Secondary Logo

Journal Logo

The role of fine-needle aspiration cytology and ultrasound elastography in predicting malignancy in thyroid nodules

Hafez, Nesreen H.a; Tahoun, Neveen S.a; Mokhtar, Omniab

doi: 10.1097/01.XEJ.0000436663.29516.1a
ORIGINAL ARTICLES
Free

Background It is crucial to have a clear diagnostic approach to ensure patients presenting with thyroid nodules are managed appropriately. Fine-needle aspiration cytology (FNAC) is the best preoperative test used for differentiating malignant from benign nodules. Cytology has certain limitations that restrict its use as a single diagnostic test. Ultrasound elastography (USE) is a newly developed technique that uses ultrasound to provide an estimation of tissue stiffness under application of external force. The aim of this study was to assess whether USE improves the accuracy of FNAC in predicting malignancy in thyroid nodules.

Patients and methods This prospective study included 96 patients with a solitary thyroid nodule or dominant nodule in the setting of multinodular goiter from January 2011 to July 2012. They were examined using USE at the Radiology Department, National cancer Institute, Cairo University. Tissue stiffness was scored from 1 (low stiffness over the entire nodule) to 5 (high stiffness over the entire nodule and surrounding tissue). All patients were referred for FNAC at the Cytology Unit, Pathology Department. Cytological specimens were evaluated blindly (without knowing the scores of USE) and classified according to the recent Bethesda classification. All the studied patients underwent surgery and were evaluated for the final histopathological diagnoses.

Results On FNAC, 46 of 61 cases (75.4%) with a final histopathological diagnosis of benign nodule had a negative diagnosis, whereas 34 of 35 cases (97.1%) with a final histopathological diagnosis of carcinoma had a positive diagnosis (P<0.005), with a sensitivity of 97.1%, a specificity of 75.4%, a positive predictive value (PPV) of 69.4%, a negative predictive value (NPV) of 97.9%, and diagnostic accuracy of 83.3%. On real-time USE, 57 of 61 benign nodules had a score of 1 or 2, whereas all malignant nodules had a score of 3–5 (P<0.0001), with a sensitivity of 100%, specificity of 93.4%, PPV of 89.7%, NPV of 100%, and diagnostic accuracy of 95.8%. In the 18 nodules with a cytological report of follicular neoplasm, USE was efficient in differentiating benign from malignant nodules (P<0.0005).

Conclusion USE is a promising imaging technique that can assist in the differential diagnosis of thyroid nodules with higher sensitivity, specificity, PPV, NPV, and diagnostic accuracy compared with FNAC. Therefore, it may be useful to introduce it into the routine clinical practice in association with FNAC.

aDepartment of Pathology, Cytopathology Unit

bDepartment of Radiodiagnosis, National Cancer Institute (NCI), Cairo University, Cairo, Egypt

Correspondence to Nesreen H. Hafez, MD, Department of Pathology, Cytopathology Unit, National Cancer Institute (NCI), Cairo University, Cairo, Egypt Tel: +20 222 607 376/+20 010 495 9148; e-mail: nesreennci@hotmail.com

Received January 15, 2013

Accepted January 30, 2013

Back to Top | Article Outline

Introduction

Thyroid nodules are a common clinical presentation. They can be detected by palpation in 10% of women and 2% of men. The prevalence of thyroid nodules may approach 60% or more when ultrasonography is used. These figures would have been much higher in areas of iodine deficiency (Bongiovanni et al., 2012).

Thyroid nodules may be either a true solitary nodule or multinodular. By far, the multinodular goiter is the most common thyroid disease encountered in clinical practice (Perlmutter and Slater, 1956). Up to 10–15% of cases of multinodular goiter may harbor a neoplasm, but of these 90% are benign and 10% are malignant, resulting in an overall incidence of malignancy in multinodular goiter of 1–2% (Gharib et al., 2008).

In contrast, solitary thyroid nodule is far more likely to be a neoplasm; 90% of them were proved to be benign and 10% to be malignant. About 50% of the clinically apparent solitary thyroid nodules turn out to be dominant nodules of multinodular goiter by radiological studies (Rago et al., 2010). Solitary toxic nodules account for 4–7% of all solitary thyroid nodules. The percentage of malignancy is rare in hot nodules. The overall incidence of malignant hot nodules is 0.2% (Tareq et al., 2010).

It is crucial to have a clear diagnostic approach to ensure patients presenting with thyroid nodules are managed appropriately and are not overtreated or undertreated. In the management of thyroid nodules, the primary challenge is to separate benign nodules (the majority) from malignant lesions (the minority) (Rago et al., 2010).

In the assessment of thyroid nodules, clinical evaluation is very important. One of the key features of thyroid gland malignancy at palpation is the degree of firmness; malignant lesions tend to be much firmer or harder than benign ones. However, this clinical parameter is highly subjective and dependent on the experience of the examiner. Thus, it may be misleading. Some papillary carcinomas are cystic and soft. Follicular carcinomas can be hemorrhagic and soft, and dystrophic calcification may render the benign nodules hard (Lyshchik et al., 2005).

Fine-needle aspiration cytology (FNAC) is the best single preoperative test used for differentiating malignant from benign thyroid nodules. FNAC has a sensitivity ranging from 65 to 98% and specificity ranging from 72 to 100%. Hence, it distinguishes between patients who require clinical management or surgical excision and assists in deciding the appropriate surgical procedure when necessary (Eng et al., 2010). Cytology has certain limitations that restrict its use as a single diagnostic test in case of thyroid nodules. The major limitation of FNAC is that 10–20% of specimens are classified cytologically as indeterminate – that is, it is unable to differentiate hypercellular follicular adenoma from well-differentiated follicular carcinoma. This distinction is based on capsular or vascular invasion in histopathologic sections rather than on cytologic criteria. Thus, a substantial proportion of nodules are not correctly diagnosed even at the hands of an experienced cytopathologist, and these nodules still need to be excised for histolopathological final diagnoses (Bonzanini et al., 2011).

