A fine-needle aspiration biopsy specimen suspicious for papillary thyroid cancer (original magnification ×510) demonstrating marked cellularity (A), papillary formation (B), and nuclear inclusions and grooves (arrows) (C) from a patient with a final pathologic diagnosis of papillary carcinoma.
An indeterminate fine-needle aspiration biopsy specimen that was interpreted as “papillary thyroid cancer cannot be excluded” (original magnification ×510) based on the presence of abundant colloid (lavender background material) (A) and sheets of follicular epithelial cells with rare intranuclear inclusions (arrows) (B) from a patient with a final pathologic diagnosis of benign adenomatous goiter.
Algorithm for treatment of a patient with nodular thyroid disease and a fine-needle aspiration biopsy specimen (FNAB) suspicious for papillary thyroid cancer (PTC). FSE indicates frozen section examination.
Mittendorf EA, Khiyami A, McHenry CR. When Fine-Needle Aspiration Biopsy Cannot Exclude Papillary Thyroid CancerA Therapeutic Dilemma. Arch Surg. 2006;141(10):961–966. doi:10.1001/archsurg.141.10.961
Clinical or cytologic factors predictive of malignancy can be identified and incorporated into a treatment algorithm for patients with a fine-needle aspiration biopsy (FNAB) specimen interpreted as “suspicious for” papillary thyroid cancer (PTC).
Retrospective review of a prospectively maintained database.
University-affiliated tertiary care hospital.
Seven hundred thirty-eight patients with nodular thyroid disease evaluated between 1990 and 2004.
Patients with an FNAB specimen suspicious for PTC were identified. The frequency of carcinoma was determined. Clinical features were reviewed. The FNAB specimens suspicious for PTC were examined in a blinded fashion to determine if specific cytologic features were important in distinguishing benign vs malignant disease.
Main Outcome Measures
The presence of specific clinical and cytologic features was correlated with the incidence of carcinoma. A secondary outcome measure was to determine the value of frozen section examination in establishing the extent of thyroidectomy.
Forty-five patients (7%) had an FNAB specimen suspicious for PTC; 18 (40%) of these patients had carcinoma. Prominent nuclear inclusions and/or grooves, papillary formations, and the absence of colloid were features associated with PTC (P<.05). No clinical features reliably identified malignant disease. Frozen section examination results altered treatment in 15 (56%) of 27 patients.
An FNAB specimen suspicious for PTC is associated with a 40% incidence of carcinoma. Extensive nuclear inclusions and/or grooves, papillary formations, and the absence of colloid are predictive of carcinoma. Rare intranuclear inclusions and/or grooves alone in an otherwise benign-appearing specimen are uniformly associated with benign disease. Frozen section examination is of value in determining the extent of thyroidectomy.
Fine-needle aspiration biopsy (FNAB) is an accurate method for differentiating benign from malignant thyroid nodules. Its routine use has decreased surgical intervention by 25% and has increased the yield of cancer in surgical specimens from 15% to more than 30%.1 With reported sensitivity rates exceeding 90%,1- 3 FNAB cytologic features are particularly accurate in the detection of papillary thyroid cancer (PTC), which accounts for 70% to 80% of all thyroid malignancies.4
The cytologic criteria that are necessary to make a definitive diagnosis of papillary carcinoma on an FNAB specimen include large monolayer sheets of follicular epithelial cells with enlarged nuclei containing fine powdery chromatin, intranuclear cytoplasmic inclusions and nuclear grooves, and papillary structures with or without tall columnar cells.5 If these cytologic criteria are met, definitive operative management can be recommended solely on the basis of the FNAB result without further testing. In most patients with papillary carcinoma, total thyroidectomy is preferable, although thyroid lobectomy is an acceptable alternative for patients with “low-risk” cancers.4
If some but not all of the cytologic criteria for PTC are present with otherwise benign cytologic features, a PTC cannot be excluded. There is no consensus on what constitutes appropriate treatment for these patients. Chen and colleagues6 reported a malignancy rate of 54% in patients with an FNAB specimen suspicious but not diagnostic for PTC. They advocated lobectomy and isthmusectomy with intraoperative frozen section examination (FSE). If FSE provided histologic confirmation of the diagnosis, then a total or near-total thyroidectomy was performed. Sclabas et al7 reported a malignancy rate of 82% in patients with an FNAB specimen suspicious for PTC. They considered these patients to be a high-risk subgroup among those with an indeterminate FNAB specimen and recommended total thyroidectomy. This is a change in their institutional approach because earlier reports from the M. D. Anderson Cancer Center recommended the use of FSE before committing to total thyroidectomy.8
The purpose of this study was to review our experience with patients with nodular thyroid disease and an FNAB specimen that was interpreted as suspicious for PTC. We sought to determine the frequency of an FNAB specimen that was suspicious but not definitive for PTC, the frequency of malignancy in patients with an FNAB specimen suspicious for PTC compared with patients with an FNAB specimen definitive of PTC, and the role of FSE in the treatment of patients with nodular thyroid disease and an FNAB specimen suspicious for PTC. Importantly, we sought to identify clinical factors and specific cytologic features that could help distinguish benign from malignant disease in patients with an FNAB specimen suspicious for PTC.
