The association of colour flow Doppler sonography and conventional ultrasonography improves the diagnosis of thyroid carcinoma.

OBJECTIVE

In the present study, we compared the results of conventional ultrasonography (US) and colour flow Doppler sonography (CFDS) with those of US guided fine needle aspiration biopsy (FNAB) and of pathologic staging of resected thyroid nodules, to assess the relative importance of US and CFDS in discriminating malignant thyroid nodules.

SUBJECTS AND STUDY DESIGN

We retrospectively reviewed records of 230 patients submitted to US-guided FNAB before surgery for solitary, not hot thyroid nodules. Before US guided FNAB, they were examined with conventional US and CFDS. Conventional US evaluated nodule size, echogenicity, presence of halo sign and microcalcifications. CFDS evaluated the vascular pattern classified as types I, II and III. Twenty-seven patients with inadequate cytology were excluded from this study (11.7%).

RESULTS

Two hundred and three patients underwent surgery. At histology a thyroid carcinoma was found in 36 patients (17.7%) and a benign nodule was observed in 167 patients (82.3%). We did not find any difference in cancer prevalence between nodules with a primary tumour size < or =1 cm and those >1 cm (17.6 vs. 17.7%; p = 0.99). A solid echo texture was not statistically significant to suggest malignancy (p = 0.32). Microcalcifications were seen in 83.3% (30/36) of malignant nodules and in 33.5% (56/167) of benign nodules. These results were statistically significant (p < 0.0001). The type III flow as determined by CFDS was a statistically significant criterion to suggest malignant disease (p < 0.005). The most predictive findings of malignancy on conventional US was the combination of microcalcifications plus the absence of halo sign (sensitivity 75%, specificity 71.9%, p < 0.0001). The combination of an absence of halo sign on conventional US and a type III pattern on CFDS presented the higher sensitivity (83.3%) for malignancy with a specificity of 43.7%. Microcalcifications on US in combination with a type III CFDS pattern showed a lesser sensitivity (80.6%) with an improved specificity (75.4%). In our opinion, the better balanced combination of US and CFDS features was the absence of halo sign plus microcalcifications and a type III CDFS pattern (sensitivity 72.2%, specificity 77.2%).

CONCLUSIONS

The combination of conventional US and CFDS provides benefits in increasing the screening sensitivity and accuracy in distinguishing malignant thyroid nodules.