Effect of Thyroid Gland Volume in Preoperative Detection of Suspected Malignant Thyroid Nodules in a Multinodular Goiter

Hypothesis  The detection of suspected malignant thyroid nodules by ultrasonography is associated with thyroid gland volume and tumor size.

Design  Prospective clinical trial.

Setting  A tertiary referral center.

Patients  Three hundred sixty-five patients with a multinodular goiter with coexistent dominant nodules.

Main Outcome Measures  The correlation between thyroid gland volume and tumor size and the detection of suspected malignant thyroid nodules by ultrasonography. The incidence of thyroid carcinoma inside and outside the dominant nodules and thyroid gland volume and tumor size in suspected or unsuspected malignant thyroid nodules by ultrasonography were determined. Receiver operating characteristic analysis was used to identify the cutoffs of the tumor size and thyroid gland volume.

Results  One hundred thyroid carcinomas were found in 69 (18.9%) patients. Forty-one of these carcinomas were inside the dominant nodule, whereas 59 were outside the dominant nodule. Only 9 of the 59 thyroid carcinomas outside the dominant nodules were suspected of being malignant by ultrasonography. Thyroid gland volume less than 38 mL and tumor size larger than 7 mm had 48-fold (odds ratio, 48; P < .001) and 21.5-fold (odds ratio, 21.5; P < .001) increased rates, respectively, of detecting suspected malignant thyroid nodules by ultrasonography.

Conclusions  Thyroid gland volume and tumor size were significantly associated with detection of suspected malignant thyroid nodules by ultrasonography. Small thyroid gland volume was associated with detection of suspected malignant thyroid nodules in multinodular goiters.

Thyroid nodules occur frequently, especially in middle-aged to elderly women.^1,2 The prevalence of thyroid nodules rises from 5% using palpation alone to 30% to 50% using ultrasonography. However, only 5% of these thyroid nodules have been found to be malignant.^3-6 Ultrasonography is a useful tool for tumor evaluation because of its safety, noninvasiveness, lack of radiation, and effectiveness.^7-9 Using a high-resolution probe, it not only detects the presence, site, number, and size of thyroid nodules, but also clearly documents their characteristics. Ultrasonography has a high sensitivity for detecting nodules as small as 2 to 3 mm. Incidental thyroid nodules are an increasingly common finding because of the expanding use of high-resolution ultrasonography. However, it is important to differentiate malignant nodular lesions from benign nodules to avoid unnecessary thyroidectomies.^9,10 The microcalcification, hypoechogenicity, and intranodular vascularity detected by ultrasonography are generally accepted as the most reliable indicators of malignancy; however, the overall sensitivity is low.^8-10

Fine-needle aspiration biopsy (FNAB) is considered the most reliable test for the diagnosis of malignant nodular lesions.^11-15 Although most clinicians recommend FNAB for a single nodule,^16,17 this consensus does not include multinodular goiters (MNGs).^18-20 Puncturing all of the palpable nodules is not practical when the clinical suspicion of malignancy is low.^20-23 The problem for clinicians is selecting nodules for FNAB, because FNAB is not useful for differentiating an MNG with a malignant nodular lesion from a benign MNG.^23,24

The aims of this prospective clinical study were (1) to evaluate the utility of preoperative FNAB on the dominant nodule and nodules sonographically suspected of being malignant in the MNG and (2) to investigate whether thyroid gland volume and tumor size might affect the detection of suspected malignant nodular lesions using ultrasonography. To our knowledge, this is the first study performed to evaluate the utility of thyroid gland volume in detecting suspected malignant thyroid nodules in MNGs.

We performed a prospective clinical study in 402 consecutive patients undergoing an operation for an MNG at the Istanbul Medical Faculty between September 2005 and September 2006. This study included 365 patients with an MNG with coexistent dominant nodules. Thirty-seven patients were excluded owing to a history of head and neck irradiation; a family history of thyroid carcinoma; suspicious symptoms; or clinical findings that would suggest thyroid malignancy, such as a rapidly enlarging nodule, throat hoarseness, or palpable lymph nodes. None of the patients refused to participate. Of 365 patients, 336 (92.1%) had a euthyroid MNG and 29 (7.9%) had a toxic MNG. All patients had a single dominant nodule with multiple nodules. A dominant nodule was defined as a nodule larger than 1.5 cm in a multinodular thyroid gland. If 2 or more nodules were larger than 1.5 cm, the largest was considered the dominant nodule. If the thyroid carcinoma rose from the dominant nodule, we defined it as a carcinoma inside the dominant nodule. If the carcinoma was in another nodular lesion, it was defined as a carcinoma outside the dominant nodule. The indications for surgical treatment were a large goiter with compression effect (n = 268), hyperthyroidism (n = 29), or a suspicious or positive (malignant) result on FNAB (n = 44). All of our patients were from an iodine-inadequate geographic area and underwent physical examinations, thyroid function tests, and thyroid ultrasonography. The study design was reviewed and approved by our institutional ethical committee; informed consent was obtained for all patients.

Total thyroid gland volume was evaluated by ultrasonography. A radiologist (A.S.) experienced in sonography conducted the examinations using different sonographic equipment (SI 400, Siemens, Erlangen, Germany; Logic 7, General Electric, Milwaukee, Wisconsin; Sonoline Antares, Siemens) with high-frequency (13 MHz) linear probes. Volumetric measurements were made from remnant tissue in 3 perpendicular planes using axial and sagittal images and volume-calculation software available within the equipment. Thyroid volume was estimated through the following equation: volume = length × width × depth × 0.5233.^25,26

A single radiologist (A.S.) performed ultrasonography FNAB on all patients by using a broad-band linear transducer (VFX 13-5, Sonoline Antares, Siemens). Ultrasonography FNAB was routinely employed for dominant nodules whether or not they were palpated. Fine-needle aspiration biopsy was also performed when ultrasonography had suspicious findings (ie, a hypoechoic nodule with punctuate calcifications and/or irregular borders). Experienced endocrine cytopathologists performed cytologic examinations on the FNAB specimens. During the procedure, the patient was kept in the supine position with a slight hyperextension of the neck. Local anesthesia was routinely applied. After aspiration, samples were placed on slides and dried. One to 3 slides from each patient were stained with hematoxylin-eosin to confirm the presence of thyroid follicular cells. If the number of follicular cells was insufficient, the procedure was repeated.

The nodules were classified as benign (typical follicular cells), malignant (atypical follicular cells with malignant nuclear features), indeterminate (follicular neoplasms or suspicious for malignancy), or nondiagnostic (< 6 clusters of follicular cells visualized on ≥ 2 samples). Suspicious or indeterminate FNAB, suggesting follicular proliferation, was considered positive. All thyroidectomy specimens underwent histopathologic examination. On gross examination, all occult lesions were sampled. Tissue samples were embedded into paraffin blocks, from which ultrathin sections of 3 to 5 μm were obtained.

The number of true-positive, true-negative, false-positive, and false-negative results were calculated. The suspicious/malignant FNAB results were considered true-positives in cases in which histological examination revealed a malignancy, and they were considered false-positives when no malignancy was found. Benign FNAB results were considered true-negatives if the histological finding was benign, and false-negative for cases of histologically proven malignancy. The sensitivity, specificity, diagnostic accuracy, positive predictive values, and negative predictive values of ultrasonography FNABs were determined according to the following formulas:

• Positive Predictive Value (%) = (True-Positive/ [True-Positive + False-Positive]) × 100

• Negative Predictive Value (%) = (True-Negative/ [True-Negative + False-Negative]) × 100

• Sensitivity (%) = (True-Positive/[True-Positive + False-Negative]) × 100

• Specificity (%) = (True-Negative/[True-Negative + False-Positive]) × 100

• Accuracy (%) = ([True-Positive + True-Negative]/ [True-Positive + True-Negative + False-Positive + False-Negative]) × 100

Data were analyzed using SPSS, version 11.0 (SPSS Inc, Chicago, Illinois). Results were expressed as mean (SD). Comparisons of data were carried out using Mann-Whitney U and χ² tests. Receiver operating characteristic analysis was used to identify the cutoff values of the tumor size and thyroid gland volume. The Spearman test was used for correlation analyses. Results were considered statistically significant at P < .05 (2-tailed).

The median age of the patients was 49 years (range, 17-75 years), with a female to male ratio of 7.1:1 (n = 320:45). The mean dominant nodule size was 33.4 (6.7) mm (range, 15-80 mm). Total and near-total thyroidectomies were performed in all patients. The mean thyroid volume was 66.6 (46) mL (range, 10-365 mL).

Of 365 patients, 69 (18.9%) patients had thyroid carcinoma and 296 (81.1%) patients had benign MNGs. There were no significant differences in age or sex between patients with thyroid carcinoma and patients with MNG (48.4 [13.7] years, 58 women and 11 men vs 47.7 [12.6] years, 262 women and 34 men, respectively, P > .05). The mean dominant nodule size and thyroid volume in patients with thyroid carcinoma were significantly lower than in patients with benign pathology (22.19 [14] mm and 38.05 [18] mL vs 30.1 [12] mm and 73.3 [48] mL, respectively, P < .001).

