Repeat Thyroid Nodule Fine-Needle Aspiration in Patients With Initial Benign Cytologic Results

Anatomic Pathology / REPEAT THYROID FINE-NEEDLE ASPIRATION Repeat Thyroid Nodule Fine-Needle Aspiration in Patients With Initial Benign Cytologic Results Melina B. Flanagan, MD, MSPH, 1 N. Paul Ohori, MD, 1 Sally E. Carty, MD, 2 and Jennifer L. Hunt, MD 1 Key Words: Thyroid; Fine-needle aspiration; FNA; Repeat testing Abstract Fine-needle aspiration (FNA) is the standard of care for the initial workup of thyroid nodules, but there is no consensus algorithm to manage patients with benign results. We examined performance characteristics of initial and repeat satisfactory FNAs for all 402 patients who underwent thyroid surgery during a recent 22-month period. Of these patients, 267 had at least 1 satisfactory FNA and 70 had 2 or more. After an initial benign FNA, 1 repeat FNA correctly identified an unsuspected malignancy in 2 of 70 patients and was indeterminate in 17; of these, 7 of 17 were identified as malignant in the final pathologic diagnosis. Overall, the use of 1 repeat FNA increased the sensitivity for malignancy from 81.7% to 90.4% and decreased the false-negative rate from 17.1% to 11.4%. With more than 1 repeat FNA, there was no improvement in performance characteristics. These data make a strong argument for 1 repeat FNA following an initial benign FNA diagnosis. A clinically palpable thyroid nodule is present in 4% to 7% of the adult population in North America 1 ; however, it has been estimated that only 5% to 10% of these are malignant. 1,2 Fineneedle aspiration (FNA) has become the standard of care for the initial diagnostic workup of a thyroid nodule. Although diagnostic categories for reporting thyroid cytology results are not entirely standardized, most cytopathologists use 4 categories: malignant, indeterminate, benign, and unsatisfactory. 3 At our institution, we subclassify the indeterminate category with descriptions that include follicular lesion, Hürthle cell lesion, atypical cells present, and suspicious for malignant cells. Most clinicians use the results of a thyroid FNA in conjunction with clinical findings to guide patient management. Patients with a malignant or an indeterminate FNA diagnosis will be counseled to undergo surgery, whereas benign diagnoses in stable nodules do not usually warrant surgical intervention. 2 Use of FNA to manage thyroid nodules has been associated with a considerable decrease in the need for diagnostic thyroidectomy and an increase in the percentage of thyroid cancer found in surgical specimens. 1 As with any screening test, the usefulness of thyroid FNA depends on its accuracy. Estimates of FNA sensitivity range from 65% to 98% and specificity from 72% to 100%. The false-positive rate reportedly ranges from 0% to 7.7% and is considered low. The false-negative rate is more concerning, however, because it has been estimated to range from 1.3% to 11.5% 1 and has been shown to be associated with a significant delay in appropriate treatment. 4 Although there is general agreement on the management of patients with a malignant or indeterminate FNA diagnosis, there is no consensus on clinical management of patients with benign FNA results. Several studies have examined the usefulness of repeat thyroid FNA after 698 Am J Clin Pathol 2006;125:698-702 Downloaded 698 from https://academic.oup.com/ajcp/article-abstract/125/5/698/1759839

Anatomic Pathology / ORIGINAL ARTICLE an initial benign result. Although some authors suggest that repeat FNA reduces the false-negative rate and is, therefore, a valuable tool, 5-9 others find that the routine repeat aspiration of cytologically benign nodules does not improve the detection of malignancy and is not warranted. 10-12 In most studies, however, the patients generally had only 1 repeat FNA, study sizes were limited, and only small subsets were correlated with surgical specimens. The present study examined the role of 1 or more repeat thyroid FNAs in a large cohort of patients undergoing thyroidectomy with an initial benign FNA diagnosis. Materials and Methods The pathology records of the University of Pittsburgh Medical Center, Pittsburgh, PA, were reviewed to identify all patients who underwent thyroid surgery for nodular thyroid disease for the 22-month period from January 2001 through October 2002. For each patient, diagnoses from the final surgical pathology report and all preceding cytopathology reports were retrieved. The final histologic permanent section diagnoses were categorized as benign (including nodular goiter, dominant hyperplastic nodule, chronic lymphocytic thyroiditis, follicular or Hürthle cell adenoma, and Graves disease) or malignant (including follicular, papillary, Hürthle cell, and medullary carcinomas). Similarly, the preoperative cytologic diagnoses were categorized as benign, indeterminate, malignant (papillary and medullary carcinomas), or unsatisfactory. The indeterminate category included diagnoses of follicular lesion, Hürthle cell lesion, atypical, and suspicious for malignant cells (listed in increasing probability of malignant outcome). At our institution, FNA specimens that do not contain sufficient diagnostic material (usually owing to a paucity of follicular epithelial groups or to blood contamination) are given the qualifier unsatisfactory. Because unsatisfactory FNAs will be repeated, these samples were excluded from further statistical analysis in the present study. Pathologically identified incidental microscopic papillary thyroid carcinomas were not included, and for the purposes of this study, only index lesions were examined for FNA predictive ability. All FNAs were performed by University of Pittsburgh physicians (including endocrine surgeons, ultrasound radiologists, and thyroid-specialized endocrinologists), all with substantial experience in the technique. Each specimen was divided between air-dried smears stained with rapid Romanowsky stain and alcohol-fixed smears for Papanicolaou staining. All thyroid FNA specimens were interpreted by experienced staff cytopathologists. All surgical specimens were processed according to standard procedures, with extensive sampling of gross nodules (ie, fully submitted or at least 20 cassettes). Histologic sections were interpreted by experienced staff surgical pathologists. Performance characteristics (sensitivity, specificity, and false-positive and false-negative rates) of FNA for the final histologic diagnoses were calculated. Because patients with an indeterminate or malignant result usually are counseled to undergo surgery, patients with these FNA diagnoses were grouped as positive for the purposes of this analysis. All performance characteristics (sensitivity, specificity, and false-negative and false-positive rates) were calculated for the initial FNA and then for each repeat aspirate before surgery. For calculations of characteristics on repeat aspirates, the most worrisome diagnosis was used. The false-negative rate was defined as the number of patients with benign FNA diagnoses who at surgery were found to have malignancy divided by the total number of patients with benign FNA diagnoses. Again, unsatisfactory samples were not included in the calculations. Statistics were performed using SPSS version 11.5 (SPSS, Chicago, IL). This study was approved by the University of Pittsburgh Institutional Review Board (IRB number 0307076). Results During the 22-month period from January 2001 through October 2002, 402 patients underwent thyroid surgery at our institution for nodular thyroid disease. Patients had 0 to 5 FNAs preceding surgery, for a total of 394 FNAs Figure 1. 0 FNAs (135) FNA #1 (267) FNA #2 (70) FNA #3 (19) FNA #4 (9) FNA #5 (2) (135) (197) (51) (10) (7) (2) Surgery (402 patients, 367 satisfactory FNAs) Figure 1 From January 2001 through October 2002, 402 patients underwent thyroid surgery at our institution for nodular thyroid disease; 0-5 FNAs were performed preceding surgery for a total of 367 FNAs. Numbers indicate how many patients had a certain number of aspirates and then proceeded to surgery or a repeat aspirate. FNA, fine-needle aspiration. Downloaded from https://academic.oup.com/ajcp/article-abstract/125/5/698/1759839 Am J Clin Pathol 2006;125:698-702 699 699 699