During the last few years, a novel dynamic technology called ultrasound elastography (USE) has been added to the diagnostic investigation. It is used in combination with ultrasound and is based on the principle that the softer parts of tissues deform easier than the harder parts under compression, thus allowing a semiquantitative determination of tissue elasticity. This technology involves placing the probe and applying a uniform and slight pressure on the neck. A box, covering the target nodule and the immediate surrounding normal tissue, is highlighted on the scan image. Two ultrasound images, before and after tissue compression, are obtained. Dedicated software tracks the tissue displacement and displays this using a color scale ranging from red (highest elasticity corresponding to the lowest risk of malignancy) to green (intermediate elasticity) to blue (lowest elasticity corresponding to the highest risk of malignancy). The results of this technique are scored by measuring the degree of distortion of the ultrasound (US) beam under the application of an external force. USE has the advantage of being noninvasive and gives immediate information (Rago et al., 2007). This technique has been extensively studied in breast, prostate, pancreas, and lymph node lesions. Recent studies have evaluated the use of USE for detecting malignant thyroid nodules (Merino et al., 2011).

The aim of this study was to analyze whether real-time USE (ultrasound strain imaging) improves the diagnostic accuracy of FNAC in predicting malignancy in thyroid nodules to decrease the number of surgical operations performed for benign nodules. The statistical analysis was based on the histopathological examination obtained after surgery (the gold standard).

Back to Top | Article Outline

Patients and methods

This prospective study included 96 patients with painless thyroid nodules, either solitary nodules or dominant nodules, in the sitting of multinodular goiter who presented themselves at the Radiology Department, National cancer Institute, Cairo University, from January 2011 to July 2012.

Conventional ultrasound was performed by using a digital electronic scanner with a frequency of 7.5–13 MHz (EUB-7500; Hitachi Medical Corporation, Tokyo, Japan) to obtain B-mode. During B-mode ultrasound, thyroid gland lesions were identified and a region of interest for elastography was identified. After B-mode ultrasound, real-time strain imaging in the elasticity imaging mode, which was implemented on the same ultrasound scanner and probe modified for research, was performed for each region of interest selected during B-mode ultrasound.

The ultrasound probe was placed on the fully extended neck and the usual ultrasonographic study was carried out. Then the examined nodule was pressed with a stable light pressure, and a box was highlighted by the operator that included the nodule to be evaluated as well as sufficient surrounding thyroid tissue. The principle of USE is to acquire two ultrasonic images (before and after tissue compression by the probe) and track tissue displacement by assessing the propagation of the imaging beam. Dedicated software able to provide an accurate measurement of tissue distortion was used. The elasticity software includes a scale for pressure monitoring. Optimally, to obtain images that were appropriate for analysis, we applied the probe with only light standardized pressure ranging from 3 to 4 (Rago et al., 2007). We avoided using higher levels of pressure as the pattern of the elasticity image started to change markedly as the pressure increased. It is important that the level of pressure is maintained constant throughout the examination.

B-mode images and elasticity images were displayed simultaneously side-by-side on the same screen to aid in lesion identification.

For each lesion examined, multiple frames of elasticity images, which were acquired during compression and relaxation by continuously moving the probe, were stored in the cine-buffer memory of the scanner. We then selected the best-fit B-mode sonogram–elastogram image pairs for examination.

The elasticity image was matched with an elasticity color scale displayed on the computer monitor. The scale of the elasticity images ranged from red for components with the greatest strain (i.e. softest components) to blue for those with no strain (i.e. hardest components). Green indicated average strain. Color images were constructed automatically with the same image processing settings throughout the study.

To classify elasticity images, we evaluated the color pattern in the nodule. On the basis of the overall pattern, each image was given a score of five elasticity scores using the Ueno and Ito elasticity score (Ueno and Ito, 2004): Score 1: the nodule is relatively homogenous and colored green (indicating elasticity in the whole nodule); score 2: the nodule is colored a mixture of green and blue (indicating elasticity in large parts in the nodule); score 3: the center of the nodule is colored blue and its periphery is colored green (indicating elasticity only at the peripheral part of the nodule with sparing of the center, which is hard); score 4: the entire nodule is hard and displayed in blue (indicating no elasticity in the nodule); and score 5: no elasticity in the nodule and in the surrounding tissue, which are displayed blue. This technique is easy to perform and requires no more than 3–5 min.

The patients were referred to the Cytology Unit, Pathology Department, National cancer Institute, Cairo University, for cytopathological diagnosis (as occurs in real clinical practice). FNAC was performed using 22–23-G disposable needles fitted to 5 ml plastic syringes after correct positioning of the patient and cleaning of skin with antiseptic solution. At least two passes of the thyroid nodule should be performed to achieve a detailed cytologic picture of the nodule. The aspirated contents of the needle were expelled on to clean glass slides. At least four slides were made for each case and were immediately fixed in 95% ethyl alcohol at room temperature for about 15 min. The slides were stained with the modified Papanicolaou stain (Gill et al., 1974) and examined with a light microscope by the cytopathologists. The adequacy of aspirates was defined according to the guidelines of the Papanicolaou Society (Suen, 1996).

Cytological specimens were evaluated blindly (without knowing the scores of USE) and classified according to the recent Bethesda classification (Theoharis et al., 2009) as follows: insufficient for diagnosis (these cases were excluded from the beginning of the work); benign; follicular lesion of undetermined significance; follicular neoplasm; suspicious for malignancy; and malignant sampling.

All the studied patients underwent surgery, either because of compressive symptoms, cancer phobia due to family history of malignancy, or because of malignancy or suspicion of malignancy on FNAC and/or USE. Patients with follicular neoplasm (indeterminate cyologic results) also underwent surgery, as the diagnostic accuracy of these thyroid lesions cannot be relied solely on cytologic criteria. All these cases were evaluated for final histopathological diagnoses. The final diagnosis was based on the histopathologic examination of an excised thyroid nodule.

