Assessment of fine-needle aspiration sampling technique in thyroid nodules

BACKGROUND: A prospective two-yr study was undertaken to assess the heterogeneity of thyroid nodules using the fine-needle aspiration (FNA) technique of systematic regional sampling. In addition, we determined the number of regions to be sampled to minimize non-diagnostic results, and to optimally characterize thyroid nodules. DESIGN: FNA was performed on 74 nodules > or = 1.5 cm. in diameter in five distinct regions in sequence (center, then four quadrants starting at 12:00, clockwise). Slides from each region were coded, randomized, subjected to blind review, and categorized as non-diagnostic (ND), benign (B), indeterminate (ID), suspicious/neoplastic (S/N), or malignant (M). Final cytologic diagnosis (CD) was made from all slides of each nodule. RESULTS: The ND rate for center FNAs alone was 16%, but addition of the 12:00 region decreased it to 5.3%. With 3, 4, or 5 sequential sites the nondiagnostic rates were 4, 2.6, and 2.6%. The center region diagnosis was identical to the final CD in 71% of the cases. Addition of the 12:00 region increased the concordance to 88%. Three sequential regions equaled the CD in 93% of cases, and 4 regions equaled the CD in 99% of cases. All nodules characterized as M or S/N were resected as were 76% of the ID nodules. Of the 43 nodules characterized as B, 3 were resected, 24 involuted, 6 were unchanged, and 10 were lost to follow-up. All 3 M nodules proved malignant by histology, as did 7/10 S/N, 0/17 ID, and 0/43 B nodules; 3/10 S/N, 1/17 ID and 1/43 B were adenomas. Likelihood ratios for diagnosing neoplasia were ND:0, B:0.10, ID:0.21, S/N:infinity, M:infinity. CONCLUSIONS: Sampling of at least four distinct regions accurately assesses thyroid nodules while minimizing ND results. Regional sampling also addresses intranodular heterogeneity.