After approval by the MetroHealth Medical Center institutional review board, data were obtained from a prospectively maintained, institutional review board–approved database of patients with nodular thyroid disease evaluated by a single surgeon since 1990. Patients with an FNAB result reported as “suspicious for” or “cannot exclude” PTC were identified and formed the study population. Although patients with an FNAB specimen interpreted as follicular neoplasm may have papillary carcinoma or a follicular variant of papillary carcinoma on final pathologic examination, they were not the subject of this study and thus were not included. An FNAB specimen suspicious for PTC was defined as a specimen containing follicular epithelial cells with a variable presence of nuclear grooves and/or inclusions, papillary formations, and/or psammoma bodies in an otherwise benign-appearing specimen. The FNAB specimens were retrospectively examined by a single cytopathologist, who was blinded with respect to the final pathologic diagnosis, to determine if specific cytologic features were important in distinguishing benign and malignant disease. The FNAB specimens were evaluated for cellularity, nuclear inclusions and/or grooves, papillary formations, psammoma bodies, colloid, atypia, and pleomorphism and were graded on a scale of 0 to 4 with 0 being absent; 1, rare; 2, few or mild; 3, moderate; and 4, extensive. Colloid was further characterized as thin or thick (“chewing gum”) in character. The final pathologic results for all patients with an FNAB specimen that was suspicious for PTC were reviewed to determine the incidence of carcinoma and the spectrum of benign disease. The frequency of carcinoma for patients with an FNAB specimen suspicious for PTC was compared with patients with a benign vs a malignant FNAB result.
Other data that were obtained included patient age and sex, a history of head or neck irradiation, size of the thyroid nodule, serum thyrotropin level, results of thyroid scintigraphy, results of intraoperative FSE, and extent of thyroidectomy. The results of patient age, sex, history of head or neck irradiation, nodule size, serum thyrotropin level, and iodine 123 (123I) scintigraphy were compared in patients with benign disease and patients with malignant disease. The role of FSE in determining the extent of thyroidectomy was also examined.
Statistical significance between subgroups was determined using the χ2 test. All P values were calculated using 2-sided significance testing, and P values <.05 were considered statistically significant. Statistical analysis was performed using SPSS software (SPSS Inc, Chicago, Ill).
During the period from 1990 to 2004, 738 patients were evaluated for nodular thyroid disease, including 622 who had 1 or more FNABs performed on a dominant thyroid nodule. Among the 622 patients who had an FNAB performed, the results were malignant in 36 (6%), consistent with a follicular or Hürthle cell neoplasm in 164 (26%), benign in 308 (50%), nondiagnostic in 67 (11%), and suspicious for cancer in 47 (7%). The reasons for not performing an FNAB in 116 patients with a thyroid nodule included patient refusal, nonpalpable nodules less than 1 cm, palpable associated lymphadenopathy on which a biopsy was performed, or an indication to proceed directly with thyroidectomy, such as history of head or neck irradiation, progressive nodule enlargement, or compressive symptoms.
