetuate disparities in cardiac care, particularly in underserved communities. Continuous auditing of model LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 75 performance across diverse populations is necessary to minimize harm. Education and Professional Development LLMs can also serve as educational partners for medical students and clinicians. ChatGPT can simulate patient cases, quiz learners on diagnostic criteria, or explain pharmacological mechanisms in simple terms. A student studying arrhythmias could ask the model to describe the difference between atrial flutter and fibrillation, or to list common antiarrhythmic drugs and their mechanisms. For continuing education, clinicians can use ChatGPT to review updates to guidelines, such as new recommendations for lipid management or anticoagulation. The model can summarize lengthy publications into digestible points, allowing professionals to stay informed without dedicating excessive time to literature review. In multidisciplinary teams, AI systems can facilitate communication by producing summaries that bridge specialties. For instance, a cardiologist, surgeon, and anesthesiologist can all access a unified report generated from the same data, reducing misunderstandings. Integration into Health Systems Integrating LLMs into cardiac care requires coordination between clinicians, engineers, and AMIR TAHAVVORI 76 administrators. Hospitals can deploy locally trained models that align with their workflow and terminology. These systems could interact with existing electronic health record platforms to deliver real-time assistance. For example, during ward rounds, a physician could query the model to summarize trends in troponin levels or compare current medication doses to prior ones. The model could generate a concise summary that fits into the clinical discussion. In outpatient settings, integration with scheduling and messaging systems can enhance efficiency. ChatGPT could send reminders for follow-up appointments or lab tests, improving continuity of care and reducing missed visits. Limitations and Future Prospects Despite their promise, LLMs face several limitations. They cannot replace human intuition or clinical experience. Their understanding of medicine is based on statistical associations rather than reasoning grounded in biology. They may produce errors that sound convincing, and such inaccuracies can be dangerous in clinical contexts. Access to current, verified data is another limitation. Without continuous updates from reliable medical databases, an LLM’s information can become outdated. Collaboration between medical societies and AI developers could help ensure accuracy and relevance. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 77 The future likely lies in hybrid models that combine LLMs with structured medical data and imaging analysis. These multimodal systems could offer deeper insights, integrating text, numbers, and visuals into unified clinical tools. As transparency improves and ethical frameworks mature, AI will likely become a standard complement to cardiovascular medicine rather than a replacement for human expertise. Conclusion Large language models and ChatGPT hold significant potential for the management of cardiac diseases. They can assist in diagnosis, documentation, education, and patient engagement while supporting research and preventive care. When integrated responsibly, they can help clinicians navigate the increasing complexity of cardiovascular medicine with greater clarity and efficiency. These tools must be implemented thoughtfully, respecting privacy, accuracy, and equity. The relationship between physicians and technology should remain collaborative, where machines augment rather than substitute clinical judgment. With careful oversight, large language models could help transform cardiac care into a more precise, informed, and patient-centered practice that combines human compassion with the analytical power of artificial intelligence. AMIR TAHAVVORI 78 LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 79 5. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF DERMATOLOGICAL DISEASES Background Skin diseases are among the most visible and psychologically burdensome conditions in medicine. They affect people across every age and population, shaping quality of life and selfimage. The skin serves as a reflection of both internal health and environmental exposure, and its disorders range from mild rashes to severe autoimmune and malignant diseases. Dermatology, while deeply visual, also depends on complex clinical reasoning, patient history, and communication. The field generates vast amounts of textual and image-based information, including clinical notes, biopsy reports, and patient inquiries. In recent years, artificial intelligence has entered this space, offering new ways to manage dermatological disorders through data integration, diagnostic support, and patient education. Among these tools, large language models such as ChatGPT stand out for their ability to handle unstructured text, interpret natural language, and facilitate understanding between patients and professionals. Large language models, often abbreviated as LLMs, 80 are advanced computational systems trained on enormous volumes of written data. They learn patterns in how language is used and can generate coherent, context-aware text in response to queries. ChatGPT, developed by OpenAI, is one of the most well-known examples. It interacts conversationally, answering questions, drafting documents, and summarizing information. In the context of dermatology, where clear communication and quick access to knowledge are essential, these capabilities present practical opportunities. They can enhance the efficiency of clinical workflows, help patients understand complex diagnoses, and support researchers in managing expanding scientific literature. Foundations of LLM Use in Dermatology The practice of dermatology blends visual recognition with narrative reasoning. Clinicians rely on both image interpretation and patient history to arrive at accurate diagnoses. LLMs, though not designed to interpret images directly, can complement visual AI systems by processing the textual aspects of dermatologic care. They can summarize descriptions of lesions, extract key findings from reports, and integrate diagnostic impressions with patient histories. A dermatologist managing multiple cases each day must often review a series of notes that describe similar symptoms. ChatGPT could generate LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 81 AMIR TAHAVVORI 94 6. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF ONCOLOGICAL CONDITIONS Background Cancer remains one of the major global health challenges, causing approximately 10 million deaths each year. Managing oncological diseases requires an integrated approach that involves timely detection, accurate assessment, personalized treatments, and continuous patient care. Conventional oncology practices depend heavily on professional expertise; however, they face challenges such as limited time, overwhelming data from rapidly expanding research, and the difficulty of integrating diverse types of information, including genetic profiles, imaging results, and patient records. In this context, advancements in artificial intelligence, particularly large language models (LLMs) and tools such as ChatGPT, provide new opportunities to enhance medical workflows and improve clinical decision-making. The application of LLMs in oncology arises from their capacity to process and generate natural language, a capability that has advanced rapidly since the release of models like GPT-3.5 in 2022. Within clinical environments, these 95 models act as assistants that can compile information from medical records, research literature, and established treatment guidelines to support therapeutic recommendations. Studies have shown that they can analyze clinical cases, propose treatment plans aligned with established protocols such as those from the American Cancer Society, and offer reasoning behind their suggestions, potentially reducing the workload for healthcare professionals. In precision oncology, where understanding genetic mutations and selecting targeted therapies are critical, customized LLMs integrated with data retrieval methods demonstrate strong agreement with expert tumor boards in recommending individualized treatment strategies. This capability is especially valuable for rare or advanced-stage cancers, where rapidly organizing scattered information can help translate genetic findings into actionable interventions. Beyond assisting physicians, LLMs play an important role in patient-centered oncology care. They provide accessible and comprehensible sources of information, explaining disease mechanisms, treatment effects, and survivorship care in conversational formats. Early studies show that models such as ChatGPT-4 perform well when responding to patient questions about cancers like prostate or colorectal cancer, frequently matching expert advice or official medical guidelines. AMIR TAHAVVORI 96 This function can empower patients to better understand their conditions, adhere to treatment plans, and receive emotionally supportive communication. In primary care and community health settings, LLMs can assist in symptom triage and risk assessment, thereby reducing diagnostic delays that negatively affect outcomes, particularly in low-resource regions. By filtering out unlikely cases based on patient history and symptom patterns, these tools can help expedite specialist referrals, minimize unnecessary testing, and improve resource efficiency. In research and data management, LLMs facilitate the extraction of structured information from unstructured medical texts, including laboratory reports and clinical trial protocols, thereby accelerating participant selection and result verification in cancer research. Reviews have highlighted their usefulness across multiple cancer types, where adapted models such as GPT have achieved reliable accuracy in identifying key factors such as disease stage or biomarkers from digital health records. This capability supports population-based studies and real-time tracking of treatment outcomes, which can contribute to the development of adaptive clinical trial designs and evidence-based oncology practices. Nevertheless, integrating LLMs into oncology presents significant challenges. Their performance can vary across models and LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 97 contexts, influenced by phrasing of inputs, biases embedded in training data, and random variability in generated outputs. Instances of fabricated or inaccurate information pose risks to patient safety and underscore the necessity for rigorous verification and continuous human oversight. Ethical concerns, including data privacy protection, accountability for incorrect recommendations, and equitable access to avoid deepening health disparities, also require careful attention. Moreover, while LLMs excel in textbased tasks, their limited ability to process multimodal data such as medical images or realtime interactions highlights the need for hybrid systems that combine LLMs with other AI technologies. Moving forward, addressing these challenges through standardized evaluation frameworks, continual model refinement with domain-specific data, and integration into multimodal AI systems will be crucial for unlocking the full potential of LLMs in oncology. Future directions should include comparative studies with traditional methods, external validation in diverse populations, and the establishment of ethical guidelines to ensure responsible and trustworthy implementation. This section outlines the foundational concepts of LLMs and ChatGPT, examines their specific applications in cancer management, reviews evidence from recent AMIR TAHAVVORI 98 The Role of LLMs in Oncology Data Management Oncology generates an extraordinary amount of textual and numerical information. Pathology reports, radiology findings, operative notes, and genomic data are recorded in varied formats. Integrating these sources requires time and expertise. LLMs can organize this material by extracting relevant information and summarizing it into coherent narratives. For example, an oncologist reviewing multiple test results might use ChatGPT to summarize key findings from several documents. The model could identify tumor type, staging details, biomarker status, and previous treatment responses, presenting them in a structured report. This ability reduces the time spent searching through records, allowing specialists to focus on decisionmaking rather than administrative work. In cancer registries, which track patient demographics and outcomes, LLMs can assist in data curation. They can recognize cancer-related terminology, standardize variations in report wording, and classify cases accurately. This not only enhances data quality but also accelerates the production of research-ready datasets. Genomic oncology represents another field research, and discusses strategies to mitigate associated risks, aiming to guide their careful and informed adoption in clinical oncology. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 99 where text-based AI models hold potential. As personalized medicine grows, clinicians must interpret reports containing genetic mutations, molecular profiles, and targeted therapy recommendations. ChatGPT can summarize such findings, highlighting clinically relevant mutations and linking them to approved or experimental treatments. Although it cannot replace genetic counselors or molecular pathologists, it can serve as a helpful companion in translating technical data into practical guidance. Diagnostic Support and Clinical Decision-Making Diagnosis in oncology involves combining clinical evaluation with imaging, pathology, and laboratory analysis. LLMs can support this process by synthesizing available information and providing structured insights. A clinician faced with a complex case—such as a metastatic tumor of unknown origin—could consult ChatGPT for a summary of potential differential diagnoses and a reminder of standard diagnostic algorithms. In situations where guidelines frequently change, such as breast or lung cancer staging, ChatGPT can provide updates based on current standards. It can recall recent recommendations for imaging modalities, biopsy protocols, or molecular testing. When used carefully, these models help clinicians stay aligned with evolving evidence. AMIR TAHAVVORI 100 During tumor board meetings, where multiple specialists review cases, LLMs can serve as real-time assistants. They can summarize prior discussions, extract key data from records, and prepare concise case briefs. This reduces redundancy and ensures that important details are not overlooked. While LLMs cannot interpret medical images directly, they can complement image analysis software. For example, after an AI-based imaging system identifies a suspicious lung nodule, ChatGPT could generate a descriptive summary of findings and suggest standard next steps, such as PET scanning or biopsy. The text-based model translates visual insights into actionable reports, creating continuity between imaging and clinical documentation. Patient Education and Communication Cancer patients often face overwhelming amounts of information about their diagnosis and treatment. Many struggle to understand medical terminology or interpret test results. ChatGPT’s conversational format allows it to act as an accessible source of information. Patients can ask questions such as, “What does stage III colon cancer mean?” or “How do chemotherapy side effects work?” The model can explain these concepts in clear, empathetic language without assuming prior knowledge. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 101 For patients undergoing treatment, ChatGPT could generate reminders about medication schedules, hydration, or managing common side effects. It could explain what symptoms require immediate medical attention and when to contact their healthcare team. While these functions should operate under medical supervision, they have the potential to improve adherence and reduce anxiety. Caregivers also benefit from clear information. ChatGPT can help families understand complex care instructions, such as handling feeding tubes, recognizing infection signs, or supporting emotional well-being. When integrated into clinical portals, the model can produce personalized summaries after consultations, ensuring patients leave with understandable explanations of what was discussed. In palliative care, where communication and compassion are central, ChatGPT can assist in generating sensitive, patient-centered materials. It can help healthcare providers craft messages that balance honesty with empathy, promoting understanding without overwhelming patients or families. Documentation and Administrative Support The administrative workload in oncology is immense. From clinical trial documentation to insurance forms, much of the physician’s time AMIR TAHAVVORI 102 is consumed by paperwork. LLMs can automate several aspects of this process. During clinical visits, ChatGPT could generate drafts of encounter notes, capturing symptoms, assessments, and treatment plans. It could highlight new developments, such as treatment responses or emerging side effects, and suggest structured templates for record consistency. After verification, these notes can be entered directly into electronic health records. In cancer centers that manage large numbers of patients, LLMs can standardize discharge summaries, referral letters, and patient instructions. This reduces errors and maintains clarity across multidisciplinary teams. For clinical research, language models can assist in formatting case report forms, checking for missing data, and ensuring that descriptions align with protocol requirements. They can also generate narratives summarizing adverse events or deviations, streamlining communication with regulatory bodies. Research, Literature Review, and Knowledge Synthesis Cancer research evolves quickly. Thousands of studies are published each year, making it nearly impossible for clinicians to stay current on every topic. LLMs can help researchers and practitioners keep up by summarizing new findings and identifying patterns across studies. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 103 A researcher interested in immunotherapy for melanoma might ask ChatGPT to summarize the most recent trials on checkpoint inhibitors. The model could produce an overview that includes study size, key outcomes, and safety data. This synthesis saves time and helps identify promising directions for further reading. For meta-analyses and systematic reviews, LLMs can screen abstracts, extract data points, and categorize findings according to inclusion criteria. Although human validation is required, this automation significantly speeds up the early stages of evidence synthesis. In translational oncology, where bench research connects to clinical application, ChatGPT can help bridge language between disciplines. It can translate molecular biology findings into clinically relevant summaries, helping researchers identify potential therapeutic implications. Scientific writing also benefits from these systems. ChatGPT can assist in drafting grant proposals, editing manuscripts, or summarizing reviewer feedback. By improving clarity and coherence, it allows researchers to focus on scientific content rather than formatting. Personalized Oncology and Precision Medicine Cancer treatment is shifting toward personalization, guided by molecular characteristics and patient-specific factors. This AMIR TAHAVVORI 104 approach demands interpretation of vast datasets, including genomic profiles, proteomic signatures, and treatment responses. LLMs can organize this information and relate it to known therapeutic options. For example, when presented with a genetic report showing mutations in EGFR or KRAS, ChatGPT can summarize which targeted therapies are relevant and what resistance mechanisms are known. It can also list ongoing clinical trials that match the patient’s profile, serving as a research aid for oncologists. In multidisciplinary meetings, where geneticists, pathologists, and oncologists collaborate, LLMs can act as interpreters of technical data. They can convert molecular descriptions into concise explanations that facilitate discussion and planning. As pharmacogenomic databases expand, LLMs could eventually help predict responses to specific drugs or identify potential toxicities. By integrating laboratory data, previous treatment outcomes, and literature references, they could offer insights that guide personalized care decisions. Predictive Modeling and Outcome Monitoring Predicting cancer outcomes requires integrating diverse information sources. LLMs, when linked to structured databases, can support predictive LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 105 analytics by identifying relationships between textual patterns and patient outcomes. For instance, the model might detect that specific phrases in clinical notes, such as “rapidly enlarging mass” or “persistent weight loss”, are correlated with advanced disease or poor prognosis. These associations can inform risk stratification tools. In survivorship care, ChatGPT could assist in monitoring long-term health. It can analyze follow-up notes to detect early signs of recurrence or late treatment complications. By generating reminders for screening and symptom tracking, it supports continuity of care even years after treatment ends. Integration with wearable devices and mobile health apps could further expand this potential. Patients could describe symptoms through natural language inputs, and ChatGPT could interpret them in relation to cancer history, alerting clinicians if patterns suggest concern. Ethical and Practical Challenges Despite its potential, introducing LLMs into oncology raises ethical and practical concerns. Data privacy is paramount. Cancer records often include genetic information that is uniquely identifiable. Systems using language models must operate within secure environments and comply with strict data protection regulations. Accuracy is another issue. LLMs may generate AMIR TAHAVVORI 106 text that sounds convincing but contains factual inaccuracies. In medicine, such errors can have serious consequences. Rigorous validation and human oversight are essential before AI-generated recommendations are used clinically. Bias in training data presents further risks. If models are trained on literature dominated by studies from certain regions or populations, they may produce guidance less applicable to underrepresented groups. This imbalance could exacerbate global disparities in cancer care. Accountability must also be defined. If a model provides a misleading summary that influences a treatment decision, responsibility must remain with the clinician. LLMs should be viewed as supportive tools rather than autonomous decision-makers. Education and Training Medical education in oncology involves constant learning. LLMs can supplement this process by offering accessible explanations, quick references, and interactive learning experiences. Students can query ChatGPT about cancer biology, treatment principles, or clinical case examples, receiving explanations tailored to their level of knowledge. In residency programs, ChatGPT can simulate patient interactions, allowing trainees to practice breaking bad news or explaining treatment plans. It can generate different patient personalities, helping learners develop empathy LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 107 and adaptability in communication. For practicing oncologists, language models can summarize updates in clinical guidelines or highlight new therapeutic approvals. They can also assist in generating educational materials for patients and colleagues. Academic institutions might integrate LLMs into teaching platforms, where students discuss complex cases with AI assistance. Such environments encourage critical thinking and digital literacy, skills increasingly important for modern clinicians. Global and Public Health Perspectives Cancer care