Normal variance in emphysema index measurements in 64 multidetector-row computed tomography

The purpose of this study was to identify the normal variance of emphysema index (EI) measured in examinations acquired with 64 multidetector-row computed tomography (64-MDCT). A longitudinal, noninterventional study was performed retrieving all patients in our institution who are currently registered in our lung nodule protocol. All patients with clinical, functional, or significant radiological changes were excluded. We assumed that EI should remain unchanged within a short period of time. We reviewed 475 MDCTs in order to select 50 clinically stable patients who had two sequential chest MDCTs performed within a time interval of less than three months, and who presented at least one lung free of abnormalities but emphysema. CT densitovolumetry was used to calculate EI with thresholds set at -950 Hounsfield units (HUs) (EI-950) and -970 HUs (EI-970); on both studies from each patient. We observed the variation of total lung volume (TLV), mean lung density (MDL), and EI for measurements at the baseline and at follow-up scans. Differences observed between baseline and follow-up measurements were: TLV μ= 149 ml; IC = μ +1.96 (133); EI-950 μ = 0.02%; p 95 = 0.89%; EI-970 μ = 0.04%; p 95 = 0.23% and MLD μ = 15 HU; IC = μ +1.96 (18). The correlations obtained were the following: TLV r = 0.96, EI-950 r = 0.79, EI-970 r = 0.85. Accepting that emphysema would remain unchanged within three months on stable patients, differences of less than 0.89% for EI-950 and of less than 0.23% for EI-970 are within the variance of the method.