Histogram-based comparison between dynamic and static lung perfused blood volume images using dual energy CT.

ABSTRACT

OBJECTIVE: The purpose of this study was to compare the results of a histogram-based analysis of static and dynamic lung perfused blood volume (LPBV) images. METHODS: Sixty-five patients (mean age 61.3 years, 36 male) underwent dynamic and static LPBV for evaluation of pulmonary vascular diseases (n = 11), lung carcinoma (n = 27) or pulmonary thromboembolism (PTE n = 27). Seven sets of dynamic sequential scans were performed at the pulmonary trunk using dual-energy technique before the static LPBV scan. The image of lung parenchyma that showed the greatest mean attenuation in dynamic series was defined as the peak dynamic LPBV image. The differences and correlations in the mean attenuation, image noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), histogram skewness and histogram kurtosis were evaluated according to the type of disease in static and dynamic LPBV images. RESULTS: Static LPBV images showed significantly larger mean attenuation (Rt24.2, Lt 24.2), SNR (Rt2.31, Lt2.30), and CNR (Rt2.40, Lt2.39), and smaller kurtosis values (Rt1.06, Lt0.61) values in comparison to dynamic LPBV images (p < 0.001); however, with the exception of kurtosis of the left lung (r = 0.17), these values were well-corrected with that of the dynamic LPBV images in these values (r = 0.4-0.77, p ≤ 0.001) without kurtosis of left lung (r = 0.17) in all patients. The histogram kurtosis of static LPBV image showed a good correlation with that of dynamic LPBV (r = 0.41-0.77, p < 0.05), especially in patients with PTE. CONCLUSION: In patients with PTE, the static LPBV image valueswere well correlated with the peak dynamic LPBV images which demonstrated pulmonary artery-dominant flow.