A review of the 47 patients with an FNAB specimen categorized as suspicious for cancer revealed 2 patients with features that were suggestive of a lymphoproliferative disorder. These patients were excluded from this review. This left 45 patients who had an FNAB specimen that was suspicious for PTC. The extent of thyroidectomy included lobectomy and isthmusectomy in 10 (22%), near-total thyroidectomy in 4 (9%), and total thyroidectomy in 31 patients (69%). The final pathologic results for the 45 patients with an FNAB specimen suspicious for PTC were carcinoma in 18 (40%), adenomatous goiter in 14 (31%), follicular adenoma in 9 (20%), and thyroiditis in 4 (9%). Associated thyroiditis was also present in 4 of the patients with carcinoma and 2 of the patients with benign disease. In comparison, all 36 patients (100%) with a malignant FNAB specimen and 11 (3.6%) of 308 patients with a benign FNAB specimen, operated on because of compressive symptoms, tracheal compression, or progressive nodule enlargement, had a carcinoma. In 7 of the 11 patients with a benign FNAB specimen, the carcinoma was an incidentally discovered occult papillary microcarcinoma less than 10 mm.
A comparison of sex distribution, age, history of head or neck irradiation, mean nodule size, mean serum thyrotropin level, and results of 123I thyroid scintigraphy for patients with malignant vs benign disease is detailed in Table 1. There was a significant difference in the mean age of patients found to have a carcinoma, 40 years vs 52 years for patients with benign disease (P<.05). There were no other significant differences observed between the 2 groups of patients. Iodine 123 scintigraphy was performed in 21 patients, revealing a solitary hypofunctioning nodule in 18 (86%), heterogeneous uptake in 2 (9%), and a solitary isofunctioning nodule in 1 patient (5%).
The specific cytologic features and their frequency in patients with benign vs malignant disease are detailed in Table 2. Analysis of our data demonstrates that the presence of moderate or extensive nuclear inclusions and/or grooves, any papillary formations, and rare or absent colloid were significantly more likely to be associated with malignant disease (P<.05) (Figure 1). There was no significant difference in the presence of psammoma bodies, cellular atypia, or cellular pleomorphism between the patients with benign vs malignant disease. Thirteen patients had only rare intranuclear inclusions and/or grooves (grade 1) without papillary formations in a background of significant colloid (grade 2-4) (Figure 2), and all of them had benign disease, 7 with a colloid nodule or adenomatous goiter, 2 with a follicular adenoma, and 1 with thyroiditis.
Frozen section examination of the dominant thyroid nodule was performed in 27 of the 45 patients, and it was helpful for intraoperative decision making in 15 patients (56%). Frozen section examination was definitive in identifying benign disease in 8 patients, and this was important in limiting the extent of thyroid resection to a less than total thyroidectomy. Frozen section examination made a definitive diagnosis of carcinoma in 6 patients, which led to the performance of a total thyroidectomy and removal of adjacent lymph nodes. In the remaining patients who had an FSE performed, the interpretation was follicular neoplasm, and definitive diagnosis was deferred until all of the permanent sections could be reviewed postoperatively. As a result, an FSE diagnosis of a follicular neoplasm was not helpful in intraoperative decision making. In 18 patients, a decision was made to forego FSE and perform a total thyroidectomy based on the presence of bilateral nodular disease or to perform a completion thyroidectomy in a patient who had undergone a prior lobectomy.
Fine-needle aspiration biopsy is the most important diagnostic modality used for the evaluation of patients with nodular thyroid disease. Fine-needle aspiration biopsy has an accuracy rate approaching 95%.9 Cytologic results from FNAB can be reported in 1 of 4 categories: benign, malignant, indeterminate for malignancy, or inadequate. Patients with a benign FNAB specimen can be followed up clinically. The false-negative rate for a benign FNAB specimen is approximately 5%.10 If the nodule increases in size or becomes symptomatic, thyroidectomy is recommended.
An FNAB specimen that is malignant may be either a papillary, medullary, or anaplastic carcinoma. These lesions are differentiated based on characteristic cytologic features. The false-positive rate for patients with a malignant FNAB result is less than 2%.11 As a result, definitive therapy is recommended based on the cytologic result alone. In our series, there were no false-positive FNAB results among our 36 patients with an FNAB specimen definitive for PTC. We advocate total thyroidectomy for any patient with a PTC nodule larger then 1 cm. Total thyroidectomy for these patients has been shown to decrease the recurrence rate and facilitate the use of 123I for ablation of residual uptake in the thyroid bed and for diagnosis and treatment of metastatic disease.12,13 Total thyroidectomy also allows for the effective use of serum thyroglobulin for detection of recurrent disease.
