Jason Akulian, MD, PhD, associate professor of medicine in the UNC Division of Pulmonary & Critical Care Medicine, will lead research efforts aimed to advance AI-driven lung cancer detection through bronchoscopy.
Researchers at the University of North Carolina will study an AI-based image analysis module for lung cancer through the ON-SITE study. The multicenter study in bronchoscopy combines Stimulated Raman Histology with Artificial Intelligence to enable rapid lung cancer detection and has enrolled its first patient.
The ON-SITE study aims to develop and validate an AI-based image analysis module for the NIO® Laser Imaging System, created by Invenio Imaging, that is intended to assist physicians in the detection of cancer in bronchoscopic lung biopsies in situations where rapid-on-site tissue evaluation (ROSE) is not available for the sample type.
The prevalence and mortality of lung cancer have led to large-scale screening efforts in high-risk patients. Despite significant investment in minimally-invasive biopsy technology, obtaining adequate tissue for biomarker and treatment determination remains a challenge. For this reason, bronchoscopy guidelines recommend ROSE for lung biopsies.
“ROSE requires that a cytologist or highly trained cytotechnician be physically present in the procedure room, and thus it is not available at many centers performing lung biopsy due to resource limitations,” said Jason Akulian, MD, PhD, UNC site PI.
The new technology will allow biopsies to be completed in the treatment room by OR-staff, as sample preparation will not require staining or section. Additionally, the NIO® Slides are also able to be recalled for later analysis.
“While still investigational, the promise of fast, in-room, accurate identification of tissue that is suspicious for cancer has the potential to ultimately lead to improved outcomes, a beneficial cost/benefit profile, and personalized treatments,” said Gustavo Cumbo-Nacheli, MD, pulmonologist at Corewell Health and one of the site PIs for the ON-SITE study.
The hope is the study will lead to FDA approval of the image analysis module for assisting physicians in identifying cells or tissue morphology that is suspicious for cancer in lung biopsies.
Other participating sites: The University of Texas MD Anderson Cancer Center, Corewell Health,and Memorial Sloan Kettering Cancer Center.