Objectives: To compare the diagnostic performance of ultrasound, contrast-enhanced computed tomography (CT) and (18)F-FDG positron emission tomography (PET)/CT for detecting recurrent differentiated thyroid cancer in the neck.

Methods: Twenty patients who had undergone previous surgery for differentiated thyroid cancer (19 papillary carcinomas; 1 medullary carcinoma) and presented with pathologically proven recurrence in the neck were included. All patients had undergone ultrasound, CT and PET/CT in the 2 months before further surgery. In each patient, ultrasound, CT and PET/CT images were retrospectively reviewed to determine the presence of loco-regional recurrence by level-by-level analysis. Imaging results were correlated with the histological evaluation of the neck dissection as a standard of reference.

Results: Recurrences were found at 52 out of 110 cervical nodal levels surgically explored. The sensitivity, specificity and accuracy were 69.2 %, 89.7 % and 80.0 % for ultrasound; 63.5 %, 94.8 % and 80.0 % for CT; and 53.8 %, 79.3 % and 67.3 % for PET/CT, respectively. ROC analysis revealed higher diagnostic performance with ultrasound than with PET/CT for detecting recurrent tumour.

Conclusions: Although no significant difference was found among the three techniques, the sensitivity and specificity of ultrasound and CT were higher than those of PET/CT for the evaluation of cervical recurrence in patients with differentiated thyroid cancer.

Key points: • Ultrasound, CT and ( 18 ) F-FDG PET/CT can all detect recurrent thyroid cancer. • Ultrasound and CT have higher sensitivity and specificity. • Ultrasound, CT and ( 18 ) F-FDG PET/CT frequently demonstrated discordant findings.