CREATING A MODEL TO PREDICT MALIGNANT PULMONARY NODULES CALIBRATED FOR THE INTERVENTIONAL PULMONOLOGIST: THE ASSURE MODEL

TOPIC: Lung Cancer TYPE: Original Investigations PURPOSE: A number of predicting models have been developed to estimate the pretest probability malignancy in indeterminate pulmonary nodules (IPN), but none are specifically calibrated for patient population referred lung nodule clinics consideration minimally invasive biopsy. We aimed identify independent predictors associated with this population. prediction model based on these factors will inform shared decision making regarding common clinical problem. METHODS: Consecutive patients who underwent electromagnetic navigational bronchoscopy (SuperDimension, Medtronic, Minneapolis, MN) diagnosis peripheral lesions between 2017-2019 at a large tertiary referral center were retrospectively identified. Demographic, radiographic, and procedural abstracted verified by two physicians (KL RJL). Two-year follow-up was documented all not diagnosed as malignant confirm benign disease. Bronchoscopic biopsy deemed diagnostic if lesional tissue acquired during procedure (defined histological findings which convincingly explain presence nodule) no evidence through years those initially pathology. Descriptive statistics performed Wilcoxon signed-rank Pearson chi-square tests used compare nondiagnostic groups continuous categorical variables, respectively. RESULTS: total 450 biopsied 394 included analysis. The median lesion size 18 mm (IQR, 14-28). Mean age 64 ±12 years, 44% men, 75% current or former smokers. prevalence 61% (276 lesions). Factors older (mean 66 ± 11 vs. 61 13 p <0.01), smoker (OR 2.1, 95% CI 1.2 3.4, = 0.01), upper lobe 1.5, 1.0 2.2, 0.03), spiculation edge 2.7, 1.8 4.0, < lobulation 1.8, 2.9, radiographic emphysema 2.0 4.4, 0.01). benignity multiple macronodules 0.5, 0.3 0.9, CONCLUSIONS: cohort clinic subsequent higher than screening populations lower resection. Existing from cohorts resection therefore likely perform sub-optimally. CLINICAL IMPLICATIONS: These be create predictive tool characteristics clinic. DISCLOSURES: No relevant relationships Sameer Avasarala, source=Web Response Heidi Chen, Sheau-Chiann Joyce Johnson, James Katsis, Robert Lentz, Kaele Leonard, PI ofan investigator-initiated study relationship Medtronic Please note: >$100000 Fabien Maldonado, Response, value=Grant/Research Supportco-I studies Janssen $20001 - $100000 Support Removed 04/19/2021 ResponsePI Erbe $5001 $20000 Consulting value=Honorariaco-I industry-sponsored trial Therapeutics Supportspeaker Longitudinal fellows program $1001 $5000 value=Honoraria Board director member AABIP $1-$1000 value=Travel Consultant Medtronic/Covidien Otis Rickman, value=Consulting fee Advisory Committee Member Olympus 04/29/2021 Intuitive Surgical Auris Health Abbvie Clinical Trial Spiration Becton, Dickinson Lance Roller,