Clinical Outcomes for “Suspicious” Category in Thyroid Fine-Needle Aspiration Biopsy: Patient's Sex and Nodule Size Are Possible Predictors of Malignancy

Palpable thyroid nodules are relatively common, being reported in 4% to 7% of the adult population. Most of these lesions are benign, with a malignancy rate of less than 5%.1 Therefore, surgery with its attendant costs, risks, and complications is an unnecessary procedure for many patients.

Fine-needle aspiration (FNA) is widely accepted as the most accurate and cost-effective tool for the initial screening and triage of thyroid nodules.2–4 According to the “National Cancer Institute (NCI) Thyroid Fine Needle Aspiration State of the Science Consensus Conference,” every patient with a palpable thyroid nodule is a candidate for FNA procedure.5 The introduction of fine-needle aspiration has decreased the percentage of patients undergoing thyroidectomy by 25% and has increased the yield of surgery (ie, finding malignancy in thyroidectomy specimens) from 15% to at least 30%.1,6 

In the recent NCI Consensus Conference, a 7-tier diagnostic scheme was proposed consisting of the following categories: benign, atypia of undetermined significance, suspicious for follicular, suspicious for Hürthle cell neoplasm, suspicious for malignancy, malignant, and insufficient for diagnosis. The aspirates that are “suspicious for follicular or Hürthle cell neoplasm” are cellular smears consisting of follicular/Hürthle cells in a scant colloid background. Because the criteria for malignancy in both follicular neoplasm (FN) and Hürthle cell neoplasm (HN) requires vascular or capsular invasion, “suspicious” diagnosis on FNA usually warrants surgical resection. In reported series, the rate of malignancy in thyroidectomy specimens of patients with initial cytologic diagnoses of “suspicious for FN” varies from 15% to 47% and from 20% to 45% for “suspicious for HN.” Some studies suggest that a Hürthle cell versus follicular cell morphology, patient's age or sex, and/or size of nodule can be predictors of a malignant outcome.7–15 In contrast, other reports state that clinical factors are not helpful in predicting carcinoma in patients with suspicious FNA findings.12 These results are contradictory and confusing. In this study, we report our experience with thyroid FNA diagnoses of “suspicious” in a series of patients who underwent subsequent thyroidectomy. Age, sex, and size of the nodule were examined as possible predictors of malignancy. We specifically focused on the FNA diagnosis of “suspicious for FN” and “suspicious for HN” to compare the rate of malignancy in these 2 groups.

Cytopathology reports of all thyroid FNAs performed at The Methodist Hospital in Houston, Texas, from 2000–2005 were reviewed. The total number of thyroid FNAs during this period was 3997. Thyroid FNA diagnoses were grouped into 4 major categories: benign, malignant, suspicious, and nondiagnostic. The “suspicious” category was subclassified to “suspicious for follicular neoplasms,” “suspicious for Hürthle cell neoplasms,” or “suspicious for malignancy.” Cellular adenomatous nodules that were worrisome for adenoma or carcinoma were reported under the “suspicious” category and comprised less than 5% of cases in this category. The cases with mild cytologic atypia and abundant colloid were reported as “negative.”

Cytopathology slides and reports of all thyroid FNAs interpreted as “suspicious” were reviewed. Four hundred and two cases were suspicious for neoplasm. Surgical follow-up information was available for 180 of these 402 patients.

All FNA biopsies were performed at the Methodist Hospital, Houston. The aspirations were done with or without ultrasound guidance. Thyroid nodules that were easily palpable were aspirated by a board-certified cytopathologist and nonpalpable thyroid nodules were aspirated under ultrasound guidance by a radiologist. Less than 10% of FNA biopsies were performed by endocrinologists. All aspirations were performed with 23- to 25-gauge needles attached to a 10- or 20-mL syringe with or without a syringe holder. Each nodule was aspirated 3 to 5 times. Air-dried smears were immediately stained with the Diff-Quik methodology (Siemens Healthcare Diagnostics, Deerfield, Illinois) for adequacy check, and alcohol-fixed aspirates were stained according to the Papanicolaou methodology. The needles were rinsed in Roswell Park Memorial Institute (RPMI) media or CytoLyt (Cytyc Corporation, a Hologic Company, Boxborough, Massachusetts). These were used for making liquid-based preparations (ThinPrep) (Cytyc). An adequate FNA specimen was defined as containing 8 to 10 clusters of at least 6 to 8 well-preserved and well-visualized follicular cells.16 

Histologic diagnoses were available in 180 cases. In most cases, lobectomy and isthmectomy were performed with intraoperative frozen-section evaluation.

The reports of all patients who underwent thyroid FNA biopsy and thyroidectomy with histologic evaluation were reviewed in accordance with institutional review board guidelines. The FNA biopsy results were analyzed for a possible association between malignancy rate and preoperative factors such as patient's age, patient's sex, size of nodule, and/or morphology (FN vs HN). Positive predictive values (PPV) were calculated for each group. Student t and Fisher exact tests were used for comparison and evaluation of variables. P values less than .05 were considered statistically significant.