All patients gave their written informed consent to participate in this study. Patients not willing to undergo elastography, FNAC, or were unfit for surgery, as well as cases with insufficient cytological specimens, were excluded from the study. No patient included in this series had a history of neck irradiation, previous diagnosis of thyroid malignancy, previous thyroid surgery, or toxic thyroid nodule confirmed by clinical evaluation.

The results of elastography, FNAC, and histopathology were statistically compared. Histopathology was taken as the gold standard. Calculation of sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy was done for both FNAC and USE.

Statistically, fine-needle aspiration specimens interpreted as benign or follicular lesion of undetermined significance were considered to be negative samples as both have nearly the same clinical impact. The false-negative cases included those diagnosed as negative on FNAC but confirmed as malignant upon surgical excision. The remaining categories – follicular neoplasm (indeterminate), suspicious for malignancy, and malignant – were considered to be positive samples because they led to a recommendation for surgery. The false-positive category included those cases that were diagnosed as positive on FNAC and confirmed to be benign on histopathology. The use of the term ‘positive’ is for statistical purposes only and does not indicate malignancy.

Statistically, USE scores 1 and 2, describing high elasticity, were strongly predictive of benignity with no need for surgical intervention (negative samples), whereas scores 3, 4, and 5, describing nodules with low elasticity, were strongly predictive of malignancy (positive samples).

Back to Top | Article Outline

Results

This study included 96 patients with painless thyroid nodules. A total of 59 patients (61.5%) had a solitary thyroid nodule and 37 patients (38.5%) had a dominant nodule in the setting of multinodular goiter.

Eighteen patients (18.8%) were male and 78 (81.2%) were female with a male-to-female ratio of about 1 : 4.3. Their ages ranged from 15 to 66 years with a mean age and SD of 37.64±13.19 years. All patients had normal serum levels of thyroid hormones.

Among the 96 studied cases, 35 (36.5%) had a final histopathological diagnosis of malignancy: 31 were papillary carcinoma, three were invasive follicular carcinoma, and one was medullary carcinoma. The remaining 61 cases (63.5%) had a benign diagnosis on histopathological examination: 39 were hyperplastic nodules (nodular goiter) and 22 were follicular adenoma.

FNAC results were interpreted as benign in 34 cases (35.4%), follicular lesion of undetermined significance in 13 cases (13.5%), follicular neoplasm in 18 cases (18.8%), suspicious for malignancy in six cases (6.3%), and positive for malignancy in 25 cases (26.0%) (Table 1).

Table 1

Table 1

In thyroid nodules interpreted cytologically as benign, the cytological features revealed abundant colloid in the background with follicular cells arranged predominantly in monolayer sheets and microfollicles (Fig. 1). Degenerative changes and stromal fragments were also detected. When these cases were compared with their corresponding histopathological diagnosis, 32 cases (94.1%) were confirmed as nodular goiter, one case (2.9%) was follicular adenoma, and one case (2.9%) was papillary carcinoma (Table 1). The papillary carcinoma case was 1.5 cm in diameter and cystic, yielding 4 ml of dark brown liquid with near total collapse of the clinically felt nodule. The case showed monolayered sheets of enlarged follicular cells and multiple hemosidrin-laden macrophages. The case demonstrated no cytologic evidence of papillary carcinoma on retrospective review of cytological material. The risk of malignancy in the studied benign cytological diagnosis was 2.9%.

Fig. 1

Fig. 1

The 13 cases that were reported cytologically as follicular lesion of undetermined significance showed borderline cellularity with a predominance of microfollicles and moderate to scant amount of colloid in the background. Ten of these 13 cases (76.9%) were diagnosed histopathologically as follicular adenoma and three cases (23.1%) were nodular goiter (Table 1). Therefore, this category at our institution has no risk of being malignant (0%).

The 18 thyroid smears that were interpreted on FNAC as follicular neoplasm revealed marked cellularity with predominantly microfollicular architectures, tridimensional clusters with cellular crowding, single cells and scant to absent colloid in the background (Fig. 2). Nuclear atypia was observed in four cases. Eleven cases (61.1%) were diagnosed histopathologically as follicular adenoma, three cases (16.7%) as nodular goiter, three cases (16.7%) as invasive follicular carcinoma, and one case (5.6%) as follicular variant of papillary carcinoma (Table 1). A review of the papillary carcinoma smear revealed that the nuclear features of papillary carcinoma are subtle and focal, not widespread throughout the follicular cell population, and hence it was difficult to be differentiated from a follicular neoplasm during interpretation. The lesion may be incompletely sampled and thus may yield only a small number of abnormal cells. The risk of malignancy in the examined follicular neoplasm was 22.2%.

Fig. 2

Fig. 2

Five of the six smears (83.3%) that had cytologic features suspicious for malignant neoplasm were confirmed after resection as being malignant, papillary carcinoma in four cases (66.7%) and medullary carcinoma (Fig. 3) in one case (16.7%). The remaining case (16.7%) was proved histopathologically as a papillary hyperplastic nodule. The aspirate of the papillary hyperplastic nodular case showed some suspicious findings such as hypercellularity with no colloid in the background and showed predominantly papillae with peripheral columnar cells, trabeculae, large cells with hyperchromatic (smudged) chromatin and irregular nuclear membrane, and focal Hurthle cell changes (Fig. 4). However, it also revealed scattered hemosidrin-laden macrophages and multinucleated giant cells. The smear of medullary carcinoma revealed predominantly dispersed cells with moderate to abundant cytoplasm, nuclear hyperchromasia, and variation in nuclear size. The case reported as suspicious for malignancy with two possibilities of being Hurthle cell neoplasm or medullary carcinoma. The risk for malignancy in suspicious cytological diagnosis was 8.3%.

Fig. 3

Fig. 3

Fig. 4

Fig. 4

The 25 cases (100%) that were interpreted cytologically as papillary carcinoma were confirmed on histopathological examination (Table 1, Fig. 5). The risk for malignancy in malignant cytological diagnosis was 100%.

