[OA082] Towards normalization of contrast enhancement of pulmonary parenchyma on subtraction CT imaging

Purpose To determine the impact of patient and acquisition parameters on the quantification of the enhancement of pulmonary parenchyma in subtraction computed tomography pulmonary angiography (CTPA) imaging in patients without pulmonary pathology. Methods 2515 patients underwent clinical CTPA with subtraction imaging on a 320-multidetector row CT scanner (Aquilion ONE VISION or GENESIS, Canon Medical Systems Corporation, Japan) between January 2013 and August 2017 for suspicion of pulmonary embolism. In subtraction CT, an unenhanced chest CT is subtracted from an enhanced CT after motion correction, resulting in an iodine map representing pulmonary perfusion. After examination of patient records and images, 74 patients (2.9 %) without any apparent pulmonary or systemic pathology were selected. The mean enhancement in circular regions of interest (ROI) in all lung lobes of the iodine maps, placed to avoid vessels depicted in the CTPA images, were calculated. Patient age, sex, BMI, and smoking habits, as well as CT scan time, diaphragm position and lung volume changes between the unenhanced and enhanced CT, type and volume of contrast, and enhancement in the aortic arch, pulmonary trunk and left atrium were collected. Linear regression analysis was used to evaluate the correlation between the parameters and the measured enhancement. Results The mean enhancement of the ROIs throughout the whole lung was 32.2 ± 21.6 HU. The enhancement in the lower lobes (39.4 ± 24.3 HU) was significantly higher than in the upper lobes (31.7 ± 22.3 HU, p  r 2 = 0.54 p 0.001 ) , and with lung volume changes ( r 2 = 0.73 p 0.001 ) between the scans. No correlation was found between the mean enhancement and all other variables studied. Conclusions The enhancement of the pulmonary parenchyma on subtraction CT imaging in patients without pulmonary pathology varies widely. However, normal enhancement values correlate strongly with changes in diaphragm position and in pulmonary volume between scans. This implies that normalization of lung enhancement based on lung volume changes could improve clinical interpretation and would be the first step towards quantification of enhancement in subtraction CT imaging.