Retrospective Study of Carebot AI CXR Performance in Preclinical Practice

The performance of the trained and internally validated Carebot AI CXR software is tested on a set of 127 CXR images from target population. This is compared to common clinical practice, i.e., image assessment by a radiologist in a hospital. Patients may have a variety of findings; at this stage of the evaluation, an abnormal finding is considered to be an abnormality in any of the defined classes. False negative images incorrectly predicted by Carebot AI CXR software result in a clinical impact determination. To collect the CXR data for retrospective study, investigators addressed a municipal hospital in the Czech Republic that provides healthcare services to up to 130,000 residents of a medium-sized city (approximately 70,000 inhabitants) and the surrounding area. 127 anonymized CXR images were collected between August 15 and 17, 2022, and subsequently submitted to five independent radiologists of varying experience for annotation. The selected radiologists were asked to assess whether the CXR image shows any of the 12 pre-selected abnormalities. Pediatric CXR images (under 18 years of age), scans with technical problems (poor image quality, rotation), and images in lateral projection were excluded from the dataset.