Comparison of Color Flow Doppler Sonography (CFDS) and Immunohistologic Detection of Microvessels for the Assessment of the Malignancy of Thyroid Nodules

Abstract

The assessment of tumor vascularization by color flow Doppler sonography (CFDS) has been suggested for the distinction between benign and malignant thyroid nodules. Our objective was to investigate if the CFDS results reflect the percentage of histologically determined microvessels in adenomas (As), adenomatous nodules (ANs), and papillary carcinomas (PCs). Tissue sections from 10 adenomas, 8 ANs and 13 PC and surrounding tissue of 10 PCs and 2 benign nodules were immunostained for CD34. A computerized image analysis was used to determine the microvessel density in four hot spots and ten systematically selected fields. Preoperatively CFDS was performed and classified according to Frates et al. We found a consistent percentage increase of CD34 stained microvessels in PCs (83 and 96%) as compared to adenomas and ANs (38 and 49%) determined by the hot spot analysis and systematic field analysis. A ROC analysis on the basis of the histologically determined number of microvessels demonstrated 70% microvessels as an optimal cut point for the diagnosis of PC with the highest sensitivity of 92% and highest specificity of 89%. The analysis of the CFDS-classification IV for the distinction between PCs and adenomas and ANs showed a sensitivity of 62% with a specificity of 100%. The lower sensitivity of the CFDS classification as compared with the immunohistologic determination of the microvessel density indicates that the CFDS classification detects the pathognomonic intranodular microvessels only incompletely. The higher CFDS specificity is most likely due to the detection of other vascular aspects of malignancy in addition to intranodular microvessels.