QUANTITATIVE LUNG FUNCTION IMAGING USING NOVEL INTEGRATED JACOBIAN VENTILATION METHOD IN A HEALTHY COHORT WITH NO RESPIRATORY SYMPTOMS

Abstract

SESSION TITLE: Radiology Done by the Nonradiologist SESSION TYPE: Original Investigations PRESENTED ON: 10/23/2019 10:45 AM - 11:45 AM PURPOSE: Regional ventilation differences recovered from changes in local tissue volumes at inspiration and expiration CT scan would be useful to delineate and quantify longitudinal parenchymal and airway related changes. However, previous studies using deformable image registration (DIR) has shown to be unstable and non-reproducible. We recently developed a method with quantified controllable levels of uncertainty – Integrated Jacobian Ventilation (IJV), which has robust numerical stability. In this study, we studied the quantitative lung function imaging (QLFI) differences using IJV and correlated this to concurrent pulmonary function study (PFT) from a well-characterized, healthy cohort with no respiratory symptoms. We hypothesized QLFI would correlate well with pulmonary function study. METHODS: We included 81 available CT scans from the NORM dataset. Patients were older than 18 years and had no respiratory symptoms or diseases. All subjects had normal pulmonary function. For each case, the lungs were segmented using an automated machine learning method (convolutional neural network) and DIR was applied to estimate inhale-to-exhale lung motion. IJV was then computed according to the segmented lung volumes and the DIR-recovered spatial transformation. We classified participants into smokers versus non-smokers. RESULTS: There were no significant differences in baseline characteristics between the 46 smokers and 35 non-smokers . Both PFT and QLFI based total ventilation were not different in this cohort with no respiratory symptoms. Pearson correlation between QLFI derived total lung capacity and PFT showed an 80% correlation (p<0.0001). However, on regional analysis of QLFI lung map there were several cases with abnormal ventilation showing “cold spots” with decreased ventilation score, although there was no visible corresponding radiographic abnormality on High resolution CT scan. On multivariate linear regression model after adjusting for sex, smoking history and abnormal CT scan, age (β coefficient 0.01877, 95% CI: 0.00445 – 0.03310, p 0.01) and height (β coefficient 0.08933, 95% CI: 0.05489 – 0.12377, p <0.0001) were significantly associated with QLFI. CONCLUSIONS: - Our results indicate that there is a significant correlation between QLFI and concurrent pulmonary function values. - Age and Height are significant predictors for QLFI after adjusting for all other parameters. CLINICAL IMPLICATIONS: Future studies should distinguish QLFI based mapping to distinguish regional differences in lobar ventilation. QLFI can be used as a valuable imaging tool for assessing disease progression and quantify effects of intervention in patients with diffuse lung disease. DISCLOSURES: No relevant relationships by Edward Castillo, source=Web Response No relevant relationships by Aleksa Fortuna, source=Web Response other relationship with Imbio, LLC Please note: $1001 - $5000 Added 01/17/2019 by Craig Galban, source=Web Response, value=Royalty<br Speaker/Speaker's Bureau relationship with Boehringer Ingelheim Please note: $1-$1000 Added 03/13/2019 by Girish Nair, source=Web Response, value=Honoraria