Context: Thyroid nodules are selected for biopsy on the basis of clinical and ultrasound (US) findings. Ultrasonography detects nodules at risk of malignancy, but its diagnostic accuracy does not rule out with certainty the possibility of cancer in lesions without suspicious findings.
Objective: The objective of the study was to evaluate the diagnostic accuracy of real-time elastography (RTE) in thyroid nodules and to assess the improvement provided by combination of RTE, B-mode US, and color flow Doppler (CFD).
Design: This was a prospective multicenter study.
Patients: A consecutive series of 498 thyroid nodules was blindly evaluated by US, CFD, and RTE before biopsy or surgery. Nodules were classified at RTE by four-class color scale. Patients with benign cytology underwent follow-up over 12 months, whereas patients with indeterminate, suspicious, or malignant cytology were surgically treated.
Results: At follow-up, 126 nodules were malignant and 372 benign. RTE classes III-IV showed 81% sensitivity and 62% specificity. The presence of at least one US risk factor (hypoechogenicity, microcalcifications, irregular margins, intranodular vascularization, and taller than wide shape) had 85% sensitivity and 91% negative predictive value. When RTE was combined with US, the presence of at least one of the six parameters had 97% sensitivity and 97% negative predictive value, with an odds ratio of 15.8 (95% confidence interval 5.7-43.8).
Conclusions: RTE is a valuable tool for detecting malignant thyroid lesions with a sensitivity similar to traditional US and CFD features. By adding RTE evaluation, the sensitivity for malignancy of US findings is markedly increased and the selection of nodules that do not need cytology is made more reliable.