Smith C W, Winfield A C, Price R R, Harding D R, Tucker S W, Witt W S, Hollifield J W
Radiology. 1982 Jul;144(1):51-4. doi: 10.1148/radiology.144.1.7089265.
Thirty-two patients being evaluated for hypertension by angiography were also studied by digital video subtraction angiography (DVSA). Twenty-three of the 76 renal arteries were found by conventional angiography to have significant lesions. Two experienced angiographers evaluated the DVSA studies without knowledge of the angiographic results. The accuracy of DVSA for evaluation of renal arteries was 87% for Observer I and 80% for Observer II. Sensitivities were 87% and 83% and specificities 87% and 79% for the two observers. Of the 13 patients with significant lesions, Observer I identified at lest one lesion in all 13 while Observer II identified a lesion in 12 of the 13. The high false-positive rate (26% for Observer I and 37% for Observer II) was thought to be caused by subtraction artifacts, quantum noise, relatively low spatial resolution, and the Mach effect.
对32例通过血管造影评估高血压的患者也进行了数字视频减影血管造影（DVSA）研究。在76条肾动脉中，传统血管造影发现23条有明显病变。两名经验丰富的血管造影师在不了解血管造影结果的情况下评估了DVSA研究。观察者I评估肾动脉的DVSA准确性为87%，观察者II为80%。两名观察者的敏感性分别为87%和83%，特异性分别为87%和79%。在13例有明显病变的患者中，观察者I在所有13例中至少识别出一个病变，而观察者II在13例中的12例中识别出病变。高假阳性率（观察者I为26%，观察者II为37%）被认为是由减影伪影、量子噪声、相对较低的空间分辨率和马赫效应引起的。
