Given the rapidly expanding capabilities of generative AI models for radiology, there is a need for robust metrics that can accurately measure the quality of AI-generated radiology reports across diverse hospitals. We develop ReXamine-Global, a LLM-powered, multi-site framework that tests metrics across different writing styles and patient populations, exposing gaps in their generalization. First, our method tests whether a metric is undesirably sensitive to reporting style, providing different scores depending on whether AI-generated reports are stylistically similar to ground-truth reports or not.
View Article and Find Full Text PDFTreatment of patients with lung cancer during the current COVID-19 pandemic is challenging. Lung cancer is a heterogenous disease with a wide variety of therapeutic options. Oncologists have to determine the risks and benefits of modifying the treatment plans of patients especially in situation where the disease biology and treatment are complex.
View Article and Find Full Text PDF