Fig. 5

Fig. 5

Of the 47 cases cytologically diagnosed as negative, 46 (97.9%) were proved to be benign on histopathology (true negative) and one case (2.1%) was papillary carcinoma (false negative). Of the remaining 49 cytologically positive samples, 34 (69.4%) were proved to be malignant (true positive) and 15 (30.6%) were diagnosed as benign on histopathological reports (false positive) (Table 2).

Table 2

Table 2

Thus, FNAC achieved a sensitivity of 97.1%, specificity of 75.4%, PPV of 69.4%, NPV of 97.9%, and total diagnostic accuracy of 83.3%. The false-negative rate and false-positive rate were 2.9 and 24.6%, respectively.

Forty-six of 61 cases (75.4%) with a final histopathological diagnosis of being a benign nodule had a negative cytological diagnosis, whereas 34 of 35 cases (97.1%) with a final histopathological diagnosis of carcinoma had a positive cytological diagnosis (P<0.005).

Among the 39 cases with histopathological diagnosis of nodular goiter, 35 (89.7%) were correctly diagnosed by cytology as being negative. The last four cases were overdiagnosed by cytology as being positive – follicular neoplasm in three cases (7.7%) and suspicious for malignancy in one case (2.6%) – achieving 89.7% diagnostic accuracy. With regard to the 22 follicular adenoma cases, 11 (50%) revealed negative cytological result and 11 cases (50%) showed positive results, achieving 50% diagnostic accuracy. Of the 31 nodules with a final histopathological diagnosis of papillary carcinoma, 30 cases (96.8%) were positive on cytology and one case (3.2%) was negative, achieving 96.8% diagnostic accuracy. All the three nodules (100%) with a final diagnosis of invasive follicular carcinoma gave positive cytological reports. The only medullary carcinoma case was positive on cytology (Table 1).

On USE, score 1 was found in six cases (6.3%), score 2 in 51 cases (53.1%), score 3 in five cases (5.2%), score 4 in 15 cases (15.6%), and score 5 in 19 cases (19.8%) (Table 3).

Table 3

Table 3

The results of USE were then compared with their corresponding histopathological diagnosis. All cases with a USE score of 1 or 2 (100%) were confirmed as benign on histopathological diagnosis (Figs 6 and 7). Two of the five cases (40%) with score 3 were malignant on histopathology (Fig. 8), whereas three cases (60%) were benign. Among the 15 cases with a score of 4, 14 (93.3%) were malignant and the remaining one case (6.7%) was benign on final histopathology. This case showed incomplete rim-like calcification on conventional ultrasonography. All cases with USE score 5 (100%) were malignant (Table 3).

Fig. 6

Fig. 6

Fig. 7

Fig. 7

Fig. 8

Fig. 8

All the 57 included cases with an elastography score of 1 or 2 (negative) were proved to be benign on histopathology (true negative) and there were no false-negative cases. The remaining 39 cases were referred for surgery because they had scores of 3, 4, or 5. Of these, 35 cases (89.7%) were proved to be malignant on histopathological reports (true positive) and four cases (10.3%) were diagnosed as benign (false positive) (Table 4).

Table 4

Table 4

Thus, USE achieved a sensitivity of 100%, specificity of 93.4%, PPV of 89.7%, NPV of 100%, and total accuracy of 95.8%. The false-negative rate and false-positive rate were 0 and 6.6%, respectively.

Fifty-seven of 61 cases (93.4%) with a final histopathological diagnosis of benign nodule had USE scores of 1 and 2, whereas all cases (100%) with a final histopathological diagnosis of malignant nodule had scores 3, 4, or 5 (P<0.0001).

The 39 cases with histopathological diagnosis of nodular goiter correctly showed a negative score (1 and 2), achieving 100% diagnostic accuracy. Among the 22 follicular adenoma cases, 18 were correctly scored 1 or 2, whereas four cases were overscored as 3 or 4, achieving 81.8% diagnostic accuracy. All the 31 nodules with a final hiatopathological diagnosis of papillary carcinoma were correctly scored as 5, 4, or 3, achieving 100% diagnostic accuracy. The three nodules with a final diagnosis of invasive follicular carcinoma were correctly scored as 4 or 5. The only medullary carcinoma case had a score of 5 (Table 3).

The USE scores were compared with the results of FNAC. In the 18 nodules with cytological report of follicular neoplasm (indeterminate cytology), score 2, describing high elasticity, was found in 10 of 14 benign nodules on histology and in no malignant nodule and scores 3, 4, and 5, describing low elasticity, were found in the four malignant nodules and in four of the 14 benign nodules (P<0.0005). The elastography scoring was efficient in differentiating benign from malignant nodules in indeterminate cytological results.

For the 31 nodules with suspicious or malignant cytological reports, score 2, describing benignity, was found in one case, which was nodular goiter (papillary hyperplastic nodule) on histopathology, whereas scores 3, 4, and 5, describing malignancy, were scored in the remaining 30 nodules, which were malignant on histopathology (P<0.0005). The elastography scoring was highly efficient in differentiating benign from malignant nodules in suspicious or malignant cytological results. Among the 47 cases with benign cytological reports, 46 nodules had scores 1 and 2 and the only case that was falsely diagnosed on cytology as benign was correctly scored 4 on elastography (P<0.0001). This case was papillary carcinoma on histopathology. The elastography scoring was highly efficient in differentiating benign from malignant nodules in benign cytological results (Table 5).

