Large Language Models for Cardiovascular Disease, Cancer, and Mental Disorders: A Review of Systematic Reviews

Abstract

Background/Objective: The use of Large Language Models (LLMs) has recently gained significant interest from the research community toward the development and adoption of Generative Artificial Intelligence (GenAI) solutions for healthcare. The present work introduces the first meta-review (i.e., review of systematic reviews) in the field of LLMs for chronic diseases, focusing particularly on cardiovascular, cancer, and mental diseases, with the goal to identify their value in patient care, as well as challenges for their implementation and clinical application. Methods: A literature search in the bibliographic databases of PubMed and Scopus was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, in order to identify systematic reviews incorporating LLMs. The original studies included in the reviews were synthesized according to their target disease, specific application, used LLMs, data sources, accuracy, and key outcomes. Results: The literature search identified 5 systematic reviews respecting our inclusion and exclusion criteria, which examined 81 unique LLM-based solutions. The highest percentage of the solutions targeted mental disease (70 studies, 86%), followed by cancer (6 studies, 7%) and cardiovascular disease (5 studies, 6%). Generative Pre-trained Transformer (GPT) models were used in most studies (45 studies, 55%), followed by Bidirectional Encoder Representations from Transformers (BERT) models (33 studies, 40%). Key application areas included depression detection and classification (31 studies, 38%), suicidal ideation detection (6 studies, 7%), question answering based on treatment guidelines and recommendations (6 studies, 7%), and emotion classification (4 studies, 5%). Studies were highly heterogeneous, with most studies (41 studies, 50%) focusing on clinical/diagnostic accuracy using metrics for correct diagnosis, agreement with guidelines or model performance, whereas other studies (34 studies, 42%) were descriptive and focused on narrative outcomes, usability, trust or plausibility. The most significant found challenges in the development and evaluation of LLMs include inconsistent accuracy, bias detection and mitigation, model transparency, data privacy, need for continual human oversight, ethical concerns and guidelines, as well as the design and conduction of high-quality studies. Conclusion: The results of this review suggest LLMs could become valuable tools for enhancing diagnostic precision and decision support in cardiovascular disease, cancer and mental health. Given the limited number of studies included in this review and their moderate quality, we urge the research community to conduct more investigations in real-world intervention settings to better demonstrate the clinical utility of LLMs.

Ähnliche Arbeiten