Delorme S, Zuna I, Huber S, Albert B, Bahner M L, Junkermann H, van Kaick G
Department of Oncologic Diagnostic and Therapy, German Cancer Research Centre, Im Neuenheimer Feld 280, D-69 120 Heidelberg Germany.
Eur Radiol. 1998;8(2):189-93. doi: 10.1007/s003300050361.
The aim of our study was to reassess the diagnostic performance of image-based, computer-assisted colour Doppler quantification under routine conditions. We used a computer-assisted protocol that quantitatively characterises a colour Doppler image by the relative amount of colour pixels (colour pixel density, CPD) and the colour hues (numerically expressed by the mean colour value, MCV) in 44 patients with breast carcinoma and 49 patients with benign breast lesions. Studies were carried out over two periods by two examiners, subsequently in charge of breast ultrasound. During the first period, the sensitivity of the MCV was 92 %, the specificity 75 %; the sensitivity of the CPD was 80 %, the specificity 81 %. During the second period, the sensitivity of the MCV was 58 %, the specificity 77 %; the sensitivity of the CPD was 68 %, the specificity 71 %. Despite measures to create uniform examination conditions, the diagnostic performance of this method may decline under routine conditions.
我们研究的目的是在常规条件下重新评估基于图像的计算机辅助彩色多普勒定量分析的诊断性能。我们采用了一种计算机辅助方案，通过彩色像素的相对数量（彩色像素密度，CPD）和色彩色调（以平均色彩值，MCV进行数值表示）对44例乳腺癌患者和49例乳腺良性病变患者的彩色多普勒图像进行定量分析。研究由两名负责乳腺超声检查的检查者分两个阶段进行。在第一阶段，MCV的敏感性为92%，特异性为75%；CPD的敏感性为80%，特异性为81%。在第二阶段，MCV的敏感性为58%，特异性为77%；CPD的敏感性为68%，特异性为71%。尽管采取了措施来创造统一的检查条件，但在常规条件下该方法的诊断性能可能会下降。
