Generative AI in Medicine

Article information
Title: ChatGPT in Liver Transplantation: Current Applications, Limitations, and Future Directions
PubMed ID: 41607825
First Author Last Name: Avramidou

Abstract:
Liver transplantation (LT) remains the optimal life-saving intervention for patients with end-stage liver disease. Despite the recent advances in LT several barriers, including organ allocation, donor-recipient matching, and patient education, persist. With the growing progress of artificial intelligence, particularly large language models (LLMs) like ChatGPT, new applications have emerged in the field of LT. Current studies demonstrating usage of ChatGPT in LT include various areas of application, from clinical settings to research and education. ChatGPT usage can benefit both healthcare professionals, by decreasing the time spent on non-clinical work, but also LT recipients by providing accurate information. Future potential applications include the expanding usage of ChatGPT and other LLMs in the field of LT pathology and radiology as well as the automated creation of discharge summaries or other related paperwork. Additionally, the next models of ChatGPT might have the potential to provide more accurate patient education material with increased readability. Although ChatGPT usage presents promising applications, there are certain ethical and practical limitations. Key concerns include patient data privacy, information accuracy, misinformation possibility and lack of legal framework. Healthcare providers and policymakers should collaborate for the establishment of a controlled framework for the safe use of ChatGPT. The aim of this minireview is to summarize current literature on ChatGPT in LT, highlighting both opportunities and limitations, while also providing future possible applications.

Annotation
The Finding:
This narrative review summarizes the emerging literature on the use of ChatGPT and other large language models in liver transplantation. The authors found that current applications span clinical care, patient education, research, and medical education. Existing studies suggest that ChatGPT may reduce administrative burden, assist with clinical information delivery, improve patient education materials, and support research activities. However, performance remains variable, with the highest accuracy seen for general informational tasks and lower reliability for complex treatment and prognostic questions. The authors conclude that ChatGPT has substantial potential as an adjunctive tool in transplantation but requires careful human oversight and a structured regulatory framework.

Strengths and Weaknesses:
A major strength of this article is that it provides a comprehensive and balanced overview of the rapidly expanding literature on artificial intelligence in liver transplantation. The authors review applications relevant to clinicians, patients, researchers, and educators while also discussing important ethical, legal, and practical limitations. The review is particularly useful because it synthesizes evidence from multiple studies and highlights both successes and failures of ChatGPT in transplantation-related tasks. However, the article is a narrative review rather than a systematic review, which introduces the possibility of selection bias. Much of the evidence cited comes from feasibility studies and evaluations of chatbot responses rather than studies examining real-world clinical outcomes. In addition, because the field is evolving rapidly, some of the conclusions may become outdated as newer AI models are developed and validated.

Relevance:
Although focused on liver transplantation, this article is highly relevant to consultation-liaison psychiatry because many of the proposed applications mirror challenges encountered in general psychiatric consultation practice. These include patient education, documentation, discharge communication, and clinical decision support. The discussion of ethical concerns, privacy, misinformation, hallucinations, and the limitations of AI-generated clinical recommendations is particularly applicable to psychiatric practice, where nuanced clinical judgment and the therapeutic relationship remain central. The review provides a thoughtful framework for considering how generative AI may be integrated into complex multidisciplinary medical settings while emphasizing that such tools should augment, rather than replace, clinician expertise.

Article information
Title: A Multiagent Large Language Model-Based System to Simulate the Liver Transplant Selection Committee: A Retrospective Cohort Study
PubMed ID: 41951492
First Author Last Name: Hasjim

Abstract:
Background:
Transplantation is one of the few areas in medicine in which the definitive treatment is rationed. Subjective decision making poses challenges for the transplantation selection process. Large language models (LLMs), through autonomous artificial intelligence (AI) agents, have implications in objective decision making for complex problems; however, there are gaps in the literature regarding their clinical applications. We aimed to examine the performance of a multidisciplinary committee of AI agents in the liver transplantation selection process as a proof of concept towards improving the objectivity of decision making.

Methods:
This retrospective, hybrid cohort study incorporated both real-world observational data and artificially generated data derived from the Scientific Registry of Transplant Recipients (SRTR) database in the USA. We retrospectively analysed adult patients (aged ≥18 years) who received liver transplants from Jan 1, 2004 to June 1, 2024, and generated a hypothetical cohort of patients with standard absolute contraindications to liver transplantation from a random subset of the cohort. An AI selection committee was configured, consisting of four LLM-based AI agents using GPT-4o-2024-08-06 prompted with specialised roles: transplant hepatologist, transplant surgeon, cardiologist, and social worker. This committee was instructed to assess clinical vignettes generated from the SRTR tabular data (eg, demographics, social determinants of health, end-stage liver disease, comorbidities, and laboratory measurements), placing candidates on the transplantation waitlist if liver transplantation would likely offer a 6-month or 1-year survival benefit or rejecting candidates for transplantation if contraindications were present or if such a survival benefit was unlikely. The primary objective was to assess the accuracy, sensitivity, and specificity of the AI selection committee in predicting 1-year post-transplantation survival. The secondary objectives were the accuracy, sensitivity, and specificity of identifying contraindications to liver transplantation and predicting 6-month post-transplantation survival, and the numbers of false-negative cases (ie, patients who were rejected by the AI selection committee but did receive a transplant) and false-positive cases (ie, patients who were placed on the waitlist by the AI selection committee but had assigned contraindications to transplantation).