varies widely across regions. In many countries, access to oncologists and diagnostic tools is limited. LLMs can help bridge these gaps by providing guidance to primary care physicians in resource-constrained settings. A clinician in a rural area could describe a patient’s symptoms and receive a summary of likely conditions, initial management steps, and referral guidelines. While not a substitute for specialist evaluation, this can improve early detection and triage. Language diversity also poses barriers in oncology. ChatGPT’s multilingual capabilities can help translate patient education materials into local languages while retaining accuracy. This supports global equity in cancer literacy. AMIR TAHAVVORI 108 Public health agencies can use LLMs to monitor patterns in cancer-related communications. By analyzing patient inquiries or online discussions, they can identify misconceptions or rising concerns, informing educational campaigns. Integration with Health Systems To maximize impact, LLMs must integrate smoothly with healthcare infrastructure. Hospitals could deploy customized versions trained on their internal data, ensuring contextspecific accuracy. These localized models would align outputs with institutional policies, preferred terminologies, and treatment protocols. Integration with electronic health records allows real-time assistance. When clinicians enter new data, ChatGPT could suggest appropriate orders, flag missing documentation, or remind users about protocol requirements. Such automation enhances workflow efficiency without disrupting established practices. For clinical trials, integrated LLMs can match patients to ongoing studies based on eligibility criteria derived from their records. This function accelerates recruitment and supports equitable access to experimental treatments. The Human Element in AIAssisted Oncology Despite technological progress, the human element in cancer care remains irreplaceable. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 109 Empathy, moral judgment, and nuanced understanding cannot be automated. LLMs can handle data, but they cannot grasp the emotional dimensions of illness. The physician’s role as guide and advocate persists at every stage of care. ChatGPT’s greatest value lies in extending human capability rather than replacing it. By managing information overload, it allows clinicians to devote more attention to listening and counseling. When used thoughtfully, AI can strengthen the human connection at the core of oncology. Future Directions As research continues, LLMs will become more specialized for medical use. Fine-tuned models trained on oncology-specific corpora will achieve greater accuracy in terminology, treatment recommendations, and contextual reasoning. Integration with multimodal systems that analyze both text and images will make AI assistants even more versatile. Continuous collaboration between clinicians, data scientists, and ethicists will shape responsible innovation. Establishing transparent auditing processes, maintaining dataset diversity, and ensuring human oversight will determine whether these systems become trusted allies in cancer care. If used wisely, large language models could help build a future where oncologists spend less time buried in paperwork and more time with their AMIR TAHAVVORI 110 patients, where research progresses faster, and where knowledge is shared more widely and equitably. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 111 AMIR TAHAVVORI 112 7. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF ORAL AND DENTAL DISEASES Background The integration of artificial intelligence (AI) into healthcare has emerged as one of the most transformative developments of the 21st century, reshaping various fields such as diagnostics, education, patient engagement, and decision-making. Large language models (LLMs), especially those based on generative transformer architecture, such as OpenAI’s ChatGPT, have garnered significant attention for their ability to interact with users in natural language, provide evidence-based reasoning, and emulate a human-like understanding of medical knowledge. Dentistry, a field traditionally reliant on visual diagnosis, tactile skills, and interpersonal communication, is on the verge of a paradigm shift with the incorporation of these languagebased AI tools into daily practice. Since the release of ChatGPT-3.5 and later ChatGPT-4, LLMs have demonstrated remarkable advancements in synthesizing clinical information, understanding dental terminology, and providing informed responses to a wide range of questions from both patients and 113 Ethical and Privacy Considerations The integration of AI tools into dental practice must prioritize ethical principles and data protection. Patient records contain sensitive personal information, including medical histories, images, and biometric data. Systems using LLMs must operate within strict privacy frameworks, ensuring that no identifiable information is shared or stored insecurely. Another concern is the accuracy of AIgenerated content. While ChatGPT can provide useful explanations, it may occasionally produce errors or outdated information. Therefore, human oversight remains mandatory. Dentists should verify all AI-generated outputs before applying them clinically or distributing them to patients. Bias in training data is also a potential problem. If models are trained on data that underrepresent certain populations, their outputs may be less reliable for diverse groups. Dentistry already faces disparities in access and outcomes; LLM development must consciously avoid reinforcing these inequities. Regulation and accountability will evolve as AI becomes more common in healthcare. Professional bodies and educational institutions must establish guidelines for ethical use, ensuring that innovation aligns with patient safety and trust. AMIR TAHAVVORI 126 Education and Professional Development LLMs have significant potential in dental education. They can serve as tutors, providing explanations, generating case scenarios, and testing students’ knowledge. A student preparing for an exam could ask ChatGPT to create sample questions on restorative principles or oral anatomy. The model can adapt the difficulty level and provide feedback based on responses. Faculty members can use ChatGPT to assist in preparing lectures or summarizing new research findings for classroom discussion. This saves time while ensuring that material remains current and comprehensive. For continuing education, practitioners can use language models to stay informed about updates in materials, techniques, and regulations. They can request summaries of new studies or clinical guidelines without manually searching through databases. By offering quick access to curated information, LLMs support lifelong learning, a critical aspect of professional growth in dentistry. Integration with Digital and Clinical Systems Dentistry is moving toward digital integration, with electronic records, intraoral scanners, and 3D printing becoming standard tools. LLMs fit LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 127 naturally into this ecosystem by connecting the textual elements of digital workflows. When paired with imaging systems, ChatGPT can generate structured reports describing radiographic findings. It can assist in tracking follow-up appointments, reminding clinicians when recall exams are due, or summarizing changes between visits. In dental laboratories, AI can facilitate communication by automatically converting design specifications into standardized orders. This reduces misinterpretation and speeds up production. Future dental software could integrate ChatGPT directly into chairside applications, allowing clinicians to access reference material, treatment guidelines, or patient education content instantly during appointments. Limitations and Challenges Despite its promise, ChatGPT has clear limitations. It cannot perceive visual cues, evaluate tactile findings, or perform physical examinations. Its knowledge depends on the data it was trained on, which may not always reflect the latest evidence. Overreliance on AI could risk diminishing critical thinking if practitioners accept generated content uncritically. Therefore, training on appropriate AI use should become part of dental education. There are also practical issues related to cost, infrastructure, and user training. Implementing AMIR TAHAVVORI 128 secure and reliable AI systems requires investment in technology and staff readiness. Ensuring compliance with legal and ethical standards adds further complexity. Finally, while LLMs can simulate empathy through tone, they lack genuine understanding. The human element in dental care, trust, reassurance, and compassion—remains irreplaceable. Future Directions The next generation of LLMs will likely combine text with visual data, creating systems capable of analyzing radiographs, intraoral photos, and 3D scans alongside written information. Such multimodal models could identify lesions, suggest possible diagnoses, and generate complete clinical summaries in real time. Personalization will continue to evolve. AI systems may learn from individual patient data to predict risk for caries or periodontal disease, suggesting preventive strategies before problems arise. Collaboration between clinicians, data scientists, and educators will guide this progress. Transparent development, inclusive datasets, and consistent validation will ensure these technologies serve patients equitably and responsibly. Conclusion Large language models and ChatGPT represent a significant step forward in the digital LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 129 transformation of oral healthcare. They assist in diagnosis, documentation, education, research, and communication, all while easing the burden of administrative tasks. By making complex information understandable and accessible, they strengthen the connection between patients and providers. Challenges regarding accuracy, privacy, and equity must be addressed carefully, but the overall direction is promising. As dentistry continues to embrace technology, these language-based tools will become integral to daily practice, supporting clinicians in delivering more efficient, informed, and compassionate care. AMIR TAHAVVORI 130 LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 131 Background Infectious diseases continue to rank among the leading causes of illness and death worldwide. They affect people of all ages, across all regions, and place enormous strain on health systems. From influenza and tuberculosis to HIV and COVID-19, infectious conditions test the limits of clinical practice, public health, and scientific understanding. Their unpredictable nature and rapid spread require constant vigilance and communication. Managing them effectively depends on accurate diagnosis, timely reporting, and collaboration among clinicians, researchers, and policymakers. Recent advances in artificial intelligence have introduced new tools that can support these efforts. Among them, large language models, often referred to as LLMs, have shown particular promise. These systems use complex algorithms to understand and generate human language based on vast collections of text. ChatGPT, developed by OpenAI, is a prominent example. It can interpret questions, summarize knowledge, and produce coherent responses that resemble 8. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF INFECTIOUS DISEASES 132 human conversation. Such capabilities make it valuable for managing the complex flow of information that characterizes infectious disease control. LLMs can support physicians by retrieving clinical guidelines, summarizing research, or suggesting possible differential diagnoses. They can assist public health workers in preparing reports, analyzing outbreaks, or communicating with the public. While they are not replacements for expert judgment, they serve as companions in organizing knowledge and improving efficiency. The Role of Data and Language in Infectious Disease Management Effective management of infectious diseases depends heavily on information. Clinicians must analyze patient histories, laboratory results, and environmental factors, while public health agencies monitor cases across regions. Much of this information exists as unstructured text in clinical notes, research articles, or field reports. Traditional systems often struggle to organize or interpret such material quickly. LLMs can process large quantities of unstructured data in seconds. They identify patterns in text and translate them into useful summaries or insights. In hospitals, they can assist infection control teams by scanning clinical records to detect unusual patterns of fever, antibiotic use, or microbiology results. These insights may LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 133 signal the emergence of resistant infections or nosocomial outbreaks. For health authorities, LLMs can aggregate information from different regions, summarizing it in clear terms for decision-makers. During an outbreak, they can condense thousands of reports into a coherent overview, helping officials understand trends and allocate resources efficiently. The ability to interpret language also allows LLMs to bridge communication gaps between technical experts and the general public. They can translate scientific findings into accessible language without oversimplifying the facts, which is essential in controlling misinformation during epidemics. Diagnostic Assistance and Clinical Applications Infectious diseases often present with overlapping symptoms, making diagnosis difficult. Fever, fatigue, and cough can point to dozens of different pathogens. While diagnostic tests remain the gold standard, physicians frequently rely on clinical reasoning supported by background information. LLMs can serve as tools that expand this reasoning process. When a clinician provides a description such as “a patient with prolonged fever, rash, and arthralgia after recent travel,” ChatGPT can recall diseases with similar presentations. It can list AMIR TAHAVVORI 134 dengue, chikungunya, or Zika infection, along with their key differences in incubation period and associated findings. This process encourages a broader consideration of possibilities and helps prevent premature closure in diagnostic reasoning. In addition, LLMs can explain laboratory findings. When given a report mentioning elevated Creactive protein and leukocytosis, ChatGPT can discuss likely infectious causes or guide further investigations. It can also summarize the principles of isolation precautions or suggest when a case should be reported to health authorities. These capabilities are not meant to replace expert interpretation but to enhance it. When clinicians use LLMs as reference companions, they can confirm their reasoning, explore alternative perspectives, or recall rarely encountered conditions that might otherwise be missed. Antimicrobial Stewardship and Rational Therapy Antimicrobial resistance remains a growing global threat. Misuse and overuse of antibiotics accelerate the spread of resistant strains. LLMs can help promote rational prescribing by checking compatibility between diagnosis, culture results, and chosen antibiotics. In a hospital setting, ChatGPT could review prescriptions and identify when a broad-spectrum LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 135 antibiotic is being used unnecessarily. It could remind the prescriber of relevant guidelines and suggest narrower options. If integrated with local microbiology data, it could summarize resistance trends in real time, supporting evidence-based decisions. For education, LLMs can generate casebased exercises for students or junior doctors to practice antimicrobial selection. They can explain pharmacological mechanisms and discuss potential interactions. By serving as an accessible source of updated information, ChatGPT can help maintain adherence to stewardship principles, reducing inappropriate use and contributing to the longterm fight against resistance. Public Health Surveillance and Epidemic Intelligence Public health agencies depend on continuous surveillance to detect outbreaks early. Reports from hospitals, laboratories, and even social media can indicate changes in disease patterns. However, these data streams are often fragmented and filled with unstructured text. LLMs can analyze such reports for specific terms or clusters of symptoms. They can identify unusual increases in mentions of respiratory illness or gastrointestinal symptoms in a particular region. This automated detection allows officials to focus on areas where intervention may be needed. AMIR TAHAVVORI 136 When an outbreak occurs, ChatGPT can assist in generating situation reports. It can compile case counts, summarize containment measures, and describe the affected populations. These summaries can then be adapted for policymakers, clinicians, or the public, depending on the level of detail required. During recovery, LLMs can review response documents, extract lessons learned, and produce concise summaries for future planning. Their capacity to process large archives quickly supports institutional memory, helping countries learn from previous experiences. Communication with the Public During infectious disease emergencies, misinformation spreads faster than the pathogens themselves. Fear, confusion, and distrust can undermine public health measures. Clear and consistent communication is therefore essential. ChatGPT can assist in producing accurate, accessible explanations of disease prevention and treatment. It can craft responses to common questions about vaccination, hygiene, or symptoms. When integrated into official health websites or messaging systems, it can provide immediate answers without requiring human operators to manage every inquiry. In multilingual settings, LLMs can translate health information accurately while preserving meaning. This capability is particularly useful in regions LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 137 with linguistic diversity, where communication barriers can delay response efforts. For example, during a dengue outbreak, ChatGPT could help create educational materials about mosquito control, symptoms, and when to seek care. The tone could be adjusted to suit community leaders, school programs, or public service announcements. Such uses not only enhance understanding but also build trust. When information is delivered clearly and consistently, people are more likely to comply with prevention and treatment recommendations. Patient Education and Adherence Individual patients benefit from clear explanations of their diagnoses and treatments. Infectious diseases often require strict adherence to medication schedules and follow-up visits. LLMs can produce personalized educational material that helps patients understand their conditions. A person diagnosed with tuberculosis, for example, might receive an AI-generated guide summarizing the importance of completing therapy, potential side effects, and lifestyle adjustments. The tone and complexity can be adjusted based on literacy level, ensuring comprehension. For those managing chronic infections such as HIV, ChatGPT can provide reminders about AMIR TAHAVVORI 138 medication timing, laboratory testing, or safe practices. It can also address emotional aspects of disease management through empathetic conversation, although such support must always remain under professional supervision. By reinforcing clinician guidance and offering accessible information, LLMs can improve adherence and reduce relapse or transmission. Research and Scientific Discovery The field of infectious diseases evolves rapidly, with thousands of new studies published every year. Keeping up with this expanding body of knowledge is nearly impossible for individual researchers. LLMs offer a solution by summarizing and organizing information efficiently. ChatGPT can review a collection of research abstracts and extract the main findings. It can highlight emerging topics such as new antiviral compounds, resistance mechanisms, or vaccine development. For systematic reviews, it can assist in identifying relevant articles and grouping them by study type or region. When scientists write papers or grant proposals, LLMs can help structure arguments, suggest phrasing, or generate summaries. They can also convert technical writing into lay summaries for public dissemination. By accelerating literature review and improving communication, language models can shorten the time between discovery and application, which is LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 139 crucial in responding to new infectious threats. Global Collaboration and Health Equity Infectious diseases do not respect borders. Effective control requires cooperation among nations, institutions, and organizations. LLMs can facilitate such collaboration by improving communication and sharing of information. They can translate reports, harmonize terminology, and summarize regional updates. When health agencies in different countries describe cases differently, ChatGPT can align their definitions to ensure consistent reporting. For global organizations such as the World Health Organization, language models can draft policy briefs, summarize field reports, and generate multilingual updates. This streamlines coordination during emergencies. In low-resource settings, where specialists may be scarce, LLMs can provide accessible training materials for healthcare workers. They can explain disease recognition, sample collection, or isolation procedures in simple language. Such support enhances capacity building and reduces dependency on external experts. Integration with Digital Systems As hospitals and public health agencies adopt electronic systems, integrating LLMs becomes increasingly practical. ChatGPT can be embedded AMIR TAHAVVORI 140 within medical record platforms, laboratory information systems, or telemedicine portals. When combined with structured data sources, it can produce comprehensive summaries that include both numerical and narrative components. For instance, a clinician could request a patient overview that combines lab results, vital signs, and recent clinical notes. The model could generate a concise report for morning rounds or handover meetings. In telemedicine, ChatGPT can support remote consultations by collecting preliminary information from patients before connecting them with a physician. It can guide them through symptom checklists, ensuring that important details are recorded accurately. In surveillance databases, LLMs can tag reports with relevant categories, simplifying searches and statistical analysis. This integration transforms raw data into actionable knowledge. Ethical and Legal Issues As artificial intelligence becomes more involved in infectious disease management, ethical considerations grow in importance. Protecting privacy is a central concern. Patient data, especially in infectious diseases, can reveal sensitive details about behavior, travel, or social contact. LLMs must be used under strict security and de-identification protocols. Accuracy is another challenge. ChatGPT and LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 141 similar models sometimes produce incorrect information with a confident tone. In clinical or public health contexts, such errors could lead to harm. Human oversight is therefore essential, and institutions must establish systems for review and accountability. Bias in training data can also influence model behavior. If most information originates from certain regions or populations, the AI may generalize poorly to others. In infectious disease control, this could result in unequal recommendations or misrepresentation of local realities. Legal frameworks for AI in healthcare are still evolving. Questions about responsibility for AIgenerated advice, consent for data use, and transparency in model operation require careful attention. Ethical guidelines and continuous auditing will be necessary to maintain trust. Education and Workforce Development For LLMs to reach their full potential, healthcare workers need to understand how to use them effectively. Training programs can introduce the principles of AI, emphasizing both its strengths and its limitations. Medical schools can include practical exercises where students use ChatGPT to interpret data or summarize guidelines, learning how to evaluate responses critically. For public health AMIR TAHAVVORI 142 professionals, workshops can demonstrate how to use LLMs for report writing, communication, and outbreak analysis. By building literacy in AI tools, institutions ensure that clinicians remain in control while benefiting from technological support. Education helps prevent misuse and encourages collaboration between health experts and data scientists. Future Directions The next generation of LLMs will likely integrate multiple data types. Combining textual, numerical, and visual information could produce systems capable of comprehensive assessment. For infectious diseases, such models could analyze laboratory values, patient narratives, and images of rashes or microscopic findings simultaneously. Predictive modeling will also improve. LLMs may learn to anticipate outbreak trends by combining textual news reports with structured case data. Their natural language outputs will allow experts to interpret predictions more intuitively. In personalized medicine, AI could tailor infectious disease management to individual risk profiles, guiding vaccination schedules or prophylactic treatments. Future developments must remain transparent and ethically grounded. Collaboration among governments, universities, and industry partners will ensure that innovation serves public interest rather than commercial goals alone. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 143 Conclusion Large language models and ChatGPT have introduced new ways of managing infectious diseases by improving how information is processed, interpreted, and shared. They assist clinicians in diagnosis, support public health surveillance, enhance communication, and strengthen research capacity. Their ability to bridge technical and human understanding makes them valuable allies in both clinical and community settings. Challenges remain. Accuracy, privacy, and bias must be managed carefully. The success of these tools depends on responsible integration with human expertise, not replacement of it. As AI continues to evolve, it offers a vision of infectious disease management that is faster, more connected, and better informed than before. Through collaboration and critical oversight, large language models can help reshape global health into a more responsive and equitable system. AMIR TAHAVVORI 144 LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 145 Implementing AI systems requires secure servers, technical support, and staff training, which can be challenging for smaller institutions. Future Prospects Future LLMs are expected to integrate multimodal data, combining text with imaging, genomics, and laboratory information. This development will enhance their role in diagnostics and personalized medicine. For example, a future version of ChatGPT might interpret ultrasound reports, correlate them with hormonal levels, and produce a unified summary for gynecological evaluation. Personalization will also improve. Models could adapt to the preferences of individual clinicians or institutions, producing reports that match their documentation style. They may also learn from feedback to refine recommendations over time. Global collaboration will expand as well. Shared LLM frameworks could allow institutions in different countries to contribute de-identified data, promoting collective learning and equitable access to AI technology. Ethical governance will remain a central concern. Regulatory bodies will need to develop standards for accuracy, transparency, and accountability. Collaborative oversight between medical professionals, computer scientists, and ethicists will help maintain trust. Conclusion AMIR TAHAVVORI 158 Large language models such as ChatGPT represent a transformative development in the management of urological and gynecological disorders. They enhance diagnostic reasoning, streamline documentation, facilitate patient communication, and support education and research. Their strength lies in their ability to process language and extract meaning from complex medical data, bridging gaps between technical information and human understanding. Yet their use must remain cautious. They should complement rather than replace clinical expertise. With appropriate safeguards, ongoing evaluation, and ethical use, these tools can contribute significantly to improving healthcare delivery in urology and gynecology. The future likely holds more integration, greater personalization, and wider accessibility, leading to more informed and efficient patient care. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 159 AMIR TAHAVVORI 160 10. LARGE LANGUAGE MODELS AND CHATGPT FOR THE MANAGEMENT OF OTHER DISEASES Large Language Models and ChatGPT for the Management of Autoimmune Diseases Background Autoimmune diseases represent a vast group of conditions in which the body’s immune system mistakenly targets its own tissues. This complex dysfunction leads to chronic inflammation, tissue destruction, and long-term health consequences. Disorders such as rheumatoid arthritis, systemic lupus erythematosus, multiple sclerosis, type 1 diabetes, and inflammatory bowel disease account for a large portion of chronic morbidity worldwide. Their management requires early diagnosis, careful monitoring, and continuous adjustment of therapy. Given their multifaceted nature, autoimmune conditions demand constant interpretation of laboratory data, clinical findings, and patient-reported symptoms, all of which produce enormous volumes of unstructured information. Recent advances in artificial intelligence, 161 particularly large language models, have begun to transform how this information is processed. LLMs such as ChatGPT possess the ability to interpret and generate human language, a feature that allows them to handle medical text with increasing sophistication. These models are trained on vast data sources, including scientific literature, clinical notes, and general text, enabling them to respond to natural language questions and summarize complex ideas. Their integration into the management of autoimmune diseases provides opportunities to assist physicians, educate patients, and support research in ways that were previously not possible. While still in early stages of clinical application, LLMs can act as intelligent companions for healthcare professionals. They help identify potential diagnoses, interpret laboratory values, summarize patient histories, and assist in therapeutic decision-making. Their potential extends beyond clinical use, offering valuable contributions to patient communication, education, and population-level analysis. Complexity of Autoimmune Disease Management Managing autoimmune disorders is inherently complex because these conditions are heterogeneous and can affect multiple organ systems. A patient with lupus, for example, may experience renal inflammation, neurological AMIR TAHAVVORI 162 symptoms, and dermatologic manifestations simultaneously. Another with multiple sclerosis may show unpredictable progression that depends on genetic, environmental, and immunological interactions. Clinicians must integrate a wide range of information sources, from serologic markers and imaging to evolving clinical symptoms and treatment responses. The data volume can be overwhelming. Electronic medical records contain text entries that vary in style, detail, and completeness. Laboratory results may be repeated over years, and treatment protocols evolve rapidly as new therapies become available. LLMs are designed to navigate this kind of complexity. Their ability to parse language allows them to extract structured meaning from unstructured text, identifying patterns or inconsistencies that human readers might miss. In autoimmune disease clinics, this capacity can support data organization, summarization, and prioritization, improving efficiency and accuracy in care delivery. Diagnostic Support Autoimmune diseases are often difficult to diagnose because their symptoms overlap with many other conditions. Early manifestations may be vague, such as fatigue or joint pain, and laboratory tests can produce inconclusive results. LLMs can assist in this diagnostic process by LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 163 evaluating patient descriptions and correlating them with established diagnostic criteria. When provided with a detailed case summary, ChatGPT can suggest possible conditions for consideration. For instance, it may highlight systemic lupus erythematosus, mixed connective tissue disease, or vasculitis based on combinations of features like rash, anemia, and proteinuria. It can also recall standard diagnostic frameworks, such as the ACR/EULAR criteria for rheumatoid arthritis or lupus classification systems. In practical terms, such support acts as a reminder system, ensuring that rare or atypical conditions are not overlooked. When integrated into decision support software, LLMs can flag unusual symptom clusters or laboratory findings that warrant specialist review. While these outputs require human verification, they can significantly reduce diagnostic delays. Early and accurate diagnosis improves longterm outcomes and prevents unnecessary investigations. Interpretation of Laboratory and Imaging Data Autoimmune diseases rely heavily on laboratory monitoring. Tests such as ANA titers, antidsDNA, rheumatoid factor, complement levels, and inflammatory markers guide diagnosis and treatment adjustment. Each of these results carries contextual meaning that depends on AMIR TAHAVVORI 164 clinical presentation and disease phase. ChatGPT can interpret such data by correlating patterns. For instance, declining complement levels alongside rising anti-dsDNA titers in a lupus patient might suggest a disease flare. Similarly, persistent elevation of C-reactive protein in rheumatoid arthritis could indicate inadequate therapeutic control. LLMs can also summarize radiology or ultrasound reports related to joint inflammation, synovitis, or organ involvement. When linked with electronic health records, these models can create dynamic summaries showing trends over time. They may highlight discrepancies, such as a mismatch between clinical improvement and persistently abnormal lab results, prompting re-evaluation. This interpretative assistance is especially useful for early-career clinicians or those managing complex multisystem cases. By contextualizing laboratory data, LLMs help reduce the cognitive load involved in synthesizing diverse information. Personalized Treatment Planning Treatment of autoimmune diseases requires balancing efficacy and safety. Therapies such as corticosteroids, immunosuppressants, and biologics carry risks that must be weighed against their potential benefits. Large language models can assist in reviewing clinical guidelines and summarizing best practices for particular patient profiles. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 165 A physician managing a patient with severe rheumatoid arthritis could ask ChatGPT for a summary of recommended biologic therapies based on current guideline updates. The model might outline the rationale for TNF inhibitors, JAK inhibitors, or IL-6 receptor blockers, along with common contraindications and monitoring requirements. For patients with systemic lupus erythematosus, ChatGPT could review the indications for hydroxychloroquine, corticosteroids, and immunosuppressants like mycophenolate mofetil. It can also describe potential adverse effects and the importance of regular ophthalmologic monitoring. Such functions save time and enhance evidencebased practice. They also support shared decision-making by providing clear, accessible explanations that patients can understand. When presented with choices, individuals are more likely to adhere to treatment if they comprehend its reasoning. Patient Engagement and Self-Management Autoimmune conditions are long-term illnesses that require active patient participation. Daily routines, diet, exercise, and medication adherence influence outcomes. Many patients search for information online, but sources often vary in reliability. ChatGPT can bridge this gap AMIR TAHAVVORI 166 by providing trustworthy, easily digestible explanations. Patients with multiple sclerosis might use ChatGPT to learn about fatigue management, exercise adaptations, or coping strategies for cognitive symptoms. Those with type 1 diabetes can seek clarification on carbohydrate counting, insulin timing, or recognizing hypoglycemia. LLMs can also generate personalized summaries after clinic visits, reinforcing the key points discussed. This improves recall and adherence. They can answer follow-up questions that arise between appointments, reducing anxiety and misinformation. For emotional support, while ChatGPT cannot replace psychological care, it can offer empathetic language that validates patients’ experiences. Such engagement helps build confidence and resilience, both crucial for chronic disease management. Clinical Documentation and Workflow Efficiency Autoimmune disease management involves detailed documentation. Each visit includes updates on symptoms, physical findings, lab results, and medication adjustments. Writing and organizing this information consumes significant clinician time. LLMs can automate parts of this process. By transcribing conversations and summarizing LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 167 key points, ChatGPT can generate draft notes for review. It can create structured templates that capture relevant details systematically. In multidisciplinary teams, it can synthesize communications from different specialists into unified summaries. When implemented properly, these functions reduce administrative workload and allow clinicians to focus more on patient interaction. They also standardize records, which facilitates quality improvement and research. Research and Data Analysis Autoimmune disease research depends on analyzing large datasets and identifying patterns within clinical text, laboratory results, and patient reports. LLMs are well suited for such tasks because of their natural language understanding capabilities. ChatGPT can extract variables from unstructured notes, categorize them, and summarize findings for statistical analysis. This accelerates cohort identification for retrospective studies. In genetic and biomarker research, LLMs can review literature, summarize associations, and highlight emerging hypotheses. In clinical trials, they can assist in drafting protocols, informed consent forms, and progress summaries. Their ability to generate coherent text reduces delays during documentation stages. For systematic reviews, ChatGPT can screen AMIR TAHAVVORI 168 abstracts and summarize study characteristics, helping researchers focus on high-value articles. The ultimate advantage lies in accelerating knowledge translation. By rapidly processing vast amounts of text, LLMs shorten the gap between discovery and clinical application. Integration with Multimodal Systems Autoimmune diseases often require integration of data beyond text. Imaging, genomics, and wearable sensor data all contribute to understanding disease activity. Future iterations of LLMs are expected to work alongside multimodal systems that combine textual and numerical information. Imagine an AI platform where ChatGPT interprets narrative clinical notes while another module processes MRI findings or gene expression data. Together, they generate comprehensive assessments of disease activity, predict flares, or suggest personalized interventions. Such integration could revolutionize early intervention. For instance, subtle linguistic changes in patient-reported fatigue or pain descriptions might correlate with biomarkers of inflammation, enabling proactive therapy adjustments. Public Health and Epidemiology From a population health perspective, LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 169 autoimmune diseases pose a growing burden. Tracking prevalence, treatment trends, and outcomes requires analyzing vast public health datasets. LLMs can assist by extracting structured data from surveillance reports and scientific publications. They can identify regional disparities in access to care or summarize trends in medication utilization. During public health crises, such as pandemics, LLMs can help evaluate the impact of infections or vaccines on autoimmune disease outcomes. By automating synthesis of epidemiological evidence, ChatGPT contributes to policy development and resource planning. This helps governments and organizations allocate funding more effectively for prevention, research, and treatment. Ethical and Privacy Challenges Despite their benefits, LLMs raise ethical questions that must be addressed before widespread adoption in clinical settings. Patient data privacy is the foremost concern. Autoimmune diseases often involve sensitive personal information, and any use of AI must comply with strict confidentiality standards. Bias presents another risk. If an LLM is trained on data that underrepresents certain populations, its outputs may perpetuate inequalities. For instance, it may perform less accurately for autoimmune AMIR TAHAVVORI 170 presentations more common in women or minority groups. There is also the problem of false confidence. ChatGPT can generate plausible yet inaccurate statements, a phenomenon that could lead to misinterpretation if clinicians rely on it without verification. Transparent oversight and human review are therefore essential. Ethical frameworks must ensure accountability and informed consent. Patients should know when AI contributes to their care and how their data are used. Continuous evaluation and regulation will help balance innovation with safety. Training and Professional Adaptation Integrating AI tools like ChatGPT into healthcare requires appropriate training for clinicians. Understanding how these systems work, what they can and cannot do, and how to interpret their outputs responsibly is critical. Medical schools and continuing education programs can incorporate AI literacy into their curricula. Clinicians should practice using LLMs to interpret data, draft notes, and review literature while learning to identify potential errors or biases. For specialists in autoimmune diseases, training can focus on leveraging AI to handle longitudinal data, patient communication, and LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 171 multidisciplinary coordination. As familiarity grows, these tools will become natural extensions of professional practice rather than disruptive novelties. Limitations of Current Models Despite their versatility, current LLMs have clear limitations. They lack deep domain-specific training unless fine-tuned on medical data. Their understanding of context remains limited to textual input, which restricts performance in cases that depend heavily on visual or numeric interpretation. Memory constraints also hinder continuity. ChatGPT cannot retain long-term information about a patient across sessions unless integrated with secure data storage systems. Without such integration, its usefulness for longitudinal management remains partial. Another limitation involves uncertainty quantification. Unlike statistical models that produce confidence intervals, LLMs provide qualitative answers without explicit measures of reliability. This makes it difficult for clinicians to gauge how much trust to place in a given response. Continued research is needed to refine these aspects. Improved transparency, explainability, and integration with domain-specific datasets will gradually expand their role in clinical care. Future Directions AMIR TAHAVVORI 172 The evolution of large language models is progressing rapidly. Future iterations will likely incorporate real-time data access, allowing continuous updates from recent medical literature. They may include modules capable of understanding visual data, enabling direct interpretation of pathology slides or MRI images. For autoimmune disease management, predictive modeling will become increasingly valuable. AI could anticipate disease flares based on subtle linguistic cues in patient communications or changes in reported symptoms. Combined with wearable data, this might allow early interventions before major exacerbations occur. Another direction involves collaborative AI systems. Rather than working independently, multiple models may interact to cross-validate findings, reducing the likelihood of hallucinations or bias. As regulations mature, health systems will establish clear frameworks for AI governance, emphasizing transparency and shared responsibility. If guided responsibly, these developments could reshape how autoimmune diseases are detected, monitored, and treated across the world. Conclusion Large language models such as ChatGPT hold significant promise for transforming the management of autoimmune diseases. Their LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 173 capacity to interpret text, synthesize knowledge, and generate coherent communication makes them valuable companions for both clinicians and patients. They assist in diagnosis, treatment planning, documentation, research, and education, improving efficiency and accessibility across multiple dimensions of care. While limitations persist regarding accuracy, bias, and privacy, careful integration and oversight can mitigate these challenges. As technology evolves, LLMs will become increasingly embedded within healthcare systems, complementing human expertise rather than replacing it. For complex, chronic conditions like autoimmune diseases, where language, data, and emotion intersect, these tools offer a way to bring structure to complexity and clarity to communication. With ongoing refinement and responsible use, they may help redefine the practice of personalized, informed, and compassionate medicine. AMIR TAHAVVORI 174 LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 175 AMIR TAHAVVORI 176 REFERENCES 1. CADTH Health Technology Review. Implications of ChatGPT on Radiology Workflow: CADTH Health Technology Review. Ottawa (ON): Canadian Agency for Drugs and Technologies in Health; 2023. 2. Abujaber AA, Abd-Alrazaq A, Al-Qudimat AR, Nashwan AJ. A Strengths, Weaknesses, Opportunities, and Threats (SWOT) Analysis of ChatGPT Integration in Nursing Education: A Narrative Review. Cureus. 2023;15(11):e48643. 3. Alhaidry HM, Fatani B, Alrayes JO, Almana AM, Alfhaed NK. ChatGPT in Dentistry: A Comprehensive Review. Cureus. 2023;15(4):e38317. 4. Alotaibi SS, Rehman A, Hasnain M. Revolutionizing ocular cancer management: a narrative review on exploring the potential role of ChatGPT. Front Public Health. 2023;11:1338215. 5. Arshad HB, Butt SA, Khan SU, Javed Z, Nasir K. ChatGPT and Artificial Intelligence in Hospital Level Research: Potential, Precautions, and Prospects. Methodist Debakey Cardiovasc J. 2023;19(5):77-84. 6. Barrington NM, Gupta N, Musmar B, Doyle D, Panico N, Godbole N, et al. A Bibliometric Analysis of the Rise of ChatGPT in Medical 177 material: A comparison with the leading national cardiology institutes. Curr Probl Cardiol. 2024;49(11):102797. 85. Arif F, Safri MK, Shahzad Z, Yasmeen SF, Rahman MF, Shaikh SA. Exploring the application of CHATGPT in plastic surgery: a comprehensive systematic review. J Pak Med Assoc. 2024;74(4 (Supple-4)):S17-s28. 86. Bektaş M, Pereira JK, Daams F, van der Peet DL. ChatGPT in surgery: a revolutionary innovation? Surg Today. 2024;54(8):964-71. 87. Bellanda VCF, Santos MLD, Ferraz DA, Jorge R, Melo GB. Applications of ChatGPT in the diagnosis, management, education, and research of retinal diseases: a scoping review. Int J Retina Vitreous. 2024;10(1):79. 88. Bhattacharya M, Pal S, Chatterjee S, Lee SS, Chakraborty C. Large language model to multimodal large language model: A journey to shape the biological macromolecules to biological sciences and medicine. Mol Ther Nucleic Acids. 2024;35(3):102255. 89. Birkun AA, Gautam A. Large Language Model-based Chatbot as a Source of Advice on First Aid in Heart Attack. Curr Probl Cardiol. 2024;49(1 Pt A):102048. 