Institute of Diagnostic and Interventional Radiology, University Hospital Zurich, Raemistrasse 100, CH-8091 Zurich, Switzerland.
AJR Am J Roentgenol. 2012 Jan;198(1):138-44. doi: 10.2214/AJR.11.6957.
The purpose of our study was to evaluate the accuracy of virtual unenhanced images reconstructed from contrast-enhanced dual-energy CT for the differentiation of incidental adrenal masses in comparison with standard unenhanced CT.
One hundred-forty patients (mean age, 74±9 years) underwent unenhanced and contrast-enhanced CT of the abdomen, the latter acquired with dual-energy for reconstruction of virtual unenhanced images. Two blinded and independent readers (R1 and R2) measured attenuation of each incidental adrenal mass on standard unenhanced and virtual unenhanced images using an optimized dual-energy three-material decomposition algorithm.
Fifty-one incidental adrenal masses were found in 42 of 135 patients (31%); 39 incidental adrenal masses were ≥1 cm. On the basis of unenhanced CT, 29 of 51 incidental adrenal masses and 22 of 39 incidental adrenal masses≥1 cm were classified as benign (HU<10). Virtual unenhanced image quality was rated as good or with mild impairment (2.45±0.83 for R1, 2.45±0.99 for R2). Image noise was 12.7±3.6 HU in unenhanced images and 8.8±2.0 HU in virtual unenhanced images (p<0.001). There was no significant difference in incidental adrenal mass attenuation between unenhanced and virtual unenhanced images for all incidental adrenal masses (5.9±21.0 HU vs 7.0±20.6 HU, p=0.48) and for those≥1 cm (6.6±18.5 HU vs 7.9±18.3 HU, p=0.87). Sensitivity, specificity, and accuracy of virtual unenhanced images for the characterization of incidental adrenal masses as probably benign were 76%, 82%, and 78% for R1 and 79%, 95%, and 86% for R2, respectively. For incidental adrenal masses≥1 cm, sensitivity, specificity, and accuracy increased to 95%, 100%, and 97% for R1 and 91%, 100%, and 95% for R2.
Reconstruction of virtual unenhanced images from contrast-enhanced dual-energy CT of the abdomen allows the characterization of the incidental adrenal masses with a good accuracy compared with standard unenhanced CT, with the most favorable results in incidental adrenal masses measuring≥1 cm.
本研究旨在评估对比增强双能 CT 重建的虚拟非增强图像在区分偶发肾上腺肿块方面的准确性，并与标准非增强 CT 进行比较。
140 例患者（平均年龄 74±9 岁）接受了腹部非增强和增强 CT 检查，后者采用双能采集进行虚拟非增强图像重建。两位盲法和独立的读者（R1 和 R2）使用优化的双能三物质分解算法，分别在标准非增强和虚拟非增强图像上测量每个偶发肾上腺肿块的衰减值。
在 135 例患者中的 42 例（31%）中发现了 51 个偶发肾上腺肿块；39 个偶发肾上腺肿块≥1cm。基于非增强 CT，51 个偶发肾上腺肿块中的 29 个和 39 个≥1cm 的偶发肾上腺肿块被归类为良性（HU<10）。虚拟非增强图像质量被评为良好或轻度受损（R1 为 2.45±0.83，R2 为 2.45±0.99）。非增强图像的图像噪声为 12.7±3.6HU，虚拟非增强图像的噪声为 8.8±2.0HU（p<0.001）。对于所有偶发肾上腺肿块（5.9±21.0HU 与 7.0±20.6HU，p=0.48）和≥1cm 的偶发肾上腺肿块（6.6±18.5HU 与 7.9±18.3HU，p=0.87），非增强图像与虚拟非增强图像的偶发肾上腺肿块衰减值之间无显著差异。R1 为 76%、82%和 78%，R2 为 79%、95%和 86%，虚拟非增强图像对偶发肾上腺肿块特征为可能良性的诊断具有较高的灵敏度、特异性和准确性。对于≥1cm 的偶发肾上腺肿块，R1 的灵敏度、特异性和准确性分别提高至 95%、100%和 97%，R2 分别提高至 91%、100%和 95%。
与标准非增强 CT 相比，腹部对比增强双能 CT 重建的虚拟非增强图像能够对偶发肾上腺肿块进行准确的特征描述，对于直径≥1cm 的偶发肾上腺肿块效果最佳。