Other morphologic features including increased single cells, nuclear pleomorphism, nuclear enlargement, macronucleoli, and quality of the colloid were also investigated as possible predictors of outcome.

The rate of the “suspicious” diagnosis for thyroid FNAs in our laboratory varied between 7% and 10% during the years these cases were collected. Of the 180 cases initially diagnosed as “suspicious for neoplasm” for which surgical follow-up information was available, cytologic evaluation found 108 cases were suspicious for FN, 37 for HN, 34 for papillary carcinoma, and 1 for medullary carcinoma. The FNA cytology results were correlated with the subsequent histologic diagnoses (Table 1).

Of the 108 cases diagnosed as “suspicious for FN” on cytologic evaluation, the follow-up histologic diagnoses were malignant in 26 (24%) and benign in 82 (76%) cases. There were 77 women and 31 men in this group. The patients in this group ranged in age from 18 to 80 years with a mean age of 48.9 years.

Of the 37 cases diagnosed as “suspicious for HN,” 15 (41%) were diagnosed as malignant in the subsequent histologic material and 22 (59%) were benign. There were 31 women and 6 men in this group. The patients ranged in age from 27 to 77 years with a mean age of 51.5 years.

Of the 34 cases diagnosed as “suspicious for papillary carcinoma” on cytologic evaluation, the histologic follow-up showed malignant tumors in 28 (82%) and benign lesions in 6 (18%) patients. There were 28 women and 6 men in this group, with ages ranging from 26 to 95 years and a mean age of 48.6 years.

The PPV for a malignant neoplasm was 24% (26/108) for patients with a cytologic diagnosis of “suspicious for FN” and 41% (15/37) for a diagnosis of “suspicious for HN.” The difference was not statistically significant (P = .06). The PPV for a malignant neoplasm was 82% (28/34) for cases with cytologic diagnoses of “suspicious for papillary carcinoma.”

The rate of finding adenomas or carcinomas was 69% (74/108) in cases with a cytologic diagnosis of “suspicious for FN.” It was higher (92%, 34/37) in cases with a cytologic diagnosis of “suspicious for HN.”

Among cases with a cytologic diagnosis of FN, the rate of malignancy for female patients was 18% (14/77) as compared to 39% (12/31) for male patients. This difference was statistically significant (P = .04). When the cytologic diagnosis of “suspicious for FN or HN” was combined, the rate of malignancy for men was 43% compared to 22% for women. This difference was statistically significant as well (P = .02) (Figure).

The rate of malignancy in nodules less than 2 cm was 19% compared to 47% in nodules measuring 2 cm or larger. This difference was statistically significant (P < .001) (Table 2).

The patients with cytologic diagnoses of “suspicious for FN or HN” were further subclassified on the basis of their age. There were 72 patients less than 50 years old. In this group of patients, the rates of malignancy (PPV) for cases with the diagnosis of “suspicious for FN” or “suspicious for HN” were 30% and 33%, respectively. There were 73 patients aged 50 years or older. The PPVs for diagnoses of FN and HN were 19% and 42%, respectively, in this group. The overall PPV for patients with cytologic diagnoses of “suspicious for FN or HN” was 31% in patients younger than 50 years and 25% in patients 50 years or older. The difference was not statistically significant.

We also looked at several morphologic features including crowding, increased single cells, nuclear pleomorphism, nuclear enlargement, macronucleoli, and quality of the colloid. None of these factors correlated with surgical outcome.

Fine-needle aspiration biopsy of the thyroid gland is safe, inexpensive, minimally invasive, and highly accurate in the diagnosis of thyroid nodules1,2,6 Practice guidelines proposed by the American Thyroid Association and the National Comprehensive Cancer Network state that FNA should be used as the initial diagnostic test, before both thyroid scintigraphy and ultrasonography.4 If FNA is performed by a skilled physician and is evaluated by an experienced pathologist, false-positive or false-negative results are uncommon.

Until recently, there was no standard classification for reporting thyroid FNA findings. Thyroid FNA diagnoses were usually grouped into 4 categories: benign, malignant, suspicious, and nondiagnostic. In some cytology reports, vague and ambiguous terminology, such as “inconclusive” or “indeterminate,” was used. This terminology was confusing and interpreted differently by various treating physicians. At our institution, the “suspicious” category was usually followed by a descriptive diagnosis of “suspicious for follicular neoplasms,” “suspicious for Hürthle cell neoplasms,” or “suspicious for papillary carcinoma.” This is essentially in line with the recent 6-tier diagnostic scheme, which was proposed at the recent NCI Thyroid Fine Needle Aspiration State of the Science Consensus Conference. Most committee members and participants at this conference favored the following 7-category diagnostic scheme: benign, atypia of undetermined significance, suspicious for follicular neoplasm, suspicious for Hürthle cell neoplasm, suspicious for malignancy, malignant, and insufficient for diagnosis.5 This scheme is beneficial for triaging patients for either clinical follow-up or surgical management.17 

Patients with benign results from thyroid cytologic evaluation are usually followed-up with medical management, and malignant nodules are treated surgically.18 The nondiagnostic/unsatisfactory and atypical categories warrant repeated FNAs.

