[435] Timing of Repeat Thyroid FNA in the Management of Thyroid Nodules.

Remmi S Singh, Helen H Wang. BIDMC & Harvard Medical School, Boston, MA

Background: The Bethesda System for reporting thyroid FNA recommends repeat for an initial diagnosis of non-diagnostic or atypical cells/follicular lesion of undetermined significance. The interval recommended between the initial and repeat FNA for these patients is at least three months, however, there is limited data to support this recommendation. We investigated our own data to determine the yield of repeat FNA in relation to time interval between procedures and the original diagnosis.
Design: We retrospectively reviewed all reports on thyroid FNAs from 2006 to 2008 and identified those patients who had more than one FNA from the same lesion. Then the surgical pathology file was searched to determinine if any one had a thyroidectomy at our institution. All FNA diagnoses, time interval between procedures and the diagnoses of thyroidectomy were recorded.
Results:

Time Interval	Diagnosis of Repeat FNA (% of row total)				Total (% of column total)
	Non-Diagnostic	Sub-optimal Specimen	Atypical Diagnosis	Adequate Specimen
< 2 weeks	4(14)	7(24)	2(6.9)	16(55)	29(9.4)
2 weeks-1 month	15(26)	15(26)	0	28(48)	58(19)
1-2 months	9(14)	18(29)	1(1.6)	35(56)	63(21)
2-3 months	5(12)	15(35)	1(2.3)	22(51)	43(14)
3-6 months	7(18)	9(24)	1(2.6)	21(55)	38(12)
6-12 months	4(11)	6(17)	4(11)	21(60)	35(11)
> 1 year	8(20)	11(27)	2(4.9)	20(49)	41(13)
Total (% of row total)	52(17)	81(26)	11(3.6)	163(53)	307

P = 0.43 by chi-square

307 patients met the criteria of the study and 81 had thyroidectomy. Overall 53% of the repeat FNAs yielded an adequate specimen for a diagnosis. This percentage did not vary significantly according to the time interval between procedures (see Table 1). However, those patients who had a non-diagnostic or suboptimal initial FNA were more likely to have a second non-diagnostic or suboptimal specimen than those whose first FNA was adequate (47% vs 29%, P = 0.03). Although limited by the small number, there was no evidence to suggest that a short interval was more likely to yield a false positive diagnosis on the repeat when compared to histology. Fifty-three percent (17/32) of patients who had at least one “atypical” diagnosis had surgery and 47% had malignancy. This percentage did not change significantly with the diagnosis of the patients' other FNA(s).
Conclusions: The diagnostic yield or accuracy of the repeat FNA is not related to the time interval between procedures but is related to the original diagnosis. Patients who had at least one “atypical” diagnosis and underwent surgery had a ∼50% risk of finding malignancy regardless of the diagnosis of the patient's other FNA(s).
Category: Cytopathology

Tuesday, March 1, 2011 11:00 AM

Platform Session: Section F, Tuesday Morning