Frozen Section in a Cytological Diagnosis of Thyroid Follicular Neoplasm

The LaryngoscopeLippincott Williams & Wilkins, Inc., Philadelphia© 2003 The American Laryngological,Rhinological and Otological Society, Inc.

Frozen Section in a Cytological Diagnosis ofThyroid Follicular Neoplasm

Nuria Alonso, MD; Anna Lucas, PhD; Isabel Salinas, PhD; Eva Castella, PhD; Anna Sanmartı, PhD

Objective/Hypothesis: Fine-needle aspiration bi-opsy is the most accurate diagnostic test for thyroidnodules, its only limitation being the diagnosis of fol-licular neoplasm that does not distinguish betweenbenign and malignant follicular lesions. Study Design:To determine the utility of intraoperative frozen-section analysis in cases of a cytological diagnosis offollicular neoplasm, a retrospective review of 66 pa-tients with a solitary thyroid nodule and follicularneoplasm who underwent thyroid surgery was car-ried out. Methods: Fine-needle aspiration was classi-fied following the Papanicolaou Society of Cytopa-thology Classification, and frozen section was definedas malignant or “deferred.” If a malignant diagnosiswas made by frozen-section analysis, a total thyroid-ectomy was carried out. The extension of thyroidsurgery in the deferred cases was based on thedefinitive histological diagnosis. Results: Sixty-four cases were classified as deferred, and two assuspect for malignancy. Among the 64 deferredcases, 15 were malignant in the final pathologicalfindings, and 49 were benign. The two suspectcases were papillary carcinoma. Frozen-sectionanalysis classified 2 of 17 (11.7%) cases as follicu-lar variant of papillary carcinoma that could notbe diagnosed by cytological study. However, thesetwo cases had a strong clinical evidence of malig-nancy. Conclusion: The routine use of frozen-section analysis is useless in cases of cytologicaldiagnosis of follicular neoplasm on fine-needle as-piration biopsy, because of the low probability ofachieving the diagnosis of follicular carcinomaand the inability to provide additional informa-tion apart from the clinical and the cytologicaldata. Key Words: Frozen-section analysis, follicu-lar neoplasm, fine-needle aspiration biopsy, thy-roid nodule, follicular carcinoma.

Laryngoscope, 113:563–566, 2003

INTRODUCTIONNodular thyroid disease is a common clinical prob-

lem. Most thyroid nodules are benign, and nonsurgicalapproaches are preferred.1,2 Fine-needle aspiration (FNA)biopsy has emerged as the most accurate diagnostic testfor differentiating benign from malignant thyroid nod-ules.3–5 The reported sensitivity and specificity rates forthyroid FNA range from 65% to 99% and from 72% to100%, respectively,2 with an overall accuracy rangingfrom 85% to 100% and a false-negative rate of 1% to 11%.6

The only limitation of FNA is the cytological diagnosis offollicular neoplasm (seen in approximately 15% of all as-pirations) that does not distinguish between benign follic-ular lesions (i.e., hyperplastic nodules and adenomas) andfollicular cancer.7

Intraoperative frozen-section (FS) analysis of thyroidnodules is a common procedure and is often used to deter-mine definitive surgical management. Considering theglobal FNA biopsy results, routine use of FS procedurescan be justified only if they provide more accurate infor-mation or alter the choice of surgical approach.8 Frozen-section analysis can reliably diagnose papillary (exceptfollicular variant), medullar, and anaplastic thyroid car-cinomas, but these lesions can also be well diagnosed byFNA biopsy. The clinical usefulness of frozen-section anal-ysis in follicular thyroid lesions is debatable,9 and theopinions range from continued routine practice to ex-tremely limited use.10–12

The objective of the present study was to determinethe usefulness of intraoperative FS analysis to decide theextension of thyroid surgery in the presence of a cytolog-ical diagnosis of follicular neoplasm on FNA biopsy.

PATIENTS AND METHODSWe retrospectively reviewed all patients with a solitary

thyroid nodule and a cytological diagnosis of follicular neoplasmon FNA biopsy between 1992 and 2000 who underwent thyroidsurgery. According to the recommendations of Lowhagen,13 theminimum criterion for the FNA biopsy diagnosis of follicularneoplasm that were followed were cellular smear, relative paucityof colloid, and an abundance of follicles rather than sheets ofepithelium.

In all, 66 patients who satisfied the inclusion criteria ofhaving both an adequate FNA biopsy and FS analysis were stud-ied (52 women and 14 men with a mean age of 37 y [SD � 15 y]).Among all the patients, we reviewed the following clinical fea-

From the Endocrinology (N.A., A.L., I.S., A.S.) and Pathology Services(E.C.), Hospital Universitari “Germans Trias i Pujol,” Badalona, Catalonia,Spain.

Editor’s Note: This Manuscript was accepted for publication Septem-ber 26, 2002.

Send Correspondence to Anna Lucas, PhD, Endocrinology Service,Hospital Universitari “Germans Trias i Pujol,” Carretera Canyet s/n,08916, Badalona, Catalonia, Spain. E-mail: [email protected]

Laryngoscope 113: March 2003 Alonso et al.: Thyroid Follicular Neoplasm

563

tures: history of thyroid irradiation, presence of a solitary ordominant nodule on palpation, nodule size, a family history ofthyroid cancer, gender and other clinical parameters often asso-ciated with malignancy such as a firm nodule, vocal cord paraly-sis, enlarged regional lymph nodes, distant metastasis, fixation tosurrounding structures, and rapid growth of the nodule.1,14 Allpatients underwent FNA biopsy following the standard proce-dure,15 and four to six aspirations were usually performed in eachcase to reach a diagnosis.

Fine-needle aspiration diagnoses were performed by expe-rienced cytologists, and frozen-section analyses were evaluatedby expert pathologists who knew the previous cytological diagno-sis of follicular neoplasm. The fine-needle aspirations were clas-sified following the Papanicolaou Society of Cytopathology,16

whose recommended terms are as follows: unsatisfactory for in-terpretation, colloid nodule, cyst, thyroiditis, Hurthle cell neo-plasm, malignant (specify type), follicular neoplasm, and other.Frozen sections were classified as malignant or “deferred.” If theFS was deferred, the surgical intervention was terminated byperforming a hemithyroidectomy, but if frozen-section analysispresented a malignant diagnosis, the remaining lobe was re-moved, accomplishing a total thyroidectomy. The extension ofthyroid surgery in the deferred cases was based on the definitivehistological diagnosis, extending the thyroidectomy in the case ofmalignancy.

We examined the correlation between FS analysis and finalpathological findings in the cases of follicular neoplasm on FNAbiopsy and if the use of FS analysis contributed to provide addi-tional information apart from the clinical and cytological data todetermine the extent of surgery. Ethical committee approval wasobtained. For univariate analysis, Fisher’s Exact test was usedfor nominal variables, and an unpaired t test was employed forcontinuos variables. A P value less than .05 was considered sta-tistically significant.

RESULTSResults of FS analysis for the 66 specimens and the

correlation with the final pathological findings are shownin Table I. Sixty-four cases were classified as deferred, andtwo as suspect for malignancy (follicular variant of papil-lary carcinoma). In the final pathological findings, thesetwo cases were confirmed to be papillary carcinoma. Therewere no false-positive results in the FS analysis. Amongthe 64 deferred cases on FS analysis, 15 were found to bemalignant in the final pathological results, and 49 werefound to be benign. Overall, 15 patients had a malignancymissed by FS analysis; therefore, in all, 23.44% of malig-nancies were not diagnosed by FS analysis, which is con-sistent with a previous report.17

When we evaluated clinical parameters, a nodule sizegreater than 4 cm on physical examination was observed

in 7 of 15 patients (41.18%) with malignant lesions in thefinal pathological findings, and the rapid growth of thenodule was found in another 7 patients (41.18%). Rapidgrowth of the nodule was found to be associated with asignificantly increased risk of malignancy (P � .023). Theother clinical parameters examined were not significantlydifferent between patients with malignant or benign his-tological findings. In the two cases of follicular variant ofpapillary carcinoma diagnosed by FS analysis, the nodulespresented a rapid growth; both of them were larger than 4cm on physical examination and in one case the patienthad a history of thyroid irradiation.

