ChatGPT in Head and Neck Oncology-Opportunities and Challenges

Abstract

Head and neck oncology represents a complex and challenging field, encompassing the diagnosis, treatment and management of various malignancies affecting the intricate anatomical structures of the head and neck region. With advancements in artificial intelligence (AI), chatbot applications have emerged as a promising tool to revolutionize the field of Head and Neck oncology. ChatGPT is a cutting-edge language model developed by OpenAI that can help the oncologist in the clinic in scheduling appointments, establishing a clinical diagnosis, making a treatment plan and follow-up. ChatGPT also plays an essential role in telemedicine consultations, medical documentation, scientific writing and research. ChatGPT carries its inherent drawbacks too. ChatGPT raises significant ethical concerns related to authorship, accountability, transparency, bias, and the potential for misinformation. ChatGPT’s training data is limited to September 2021; thus, regular updates are required to keep pace with the rapidly evolving medical research and advancements. Therefore, a judicial approach to using ChatGPT is of utmost importance. Head and Neck Oncologists can reap the maximum benefit of this technology in terms of patient care, education and research to improve clinical outcomes.

Introduction

Head and neck oncology represents a complex and challenging field, encompassing the diagnosis, treatment and management of various malignancies affecting the intricate anatomical structures of the head and neck region. With advancements in technology and the growing role of artificial intelligence (AI) in healthcare, chatbot applications have emerged as a promising tool to revolutionize patient care in this specialized domain. Chatbots, powered by natural language processing and machine learning algorithms, have the potential to efficiently interact with patients, caregivers, and healthcare professionals, facilitating seamless communication and optimizing clinical workflows. In this context, ChatGPT, the revolutionary language model developed by OpenAI, has taken the internet by storm with its exceptional capabilities and versatility. While it has become ubiquitous in various fields, its potential application in head and neck oncology holds a promising future. Medical professionals and researchers specializing in head and neck cancers can adopt ChatGPT as a helpful tool. Furthermore, they can leverage their advanced natural language processing abilities to diagnose and stage tumours, analyze radiological images, and provide treatment recommendations based on the latest clinical guidelines. Careful utilization of ChatGPT by healthcare providers can improve patient outcomes, enhance personalized care, and advance knowledge and treatment options in head and neck oncology.

Artificial Intelligence and the Emergence of ChatGPT

Artificial intelligence (AI) is defined as the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings [1]. Artificial intelligence (AI) is a rapidly advancing field that focuses on creating intelligent systems capable of performing tasks that typically require human intelligence. One fascinating subfield of AI is Natural Language Processing (NLP), which involves teaching machines to understand and interact with human language. NLP refers to the branch of computer science—specifically, the branch of artificial intelligence or AI—concerned with giving computers the ability to understand text and spoken words in much the same way human beings can [2]. NLP enables computers to process and analyze vast amounts of text, extract meaning, and generate human-like responses. Within the realm of NLP, ChatGPT has emerged as a remarkable language model developed by OpenAI.

ChatGPT is a cutting-edge language model developed by OpenAI, a research organization dedicated to advancing artificial intelligence and pioneering state-of-the-art technologies for the betterment of society [3]. Powered by deep learning techniques, ChatGPT leverages a neural network architecture called Transformer, which revolutionized natural language processing (NLP) upon its introduction in 2017 [4]. One of ChatGPT’s key strengths lies in its adaptability through fine-tuning, allowing it to excel in a wide array of language-related tasks [5]. By training the model on specific tasks and additional data, such as text classification or machine translation, ChatGPT becomes a continual learner, showcasing its potential for future innovations. Its pre-training process involves exposure to vast amounts of data from the internet, including web pages, articles, books, and social media platforms in various languages [6, 7]. This training methodology helps ChatGPT develop a comprehensive understanding of language structure and meaning, generating coherent and contextually relevant text. The versatility of ChatGPT’s output lends itself to numerous applications, such as text summarization, sentiment analysis, language translation, and question-answering [8]. Notably, its remarkable ability to generate highly accurate and contextually relevant text has made it particularly popular in chatbots and virtual assistants [9]. The continuous advancements in language models like ChatGPT foreshadow a future where communication and machine interaction are more innovative and efficient. OpenAI has introduced several versions of ChatGPT, including GPT-2, GPT-3, GPT-3.5 and GPT-4, each distinguished by its size, number of parameters, and the range of languages covered during pre-training. For instance, the widely acclaimed GPT-3 was trained on an extensive collection of texts spanning 95 languages and boasted an impressive 175 billion parameters [10].