Flanagan et al / REPEAT THYROID FINE-NEEDLE ASPIRATION The indication for repeat FNA was almost always given as the persistence of a dominant lesion. Of the 394 aspirates, 367 samples (93.2%) were satisfactory and included in this analysis. Of the 402 cases, 267 patients had at least 1 satisfactory FNA, 70 patients had 2 or more, 19 had 3 or more, 9 had 4 or more, and 2 had 5 FNAs. In this study 197 (49.0%) had exactly 1 FNA, 51 (12.7%) had 2, 10 (2.5%) had 3, 7 (1.7%) had 4, and 2 (0.5%) had 5 FNAs. In Table 1, results are displayed by cytologic diagnostic categories for all preoperative satisfactory aspirates and for the final surgical pathology results. Of 267 patients with 1 or more FNAs, the first FNA was benign in 111 (41.6%), malignant in 34 (12.7%), and indeterminate in 122 (45.7%). Of the 70 patients who had a second FNA, in 44 (62.9%), the repeat FNA results were concordant with the first FNA result. To assess the importance of a second FNA after a benign diagnosis, the analysis was limited to the cases in which the initial FNA diagnosis was benign (ie, excluding cases with a malignant or indeterminate diagnosis). There were 111 patients with an initial benign diagnosis, and 57 of these had 1 or more additional aspirates. In this subset, the second FNA reaffirmed the original benign diagnosis in 34 cases (59.6%). The second FNA diagnosis was upgraded to indeterminate in 21 cases (36.8%), 7 of which were determined to be malignant by final histologic examination. Two cases (3.5%) were upgraded to malignant, and both were malignant on final histologic examination. Overall, 1 repeat FNA of thyroid nodules diagnosed as benign on initial FNA detected 9 previously undetected malignant neoplasms and, thus, changed the clinical plan in 9 (15.8%) of 57 patients. Nineteen patients with benign diagnoses on the first 2 FNAs had up to 3 additional FNAs before surgery. Five of these were upgraded to an indeterminate diagnosis, but all of these were determined to be benign by final histologic examination. Two patients were given a malignant diagnosis by final histologic examination, despite having had 3 or 4 benign FNA results. Performance characteristics for 1, 2, 3, 4, and 5 FNAs are shown in Table 2 (this analysis includes all satisfactory repeat FNAs). With the second FNA, the overall sensitivity increased from 81.7% to 90.4%, and the false-negative rate decreased from 17.1% to 11.4%. In the subset of patients with Table 1 Diagnoses for Each Fine-Needle Aspirate and for Final Surgical Specimen Histology Fine-Needle Aspirate No. Final Surgical Specimen 1 2 3 4 5 Benign Malignant Total N N N N N 110 25 135 B N N N N 48 6 54 B B N N N 13 2 15 B B B N N 7 1 8 B B B B N 4 1 5 B B B B B 1 0 1 B B B I N 2 0 2 B B B I I 1 0 1 B B I N N 2 0 2 B I N N N 14 7 21 B M N N N 0 2 2 I N N N N 63 47 110 I B N N N 1 0 1 I I N N N 7 2 9 I M N N N 0 2 2 M N N N N 0 33 33 M M N N N 0 1 1 B, benign; I, indeterminate; M, malignant; N, none. Table 2 Performance Characteristics for 1, 2, 3, 4, and 5 Fine-Needle Aspirations * Fine-Needle Aspirate 1 1 and 2 1-3 1-4 1-5 Sensitivity (%) 81.7 90.4 90.4 90.4 90.4 Specificity (%) 56.4 47.9 47.2 44.8 44.8 False-negative rate (%) 17.1 11.4 11.5 12.0 12.0 False-positive rate (%) 45.5 47.5 47.8 48.9 48.9 * The analysis includes all satisfactory repeated fine-needle aspirations. 700 Am J Clin Pathol 2006;125:698-702 Downloaded 700 from https://academic.oup.com/ajcp/article-abstract/125/5/698/1759839

Anatomic Pathology / ORIGINAL ARTICLE initially benign cytologic results, 1 repeated FNA resulted in an overall false-negative rate of 11.4%. Additional repeat FNAs did not improve sensitivity. Additional repeat FNAs also increased the overall false-negative rate. Discussion Whether FNA of thyroid nodules should be interpreted as a screening tool or a diagnostic procedure is not clear and depends on the diagnosis given. However, the ultimate goal of thyroid FNA should be to differentiate between nodules that are definitely nonneoplastic and those that definitely are malignant or that must be examined histologically for diagnosis. The ability of FNA to identify patients who will benefit from surgery is entirely dependent on the accuracy of the test. Thyroid FNA has been estimated to have a sensitivity for malignancy between 65% and 98% (mean, 83%), a specificity between 72% and 100% (mean, 92%), a false-positive rate of 0% to 7.7% (mean, 2.9%), and a false-negative rate of 1.3% to 11.5% (mean, 5.2%). 