Publications by authors named "K Sue Kehl"
Databases that link molecular data to clinical outcomes can inform precision cancer research into novel prognostic and predictive biomarkers. However, outside of clinical trials, cancer outcomes are typically recorded only in text form within electronic health records (EHRs). Artificial intelligence (AI) models have been trained to extract outcomes from individual EHRs.
View Article and Find Full Text PDFArticle Synopsis
- Researchers are merging unstructured patient data with structured health records to create the MSK-CHORD dataset, consisting of varied cancer types from nearly 25,000 patients at Memorial Sloan Kettering Cancer Center.
- This dataset allows for in-depth analysis of cancer outcomes using advanced techniques like natural language processing, revealing new relationships that smaller datasets may not show.
- Using MSK-CHORD for machine learning models, findings suggest that incorporating features from these unstructured texts can better predict patient survival than relying solely on genomic data or cancer staging.
Objective: Data extraction from the published literature is the most laborious step in conducting living systematic reviews (LSRs). We aim to build a generalizable, automated data extraction workflow leveraging large language models (LLMs) that mimics the real-world two-reviewer process.
Materials And Methods: A dataset of 10 clinical trials (22 publications) from a published LSR was used, focusing on 23 variables related to trial, population, and outcomes data.
Article Synopsis
- Current methods for identifying immune-related adverse events (irAEs) in patients undergoing immune checkpoint inhibitor (ICI) therapy are not very effective, but large language models (LLMs) show promise in improving this process.
- In a study, LLMs were compared to manual reviews and ICD codes for detecting common irAEs in hospitalized patients, demonstrating significantly higher sensitivity especially for conditions like hepatitis and myocarditis.
- The LLM was faster in analysis—averaging 9.53 seconds per chart compared to 15 minutes for manual adjudication—indicating that LLMs could be a valuable tool in clinical settings for accurately identifying irAEs.
Purpose: Eastern Cooperative Oncology Group (ECOG) performance status (PS) is a key clinical variable for cancer treatment and research, but it is usually only recorded in unstructured form in the electronic health record. We investigated whether natural language processing (NLP) models can impute ECOG PS using unstructured note text.
Materials And Methods: Medical oncology notes were identified from all patients with cancer at our center from 1997 to 2023 and divided at the patient level into training (approximately 80%), tuning/validation (approximately 10%), and test (approximately 10%) sets.