In our series, the patients with cytologic diagnoses of “suspicious for papillary carcinoma” had a high likelihood of having papillary carcinoma (82%) on subsequent histologic evaluation. This finding is consistent with other recent reports that cited a PPV ranging from 72% to 90%.6–9 The high prevalence of malignancy in this group of patients justifies surgical treatment. In most centers in the United States, these patients undergo lobectomy with intraoperative evaluation. If papillary carcinoma is confirmed, the surgeon proceeds with a concurrent completion thyroidectomy. More challenging and controversial is the management of patients with follicular or Hürthle cell neoplasms. Thyroid lobectomy with frozen-section analysis is a common approach in this group of patients as well. However, frozen-section evaluation may miss vascular or capsular invasion because of the limited number of sections and frozen-section artifacts. In these cases, final diagnosis is deferred to examination of the entire capsule on permanent sections, and a second operation may be required if the initial procedure was unilateral thyroid lobectomy.

Some investigators have tried to find out which preoperative patient or tumor characteristics can predict a malignant outcome for follicular or Hürthle cell neoplasms. The results are controversial and contradictory. Some studies have suggested that the findings more closely associated with a malignant nodule were patient's age, size of the tumor, or sex of the patient. Other studies have concluded that in patients with suspicious FNA biopsy findings, these clinical factors or morphologic features were not helpful in predicting carcinoma and thus cannot be used to reliably select patients for more extensive thyroidectomy.10–15 

Tyler et al reported that patients who were 50 years or older and had a cytologic diagnosis of “suspicious for FN or HN” on FNA were more likely to harbor an invasive carcinoma.9 On the basis of their data, for patients 50 years or older, if the frozen section showed follicular neoplasm on unilateral lobectomy, they suggested performing a completion thyroidectomy during the initial surgery. In this age group, for total thyroidectomy, they did not require a confirmation of malignancy on frozen-section evaluation. On the other hand, according to their algorithm, for patients younger than 50 years who had frozen-section confirmation of the FN or HN with no evidence of vascular or capsular invasion, no further surgery was required. In our series of 145 cases with cytology diagnoses of “suspicious for FN or HN,” there was no association between risk of malignancy and the patient's age. Indeed, rate of malignancy (PPV) in 72 patients younger than 50 years was 31% (22/73), which was slightly higher than the PPV value of 25% in the 73 patients 50 years and older. Another, more recent study by the same group, performed on a larger sample size, did not confirm the previous finding of increased risk of malignancy in patients 50 years or older.8 In their more recent study, this group reported PPVs of 15% for cytologic diagnoses of follicular neoplasms, 20% for Hürthle cell neoplasms, and 82% for papillary carcinomas. They indicated that a nodule diameter greater than 2 cm was the only factor associated with an increased risk of malignancy (P < .03). This finding is in accordance with our PPV finding for malignancy of 19% for nodules less than 2 cm as compared to 47% for nodules measuring 2 cm or larger.

Pu et al15 observed that FNA diagnoses for men and older patients with Hürthle cell neoplasm carry a higher risk of carcinoma upon subsequent thyroidectomy. In our study, we found statistically significant difference between the rate of malignancy in men with a prior cytologic diagnosis of “suspicious for neoplasm” as compared to women with similar diagnosis. Among cases with a cytologic diagnosis of “suspicious for FN,” the rate of malignancy in female patients was 18% as compared to 39% in male patients. The overall rate of malignancy in men with cytologic diagnosis of “suspicious for FN or HN” was 43% compared to 22% in women with similar diagnosis.

Some authors have reported a higher rate of malignancy for the FNA diagnosis of “suspicious for HN” compared to the diagnosis of “suspicious for FN” 14 (Table 3). In our series, the risks of having a malignant neoplasm for patients with cytologic diagnoses of “suspicious for HN” was 41%, which was higher than the 26% risk of malignancy for patients with diagnoses of “suspicious for FN.” This difference, however, was not statistically significant (P = .06).

Our data show that malignant tumors were more frequent in men with a cytologic diagnosis of “suspicious for follicular or Hürthle cell neoplasm” than for women with similar cytologic diagnoses. There was also a statistically significant difference between the size of the nodule (>2 cm) and rate of benign versus malignant diagnoses. The positive predictive value for malignancy in cases diagnosed as “suspicious for Hürthle cell neoplasm” was higher than in cases diagnosed as “suspicious for follicular neoplasm.” The difference, however, was not statistically significant. None of the morphologic factors, such as increased single cells, nuclear pleomorphism, nuclear enlargement, macronucleoli, or quality of the colloid, correlated with surgical outcome.