Application of ChatGPT in Healthcare

The integration of virtual assistants in healthcare has revolutionized patient care and management. One significant application of this technology is aiding patients in effectively managing their health. ChatGPT summarises the patient’s situation, case evaluation, and appropriate medical interventions. It has shown its potential to assist healthcare professionals in medical report writing [11], streamlining the medical recordkeeping process for healthcare professionals. With the ability to dictate their notes, medical practitioners can rely on ChatGPT to automatically extract critical details, including symptoms, diagnoses, and treatments from patient records. ChatGPT achieves impressive accuracy in clinical decision-making, with particular strengths emerging as it has more clinical information at its disposal [12]. It also excels in analyzing vast amounts of patient data to identify eligible individuals for clinical trial recruitment, thus enhancing the efficiency of the trial enrollment process. Furthermore, ChatGPT plays a vital role in assisting patients in medication management. It offers valuable support by providing reminders, dosage instructions, and discharge summaries [13].

Moreover, it furnishes patients with essential information about potential side effects, drug interactions, and other crucial medication considerations. This capability empowers patients to participate in their treatment and make informed decisions actively. Various published papers have shown the potential use of ChatGPT in different fields of medicine, such as dentistry, ophthalmology, obstetrics and gynaecology [14–17]. In the field of oncology, several studies have shown the potential of artificial intelligence chatbots to access cancer patients’ information [18], generate accurate information about cancer myths and misconceptions [19], identify guideline-based treatments for advanced solid tumours indicating a level of concordance with the NCCN guidelines [20].

ChatGPT in Head and Neck Oncology

Artificial intelligence can change the landscape of Head and Neck Oncology. ChatGPT can help the oncologist in the clinic in scheduling appointments, establishing a clinical diagnosis, making a treatment plan and follow-up. Judicious use of ChatGPT in Head and Neck Oncology will result in greater precision of diagnosis and treatment while making fewer errors. ChatGPT, a cutting-edge technology, thus has many aspirations in Head and Neck Oncology and apprehensions.

1. Screening and Diagnosis:

ChatGPT has excellent potential for automating patient education about cancer prevention and screening. In a study by Yi et al., the researchers created a set of 25 questions related to advice on getting screened for breast cancer. Three breast radiologists evaluated these responses and found they were appropriate for 22 out of 25 questions. The findings suggested the ability of ChatGPT to automate breast cancer prevention and screening processes [21]. In another study by Jeudy et al., ChatGPT provided appropriate results in lung cancer and incorporated the latest guidelines while screening [22]. Similarly, physicians can explore the ability of ChatGPT in head and neck cancer diagnosis and screening by correct input of symptoms. It should be noted, however, that ChatGPT must follow medical ethics and cannot replace doctors in diagnosis.

2. Clinical Decision Support:

Clinical decision-making is complex and influenced by clinical reasoning, physician judgment, and patient conditions. AI chatbots like ChatGPT can effectively support physicians, reducing errors and improving treatment outcomes. Rao et al. demonstrated the feasibility of ChatGPT for radiologic decision-making, with the potential to improve clinical workflow and responsible use of radiology services [23]. Similarly, Chiesa-Estomba et al. assessed Chat-GPT as a promising tool in the clinical decision-making process for patients who are candidates for sialendoscopy treatment [24].

ChatGPT has shown great potential in resolving complex clinical scenarios in head and neck surgery. Vaira et al. examined the accuracy of ChatGPT by analysing a total of 15 clinical scenarios and 144 questions spanning across 12 subspecialties within head and neck surgery. The ChatGPT chatbot exhibited satisfactory results except for such as malformative pathology, reconstructive surgery and condylar traumatology [25]. Patients of head and neck cancers planned to receive radiation therapy may have concerns regarding their treatment course, side effects and treatment outcomes. ChatGPT may help the patients by providing information about their disease status and outcome, offering reassurance and guiding them to have personalised resources. In a recent study, ChatGPT chatbots resulted in a 9.8x higher proportion of empathetic vs. very empathic responses than clinicians when answering patient queries [26].

3. Telemedicine Services:

ChatGPT has the potential to play an essential role in facilitating virtual doctor-patient interactions by providing virtual assistance and remote guidance. ChatGPT could be used to triage patients, provide initial assessments, and provide remote guidance for home care. It can inform patients about head and neck cancer, the treatment process, and what to expect during a telemedicine appointment. Thus, it will help patients feel more prepared and informed before their appointment, leading to better outcomes. Moreover, integrating ChatGPT in telemedicine services can remove language barriers that often hinder access to oncologic care, allowing health professionals to communicate effectively with patients who speak different languages.