It has previously been reported that approximately 20% of patients who undergo FNAB have a cytologic result that is interpreted as indeterminate for malignancy.8 This group includes patients with cytologic findings consistent with a follicular or Hürthle cell neoplasm or suggestive but not diagnostic of PTC. In our series, 33% of patients had an indeterminate FNAB result, 26% with a follicular or Hürthle cell neoplasm and 7% whose FNAB specimen was suspicious for malignancy. In patients with an indeterminate cytologic FNAB result, thyroidectomy is necessary to make a histologic diagnosis and for definitive treatment of patients who are subsequently diagnosed with carcinoma.
In general, the role of FSE is to differentiate a benign from a malignant lesion and establish a definitive histologic diagnosis. Prior to the widespread use of FNAB, FSE was routinely used in thyroid surgery. With the evolution of FNAB as a highly sensitive and specific diagnostic tool, the utility of routine intraoperative FSE has been questioned. In an earlier review from our institution, we prospectively evaluated 76 patients with nodular thyroid disease who had an adequate FNAB specimen and then underwent thyroidectomy with routine FSE. Frozen section examination was helpful in determining the extent of thyroid resection in only 2 patients (3%).14 These findings were comparable with those of Hamburger and Hamburger15 who noted that FSE had a decisive influence on the operative intervention in less than 1% of patients operated on for nodular thyroid disease. We no longer advocate the routine use of FSE in patients who have a definitive FNAB result.
There are situations, however, where FSE has been reported to be of benefit. Other investigators have reported that FSE can be useful in patients with an FNAB specimen that is suspicious but not definitive for PTC.6,16 In the current series, intraoperative FSE was used in 27 patients. In 15 (56%) of the 27 patients, findings on intraoperative pathologic assessment altered the extent of surgery performed. In 5 patients with a single nodule confirmed to be benign, the extent of thyroid resection was limited to a lobectomy and isthmusectomy. None of these patients had malignancy identified on permanent histologic sections. In 4 patients with bilateral nodular disease that was confirmed on FSE to be benign, a near-total instead of total thyroidectomy was performed, leaving residual thyroid tissue around the parathyroid glands to ensure their in situ preservation. Early in our experience, we did not perform FSE in patients with bilateral nodular disease because we planned to perform a total thyroidectomy regardless of the results of intraoperative pathologic analysis. We now advocate performing FSE in all patients with an FNAB specimen read as suspicious for PTC. If they have bilateral nodular disease and benign FSE results, a near-total thyroidectomy is performed, leaving a small amount of thyroid tissue when necessary to ensure in situ preservation of the parathyroid glands.
In 6 patients in this study, FSE established a diagnosis of PTC and all had a total thyroidectomy performed. In these patients, FSE obviated the need for a subsequent completion thyroidectomy, which would have been necessary had the diagnosis of malignancy not been made until after histologic confirmation from paraffin-embedded sections.
Clinical evaluation of patients with nodular thyroid disease involves identifying risk factors that may increase the probability that a nodule is malignant. Though some factors may be helpful in gauging an individual's risk for harboring malignancy, they are not reliable enough to aid in confirming the diagnosis of carcinoma. Sex and mean nodule size were not significantly different between patients with an FNAB specimen suspicious for PTC and benign vs malignant disease. Though we only had 3 patients with a history of head and neck irradiation, and 2 of these had benign disease on final pathologic assessment, because of the well-established association between prior irradiation and the development of thyroid cancer, we routinely perform total thyroidectomy in patients with such a history and an FNAB specimen interpreted as suspicious for PTC. Interestingly, there was a significant difference in the age of the patients in the 2 groups in our study. Those with benign disease were significantly older then those with malignancy (52 vs 40 years of age). This is in contrast to other reported series where thyroid cancer is more common in older patients.4 Cytologic features seen in PTC may also occur in patients with benign disease as a result of degenerative changes. The age difference observed in our patients with benign and malignant disease may reflect a greater likelihood of nodule degeneration in older patients who may have had nodular thyroid disease for a longer period and, as a result, were more likely to undergo degenerative changes.