Frozen-section analysis classified 2 of 17 (11.7%) fol-licular carcinomas, both of which were a follicular variantof papillary carcinoma that could not be diagnosed bycytological study. Thus, the sensitivity of FS analysis forthe diagnosis of malignancy in this group of tumors was11.7%. Approximately half of the malignancies (52.94%)were a follicular variant of papillary carcinoma. This find-ing is similar to that described in two previousreports.18,19

DISCUSSIONThyroid nodules are common, occurring in approxi-

mately 4% to 7% of the population with an annual inci-dence of 0.1%.20 The challenge to physicians in the eval-uation of thyroid nodules is to identify the few patientswith thyroid malignancy among the large majority whohave benign nodular disease. In recent years, FNA biopsyhas become the most accurate and cost-effective way todetermine the need for surgical resection versus observa-tion in patients with a thyroid nodule.2,20–22 One of theproblems associated with FNA biopsy is the “indetermi-nate” or “suspect” result (most cases corresponding tofollicular neoplasm). The cytological distinction of benignfrom malignant follicular neoplasia remains problematicbecause the criteria for malignancy, such as capsularand/or vascular invasion, cannot be determined.9 Somestudies have tried to identify clinical parameters thatcharacterize patients with an increased risk of having athyroid follicular cancer, whose preoperative FNA findingwas of follicular neoplasm.8,23–25 However, follicular neo-plasms are relatively uncommon, and the sample size ofindividual studies was often too small to reach statisticalsignificance. Furthermore, the varying results amongstudies have made it difficult for practitioners to apply thefindings to their clinical practice.26

Some studies have been published in relation to thediagnostic accuracy of FS analysis.27–30 Sabel et al.30

studied 44 of 95 patients with “equivocal findings” on FNAbiopsy (categorized on FNA as atypical features, follicularneoplasm, Hurthle cell neoplasm, or nondiagnostic). Six-teen of these 44 cases were diagnosed as follicular neo-plasm, and only two of these cases were malignant on thefinal pathological findings (one classified as benign andthe other as indeterminate on FS analysis). The authorssuggested that the extension of the operation should bebased on the patient’s history, the characteristics of thenodule, and the results of FNA biopsy. Other authorssupported these data and thought that the accuracy ofFNA biopsy is high enough to avoid intraoperative FS

TABLE I.Final Diagnosis in the 66 Cases Studied in Frozen Section (64

“Deferred” Cases and 2 Classified as Follicular Variant ofPapillary Carcinoma).

FrozenSectionCategory

Final Pathology

FollicularCarcinoma

Follicular Variant ofPapillary

CarcinomaFollicularAdenoma

NodularHyperplasia

Malignant 0 2 0 0

“Deferred” 8 7 19 30


564

analysis, and that a preoperative FNA, together withsound clinical judgment, should be able to replace FSanalysis.17,26,31 Malcahy et al.32 demonstrated the useful-ness of FS analysis, especially when the experience andaccuracy of FNA biopsy is in doubt. They reported that theaccuracy of FNA biopsy in the diagnosis of well-differentiated thyroid malignancy was 88% with the ap-plication of strict criteria by an experienced cytologist, andthat the accuracy of FS analysis was 92% and did notchange significantly (94%) with repeat examination. How-ever, according to those authors, when FNA biopsy yieldsa diagnosis of follicular neoplasm, FS analysis is unlikelyto change this diagnosis, which is consistent with ouropinion. Intraoperative FS evaluation for follicular lesionsis unsatisfactory because intraoperative diagnosis of fol-licular carcinoma depends on the chance of capturing cap-sular and/or vascular invasion on a single section.9,17,29,32

The meta-analysis of 19 studies recently published to de-termine whether FS analysis could differentiate follicularadenoma from follicular carcinoma showed that FS anal-ysis is a sensitive diagnostic test but not a specific one.26

However, FS classification of these studies included threecategories: benign, malignant, and “deferred,” which isdifferent from the present study and makes comparison ofthe results difficult.