Table 5

Table 5

Back to Top | Article Outline

Discussion

The utility of any diagnostic procedure mainly depends on three foundations: sensitivity, specificity, and practical success (i.e. ease in carrying out the procedure). A major advancement in the diagnosis of thyroid nodules has been achieved with the perfection and common use of FNAC, which can obviate a lot of unnecessary surgeries in thyroid lesions (Tareq et al., 2010). Before the routine use of thyroid FNA, the percentage of surgically resected thyroid nodules that were malignant was 14%. With current thyroid FNA practice, the percentage of resected nodules that are malignant surpasses 50% (Yassa et al., 2007). Treatment is decided on the basis of whether the cytologic results are benign or malignant, or at least suspicious for malignancy. However, when the cytological diagnosis is reported as atypical or indeterminate, clinicians could be confronted with some difficulties in the management of patients. Thus, the recommended treatment of these lesions remains the diagnostic surgical excision of the nodule (Chung et al., 2011).

This study was undertaken to specifically address whether USE improves the diagnostic accuracy of FNAC in patients with thyroid nodules.

In the statistical estimation of our cytological results, we considered the new category ‘follicular lesion of undetermined significance’ as negative samples because of many reasons; first, the malignancy risk of this category (∼5–15%) is not sufficient to justify immediate surgery, and the recommended treatment is repeated FNAs in the appropriate clinical context within a span of 3–6 months (Ohori et al., 2010). The second reason can be attributed to the subjective cytomorphological diagnostic features of this category. It was reported in the literature that this category may never have good interobserver reproducibility, even after pathologists familiarize themselves with the criteria in the Bethesda atlas (Crippa Mazzucchelli, 2010). Chung et al. (2011) reported that, although the Bethesda system systematized the pre-existing cytological knowledge, the clarification of criteria for this category is worthy of special attention. Some authors reported that a presumptive institutional philosophy has arisen with regard to this diagnosis (Theoharis et al., 2009). Some authors in their series re-evaluated these cases to reclassify them into other groups. Thus, no patient from their final series was included in this category (Merino et al., 2011). Third, in a previous work performed at our institute in 2012, no cases of this category turned out to malignant on histopathological follow-up (Sinna and Ezzat, 2012). This result is comparable to ours. In our unit, there is a tendency to use this category when the smears are suggestive but not conclusive for benignity, and any observed atypical feature in our cases was classified under follicular neoplasm or suspicious categories.

In our series, analysis of data revealed that FANC achieved a sensitivity of 97.1%, specificity of 75.4%, PPV of 69.4%, and NPV of 97.9%. These results were in agreement with the published data in which FNAC of the thyroid is reported to have a sensitivity range from 65 to 98%, a specificity from 72 to 100%, a PPV from 34 to 100%, and an NPV from 83 to 100% (Čáp et al., 1999). The explained factors for such wide range of differences may be related to: (a) the number of studied cases; (b) clinical presentation of the patients in terms of having multiple or solitary thyroid nodules; and (c) the included diagnostic categories and how the cytopathologists classify follicular lesions of undetermined significance. Some authors classify it as a neoplastic suspicious lesion as its risk for malignancy is 5–15%, whereas some do not as its usual management does not include surgery, and others exclude it from the statistics.

The specificity in our study was low, probably secondary to a relatively high number of indeterminately classified nodules, follicular neoplasms, in which determining malignancy or benignity was not possible with cytology, and all these cases were considered positive and subjected to diagnostic surgery.

In the current study, the overall accuracy of cytological diagnosis was 83.3%. Our result was much lower than that reported in a study carried out at our institute in 2012 (Sinna and Ezzat, 2012) in which the reported accuracy was 93.6%. The determinant factors for such a difference may be related to exclusion of the follicular lesion of undetermined significance from their statistical calculation or because follicular adenoma cases were considered among their positive results. Our diagnostic accuracy was higher than the reported figure of 79.1 and 68.8% in the studies by Mundasad et al. (2005) and Bongiovanni et al. (2012). In the former series, no follicular lesions of undetermined significance were included, and in the latter this category was considered true positive.

In the current series, there is a strong agreement between cytology and histopathological results (P<0.005). This is comparable to that of others (Theoharis et al., 2009; Tareq et al., 2010).

However, a discrepancy between cytological report and histopathological diagnosis was found in 16 of 96 cases (16.7%) (Table 1). One case was initially diagnosed as benign on cytology and turned out to be papillary carcinoma in the histopathological study (the false-negative rate was 2.9%). This may result from sampling error or coexistence of benign and malignant lesions. This case exhibited cystic properties. This may be one explanation for the poor test performance, as these properties can make diagnosis of malignancy difficult and it could explain the only missing case in the current study (Mistry et al., 2011). On secondary review, the cytology aspirate of this case demonstrated no evidence of papillary carcinoma. Multiple aspirations from different parts of the swelling or combined use of ultrasound and FNAC would aid in sampling the solid portions of cystic nodules (Chung et al., 2011). This case is of much importance as it refers to the potential of missing malignant lesions. Typically, the reported rates of false-negative cytologic diagnoses reported in the literature ranged from 1 to 16% (Nggada et al., 2006). Our result was lower than that reported by Sinna and Ezzat (2012), in which the false-negative rate was 5.8%.

In contrast, we reported one case as suspicious for malignancy on FNAC (Fig. 4) that proved to be papillary hyperplastic nodule on histopathology, and 14 of 18 cases that were reported as follicular neoplasm proved to be benign, nodular goiter in three cases and follicular adenoma in 11 cases on histological correlation (false-positive rate was 24.6%). These cases are of great concern because they underwent unnecessary surgery. The papillary hyperplastic nodule has been recently recognized as a potential pitfall in the cytologic diagnosis of papillary carcinoma (Khayyata et al., 2008). Aspiration from the three nodular goiter cases was probably carried out from the hypercellular and hyperplastic areas of nodules, which led to overdiagnosis. Cytological distinction between hyperplastic nodules and follicular neoplasm is sometimes very difficult. A possible remedy is multiple aspirations from different parts of the swelling that could demonstrate hypocellular, polymorphic, and colloid-rich areas. Demonstration of monolayered sheets of epithelial cells representing macrofollicles and degenerative changes would suggest the possibility of non-neoplastic lesions (Sinna and Ezzat, 2012). It is well recognized that certain thyroid lesions have similar cytological features that make diagnosis extremely difficult on cytology. This problem is commonly reported with follicular adenoma and carcinoma and is considered the main pitfalls of thyroid FNAC. The diagnosis of follicular and Hurthle cell carcinoma requires demonstration of capsular and/or vascular invasion, which cannot be evaluated on cytology (Mistry et al., 2011). This pitfall is responsible for most of our incorrect false-positive cytologic diagnoses. The published guidelines of the Papanicolaou Society of Cytopathology (Suen, 1996) suggested that a false-negative and a false-positive rate of less than 2 and 3%, respectively, should be achieved. In other series, the false-positive rate ranged from 0 to 8% (Basharat et al., 2011).

USE is one of the latest technologies that use ultrasound to analyze the stiffness of a nodule by measuring the amount of distortion that occurs when the nodule is subjected to external pressure (Hong et al., 2009). Many previous studies have proved that elastography is useful in differentiating malignant thyroid nodules from benign thyroid nodules (Asteria et al., 2008; Hong et al., 2009). Others have reported that elastography showed inferior performance in comparison with gray-scale ultrasound features (Moon et al., 2012).

In our study, tissue stiffness was scored from 1 to 5 on the basis of subjective analysis of the elastograms. Using this score classification, the predictivity of USE was highly rewarding. Scores 3, 4, and 5 identified 100% of thyroid cancers (Fig. 8) and scores 1 and 2 excluded malignancy (Figs 6 and 7) with a sensitivity, specificity, PPV, and NPV of 100, 93.4, 89, and 100%, respectively. Sensitivity was very high in this study with no false negatives found; hence, all nodules with high elasticity (score 1 and 2) in our study were benign on pathological results. Our results are comparable to another similar study in which scores 1 and 2 were detected in 49 cases; all were benign. Scores 4 and 5 were detected in 30 cases, and all were malignant. Score 3 was recorded in 13 cases: one was a carcinoma and 12 were benign. They considered score 3 as negative with an overall sensitivity of 97%, specificity of 100%, PPV of 100%, and NPV of 98% (Rago et al., 2007).

Most publications referring to USE in the evaluation of thyroid nodules revealed sensitivity ranges from 82 to 100%, specificity from 81 to 100%, PPV from 55 to 100%, and NPV from 93 to 100%. The explanations for these ranges were: (a) different machines used, which were either based on gray-scale patterns or color-scale patterns; (b) different score classifications of tissue stiffness – either four-point scores (Friedrich-Rust et al., 2010), five-point scores (Rago et al., 2007; Rago and Vitti, 2009), or six-point scores (Hong et al., 2009); (c) whether score 3, the indeterminate score, is considered negative or positive (Rago and Vitti, 2009); (d) the sample size; (e) the method of case selection.

In the current study, the overall diagnostic accuracy of USE was 95.8%, which agrees with the 94.3% reported by others (Fukunari, 2007). Our result was higher than that reported in other studies (Asteria et al., 2008; Rubaltelli et al., 2009).

However, in the group of nodules that scored 3, 4, and 5, four cases (10.3%) were false positive (score 3 in three cases and score 4 in one case). The score 4 case showed incomplete rim-like calcification on conventional ultrasonography. It is reported in the literature that the USE is not valuable in the presence of calcific shell or coarse calcifications because the US beam does not cross the calcifications, and no tissue strain is obtained by the probe pressure. It is reported that conventional US maintains a pivotal importance in defining which nodules are suitable for USE scoring (Rago et al., 2007). Thus, if this case had been excluded from the beginning of the work, the incorrect scoring would have been minimized. The remaining three benign nodules revealed microfollicular pattern with crowdening of the cells on histopathological examination. Fukunari (2007) reported that the blue zone recorded at the periphery of his studied follicular tumor in the elastography images more or less matches with the hypercellularity of the small irregular follicular structure formed under the capsules on pathological examination. Another previous report concluded that USE is not suitable for the diagnosis of follicular carcinoma, and four of nine follicular carcinomas in the meta-analysis published by Bojunga et al. (2010) were missed on USE. We had three cases of follicular carcinoma in our series that were correctly scored. However, statistically, these nine and three cases are considered unrepresentative.

On comparing the cytological results with the elastography scores, the sensitivity, specificity, PPV, NPV, and accuracy of USE are higher than that of FNAC. To our knowledge, only few reports had studied the relation between the thyroid cytology and elastography score (Rago and Vitti, 2009; Merino et al., 2011). In these works, FNAC was used as the reference standard for the diagnosis of benign nodules, but histopathological evaluations were performed when results suspicious for malignancy or malignant results were obtained on FNAC as well as in indeterminate lesions. Some works were focused mainly on the relation of elastography and indeterminate and nondiagnostic cytological results with the usage of histopathological diagnosis as a reference (Rago et al., 2010).

In the series of 103 patients who have been studied with elastography (gray-scale patterns) for the evaluation of 106 thyroid nodules, all patients included had been referred for the FNA procedure. They concluded that there was a significant statistical association between elasticity score and cytological results, and malignant nodules could be excluded by elastography (Merino et al., 2011). Some authors published an analysis of elastography for the differentiation of benign from malignant thyroid nodules, concluding that USE can be used with high sensitivity and might be a useful method in addition to or even instead of FNAC to select patients for surgery (Bojunga et al., 2010). However others reported that the use of USE alone could miss many malignant nodules. They recommended combined use of BRAF mutation analysis in FNAC and USE (Burger, 2012). Others concluded that USE can be employed in selecting thyroid nodules for FNAC as the clinicians rely more on the FNA procedure because it is easy, simple, acceptable to the patients, gives documented reports, and the only equipment needed is an ordinary plastic syringe. In their series, score 4 or 5 was found in 86.5% of the studied thyroid malignancy and in only 3% of benign nodules (Gietka-Czernel et al., 2010).

