Early Detection of Pulmonary Congestion Using Quantitative CT and Lung US in a Porcine Model

Purpose To assess the value of quantitative CT and lung US in detecting pulmonary congestion across successive stages in a porcine model. Materials and Methods Eleven pigs (female, 5 months old) were examined (February 2016-September 2016) every 15 minutes during the induction (steps 2-5) and treatment (steps 6-8) of pulmonary congestion. All animals were assessed with quantitative CT, lung US, an arterial pulse contour cardiac output catheter, and right heart catheterization to measure lung water content, lung attenuation and B-lines, central venous pressure, pulmonary capillary wedge pressure (PCWP), cardiac output, and global end-diastolic volume. Correlation analyses were performed, and diagnostic performance of quantitative CT and lung US for detecting pulmonary congestion was assessed using the area under the receiver operating characteristic curve (AUC). Results Nine of eleven pigs (82%) completed the full protocol. The median quantitative CT attenuation showed high correlation with PCWP, especially during the intervention stage (r = 0.99; P < .05). Compared with no congestion (step 1), milder stages (step 2) of congestion were better identified using quantitative CT (AUC, 0.92) versus US B-lines (AUC, 0.50; P < .001). However, the two modalities did not differ in ability to detect more-severe stages of congestion (P_step3 = .14; P_step4 = .44; P_step5 = .75). During recovery, quantitative CT attenuation remained different from baseline levels (P = .01), whereas lung US B-lines reverted to baseline levels more rapidly (P = .63). Conclusion Quantitative CT showed high performance for assessing pulmonary congestion in its earliest subclinical stages. US B-lines were less sensitive, first appearing in the later stages of pulmonary congestion, but were more effective for monitoring regression of pulmonary congestion. Keywords: Heart, Lung, Quantitative CT, Ultrasound, Hemodynamics/Flow Dynamics Supplemental material is available for this article. © RSNA, 2026.