Professor of Radiology and Radiological Sciences Kim Sandler, MD, has been awarded an ACR Fund for Collaborative Research in Imaging (FCRI) grant for her project title "Evaluation of Machine Learning Algorithms for Lung Cancer Risk and All-Cause Mortality Prediction in a Non-Lung Screening Cohort." Dr. Sandler will serve as co-PI with Assistant Professor of Computer Engineering Yuankai Huo, PhD.
Dr. Sandler and Dr. Huo will oversee and complete two phases of the project. They will first work to validate their existing lung cancer and mortality risk prediction models on a cohort of patients with non-screening detected pulmonary nodules, building on models previously trained and tested in lung cancer screening populations.
In this phase, one aim will be to determine if a previously validated lung cancer risk prediction model for lung screening CT maintains similar performance in patients with non-screening detected pulmonary nodules.
Another aim will be to determine if mortality risk prediction utilizing a previously validated machine learning evaluation of body composition maintains performance in a cohort of patients with non-screening detected pulmonary nodules.
In phase two, this testing will be expanded to multiple institutions under the guidance of the ACR. Here, the aim will be to validate the algorithms on a more diverse set of patient populations, ensuring robustness across different imaging vendors and healthcare settings.