In the current study, it was found that, in the case that was incorrectly diagnosed as benign on cytology but turned out to be papillary carcinoma on histopathological study, the elastography examination revealed a correct score (score 4). The case was predominantly cystic. Our elastography result is inconsistent with some publications that concluded that USE of thyroid nodules with cystic components is not reliable and these nodules should be excluded from the study (Rago et al., 2007; Bhatia et al., 2011). The explanation is that the main determinant of nodule stiffness is the fluid content and not the solid areas. The second cytologically incorrect suspicious case, which was diagnosed as papillary hyperplastic nodule on histopathology, was correctly scored on elastographic examination (score 2). It was reported that adenomatous nodules are correctly scored on elastography and their findings differ from that of malignancy (Fukunari, 2007).

Among the 18 cases with a fine-needle aspiration result of follicular neoplasm, 14 had a benign histopathological result and four had malignancy. Elastography scores 3, 4, and 5 were observed in 100% of the four malignant cases and elastography score 2 in 10 of 14 (71.4%) cases who had benign lesions (Table 5). Thus, among the 18 cytological diagnoses of follicular neoplasm (which should be referred for surgical diagnosis), USE may restrict the indications of surgical interference in 10 (55.6%), whereas four cases (22.2%) would be still unnecessarily referred for surgery and four cases (22.2%) would be necessarily referred for surgery. In the series of Rago et al. (2010) on 142 nodules with indeterminate cytological results, only one of 103 nodules classified as negative on elastography had malignancy on histopathological correlation, whereas nine out of 39 nodules with positive elastography were benign. They concluded that USE is an important tool in the presurgical stratification of thyroid cancer in nodules with indeterminate cytology. In another study on 32 nodules with indeterminate cytology, 25 had a benign follicular adenoma on histology and seven had carcinoma. Positive scores were observed in six of seven (86%) patients with carcinoma and negative scores in all 25 patients with benign lesions (Rago et al., 2007). They reported that USE seems to have great potential for the diagnosis of thyroid cancer, especially with indeterminate cytology. However, they recommended larger prospective studies to confirm their results.

In contrast, previous reports concluded that USE, similar to FNAC, is not suitable for the diagnosis of follicular carcinoma (Merino et al., 2011). A lower diagnostic accuracy was reported for the detection of follicular thyroid carcinoma by elastography (Asteria et al., 2008; Dighe et al., 2008).

Back to Top | Article Outline

Conclusion

FNAC is safe, easy, fast, and highly useful in differentiating benign from malignant thyroid nodules with 97.1% sensitivity, 75.4% specificity, and 83.3% diagnostic accuracy. The cytodiagnostic error of some cases can be minimized or avoided by giving greater consideration to the inherent diagnostic difficulties associated with cytological interpretation.

USE is a promising imaging technique that can assist in the differential diagnosis of thyroid nodules with higher sensitivity, specificity, PPV, NPV, and diagnostic accuracy compared with FNAC. USE is helpful in the identification of nodules that should be left alone and for which follow-up would be sufficient and has led to proper planning of surgery in malignant cases. Therefore, it is recommended in the workup of all thyroid nodules in association with FNAC as a multidisciplinary approach that can ensure the best results.

Back to Top | Article Outline

Limitation

There are some limitations in this study that need to be addressed. The studied cases were selected to be candidates for thyroid surgery, and in principle this could have influenced the sample size. Future studies based on a totally blinded selection will be necessary to provide conclusive evidence on the role of USE in the management of thyroid nodules. Another limitation is that only three follicular carcinoma cases and one medullary carcinoma case were included in the study group. Statistically, the numbers are considered unrepresentative. Further studies are required to evaluate the USE score for different types of thyroid cancer with a special interest in follicular carcinoma.