We also have confirmed that thyroid scintigraphy is of little help in distinguishing benign from malignant nodules and has no role in determining the extent of thyroid resection to be performed. We do not routinely obtain thyroid scintigraphy17 and the data presented in Table 1 summarize the results of studies ordered by referring physicians. Of interest, the majority of hypofunctioning or “cold” nodules in this study were benign. If we had advocated total thyroidectomy for patients with an FNAB specimen suspicious for PTC and a solitary hypofunctioning nodule on thyroid scintigraphy, 11 patients would have undergone more extensive resection than required for their disease. Previous reports have looked at the scintigraphic characteristics of thyroid nodules. Eighty-four percent are nonfunctioning (cold), 10.5% have uniform tracer uptake (warm), and 4% are hyperfunctioning (hot).18,19 Cancer has been reported in approximately 16% of cold nodules, 9% of warm nodules, and less than 1% of hot nodules.20
We also sought to determine if there were specific cytologic features that could help differentiate benign from malignant disease. The presence of moderate or extensive nuclear inclusions and/or grooves and papillary formations was significantly greater in FNAB specimens from patients ultimately found to have malignant disease. In addition, rare or absent colloid was more common in patients with malignancy. Our relatively small sample size precludes us from performing multivariate statistical analysis. However, one can hypothesize that an FNAB specimen with the variable combination of more than 1 of these features, extensive nuclear inclusions and/or grooves, papillary formations, and the absence of colloid, would be even more likely to be indicative of malignant disease. Though these features are more suggestive for malignancy when they are present within an otherwise benign-appearing cytologic specimen, they are not sufficient to make a definitive cytologic diagnosis. Therefore, we do not suggest that their presence should be used to determine the extent of thyroidectomy to be performed. We advocate performing an FSE on all patients with an FNAB specimen that is suspicious, but not definitive, for PTC. The extent of thyroid resection to perform is determined by the intraoperative pathologic assessment as outlined in the algorithm in Figure 3.
There was no significant difference in the presence of psammoma bodies, cellular atypia, or cellular pleomorphism between patients with benign or malignant disease. The presence of rare intranuclear inclusions and/or grooves in an otherwise benign specimen with abundant colloid and no papillary formations was uniformly associated with benign disease. As a result, we recommend clinical follow-up of patients with a dominant nodule and a benign FNAB result in the presence of rare intranuclear inclusions and/or grooves. We also recommend repeat FNAB to exclude the possibility of a false-negative result.
Recent investigations into biomolecular markers have identified a gene mutation that may aid in differentiating benign from malignant disease in patients with an FNAB specimen suspicious for PTC. BRAF gene mutations are common in human cancers21 and several investigators have reported a transversion mutation in 29% to 69% of PTCs.22- 29 Importantly, this mutation has been reported to be specific for PTC and has not been identified in benign thyroid neoplasms.30BRAF therefore has the potential to serve as a specific molecular marker for PTC. Xing and colleagues30 have recently reported their experience detecting the BRAF mutation in FNAB specimens. They showed that this mutation can be easily and reliably demonstrated and that it correctly identified 50% of PTCs. Although their study did not specifically look at patients with indeterminate cytologic findings on FNAB, through extrapolation of their results, one can hypothesize that the detection of the BRAF mutation on FNAB specimens could be a useful adjunctive diagnostic technique.
Correspondence: Christopher R. McHenry, MD, Department of Surgery, MetroHealth Medical Center, 2500 MetroHealth Dr, Cleveland, OH 44109 (email@example.com).
Accepted for Publication: July 22, 2005.
Author Contributions:Study concept and design: Mittendorf and McHenry. Acquisition of data: Mittendorf, Khiyami, and McHenry. Analysis and interpretation of data: Mittendorf, Khiyami, and McHenry. Drafting of the manuscript: Mittendorf. Critical revision of the manuscript for important intellectual content: McHenry. Statistical analysis: Mittendorf. Administrative, technical, and material support: Khiyami and McHenry. Study supervision: Khiyami and McHenry.