The sensitivity obtained in our study is similar tothat reported in Leteurtre et al.9 In our study, 23.44% ofmalignancies were missed on FS analysis, similar to therate of 28% observed by Ho-Sheng et al.17 None of thecited reports used the classification recommended by thePapanicolaou Society of Cytopathology16 that we followed.To our knowledge, there are no previous studies to assessthe utility of FS analysis carried out only in cases ofcytological diagnoses of follicular neoplasm on FNA biopsyfollowing the Papanicolaou Society of Cytopathology clas-sification.16 We think that the categories used in FNAdiagnoses from other studies, such as benign, indetermi-nate, atypical features, or suspect for malignancy, are lessaccurate than Papanicolaou’s classification.

In the present study of 66 follicular lesions usingPapanicolaou’s classification, the only 2 malignancies thatwere missed on FNA biopsy but identified on FS analysishad a strong clinical evidence of malignancy: the rapidgrowth and size of the nodule in both patients (�4 cm) anda previous exposure to ionizing radiation in one of them.

The routine use of intraoperative FS analysis is use-less in cases of cytological diagnosis of follicular neoplasmon FNA biopsy, because of the low probability of detectinga follicular carcinoma. We think that the detection offollicular carcinoma does not merit freezing all thyroidlesions that have a preoperative FNA diagnosis of follicu-lar neoplasm. The use of an adequate preoperative FNAbiopsy result with an accurate cytological classification ofthe lesions, together with sound clinical judgment, cansufficiently determine the extent of surgical resection.

BIBLIOGRAPHY1. Mazzaferri EL. Thyroid cancer in thyroid nodules: finding a

needle in the haystack. Am J Med 1992;93:359–362.2. Gharib H. Changing concepts in the diagnosis and manage-

ment of thyroid nodules. Endocrinol Metab North Am 1997;

26:777–800.3. Bennedbaek FN, Perrild H, Hegedus L. Diagnosis and treat-

ment of the thyroid nodule: results of a European survey.Clin Endocrinol (Oxf) 1999;50:357–363.

4. Bennedbaek FN, Hegedus L. Management of the solitarythyroid nodule: results of a North American survey. J ClinEndocrinol Metab 2000;85:2493–2498.

5. Lucas Martin AM, Alonso Pedrol N, Sanmartı A. Enfermedadnodular tiroidea. Med Clın (Barc) 2000;114:181–184.

6. Gharib H, Goellner JR. Fine-needle aspiration biopsy of thethyroid: an appraisal. Ann Intern Med 1993;118:282–289.

7. Hermus AR, Huysmans DA. Treatment of benign nodularthyroid disease. N Engl J Med 1998;338:1438–1447.

8. Tyler DS, Winchester DJ, Caraway NP, Hickey RC, EvansDB. Indeterminate fine-needle aspiration biopsy of the thy-roid: identification of subgroups at high risk for invasivecarcinoma. Surgery 1994;116:1054–1060.

9. Leteurtre E, Leroy X, Pattou F, et al. Why do frozen sectionshave limited value in encapsulated or minimally invasivefollicular carcinoma of the thyroid? Am J Clin Pathol 2001;115:370–374.

10. Hamburger JI, Husain M. Contribution of intraoperative pa-thology evaluation to surgical management of thyroid nod-ules. Endocrinol Metab Clin North Am 1990;19:509–522.

11. De May RM. Frozen section of thyroid? Just say no. Am JClin Pathol 1998;110:423–424.

12. Tworek JA, Giordano TJ. Frozen section? Just do it. Am JClin Pathol 1998;112:124–125.

13. Lowhagen T. Thyroid. In: Aspiration biopsy cytology, I: cytol-ogy of supradiaphragmatic organs. Monogr Clin Cytol1987;4:67–89.

