Narrative Feature or Structured Feature? A Study of Large Language Models to Identify Cancer Patients at Risk of Heart Failure
arXiv:2403.11425v4 Announce Type: replace Abstract: Cancer treatments are known to introduce cardiotoxicity, negatively impacting outcomes and survivorship. Identifying cancer patients at risk of heart failure (HF) is critical to improving cancer treatment outcomes and safety. This study examined machine learning (ML) models to identify cancer patients at risk of HF using electronic health records (EHRs), including traditional ML, Time-Aware long short-term memory (T-LSTM), and large language models (LLMs) using novel narrative features derived from the structured medical codes. We identified
The increasing availability of large language models and electronic health records is enabling novel applications in medical diagnosis and risk prediction.
This development showcases the growing utility of AI, specifically LLMs, in extracting valuable insights from complex medical data to improve patient outcomes and healthcare efficiency.
Traditional machine learning methods are being augmented, and potentially surpassed, by LLMs in critically important medical identification tasks, leveraging both structured and narrative health data.
- · Healthcare providers
- · AI/ML developers
- · Cancer patients
- · Pharmaceutical companies
- · Traditional clinical risk models
- · Hospitals with limited AI integration
- · Data entry specialists
More accurate and early identification of cancer patients at risk of heart failure will lead to improved treatment plans.
The successful application of LLMs in this domain will accelerate their adoption and development for other complex medical conditions and risk assessments.
The integration of AI-driven predictive analytics into standard clinical practice could reduce healthcare costs and improve patient quality of life across numerous disease states.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG