PURPOSE: Bronchoscopic lung volume reduction therapy(B-LVRT)has been emerging as a novel minimally invasive technology to improve lung function and relieve symptoms(e.g.,short of breath)in patients with severe emphysema.However,it is difficult to decide which patients could actually benefit from this surgery.The objective of this study is to develop and validate a computer tool that can aid in an efficient and accurate assessment of this surgery.METHODS: Given a CT chest exam,a computer tool was used to automatically identify a number of important anatomical landmarks,including pulmonary fissures,lobes,vasculatures,airways,and emphysema.Four types of image features were quantified:(1)the severity of emphysema in the target lobe,(2)the heterogeneity of emphysema in neighboring lobes,(3)fissure integrity,and(4)the passing of airways/vasculatures through fissures.When quantifying emphysema extent,a threshold of-950 Hounsfield unit(HU)was used.The spatial heterogeneity of emphysema was assessed by computing their difference in volume in the neighboring lobes.The completeness of the pulmonary fissures was computed by comparing the spatial discrepancies between the identified fissures and lobes.We in particular checked whether there are any vasculatures passing through the identified fissures.For validation purpose,we collected nine pre-operative CT exams that had the B-LVRT outcomes,among which three exams were labeled as poor outcome and the left ones as success.