90. Bongco EDA, Cua SKN, Hernandez M, Pascual JSG, Khu KJO. The performance of ChatGPT versus neurosurgery residents in AMIR TAHAVVORI 190 neurosurgical board examination-like questions: a systematic review and meta-analysis. Neurosurg Rev. 2024;47(1):892. 91. Briganti G. How ChatGPT works: a mini review. Eur Arch Otorhinolaryngol. 2024;281(3):1565-9. 92. Campbell WAt, Chick JFB, Shin D, Makary MS. Understanding ChatGPT for evidence-based utilization in interventional radiology. Clin Imaging. 2024;108:110098. 93. Campolo JA, Kwon DY, Henderson PW. Can ChatGPT Be Used as a Research Assistant and a Patient Consultant in Plastic Surgery? A Review of 3 Key Information Domains. Eplasty. 2024;24:e49. 94. Chen S, Cai M, Tan G, Guo R, Liang Q, Li H, et al. Case report: Intravascular large B cell lymphoma mimicking acute demyelinating encephalomyelitis after SARS-CoV-2 reinfection: diagnostic value of advanced MRI techniques and the literature review with the assistance of ChatGPT. Front Immunol. 2024;15:1478163. 95. Choueka D, Tabakin AL, Shalom DF. ChatGPT in Urogynecology Research: Novel or Not? Urogynecology (Phila). 2024;30(12):962-7. 96. Chu CP. ChatGPT in veterinary medicine: a practical guidance of generative artificial intelligence in clinics, education, and research. Front Vet Sci. 2024;11:1395934. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 191 97. Clark SC. Can ChatGPT transform cardiac surgery and heart transplantation? J Cardiothorac Surg. 2024;19(1):108. 98. D'Agostino M, Feo F, Martora F, Genco L, Megna M, Cacciapuoti S, et al. ChatGPT and dermatology. Ital J Dermatol Venerol. 2024;159(5):566-71. 99. de Souza LL, Pontes HAR, Martins MD, Fonesca FP, Corrêa F, Coracin FL, et al. ChatGPT and dentistry: a step toward the future. Gen Dent. 2024;72(4):72-7. 100. Ding Z, Wei R, Xia J, Mu Y, Wang J, Lin Y. Exploring the potential of large language model-based chatbots in challenges of ribosome profiling data analysis: a review. Brief Bioinform. 2024;26(1). 101. Drouaud A, Stocchi C, Tang J, Gonsalves G, Cheung Z, Szatkowski J, et al. Exploring the Performance of ChatGPT in an Orthopaedic Setting and Its Potential Use as an Educational Tool. JB JS Open Access. 2024;9(4). 102. Farhadi Nia M, Ahmadi M, Irankhah E. Transforming dental diagnostics with artificial intelligence: advanced integration of ChatGPT and large language models for patient care. Front Dent Med. 2024;5:1456208. 103. Fatima A, Shafique MA, Alam K, Fadlalla Ahmed TK, Mustafa MS. ChatGPT in medicine: A cross-disciplinary systematic review of ChatGPT's AMIR TAHAVVORI 192 (artificial intelligence) role in research, clinical practice, education, and patient interaction. Medicine (Baltimore). 2024;103(32):e39250. 104. García-Torres D, Vicente Ripoll MA, Fernández Peris C, Mira Solves JJ. Enhancing Clinical Reasoning with Virtual Patients: A Hybrid Systematic Review Combining Human Reviewers and ChatGPT. Healthcare (Basel). 2024;12(22). 105. Ghanem D, Zhu AR, Kagabo W, Osgood G, Shafiq B. ChatGPT-4 Knows Its A B C D E but Cannot Cite Its Source. JB JS Open Access. 2024;9(3). 106. Ghanta SN, Al'Aref SJ, Lala-Trinidade A, Nadkarni GN, Ganatra S, Dani SS, et al. Applications of ChatGPT in Heart Failure Prevention, Diagnosis, Management, and Research: A Narrative Review. Diagnostics (Basel). 2024;14(21). 107. Goglia M, Pace M, Yusef M, Gallo G, Pavone M, Petrucciani N, et al. Artificial Intelligence and ChatGPT in Abdominopelvic Surgery: A Systematic Review of Applications and Impact. In Vivo. 2024;38(3):1009-15. 108. Goktas P, Grzybowski A. Assessing the Impact of ChatGPT in Dermatology: A Comprehensive Rapid Review. J Clin Med. 2024;13(19). 109. González R, Poenaru D, Woo R, Trappey AF, 3rd, Carter S, Darcy D, et al. ChatGPT: What Every Pediatric Surgeon Should Know About LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 193 Its Potential Uses and Pitfalls. J Pediatr Surg. 2024;59(5):941-7. 110. Gravina AG, Pellegrino R, Cipullo M, Palladino G, Imperio G, Ventura A, et al. May ChatGPT be a tool producing medical information for common inflammatory bowel disease patients' questions? An evidence-controlled analysis. World J Gastroenterol. 2024;30(1):17-33. 111. Gunawan J, Aungsuroch Y, Montayre J. ChatGPT integration within nursing education and its implications for nursing students: A systematic review and text network analysis. Nurse Educ Today. 2024;141:106323. 112. Haltaufderheide J, Ranisch R. The ethics of ChatGPT in medicine and healthcare: a systematic review on Large Language Models (LLMs). NPJ Digit Med. 2024;7(1):183. 113. Horgan R, Martins JG, Saade G, Abuhamad A, Kawakita T. ChatGPT in maternal-fetal medicine practice: a primer for clinicians. Am J Obstet Gynecol MFM. 2024;6(3):101302. 114. Hu M, Qian J, Pan S, Li Y, Qiu RLJ, Yang X. Advancing medical imaging with language models: featuring a spotlight on ChatGPT. Phys Med Biol. 2024;69(10). 115. Ishida K, Hanada E. Potential of ChatGPT to Pass the Japanese Medical and Healthcare Professional National Licenses: A Literature Review. Cureus. 2024;16(8):e66324. AMIR TAHAVVORI 194 116. Jin HK, Lee HE, Kim E. Performance of ChatGPT-3.5 and GPT-4 in national licensing examinations for medicine, pharmacy, dentistry, and nursing: a systematic review and metaanalysis. BMC Med Educ. 2024;24(1):1013. 117. Kaboudi N, Firouzbakht S, Shahir Eftekhar M, Fayazbakhsh F, Joharivarnoosfaderani N, Ghaderi S, et al. Diagnostic Accuracy of ChatGPT for Patients' Triage; a Systematic Review and MetaAnalysis. Arch Acad Emerg Med. 2024;12(1):e60. 118. Kapsali MZ, Livanis E, Tsalikidis C, Oikonomou P, Voultsos P, Tsaroucha A. Ethical Concerns About ChatGPT in Healthcare: A Useful Tool or the Tombstone of Original and Reflective Thinking? Cureus. 2024;16(2):e54759. 119. Kedia N, Sanjeev S, Ong J, Chhablani J. ChatGPT and Beyond: An overview of the growing field of large language models and their use in ophthalmology. Eye (Lond). 2024;38(7):1252-61. 120. Keshavarz P, Bagherieh S, Nabipoorashrafi SA, Chalian H, Rahsepar AA, Kim GHJ, et al. ChatGPT in radiology: A systematic review of performance, pitfalls, and future perspectives. Diagn Interv Imaging. 2024;105(7-8):251-65. 121. Keykha A, Behravesh S, Ghaemi F. ChatGPT and Medical Research: A Meta-Synthesis of Opportunities and Challenges. J Adv Med Educ Prof. 2024;12(3):135-47. 122. Kıyak YS, Emekli E. ChatGPT prompts LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 195 for generating multiple-choice questions in medical education and evidence on their validity: a literature review. Postgrad Med J. 2024;100(1189):858-65. 123. Kleib M, Darko EM, Akingbade O, Kennedy M, Majekodunmi P, Nickel E, et al. Current trends and future implications in the utilization of ChatGPT in nursing: A rapid review. Int J Nurs Stud Adv. 2024;7:100252. 124. Kucukkaya A, Arikan E, Goktas P. Unlocking ChatGPT's potential and challenges in intensive care nursing education and practice: A systematic review with narrative synthesis. Nurs Outlook. 2024;72(6):102287. 125. Kumari K, Pahuja SK, Kumar S. A Comprehensive Examination of ChatGPT's Contribution to the Healthcare Sector and Hepatology. Dig Dis Sci. 2024;69(11):4027-43. 126. Law S, Oldfield B, Yang W. ChatGPT/ GPT-4 (large language models): Opportunities and challenges of perspective in bariatric healthcare professionals. Obes Rev. 2024;25(7):e13746. 127. Layton AT. AI, Machine Learning, and ChatGPT in Hypertension. Hypertension. 2024;81(4):709-16. 128. Lechien JR, Rameau A. Applications of ChatGPT in Otolaryngology-Head Neck Surgery: A State of the Art Review. Otolaryngol Head Neck Surg. 2024;171(3):667-77. AMIR TAHAVVORI 196 129. Lee JM. Strategies for integrating ChatGPT and generative AI into clinical studies. Blood Res. 2024;59(1):45. 130. Lee Y, Shin T, Tessier L, Javidan A, Jung J, Hong D, et al. Harnessing artificial intelligence in bariatric surgery: comparative analysis of ChatGPT-4, Bing, and Bard in generating clinicianlevel bariatric surgery recommendations. Surg Obes Relat Dis. 2024;20(7):603-8. 131. Lee Y, Tessier L, Brar K, Malone S, Jin D, McKechnie T, et al. Performance of artificial intelligence in bariatric surgery: comparative analysis of ChatGPT-4, Bing, and Bard in the American Society for Metabolic and Bariatric Surgery textbook of bariatric surgery questions. Surg Obes Relat Dis. 2024;20(7):609-13. 132. Leon M, Ruaengsri C, Pelletier G, Bethencourt D, Shibata M, Flores MQ, et al. Harnessing the Power of ChatGPT in Cardiovascular Medicine: Innovations, Challenges, and Future Directions. J Clin Med. 2024;13(21). 133. Li J, Dada A, Puladi B, Kleesiek J, Egger J. ChatGPT in healthcare: A taxonomy and systematic review. Comput Methods Programs Biomed. 2024;245:108013. 134. Liu HY, Alessandri-Bonetti M, Arellano JA, Egro FM. Can ChatGPT be the Plastic Surgeon's New Digital Assistant? A Bibliometric Analysis and Scoping Review of ChatGPT in LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 197 Plastic Surgery Literature. Aesthetic Plast Surg. 2024;48(8):1644-52. 135. Liu J. ChatGPT: perspectives from humancomputer interaction and psychology. Front Artif Intell. 2024;7:1418869. 136. Liu M, Okuhara T, Chang X, Shirabe R, Nishiie Y, Okada H, et al. Performance of ChatGPT Across Different Versions in Medical Licensing Examinations Worldwide: Systematic Review and Meta-Analysis. J Med Internet Res. 2024;26:e60807. 137. Lund BD, Khan D, Yuvaraj M. ChatGPT in medical libraries, possibilities and future directions: An integrative review. Health Info Libr J. 2024;41(1):4-15. 138. Madaudo C, Parlati ALM, Di Lisi D, Carluccio R, Sucato V, Vadalà G, et al. Artificial intelligence in cardiology: a peek at the future and the role of ChatGPT in cardiology practice. J Cardiovasc Med (Hagerstown). 2024;25(11):766-71. 139. Maggio MG, Tartarisco G, Cardile D, Bonanno M, Bruschetta R, Pignolo L, et al. Exploring ChatGPT's potential in the clinical stream of neurorehabilitation. Front Artif Intell. 2024;7:1407905. 140. Malik S, Zaheer S. ChatGPT as an aid for pathological diagnosis of cancer. Pathol Res Pract. 2024;253:154989. AMIR TAHAVVORI 198 141. Maniaci A, Saibene AM, Calvo-Henriquez C, Vaira L, Radulesco T, Michel J, et al. Is generative pre-trained transformer artificial intelligence (Chat-GPT) a reliable tool for guidelines synthesis? A preliminary evaluation for biologic CRSwNP therapy. Eur Arch Otorhinolaryngol. 2024;281(4):2167-73. 142. Margetts TJ, Karnik SJ, Wang HS, Plotkin LI, Oblak AL, Fehrenbacher JC, et al. Use of AI Language Engine ChatGPT 4.0 to Write a Scientific Review Article Examining the Intersection of Alzheimer's Disease and Bone. Curr Osteoporos Rep. 2024;22(1):177-81. 143. Mehta S. Exploring angina: A fascinating chat with ChatGPT. Curr Probl Cardiol. 2024;49(3):102393. 144. Meşe İ, Altıntaş Taşlıçay C, Kuzan BN, Kuzan TY, Sivrioğlu AK. Educating the next generation of radiologists: a comparative report of ChatGPT and e-learning resources. Diagn Interv Radiol. 2024;30(3):163-74. 145. Momenaei B, Mansour HA, Kuriyan AE, Xu D, Sridhar J, Ting DSW, et al. ChatGPT enters the room: what it means for patient counseling, physician education, academics, and disease management. Curr Opin Ophthalmol. 2024;35(3):205-9. 146. Moon JT, Lima NJ, Froula E, Li H, Newsome J, Trivedi H, et al. Towards inclusive biodesign and LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 199 innovation: lowering barriers to entry in medical device development through large language model tools. BMJ Health Care Inform. 2024;31(1). 147. Morya VK, Lee HW, Shahid H, Magar AG, Lee JH, Kim JH, et al. Application of ChatGPT for Orthopedic Surgeries and Patient Care. Clin Orthop Surg. 2024;16(3):347-56. 148. Mu Y, He D. The Potential Applications and Challenges of ChatGPT in the Medical Field. Int J Gen Med. 2024;17:817-26. 149. Nedbal C, Naik N, Castellani D, Gauhar V, Geraghty R, Somani BK. ChatGPT in urology practice: revolutionizing efficiency and patient care with generative artificial intelligence. Curr Opin Urol. 2024;34(2):98-104. 150. Nolin-Lapalme A, Theriault-Lauzier P, Corbin D, Tastet O, Sharma A, Hussin JG, et al. Maximising Large Language Model Utility in Cardiovascular Care: A Practical Guide. Can J Cardiol. 