14. Mazzaferri EL. Management of a solitary thyroid nodule.N Engl J Med 1993;328:553–559.

15. Goelner JR, Gharib H, Grant CS, Johnson DA. Fine-needleaspiration cytology of the thyroid, 1980 to 1986. Acta Cytol1987;31:587–590.

16. The Papanicolaou Society of Cytopathology Task Force onStandards of Practice. Guidelines of the Papanicolaou So-ciety of Cytopathology for the examination of fine-needleaspiration specimens from thyroid nodules. Diagn Cyto-pathol 1996;15:84–89.

17. Ho-Sheng L, Komisar A, Opher E, Blaugrund S. Surgicalmanagement of thyroid masses: assessing the need forfrozen-section evaluation. Laryngoscope 1999;109:868–873.

18. Rosai J, Carcangiu ML, DeLellis RA. Atlas of Tumor Pathol-ogy: Tumors of the Thyroid Gland, 3rd series, fascicle 5.Washington, DC: Armed Forces Institute of Pathology,1992:31–50.

19. Raber W, Kaserer K, Niederle B, Vierhapper H. Risk factorsfor malignancy of thyroid nodules initially identified asfollicular neoplasia by fine-needle aspiration: results of aprospective study of one hundred twenty patients. Thyroid2000;10:709–712.

20. Gharib H. Fine-needle aspiration biopsy of thyroid nodules:advantages, limitations and effect. Mayo Clin Proc 1994;69:44–49.

21. Singer PA, Cooper DS, Daniels GH, et al. Treatment guide-lines for patients with thyroid nodules and well-differentiated thyroid cancer: American Thyroid Associa-tion. Arch Intern Med 1996;156:2165–2172.

22. Feld S, Garcia M, Baskin HJ. Clinical practice guidelines forthe diagnosis and management of thyroid nodules. EndocrPract 1996;2:78–84.

23. Davis NL, Gordaon M, Germann E, Robins RE, Mc GregorGI. Clinical parameters predictive of malignancy of thyroidfollicular neoplasms. Am J Surg 1991;161:567–569.

24. Schlinkert RT, Van Heerden JA, Goellner JR, et al. Factorsthat predict malignant thyroid lesions when fine-needleaspiration is suspicious for follicular neoplasm. Mayo ClinProc 1997;72:913–916.

25. Tuttle RM, Lemar H, Burch HB. Clinical features associatedwith an increased risk of thyroid malignancy in patientswith follicular neoplasia by fine-needle aspiration. Thyroid


565

1998;8:377–383.26. Miltenburg DM, Prost HM, Graviss EA, Arem R, Michael E.

The role of frozen section, gender, age, and tumor size inthe differentiation of follicular adenoma from carcinoma: ameta-analysis. Surgery 2000;128:1075–1081.

27. Bugis SP, Young JEM, Archibald SD, Chen VSM. Diagnosticaccuracy of fine-needle aspiration biopsy versus frozen sec-tion in solitary thyroid nodules. Am J Surg 1986;152:411–416.

28. Shaha AR, Gleich L, DiMaio T, Jaffe BM. Accuracy andpitfalls of frozen section during thyroid surgery. J SurgOncol 1990;44:84–92.

29. Mc Henry CR, Walfish PG, Rosen IB. Influence of FNAB and

FS examination on the management of thyroid cancer.Am J Surg 1993;166:353–356.

30. Sabel MS, Staren ED, Gianackakis LM, Dwarakanathan S,Prinz RA. Use of fine-needle aspiration biopsy and frozensection in the management of the solitary thyroid nodule.Surgery 1997;122:1021–1027.

31. Hamburger JI, Hamburger SW. Declining the role of frozensection in surgical planning for thyroid nodules. Surgery1985;98:307–312.

32. Malcahy MH, Cohen JI, Anderson PE, Ditamasso J, SchmidtW. Relative accuracy of fine-needle aspiration and frozensection in the diagnosis of well-differentiated thyroid can-cer. Laryngoscope 1998;108:494–496.


566

Documents

Frozen Section in a Cytological Diagnosis of Thyroid Follicular Neoplasm