Back to Top | Article Outline

Acknowledgements

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline

References

Asteria C, Giovanardi A, Pizzocaro A, Cozzaglio L, Morabito A, Somalvico F, Zoppo A.US-elastography in the differential diagnosis of benign and malignant thyroid nodules.Thyroid2008;18:523–531.
Basharat R, Bukhari MH, Saeed S, Hamid T.Comparison of fine needle aspiration cytology and thyroid scan in solitary thyroid nodule.Patholog Res Int2011;2011:754041.
Bhatia KSS, Rasalkar DP, Lee YP, Wong KT, King AD, Yuen HY, Ahuja AT.Cystic change in thyroid nodules: a confounding factor for real-time qualitative thyroid ultrasound elastography.Clin Radiol2011;66:799–807.
Bojunga J, Herrmann E, Meyer G, Weber S, Zeuzem S, Friedrich-Rust M.Real-time elastography for the differentiation of benign and malignant thyroid nodules: a meta-analysis.Thyroid2010;20:1145–1150.
Bongiovanni M, Spitale A, Faquin WC, Mazzucchelli L, Baloch ZW.The Bethesda system for reporting thyroid cytopathology: a meta-analysis.Acta Cytol2012;56:333–339.
Bonzanini M, Amadori P, Morelli L, Fasanella S, Pertile R, Mattiuzzi A, et al..Subclassification of the ‘grey zone’ of thyroid cytology; a retrospective descriptive study with clinical, cytological, and histological correlation.J Thyroid Res2011;2011:251680.
Burger AG.Fine needle aspiration biopsy with BRAF analysis and elastography are slightly more efficient in diagnosing papillary thyroid cancer than FNAB and thyroid ultrasound.Clin Thyroidol2012;24:6–7.
Čáp J, Ryška A, Řehořková P, Hovorková E, Kerekes Z, Pohnětalová D.Sensitivity and specificity of the fine needle aspiration biopsy of the thyroid: clinical point of view.Clin Endocrinol (Oxf)1999;51:509–515.
Chung YS, Yoo C, Jung JH, Choi HJ, Suh Y-J.Review of atypical cytology of thyroid nodule according to the Bethesda system and its beneficial effect in the surgical treatment of papillary carcinoma.J Korean Surg Soc2011;81:75–84.
Crippa S, Mazzucchelli L.The Bethesda system for reporting thyroid fine-needle aspiration specimens.Am J Clin Pathol2010;134:343–344.
Dighe M, Bae U, Richardson ML, Dubinsky TJ, Minoshima S, Kim Y.Differential diagnosis of thyroid nodules with US elastography using carotid artery pulsation.Radiology2008;248:662–669.
Eng CY, Quraishi MS, Bradley PJ.Management of thyroid nodules in adult patients.Head Neck Oncol2010;2:11.
Friedrich-Rust M, Sperber A, Holzer K, Diener J, Grünwald F, Badenhoop K, et al..Real-time elastography and contrast-enhanced ultrasound for the assessment of thyroid nodules.Exp Clin Endocrinol Diabetes2010;118:602–609.
Fukunari N.More accurate and sensitive diagnosis for thyroid tumor with elastography. Detection and differential diagnosis of thyroid.Medix Suppl200716–19Available at: http://www.hitachi-medical.co.jp/english/products/us/medix/.
Gharib H, Papini E, Paschke R.Thyroid nodules: a review of current guidelines, practices, and prospects.Eur J Endocrinol2008;159:493–505.
Gietka-Czernel M, Kochman M, Bujalska K, Stachlewska-Nasfeter E, Zgliczyński W.Real-time ultrasound elastography – a new tool for diagnosing thyroid nodules.Endokrynol Pol2010;61:652–657.
Gill GW, Frost JK, Miller KA.A new formula for a half oxidized hematoxylin solution that neither overstains nor requires differentiation.Acta Cytol1974;18:300–311.
Hong Y, Liu X, Li Z, Zhang X, Chen M, Luo Z.Real-time ultrasound elastography in the differential diagnosis of benign and malignant thyroid nodules.J Ultrasound Med2009;28:861–867.
Khayyata S, Barroeta JE, LiVolsi VA, Baloch ZW.Papillary hyperplastic nodule: pitfall in the cytopathologic diagnosis of papillary thyroid carcinoma.Endocr Pract2008;14:863–868.
Lyshchik A, Higashi T, Asato R, Tanaka S, Ito J, Mal JJ, et al..Thyroid gland tumor diagnosis at US elastography.Radiology2005;237:202–211.
Merino S, Arrazola J, Cá Rdenas A, Mendoza M, De Miguel P, Ferná Ndez C, Ganado T.Utility and interobserver agreement of ultrasound elastography in the detection of malignant thyroid nodules in clinical care.Am J Neuroradiol2011;32:2142–2148.
Mistry SG, Mani N, Murthy P.Investigating the value of fine needle aspiration cytology in thyroid cancer.J Cytol2011;28:185–190.
Moon HJ, Sung JM, Kim E-K, Yoon JH, Youk JH, Kwak JY.Diagnostic performance of gray-scale US and elastography in solid thyroid nodules.Radiology2012;262:1002–1013.
Mundasad B, Mcallister I, Carson J, Pyper PC.Accuracy of fine needle aspiration cytology in diagnosis of thyroid swellings.Internet J Endocrinol20052: Available at: http://ispub.com/IJEN/2/2/4183.
Nggada HA, Musa AB, Gali BM, Khalil MI.Fine needle aspiration cytology of thyroid nodule (s): a Nigeriantertiary hospital experience.Internet J Pathol20065: Available at: http://ispub.com/IJPA/5/1/10574.
Ohori NP, Nikiforova MN, Schoedel KE, LeBeau SO, Hodak SP, Seethala RR, et al..Contribution of molecular testing to thyroid fine-needle aspiration cytology of ‘follicular lesion of undetermined significance/atypia of undetermined significance’.Cancer Cytopathol2010;118:17–23.
Perlmutter M, Slater SL.Which nodular goiters should be removed; a physiologic plan for the diagnosis and treatment of nodular goiter.N Engl J Med1956;255:65–71.
Rago T, Santini F, Scutari M, Pinchera A, Vitti P.Elastography: new developments in ultrasound for predicting malignancy in thyroid nodules.J Clin Endocrinol Metab2007;92:2917–2922.
Rago T, Vitti P.Potential value of elastography in the diagnosis of malignancy in thyroid nodules.Q J Nucl Med Mol Imaging2009;53:455–463.
Rago T, Scutari M, Santini F, Loiacono V, Piaggi P, Di Coscio G, et al..Real-time elastosonography: useful tool for refining the presurgical diagnosis in thyroid nodules with indeterminate or nondiagnostic cytology.J Clin Endocrinol Metab2010;95:5274–5280.
Rubaltelli L, Corradin S, Dorigo A, Stabilito M, Tregnaghi A, Borsato S, Stramare R.Differential diagnosis of benign and malignant thyroid nodules at elastosonography.Ultraschall Med2009;30:175–179.
Sinna EA, Ezzat N.Diagnostic accuracy of fine needle aspiration cytology in thyroid lesions.J Egypt Natl Canc Inst2012;24:63–70.
Suen K.Guidelines of the Papanicolaou Society of Cytopathology for the examination of fine-needle aspiration specimens from thyroid nodules.Diagn Cytopathol1996;15:84–89.
Tareq M, Iqbal MZ, Ali MZ, Anwar M, Khan RS, Irum S.FNAC of the thyroid; diagnostic accuracy of fine needle aspiration cytology of the thyroid.Prof Med J2010;17:589–597.
Theoharis CGA, Schofield KM, Hammers L, Udelsman R, Chhieng DC.The Bethesda thyroid fine-needle aspiration classification system: year 1 at an academic institution.Thyroid2009;19:1215–1223.
Ueno E, Ito A.Diagnosis of breast cancer by elasticity imaging.Eizo Joho Med2004;36:2–6.
Yassa L, Cibas ES, Benson CB, Frates MC, Doubilet PM, Gawande AA, et al..Long-term assessment of a multidisciplinary approach to thyroid nodule diagnostic evaluation.Cancer2007;111:508–516.
©2013Egyptian Journal of Pathology