B23-22 Clinical Utility of Fibresolve, an Imaging Artificial Intelligence Tool, in the Diagnosis of Idiopathic Pulmonary Fibrosis

Abstract Rationale Diagnosis of idiopathic pulmonary fibrosis (IPF) is challenging and frequently delayed by months to years. Access to expert multidisciplinary committee review, the gold standard for diagnosing interstitial lung disease (ILD), is limited, highlighting a need for clinically available diagnostic tools that may aid in challenging cases. Fibresolve is a novel, FDA-authorized artificial intelligence software designed to non-invasively detect findings consistent with IPF using computed tomography (CT) images. Previous studies of Fibresolve have demonstrated a sensitivity and specificity for IPF diagnosis of 56% and 86%, respectively, supporting the potential utility of Fibresolve to improve diagnostic confidence and obviate the need for more invasive testing. The aim of this study was to evaluate the utility of Fibresolve in active clinical use across multiple centers in the United States (US). Methods Fibresolve was utilized as an adjunct to diagnosis in cases of suspected IPF at the discretion of the evaluating physician across 34 centers in the US. Demographic data and Fibresolve test result (Positive, or likely to be IPF vs. Negative, or unlikely to be IPF) were collected on all patients. When available, additional information on subsequent diagnostic and treatment decisions following Fibresolve utilization were collected. Results Of 120 patients who underwent evaluation with Fibresolve, the median age was 74 years, 42.5% were female, and the median forced vital capacity was 84 percent predicted (Table 1). The majority of patients were not on background anti-fibrotic therapy at the time Fibresolve was utilized. Forty-six (38.3%) patients had a positive Fibresolve result, suggestive of an IPF diagnosis. Of 18 patients with available clinical follow-up data, 8 (44%) had a positive Fibresolve result. Seven (88%) of these 8 patients either began ILD-specific therapy (5 patients, 63%) or were referred for lung transplant (2 patients, 25%). Following Fibresolve utilization, 6 (33%) of the 18 patients with available follow-up data were able to avoid invasive diagnostic testing that was previously planned. Conclusion In this real-world study of patients with suspected ILD across the US, utilization of Fibresolve as a non-invasive adjunct to diagnosis suggested a diagnosis of IPF in approximately one-third of cases and obviated the need for invasive diagnostic procedures in several patients. Future studies evaluating the time to IPF diagnosis following Fibresolve utilization may strengthen the clinical utility of this accessible, FDA-approved diagnostic tool. This abstract is funded by: None