4. Medical Documentation:

Medical documentation is a critical yet arduous task for healthcare professionals, often consuming valuable time and leading to common errors [27]. Ensuring accuracy in medical records is paramount, necessitating recorders to adeptly capture and precisely document all pertinent medical information discussed during patient interviews. A pilot assessment by Ali et al. showed the possibility of ChatGPT to generate a clinic letter with a high overall correctness and humanness score. Furthermore, these letters were written at a reading level similar to real-world human-generated letters [28]. Similarly, Yam et al. demonstrated the ability of ChatGPT to generate a formal discharge summary based on a physician’s input in a few seconds [13]. Thus, by employing ChatGPT as a language assistant or utilizing provided templates, clinicians can experience remarkable time savings and elevated precision in their medical documentation process.

5. Scientific Writing and Research:

ChatGPT has emerged as a transformative and promising academic writing and research tool. Through immediate feedback, doubt clarification, and the provision of additional resources, the chatbot can cater to readers’ needs. ChatGPT can help non-native English speakers by enhancing the language barrier and improving the expression and communication of research ideas and findings [29]. Various authors have listed ChatGPT as an efficient and promising tool for conducting comprehensive literature reviews and generating computer codes, thereby saving time for the research steps that require more effort from human intelligence, ultimately enabling researchers to focus on the critical aspect of their work. ChatGPT will thus lead to accelerated publication processes, ensuring faster availability of research results.

Potential Drawbacks of using ChatGPT

While ChatGPT, like other language models, can be a valuable tool in various domains, including the head and neck oncology field, it has certain drawbacks that should be considered. Some of the limitations are:

1. Ethical Concerns:

Using ChatGPT raises significant ethical concerns related to authorship, accountability, transparency, bias, and the potential for misinformation [30]. To safeguard the integrity of health science publications, publishers may need to establish specific criteria for including ChatGPT as a co-author. However, it is crucial to acknowledge that further research is necessary to evaluate the effectiveness and ethical implications of employing AI chatbots like ChatGPT across various disciplines. Some peer-reviewed papers have recently been published with ChatGPT listed as an author [31]. However, considering the abovementioned limitations, the World Association of Medical Editors (WAME) has recommended against including ChatGPT in the list of authors for such publications [32].

2. Limited Data:

ChatGPT’s training data requires regular updates to keep pace with the rapidly evolving medical research and advancements. As ChatGPT’s training data is limited to September 2021, it is subjected to errors, primarily due to its limited human-like understanding and the absence of data input [33]. The information may need to reflect the latest developments, potentially limiting its usability in clinical practice. Ensuring the accuracy and relevancy of data is essential to address this limitation and enhance ChatGPT’s applicability in clinical settings. Regular updates will enable the model to stay current with the latest medical knowledge and contribute to more informed and reliable decision-making in healthcare.

3. Lack of Human Interaction:

Clinical examination and assessment play a significant role in Head and Neck oncology. Although AI tools offer rapid and efficient patient information access, the essence of Head and Neck oncology lies in the human touch—patience and empathy. Direct communication of patients with healthcare professionals often instils a sense of comfort and reassurance among patients. The interpretation of AI-generated chat responses regarding trust and credibility varies significantly between medical and health-related matters compared to other products and services. In healthcare, the human touch will always remain invaluable in understanding and addressing complex medical information and behaviours.

Future Recommendations

In the coming years, a significant impact of AI chatbots like ChatGPT in the healthcare domain, particularly oncology, is anticipated. It is expected that ChatGPT will be seamlessly integrated into all text editing software. To address the accuracy of generated content, sophisticated systems will be developed to detect even subtle data manipulations ChatGPT performs. Currently, editors have to inspect the output of ChatGPT meticulously, and there remains a challenge in adequately citing the resources it uses. Nevertheless, ongoing research is dedicated to overcoming these obstacles and presenting practical solutions. In order to mitigate potential misuse, scholarly journals should establish stringent guidelines for using AI in academic papers.

Conclusion

ChatGPT can be a valuable tool in head and neck oncology, offering numerous benefits and applications. Head and Neck Oncologists can leverage its capabilities for screening, diagnosis, clinical decision support, telemedicine services, medical documentation, scientific writing, and research. However, ChatGPT may prove to be a double-edged sword. Despite its numerous advantages, implementing ChatGPT in healthcare raises several issues that require careful consideration. In scientific papers, a comprehensive evaluation of authorship, accountability, transparency, bias, and the potential for misinformation is crucial before widespread acceptance. Research efforts should focus on finding innovative solutions to the ethical concerns posed by ChatGPT in the healthcare sector. Overall, a holistic approach is required to balance the advantages of ChatGPT with its limitations. By acknowledging and addressing its limitations, Head and Neck oncologists can harness the full potential of this technology while safeguarding patient well-being.