1 Estimates of these performance characteristics are highly dependent on excellent follow-up data, which can be difficult to acquire, and on the gold standard used for comparison. In reported studies, the gold standard compared with FNA is variable and has included repeat FNA, clinical impression, and histologic diagnoses from thyroidectomies. There are disadvantages to each of these outcome measures. Repeat FNA suffers from the same problems of sampling and interpretation as first FNA and, therefore, is not a reliable outcome measure. Clinical data are notoriously difficult to obtain and must include very long follow-up time because thyroid malignant neoplasms can be indolent. Histologic data generally are seen as the true gold standard for comparison but, because most patients with benign cytologic findings do not undergo surgery, such information is not readily available. The ideal study would include initial and repeat FNA data and histologic analysis of all thyroid nodules in all patients; however, this study is not possible as the standard of care has been established. It is well established that FNA has revolutionized the care of patients with thyroid nodules. The need for a diagnostic lobectomy has decreased greatly because many patients can be triaged based on their FNA results. However, the false-negative rate for initial thyroid FNA has important clinical implications. One false-negative result delays appropriate surgical treatment by more than 2 years, and additional false-negative diagnoses confer an even longer delay. This delayed treatment is associated with capsular and vascular invasion, both of which predict a poorer outcome. 4 Thus, the clinical significance of a false-negative FNA result certainly fuels the issue of how to manage patients with a thyroid nodule diagnosed as benign on the first FNA. Several studies support the view that 1 repeat FNA in cytologically benign nodules is valuable. In studies of 246, 196, and 235 patients with 1 repeat FNA and histologic followup on a limited subset, Hamburger, 5 Dwarakanathan et al, 6 and Chehade et al 7 reported that repeat FNA upgraded 18, 13, and 12 benign diagnoses, respectively, to suspicious or malignant, and that 6 of 10, 4 of 10, and 2 of 9, respectively, of these who underwent thyroidectomy had a malignancy found by histologic evaluation. These authors concluded that 1 repeat FNA reduces the false-negative rate and, therefore, is useful. Two studies examined multiple repeat FNAs and reported that either 2 or 3 FNAs are useful. Erdogan et al 8 reviewed multiple repeat aspiration specimens from 216 euthyroid patients with cytologically benign nodules (257 second FNAs, 137 third, 46 fourth, and 17 fifth). The second FNA upgraded 3 cases to malignant, and all were malignant by histologic examination; with 4 FNAs, 16 cases were upgraded to suspicious, but all 6 of these that underwent thyroidectomy were benign. The authors concluded that a second FNA in patients with initially cytologically benign thyroid nodules may detect a few previously missed malignant neoplasms but that additional repeat FNAs are not indicated routinely. More recently, Orlandi et al 9 examined up to 5 repeat FNA specimens in 306 patients and reported that repeat FNA upgraded 4 cases from benign to malignant (1 in the second FNA and 3 in the third). The authors concluded that it is useful to perform 3 FNAs, but that in the absence of clinical changes, additional cytologic follow-up after 3 FNAs with benign diagnoses is unnecessary. Conversely, several studies support the opposing view that repeat FNAs are not useful. Lucas et al 10 reported that in repeat FNAs on 116 euthyroid patients (2 in 116 patients, 3 in 19), no cases were upgraded from benign to malignant. In an analysis of 490 patients with cytologically benign nodules, Aguilar et al 11 reported that 2 of 90 patients who underwent thyroidectomy for clinical reasons were found to have a malignancy, 1 of 184 benign diagnoses was upgraded to malignant by repeat FNA and then confirmed histologically, and 216 patients remained free of cancer as assessed by clinical follow-up. Finally, in an analysis of repeat FNA specimens in 45 patients with initially benign diagnoses (41 with 1 repeat FNA, 4 with 2), Merchant et al 12 reported that all cases remained categorized as benign and 7 of 7 cases that underwent thyroidectomy for clinical reasons were histologically benign. These authors concluded that repeat FNA does not improve detection of malignant neoplasms. The present study examined the role of 1 or more satisfactory thyroid FNA specimens in 267 patients, including repeat FNA in 70 patients. In 57 patients with an initial benign diagnosis, 1 repeat FNA diagnosed 2 additional malignant neoplasms (both were malignant by final histologic examination) and 21 indeterminate cases (7 of which were malignant by final histologic examination). Thus, 2 repeat FNAs of thyroid nodules Downloaded from https://academic.oup.com/ajcp/article-abstract/125/5/698/1759839 Am J Clin Pathol 2006;125:698-702 701 701 701

