Is a five-category reporting scheme for thyroid fine needle aspiration cytology accurate? Experience of over 18 000 FNAs reported at the same institution during 1998–2007

Authors


Simonetta Piana, Department of Pathology, Arcispedale Santa Maria Nuova, Viale Risorgimento 80, 42100 Reggio Emilia, Italy
Tel.: +39 0522 295919; Fax: +39 0522 296945;
E-mail: Simonetta.piana@asmn.re.it

Abstract

S. Piana, A. Frasoldati, M. Ferrari, R. Valcavi, E. Froio, V. Barbieri, C. Pedroni and G. Gardini

Is a five-category reporting scheme for thyroid fine needle aspiration cytology accurate? Experience of over 18 000 FNAs reported at the same institution during 1998–2007

Objective:  Fine needle aspiration (FNA) has long been recognized as an essential technique for the evaluation of thyroid nodules. Although specific cytological patterns have been recognized, a wide variety of reporting schemes for thyroid FNA results have been adopted. This study reports our experience with a five-category reporting scheme developed in-house based on a numeric score and applied to a large series of consecutive thyroid FNAs. It focuses mainly on the accuracy of thyroid FNA as a preoperative test in a large subset of histologically distinct thyroid lesions.

Methods:  During the 1998–2007 period, 18 359 thyroid ultrasound-guided FNAs were performed on 15 269 patients; FNA reports were classified according to a C1–C5 reporting scheme: non-diagnostic (C1), benign (C2), indeterminate (C3), suspicious (C4), and malignant (C5).

Results:  Non-diagnostic (C1) and indeterminate (C3) FNA results totalled 2 230 (12.1%) and 1 461 (7.9%), respectively, while suspicious (C4) and malignant (C5) results totalled 238 (1.3%) and 531 (2.9%), respectively. Histological results were available in 2 047 patients, with thyroid malignancy detected in 840. Positive predictive value of FNA was 98.1% with a 49.0 likelihood ratio (LR) of malignancy in patients with a C4/C5 FNA report.

Conclusions:  This five-category scheme for thyroid FNA is accurate in discriminating between the virtual certainty of malignancy associated with C5, a high rate (92%) of malignancy associated with C4, and a 98% probability of a histological benign diagnosis associated with C2. Further sub-classifications of C3 may improve the accuracy of the diagnostic scheme and may help in recognizing patients eligible for a ‘wait and see’ management.

Ancillary