Thyroid cytology and risk of thyroid cancer: differences among indeterminate specimens.

OBJECTIVE: To assess the potential for stratification of indeterminate cytologic findings on fine-needle aspiration (FNA) of thyroid nodules in an effort to improve therapeutic strategies. METHODS: We attempted to determine the malignant risk associated with various indeterminate FNA cytologic patterns by correlation of specimens with the final histologic diagnosis. For this analysis, we identified 294 computerized medical records of surgically treated thyroid nodules during a 5-year period at our institution with the corresponding FNA cytology reports available. RESULTS: Of the 294 surgical cases, 162 with a positive or indeterminate cytologic report were selected, reviewed, and classified. Of 52 patients with positive cytologic findings on FNA, 51 (98%) had a final histologic report of a malignant thyroid nodule. Of 110 patients with indeterminate specimens, 30 (27%) had a final histologic diagnosis of thyroid carcinoma. The presence of nuclear atypia was predictive of thyroid carcinoma in 75% of patients, a Hürthle cell cytologic pattern was associated with a malignant thyroid nodule in 33%, and a hypercellular smear was suggestive of malignant involvement in 26% of cases. The lowest rate of malignant potential was associated with cytologic microfollicular and scant colloid alone subtype (6%). CONCLUSION: The results of this study show that indeterminate thyroid cytologic specimens can be subdivided into groups with different malignant risks. A microfollicular cytologic pattern in the absence of a hypercellular smear or nuclear atypia does not support a recommendation of surgical treatment. A malignant cytologic diagnosis has a high positive predictive value for detection of thyroid cancer.

Fine-needle aspiration biopsy of the thyroid: comparison between thyroid palpation and ultrasonography.

OBJECTIVE: To describe our experience with fine-needle aspiration biopsy (FNAB) of the thyroid and compare our results with direct palpation versus ultrasound scanning (USS) in an area of endemic goiter in Italy. METHODS: We considered all patients submitted to ultrasound-guided FNAB of thyroid nodules during a 10-month period at our outpatient clinic and analyzed the following: (1) clinical data (number of nodules and identification of the nodule for FNAB); (2) USS data (number of nodules and identification of the nodule for FNAB on the basis of hypoechoic pattern + blurred perinodal halo + microcalcifications or intranodal color Doppler signal indicative of blood flow); (3) cytologic specimens, categorized as suspicious, malignant, negative, or nondiagnostic; and (4) histologic final report of the cytologically positive nodules. RESULTS: The study group consisted of 348 female and 72 male patients who underwent FNAB of the thyroid at our institution. Among the 140 patients with no palpable thyroid nodules, USS showed that 106 had a single nodule and 34 had multinodular goiters. Among the 182 patients with a single palpable thyroid nodule, USS revealed that 138 had a single nodule, 42 had a multinodular goiter, and 2 had lobe enlargement without detectable nodules. All 98 patients with multinodular palpable goiter had a similar pattern on USS. Of the 420 cytologic specimens, 46(11.0%) were positive for thyroid cancer, 313 (74.5%)were negative, and 61 (14.5%) were nondiagnostic. Histologic malignant growth was confirmed in 27 cytologically positive nodules. Of these histologically malignant nodules, 12 (45%) were nonpalpable, 9 (33%) were single palpable nodules, and 6 (22%) were from a nodule with a suspicious ultrasound pattern within a multinodular goiter. CONCLUSION: Manually guided FNAB is not feasible in nonpalpable nodules and not accurate in a multinodular goiter. Both situations are clinical challenges, and USS should be performed for accurate FNAB under these circumstances. Because 52% of histologically malignant nodules in our study were found only with the aid of ultrasound-guided FNAB, this procedure should be used where multinodular goiter is endemic. Our overall rate of nondiagnostic specimens was comparable to that reported in the literature.