Zeiberg A S, Silverman P M, Sessions R B, Troost T R, Davros W J, Zeman R K
Department of Radiology, Georgetown University Hospital, Washington, DC 20007, USA.
AJR Am J Roentgenol. 1996 Feb;166(2):293-9. doi: 10.2214/ajr.166.2.8553933.
The purpose of this study was to evaluate the application of helical CT-generated three-dimensional images of the upper airway.
Thirty patients, 10 healthy and 20 with upper-airway disease, were studied with helical CT (5-mm collimation). Overlapping images at 2-mm intervals were retrospectively generated. In the group of healthy patients, two radiologists in independently compared overlapping with nonoverlapping images, ranked confidence in identifying small airway structures on a scale of 1-5, and tabulated the number of images demonstrating these structures. In the 20 patients with disease, three-dimensional (3D) surface models were rendered on an independent workstation and were reviewed by two radiologists and one otolaryngologist for image quality, appreciation of lesion morphology, and ability to judge lesion extent, using a similar scale. A phantom was used to optimize parameters for the 3D reconstructions.
Viewing of the retrospectively generated overlapping images increased by 122% the number of images in which laryngeal and hypopharyngeal structures could be identified (p < .01). Image confidence scores for the radiologists averaged 3.3 for nonoverlapping and 4.0 for overlapping (p < .05). Radiologists and otolaryngologist rated the quality of the 3D images equally. The otolaryngologist's assessment of the value of the models for understanding the lesion morphology was 3.5 compared with the radiologists assessment of 2.5; and for judging the lesion extent, the otolaryngologist's assessment was 3.8 compared with 2.7 for the radiologist, a statistical significance of p < .01.
Helical CT with the application of overlapping images and 3D reconstructions significantly assists the understanding of upper-airway disease.
本研究旨在评估螺旋CT生成的上气道三维图像的应用。
对30例患者进行了研究，其中10例健康，20例患有上气道疾病，均接受螺旋CT（5毫米准直）检查。以2毫米间隔回顾性生成重叠图像。在健康患者组中，两名放射科医生独立比较重叠图像与非重叠图像，以1-5分的量表对识别小气道结构的信心进行评分，并列出显示这些结构的图像数量。在20例患病患者中，在独立工作站上生成三维（3D）表面模型，并由两名放射科医生和一名耳鼻喉科医生使用类似量表对图像质量、病变形态的清晰度以及判断病变范围的能力进行评估。使用一个模型来优化3D重建的参数。
回顾性生成的重叠图像使可识别喉和下咽结构的图像数量增加了122%（p < 0.01）。放射科医生对非重叠图像的信心评分平均为3.3分，对重叠图像的评分为4.0分（p < 0.05）。放射科医生和耳鼻喉科医生对3D图像质量的评价相同。耳鼻喉科医生对模型在理解病变形态方面的价值评估为3.5分，而放射科医生的评估为2.5分；在判断病变范围方面，耳鼻喉科医生的评估为3.8分，放射科医生为2.7分，p < 0.01具有统计学意义。
应用重叠图像和3D重建的螺旋CT显著有助于对上气道疾病的理解。
