Reduction of Poor Contrast Enhancement of the Pulmonary Artery in Computed Tomography Angiography Using an Alternative Respiratory Maneuver Fernando Uliana Kay, MD, Ana Carolina Sandoval Macedo, MD, Rodrigo Caruso Chate, MD, Gilberto Szarf, MD, PhD, Gustavo Borges da Silva Teles, MD, Roberto Sasdelli Neto, MD, and Marcelo Buarque de Gusma ˜o Funari, MD, PhD Purpose: The aim of the study was to compare the effects of dif- ferent respiratory maneuvers in computed tomography pulmonary angiography for the diagnosis of pulmonary embolism (PE) on the contrast enhancement of pulmonary circulation and on the quality of lung window images. Materials and Methods: A retrospective analysis of 520 examina- tions, half obtained after deep inspiration followed by breath- holding and half solely during breath-holding. Subjective quality analyses and objective measurements of pulmonary arterial enhancement and lung parenchyma attenuation were performed. Results: Elimination of deep inspiration reduced suboptimal opa- cification of the pulmonary artery (PA), from 7.3% to 2.7%, with 2.7% of the deep inspiration scans having attenuation values <150 Hounsfield units (HU). The prevalence of PE was similar between the groups (19% vs. 23%, respectively), with excellent interob- server diagnostic agreement (k = 0.89 to 0.91). Lung windows were compromised in 6.9% of the studies with respiratory pause, and these examinations had a higher attenuation of the lung paren- chyma (median: 709.8 HU) compared with deep inspiration ( 794.8 HU). A positive correlation between attenuation of the PA and the ascending aorta was observed (r = 0.40 to 0.56). Conclusions: Eliminating deep inspiration before image acquisition had opposite effects with the same magnitude: it caused a reduction in inadequate PA enhancement at the cost of an increased number of nondiagnostic lung images and did not compromise diagnostic consistency for PE. Key Words: multidetector computed tomography, pulmonary embolism, quality improvement (J Thorac Imaging 2014;29:107–112) M ultidetector computed tomography pulmonary angiography (CTPA) combined with clinical risk score stratification and D-dimer measurement has become an essential step in the workup of patients with suspected pulmonary embolism (PE). 1,2 However, available data dem- onstrate that, on average, 6% (range, 1% to 9%) of CT angiography studies for PE detection can be inconclusive due to inadequate image quality. 3–6 Many of these poor-quality examinations are due to motion artifacts, flow-related arti- facts, or insufficient enhancement of the pulmonary arterial circulation. 7,8 Poor contrast enhancement of the pulmonary artery (PA) is cited as the cause of indeterminate CTPA in 40% to 55% of cases. 4,5 In these cases, variables that can affect arterial enhancement, and thus lead to insufficient vessel enhancement, include injection duration, contrast medium traveling time, 9 age, and body weight. 10 However, some studies have demonstrated an intriguing phenomenon, referred to as “transient interruption of contrast,” which occurs despite an adequate adjustment of injection parame- ters and continuous monitoring of the contrast arrival at the PA. 11,12 This phenomenon is characterized by inadequate opacification of the pulmonary circulation (PC) despite continuous arrival of contrast material at the superior vena cava. A putative role of atrial-level right to left shunts in the genesis of the phenomenon has been proposed 11 ; however, an alternate physiological mechanism is the preferred explanation—the increase in venous return through the inferior vena cava during continuous inspiration. 12–14 Nev- ertheless, we are not aware of any study to date that has directly examined whether the suppression of continuous inspiration immediately before and during image acquisition would reduce the number of CTPA studies with poor-quality opacification of the pulmonary arterial circulation or the impact that this potential reduction has on diagnostic con- sistency in clinical practice. The aim of this study was to retrospectively analyze and compare the quality of CTPAs from a large sample of patients referred for investigation of PE using 2 different imaging protocols: patients in 1 cohort were scanned during a “respiratory pause,” defined as suspended breathing without deep inspiration, and patients in the control cohort were scanned during breath-hold at the end of a maximum inspiratory effort. MATERIALS AND METHODS Subjects We retrospectively analyzed the CTPAs of consecutive patients (n = 520 scans) with clinical suspicion of acute PE and who were investigated at our institution from September 2010 to July 2011. The institutional ethics committee approved the study. Anonymity of the patients was ensured, and informed con- sent was waived due to the retrospective nature of the study. Considering previous evidence suggesting that the deep inspiration maneuver could increase the number of nondiagnostic studies, 11,12,14,15 all examinations conducted until March 5, 2011 were obtained during a “respiratory pause” (ie, patients were instructed to hold their breath From the Imaging Department, Hospital Israelita Albert Einstein, Sao Paulo, Brazil. The authors declare no conflicts of interest. Reprints: Fernando Uliana Kay, MD, Imaging Department, Hospital Israelita Albert Einstein, Av. Albert Einstein, 627/41 floor D, 05652- 901 Sao Paulo/SP, Brazil (e-mail: fernando.kay@einstein.br). Copyright r 2013 by Lippincott Williams & Wilkins ORIGINAL ARTICLE J Thorac Imaging  Volume 29, Number 2, March 2014 www.thoracicimaging.com | 107