2024;40(10):1774-87. 151. Omar M, Ullanat V, Loda M, Marchionni L, Umeton R. ChatGPT for digital pathology research. Lancet Digit Health. 2024;6(8):e595-e600. 152. Ong JCL, Seng BJJ, Law JZF, Low LL, Kwa ALH, Giacomini KM, et al. Artificial intelligence, ChatGPT, and other large language models for social determinants of health: Current state and future directions. Cell Rep Med. 2024;5(1):101356. AMIR TAHAVVORI 200 153. Parillo M, Vaccarino F, Beomonte Zobel B, Mallio CA. ChatGPT and radiology report: potential applications and limitations. Radiol Med. 2024;129(12):1849-63. 154. Pillai J, Pillai K. ChatGPT as a medical education resource in cardiology: Mitigating replicability challenges and optimizing model performance. Curr Probl Cardiol. 2024;49(12):102879. 155. Pressman SM, Borna S, Gomez-Cabello CA, Haider SA, Haider C, Forte AJ. AI and Ethics: A Systematic Review of the Ethical Considerations of Large Language Model Use in Surgery Research. Healthcare (Basel). 2024;12(8). 156. Puladi B, Gsaxner C, Kleesiek J, Hölzle F, Röhrig R, Egger J. The impact and opportunities of large language models like ChatGPT in oral and maxillofacial surgery: a narrative review. Int J Oral Maxillofac Surg. 2024;53(1):78-88. 157. Qin S, Chislett B, Ischia J, Ranasinghe W, de Silva D, Coles-Black J, et al. ChatGPT and generative AI in urology and surgery-A narrative review. BJUI Compass. 2024;5(9):813-21. 158. Rajjoub R, Arroyave JS, Zaidat B, Ahmed W, Mejia MR, Tang J, et al. ChatGPT and its Role in the Decision-Making for the Diagnosis and Treatment of Lumbar Spinal Stenosis: A Comparative Analysis and Narrative Review. Global Spine J. 2024;14(3):998-1017. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 201 159. Rao SJ, Isath A, Krishnan P, Tangsrivimol JA, Virk HUH, Wang Z, et al. ChatGPT: A Conceptual Review of Applications and Utility in the Field of Medicine. J Med Syst. 2024;48(1):59. 160. Reichenpfader D, Müller H, Denecke K. A scoping review of large language model based approaches for information extraction from radiology reports. NPJ Digit Med. 2024;7(1):222. 161. Sacoransky E, Kwan BYM, Soboleski D. ChatGPT and assistive AI in structured radiology reporting: A systematic review. Curr Probl Diagn Radiol. 2024;53(6):728-37. 162. Savithri Nandeesha DS. Empowering treatment decisions: ChatGPT in severe coronary artery disease. Curr Probl Cardiol. 2024;49(11):102789. 163. Schmidt KW, Lechner F. [ChatGPT: aid to medical ethics decision making?]. Anaesthesiologie. 2024;73(3):186-92. 164. Schukow C, Smith SC, Landgrebe E, Parasuraman S, Folaranmi OO, Paner GP, et al. Application of ChatGPT in Routine Diagnostic Pathology: Promises, Pitfalls, and Potential Future Directions. Adv Anat Pathol. 2024;31(1):15-21. 165. Shah K, Xu AY, Sharma Y, Daher M, McDonald C, Diebo BG, et al. Large Language Model Prompting Techniques for Advancement in Clinical Medicine. J Clin Med. 2024;13(17). AMIR TAHAVVORI 202 166. Sharma A, Medapalli T, Alexandrou M, Brilakis E, Prasad A. Exploring the Role of ChatGPT in Cardiology: A Systematic Review of the Current Literature. Cureus. 2024;16(4):e58936. 167. Sharma H, Ruikar M. Artificial intelligence at the pen's edge: Exploring the ethical quagmires in using artificial intelligence models like ChatGPT for assisted writing in biomedical research. Perspect Clin Res. 2024;15(3):108-15. 168. Shorey S, Mattar C, Pereira TL, Choolani M. A scoping review of ChatGPT's role in healthcare education and research. Nurse Educ Today. 2024;135:106121. 169. Shumway DO, Hartman HJ. Medical malpractice liability in large language model artificial intelligence: legal review and policy recommendations. J Osteopath Med. 2024;124(7):287-90. 170. Solano C, Tarazona N, Angarita GP, Medina AA, Ruiz S, Pedroza VM, et al. ChatGPT in Urology: Bridging Knowledge and Practice for Tomorrow's Healthcare, a Comprehensive Review. J Endourol. 2024;38(8):763-77. 171. Sumbal A, Sumbal R, Amir A. Can ChatGPT-3.5 Pass a Medical Exam? A Systematic Review of ChatGPT's Performance in Academic Testing. J Med Educ Curric Dev. 2024;11:23821205241238641. 172. Talyshinskii A, Juliebø-Jones P, Zeeshan LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 203 Hameed BM, Naik N, Adhikari K, Zhanbyrbekuly U, et al. ChatGPT as a Clinical Decision Maker for Urolithiasis: Compliance with the Current European Association of Urology Guidelines. Eur Urol Open Sci. 2024;69:51-62. 173. Tan S, Xin X, Wu D. ChatGPT in medicine: prospects and challenges: a review article. Int J Surg. 2024;110(6):3701-6. 174. Temperley HC, O'Sullivan NJ, Mac Curtain BM, Corr A, Meaney JF, Kelly ME, et al. Current applications and future potential of ChatGPT in radiology: A systematic review. J Med Imaging Radiat Oncol. 2024;68(3):257-64. 175. Tessler I, Wolfovitz A, Livneh N, Gecel NA, Sorin V, Barash Y, et al. Advancing Medical Practice with Artificial Intelligence: ChatGPT in Healthcare. Isr Med Assoc J. 2024;26(2):80-5. 176. Tzelves L, Kapriniotis K, Feretzakis G, Katsimperis S, Manolitsis I, Juliebø-Jones P, et al. ChatGPT in Clinical Medicine, Urology and Academia: A Review. Arch Esp Urol. 2024;77(7):708-17. 177. Ullah E, Parwani A, Baig MM, Singh R. Challenges and barriers of using large language models (LLM) such as ChatGPT for diagnostic medicine with a focus on digital pathology - a recent scoping review. Diagn Pathol. 2024;19(1):43. 178. Vij O, Calver H, Myall N, Dey M, Kouranloo AMIR TAHAVVORI 204 K. Evaluating the competency of ChatGPT in MRCP Part 1 and a systematic literature review of its capabilities in postgraduate medical assessments. PLoS One. 2024;19(7):e0307372. 179. Wang L, Wan Z, Ni C, Song Q, Li Y, Clayton E, et al. Applications and Concerns of ChatGPT and Other Conversational Large Language Models in Health Care: Systematic Review. J Med Internet Res. 2024;26:e22769. 180. Wang P, Zhang Q, Zhang W, Sun J. The application of ChatGPT in nursing: a bibliometric and visualized analysis. Front Med (Lausanne). 2024;11:1521712. 181. Wei Q, Yao Z, Cui Y, Wei B, Jin Z, Xu X. Evaluation of ChatGPT-generated medical responses: A systematic review and meta-analysis. J Biomed Inform. 2024;151:104620. 182. Wu CL, Cho B, Gabriel R, Hurley R, Liu J, Mariano ER, et al. Addition of dexamethasone to prolong peripheral nerve blocks: a ChatGPTcreated narrative review. Reg Anesth Pain Med. 2024;49(11):777-81. 183. Wu J, Ma Y, Wang J, Xiao M. The Application of ChatGPT in Medicine: A Scoping Review and Bibliometric Analysis. J Multidiscip Healthc. 2024;17:1681-92. 184. Wyatt KD, Alexander N, Hills GD, Liang WH, Kadauke S, Volchenboum SL, et al. Making sense of artificial intelligence and large LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 205 language models-including ChatGPT-in pediatric hematology/oncology. Pediatr Blood Cancer. 2024;71(9):e31143. 185. Xu X, Chen Y, Miao J. Opportunities, challenges, and future directions of large language models, including ChatGPT in medical education: a systematic scoping review. J Educ Eval Health Prof. 2024;21:6. 186. Ying L, Liu Z, Fang H, Kusko R, Wu L, Harris S, et al. Text summarization with ChatGPT for drug labeling documents. Drug Discov Today. 2024;29(6):104018. 187. Yu H. The application and challenges of ChatGPT in educational transformation: New demands for teachers' roles. Heliyon. 2024;10(2):e24289. 188. Zampatti S, Peconi C, Megalizzi D, Calvino G, Trastulli G, Cascella R, et al. Innovations in Medicine: Exploring ChatGPT's Impact on Rare Disorder Management. Genes (Basel). 2024;15(4). 189. Zhou Y, Li SJ, Tang XY, He YC, Ma HM, Wang AQ, et al. Using ChatGPT in Nursing: Scoping Review of Current Opinions. JMIR Med Educ. 2024;10:e54297. 190. Abid SK, Sulaiman N. Revolutionizing the future: unleashing the potential of generative AI ChatGPT in disaster management. Environ Sci Pollut Res Int. 2025. AMIR TAHAVVORI 206 191. Achanur M, Bhatt S, Maniyar RN, Sajjanar AK, Roy A, Rao V, et al. ChatGPT's Emerging Role in Dentistry: A Review. J Pharm Bioallied Sci. 2025;17(Suppl 1):S99-s101. 192. Ahmed M, Lam J, Chow A, Chow CM. A Primer on Large Language Models (LLMs) and ChatGPT for Cardiovascular Healthcare Professionals. CJC Open. 2025;7(5):660-6. 193. Ali S, Aslam A, Tahir Z, Ashraf B, Tanweer A. Advancements of AI in healthcare: a comprehensive review of ChatGPT's applications and challenges. J Pak Med Assoc. 2025;75(1):78-83. 194. Amankwaa I, Ekpor E, Cudjoe D, Kobiah E, Fuseini AJ, Diebieri M, et al. Patterns, advances, and gaps in using ChatGPT and similar technologies in nursing education: A PAGER scoping review. Nurse Educ Today. 2025;153:106822. 195. Antonie NI, Gheorghe G, Ionescu VA, Tiucă LC, Diaconu CC. The Role of ChatGPT and AI Chatbots in Optimizing Antibiotic Therapy: A Comprehensive Narrative Review. Antibiotics (Basel). 2025;14(1). 196. Aster A, Laupichler MC, RockwellKollmann T, Masala G, Bala E, Raupach T. ChatGPT and Other Large Language Models in Medical Education - Scoping Literature Review. Med Sci Educ. 2025;35(1):555-67. LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 207 197. Bae S, Jeon M, Moon HR. Text mining in MOF research: from manual curation to large language model-based automation. Chem Commun (Camb). 2025;61(60):11083-94. 198. Bellini D, Ferrari R, Vicini S, Rengo M, Saletti CL, Carbone I. Hi ChatGPT, I am a Radiologist, How can you help me? Radiol Med. 2025;130(8):1221-30. 199. Bhamidipaty V, Botchu B, Bhamidipaty DL, Guntoory I, Iyengar KP. ChatGPT for speech-impaired assistance. Disabil Rehabil Assist Technol. 2025;20(6):1575-7. 200. Bifarin OO, Yelluru VS, Simhadri A, Fernández FM. A Large Language Model-Powered Map of Metabolomics Research. Anal Chem. 2025;97(27):14088-96. 201. Bochtler M. How the technologies behind self-driving cars, social networks, ChatGPT, and DALL-E2 are changing structural biology. Bioessays. 2025;47(1):e2400155. 202. Busch D, Za'in C, Chan HM, Haryanto A, Agustiono W, Yu K, et al. A blueprint for large language model-augmented telehealth for HIV mitigation in Indonesia: A scoping review of a novel therapeutic modality. Health Informatics J. 2025;31(1):14604582251315595. 203. Chen D, Alnassar SA, Avison KE, Huang RS, Raman S. Large Language Model Applications for Health Information Extraction in Oncology: AMIR TAHAVVORI 208 Scoping Review. JMIR Cancer. 2025;11:e65984. 204. Cheng Y, Zhu L. A review of ChatGPT in medical education: exploring advantages and limitations. Int J Surg. 2025;111(7):4586-602. 205. Chokkakula S, Chong S, Yang B, Jiang H, Yu J, Han R, et al. Quantum leap in medical mentorship: exploring ChatGPT's transition from textbooks to terabytes. Front Med (Lausanne). 2025;12:1517981. 206. Chytas D, Noussios G, Paraskevas G, Vasiliadis AV, Giovanidis G, Troupis T. Can ChatGPT play a significant role in anatomy education? A scoping review. Morphologie. 2025;109(365):100949. 207. Ciudad-Fernández V, von Hammerstein C, Billieux J. People are not becoming "AIholic": Questioning the "ChatGPT addiction" construct. Addict Behav. 2025;166:108325. 208. Daccache N, Zako J, Morisson L, LaferrièreLanglois P. The applications of ChatGPT and other large language models in anesthesiology and critical care: a systematic review. Can J Anaesth. 2025;72(6):904-22. 209. de Luis D. [Generative artificial intelligence ChatGPT in clinical nutrition - Advances and challenges]. Nutr Hosp. 2025;42(4):797-806. 210. de Menezes Torres LM, de Morais EF, Fernandes Almeida DRM, Pagotto LEC, de Santana LARGE LANGUAGE MODELS AND CHATGPT IN MEDICAL S... 209 Hashemlu L, Chelan RJ, Hashemzadeh A, et al. Intelligent Patient Engagement: Education and Follow-Up through AI and Telemedicine. Kindle. 2025;5(1):1-185. 286. Gheibi M, Rajabloo Y, Alipour-Khabir Y, Azami P, Louia S, Bojnordi TE, et al. Artificial Intelligence in Biomarker Discovery: Applications Across Medical Specialties. Kindle. 2025;5(1):1-209. 287. Farrokhi M, Ghalamkarpour N, Nouri S, Babaei M, Rajabloo Y, Sattari M, et al. Innovative Vaccination: A New Era in Cancer Prevention. Kindle. 2025;5(1):1-194. 288. Babaheidarian P, Soltanattar A, Sajadi SK, Rostamian L, Foroutani L, Soleymanpourshamsi T, et al. Robotics in Healthcare. Kindle. 2025;5(1):1-178. AMIR TAHAVVORI 222 Proof