Findings:
Of 8412 patients, 7033 (83·6%) were placed on the transplantation waitlist and 1379 (16·4%) were assigned to the hypothetical cohort and had contraindications to liver transplantation. For the prediction of 1-year post-transplantation survival, the AI selection committee had an accuracy of 92·00% (95% CI 91·43–92·58), a sensitivity of 1·00 (0·99–1·00), and a specificity of 0·66 (0·64–0·68). For identifying absolute contraindications to liver transplantation and predicting 6-month survival, accuracies were 98·19% (97·90–98·44) and 94·88% (94·37–95·29), respectively; sensitivities were 1·00 (0·99–1·00) in both cases; and specificities were 0·91 (0·89–0·92) and 0·75 (0·73–0·77), respectively. The most common reason for the AI selection committee to reject patients in false-negative cases was hepatocellular carcinoma that did not meet the Milan criteria (16 [62%] of 26 as a contraindication to transplantation, 16 [62%] of 26 rejected on the basis of 6-month survival prediction, and 14 [61%] of 23 rejected on the basis of 1-year survival prediction); the most common cause of death in false-positive cases was malignancy at 1 year (137 [28%] of 481).

Interpretation:
LLMs can be leveraged through a multiagent AI system to simulate discussions of a liver transplantation selection committee and provide accurate, objective insights on patients who might or might not benefit from a liver transplant. Lessons learned from this study are a step towards making the liver transplantation selection process more equitable and objective."

Annotation
The Finding:
This study developed and evaluated a novel multiagent AI system designed to simulate a liver transplant selection committee. The system consisted of four GPT-4o-based agents assigned the roles of transplant hepatologist, transplant surgeon, cardiologist, and social worker. Using over 8,400 liver transplant cases from the Scientific Registry of Transplant Recipients, the AI committee demonstrated high accuracy in identifying absolute contraindications to transplantation (98.2%), predicting 6-month survival (94.9%), and predicting 1-year survival (92.0%). The authors conclude that multiagent LLM systems may be capable of providing objective and reproducible assessments of transplant candidacy and could potentially serve as clinical decision-support tools to reduce variability in transplant selection decisions.

Strengths and Weaknesses:
A major strength of this study is its innovative design and large sample size, incorporating 8,412 transplant recipients from a national registry and creating a multidisciplinary AI committee that mirrors real-world transplant selection processes. The investigators evaluated clinically meaningful outcomes, including post-transplant survival and recognition of contraindications, and examined fairness metrics, bias, and decision-making patterns among the individual AI agents. The study is also notable for moving beyond simple chatbot question-answering tasks and evaluating AI performance in a complex, high-stakes clinical decision-making environment.
The study also has important limitations. It was retrospective and relied on registry data rather than prospective real-world clinical implementation. The investigators created a hypothetical cohort by artificially assigning contraindications to transplant candidates, which may not fully reflect the complexity of actual patients who are denied transplantation. In addition, the AI committee was evaluated against post-transplant outcomes among patients who had already been selected for transplantation, introducing potential selection bias. The model's specificity was only moderate, particularly for 1-year survival prediction (66%), and the study does not address how AI recommendations would affect actual committee deliberations, patient autonomy, or legal and ethical accountability.

Relevance:
This article is highly relevant to consultation-liaison psychiatry because it explores the use of AI to replicate multidisciplinary decision-making processes that incorporate both medical and psychosocial variables. The inclusion of a dedicated social worker agent is particularly noteworthy, as psychosocial assessments, treatment adherence, social support, and substance use are critical components of transplant psychiatry evaluations. The study raises important questions about the future role of AI in resource allocation, capacity assessments, risk stratification, and other ethically complex decisions that are central to CL psychiatry. At the same time, it highlights concerns regarding bias, transparency, explainability, and the risk of delegating value-laden clinical judgments to algorithmic systems. For psychiatrists working within transplant programs, this study provides an early glimpse into how AI may eventually function as a decision-support tool within transplant selection committees while underscoring the continued need for human clinical judgment and ethical oversight.