Flanagan et al / REPEAT THYROID FINE-NEEDLE ASPIRATION diagnosed as benign on initial FNA detected 9 previously undetected malignant neoplasms. Overall, a second FNA decreased the false-negative rate from 17.1% to 11.4% and increased sensitivity from 81.7% to 90.4%. Unlike previous studies that have tried to address this issue, our study is based on histologic analysis of all cases. The strength of this design is that it allowed us to correlate cytologic with histologic findings for the entire sample. However, the surgical subset of patients that we studied is likely to have a greater percentage of malignant outcomes than the general population. Thus, our performance characteristics for 1 FNA and multiple FNAs should be interpreted as applicable to the subset of patients offered surgery by their clinicians, and the benefit of a second FNA in the general population might be somewhat lower. The present study further examined the usefulness of more than 1 repeat thyroid FNA. Despite extensive sampling, no further malignant neoplasms were detected after the first repeat FNA in up to 3 additional FNA specimens. Overall, the false-negative rate increased from 11.4% for 2 aspirates to 11.5% for 3 and 12.0% for 4 or 5. It is interesting that there were 2 cases in which thyroid nodules diagnosed as benign on 3 consecutive FNAs were diagnosed histologically as malignant: 1 papillary carcinoma and 1 follicular carcinoma. These cases highlight that even repeat FNAs will have false-negative results, and the ultimate decision to proceed to surgery must include clinical factors as well as FNA results. Our data strongly support the clinical usefulness of 1 repeat FNA after an initial benign aspiration. The data also highlight the fact that management decisions for patients with thyroid nodules must include the FNA result and clinical suspicion, because false-negative results are not eliminated with repeat FNA of thyroid nodules. From the Departments of 1 Pathology and 2 Surgery, University of Pittsburgh, Pittsburgh, PA. Address reprint requests to Dr Hunt: Dept of Anatomic Pathology, the Cleveland Clinic, 9500 Euclid Ave, L25, Cleveland, OH 44195. References 1. Gharib H, Goellner JR. Fine-needle aspiration biopsy of the thyroid: an appraisal. Ann Intern Med. 1993;118:282-289. 2. Lawrence W Jr, Kaplan BJ. Diagnosis and management of patients with thyroid nodules. J Surg Oncol. 2002;80:157-170. 3. The Papanicolaou Society of Cytopathology Task Force on Standards of Practice. Guidelines of the Papanicolaou Society of Cytopathology for fine-needle aspiration procedure and reporting. Diagn Cytopathol. 1997;17:239-247. 4. Yeh MW, Demircan O, Ituarte P, et al. False-negative fineneedle aspiration cytology results delay treatment and adversely affect outcome in patients with thyroid carcinoma. Thyroid. 2004;14:207-215. 5. Hamburger JI. Consistency of sequential needle biopsy findings for thyroid nodules: management implications. Arch Intern Med. 1987;147:97-99. 6. Dwarakanathan AA, Staren ED, D Amore MJ, et al. Importance of repeat fine-needle biopsy in the management of thyroid nodules. Am J Surg. 1993;166:350-352. 7. Chehade JM, Silverberg AB, Kim J, et al. Role of repeated fine-needle aspiration of thyroid nodules with benign cytologic features. Endocr Pract. 2001;7:237-243. 8. Erdogan MF, Kamel N, Aras D, et al. Value of re-aspiration in benign nodular thyroid disease. Thyroid. 1998;8:1087-1090. 9. Orlandi A, Puscar A, Capriata E, et al. Repeated fine-needle aspiration of the thyroid in benign nodular thyroid disease: critical evaluation of long-term follow-up. Thyroid. 2005;15:274-278. 10. Lucas A, Llatjos M, Salinas I, et al. Fine-needle aspiration cytology of benign nodular thyroid disease: value of reaspiration. Eur J Endocrinol. 1995;132:677-680. 11. Aguilar J, Rodriguez JM, Flores B, et al. Value of repeated fineneedle aspiration cytology and cytologic experience in the management of thyroid nodules. Otolaryngol Head Neck Surg. 1998;119:121-124. 12. Merchant SH, Izquierdo R, Khurana KK. Is repeated fineneedle aspiration cytology useful in the management of patients with benign nodular thyroid disease? Thyroid. 2000;10:489-492. 702 Am J Clin Pathol 2006;125:698-702 Downloaded 702 from https://academic.oup.com/ajcp/article-abstract/125/5/698/1759839