Remy-Jardin Martine, Campistron Philippe, Amara Assia, Mastora Ioana, Tillie-Leblond Isabelle, Delannoy Valérie, Duhamel Alain, Remy Jacques
Department of Radiology, Hospital Calmette, University Center of Lille, Lille, France.
J Comput Assist Tomogr. 2003 Mar-Apr;27(2):266-73. doi: 10.1097/00004728-200303000-00028.
To evaluate the diagnostic accuracy of coronal thin sections as an alternative to transverse high-resolution computed tomography (HRCT) scans in the diagnostic approach to infiltrative lung disease (ILD) with multislice computed tomography (MSCT).
Fifty consecutive patients referred for suspicion of ILD underwent MSCT (collimation: 4 mm x 1 mm; pitch: 1.75; scan time: 0.5 seconds; 80 mA per slice) of the entire thorax. Two sets of lung images were systematically reconstructed: 1-mm thick transverse computed tomography (CT) scans (i.e., HRCT scans) (group 1) and 1-mm thick coronal images (group 2). Both series of images were obtained at 10-mm intervals and reconstructed with a high-spatial frequency algorithm. Two observers independently analyzed the overall image quality, the presence and distribution of CT features of ILD, and the diagnostic value of group 1 and group 2 lung images.
Group 1 and group 2 images were coded as interpretable, with minimal respiratory artifacts in the lower lung zones in two cases (4%). Presence of abnormal lung infiltration was found in 38 patients in group 2 with concordant interpretation of group 1 images. No significant difference was found in the identification of CT features of ILD between group 2 and group 1 (nodules: 32% vs. 30%; lines: 14% vs. 16%; increased attenuation: 24% vs. 26%; fibrosis: 48% vs. 50%; distortion: 46% vs. 50%; and abnormal interfaces: 16% in both groups). Distribution of lung abnormalities in central, peripheral, anterior, and/or posterior lung zones was similarly recognized in group 2 and group 1. In patients with extensive lung infiltration, the vertical predominance of lung changes was more precisely assessed in group 2 (n = 12) than in group 1 (n = 4). For a mean coverage of 260 mm in this study group, the mean number of sections to be interpreted was significantly lower in group 2 (19 sections) than in group 1 (28 sections) (P < 0.01).
Coronal sections allow a diagnostic approach to ILD as precise as that provided with HRCT scans, based on the interpretation of a significantly reduced number of images.
在多层螺旋计算机断层扫描(MSCT)对浸润性肺病(ILD)的诊断中,评估冠状薄层扫描作为横向高分辨率计算机断层扫描(HRCT)替代方案的诊断准确性。
连续50例疑似ILD患者接受了全胸部的MSCT检查(准直:4mm×1mm;螺距:1.75;扫描时间:0.5秒;每层80mA)。系统重建了两组肺部图像:1mm厚的横向计算机断层扫描(CT)图像(即HRCT图像)(第1组)和1mm厚的冠状图像(第2组)。两组图像均以10mm间隔获取,并用高空间频率算法重建。两名观察者独立分析了整体图像质量、ILD的CT特征的存在和分布,以及第1组和第2组肺部图像的诊断价值。
第1组和第2组图像被编码为可解释,仅2例(4%)下肺区存在最小限度的呼吸伪影。第2组38例患者发现有异常肺浸润,第1组图像解释与之相符。第2组和第1组在识别ILD的CT特征方面无显著差异(结节:32%对30%;线条:14%对16%;密度增加:24%对26%;纤维化:48%对50%;变形:46%对50%;异常界面:两组均为16%)。第2组和第1组对肺中央、外周、前部和/或后部肺区肺部异常分布的识别相似。在广泛肺浸润患者中,第2组(n = 12)比第1组(n = 4)更能准确评估肺部改变的垂直优势。在本研究组平均覆盖260mm的情况下,第2组(19层)要解释的平均层数显著低于第1组(28层)(P < 0.01)。
基于对显著减少的图像数量的解读,冠状位扫描对ILD的诊断方法与HRCT扫描一样精确。