Histopathologic examination revealed thyroid carcinoma in 69 (18.9%) patients, follicular adenoma in 10 (2.7%) patients, and colloidal nodular goiters in 286 (78.3%) patients. Of 365 patients, 14 (4.8%) had chronic lymphocytic thyroiditis. Three patients with chronic lymphocytic thyroiditis had thyroid carcinoma; the remaining 11 patients had benign pathology. All patients with chronic lymphocytic thyroiditis had positive thyroid antibodies.

Of the 69 patients with thyroid carcinoma, 63 (91%) had papillary carcinoma, 5 (5%) had follicular carcinoma, and 1 (1%) had medullar thyroid carcinoma. The mean tumor size of thyroid carcinomas, either inside or outside the dominant nodule, was 12.4 (12) mm (range, 1-80 mm). The histopathologic criteria of thyroid carcinomas showed that thyroid capsule invasion was present in 38 patients (55%) and vascular invasion was present in 3 (4.3%) patients. Forty-one of the 69 patients (59.4%) with thyroid carcinoma had carcinoma inside and 28 (40.6%) patients had carcinoma outside the dominant nodule. Seventeen patients had thyroid carcinoma both inside and outside the dominant nodule. The carcinomas were multifocal in 12 patients (2 foci in 8 patients and 3 foci in 4 patients). One hundred foci of thyroid carcinomas, including those that were inside and outside the dominant nodule, were found in the 69 patients with thyroid carcinoma.

Of 100 thyroid carcinomas, 41 (41%) were inside the dominant nodule and 59 (59%) were outside. The incidence of thyroid carcinoma outside the dominant nodule was significantly higher than that inside (z = − 6.48, P < .01). Of 59 thyroid carcinomas outside the dominant nodule, 52 (88.1%) had tumor sizes smaller than 1 cm. The mean tumor size of thyroid carcinoma inside the dominant nodule was significantly larger than thyroid carcinoma outside the dominant nodule (22.19 [14] mm vs 5.7 [4.4] mm, respectively, z = − 7.554, P < .001). The prevalences of capsular and vascular invasions for thyroid carcinoma inside the dominant nodule was significantly higher than that of thyroid carcinoma outside the dominant nodule (53.6% vs 27.1%, P = .007, and 7.3% vs 0%, P = .03, respectively) (Table 1).

Fine-needle aspiration biopsy was performed in 389 nodules in 365 patients. The dominant nodule was sampled in all of them and in another 24 thyroid nodules owing to suspected malignancy by ultrasonography findings. Analysis of the FNAB parameters obtained and compared with the histological findings to rule out malignancy yielded a sensitivity of 76%, a specificity of 86%, and a diagnostic accuracy of 87%, with a positive predictive value of 47% and a negative predictive value of 96%.

Of 59 thyroid carcinomas outside of the dominant nodule, only 9 (15.2%) nodules were suspected of being a malignant nodular lesion by ultrasonography, thus ultrasonography FNAB was performed. The tumor size in suspected malignant nodules was significantly larger than in unsuspected malignant nodules (13.2 [5.6] mm vs 4.4 [2.3] mm, respectively, z = 4.29, P < .001). The mean thyroid volume in suspected malignant nodules was significantly smaller than in unsuspected malignant nodules (27.66 [13] mL vs 69.84 [29] mL, respectively, z = 4.31, P < .001) (Table 2).

According to receiver operating characteristic analysis, the optimal cutoff values of thyroid gland volume and tumor size were 38 mL and 7 mm, respectively. Thyroid gland volume less than 38 mL had a 48-fold increased rate (odds ratio, 48; 95% confidence interval, 6.62-347.74; P < .001) and tumor size larger than 7 mm had a 21.5-fold increased rate (odds ratio, 21.5; 95% confidence interval, 3.68-125.32; P < .001) of detecting suspected malignant thyroid nodules by ultrasonography. Powers for the thyroid gland volume and tumor size were 0.96 and 0.93, respectively. The detection of suspected malignant thyroid nodules by ultrasonography was significantly and positively correlated with the tumor size (r = 0.56, P < .001), whereas it was significantly and negatively correlated with thyroid gland volume (r = − 0.56, P < .001) (Figure).

We found that thyroid volume and tumor size were significantly associated with detection of suspected malignant thyroid nodules by ultrasonography. When the thyroid volume was less than 38 mL, the rate of detection of suspected malignant thyroid nodules by ultrasonography increased. Moreover, when the tumor size was larger than 7 mm, the rate of detection of suspected malignant thyroid nodules by ultrasonography increased.

Thyroid nodules are common in the adult population, and their prevalence is seemingly dependent on the method of detection.^1,6 Palpation findings in iodine-sufficient geographic areas revealed a nodule prevalence of 5%.^6,27 Even in iodine-sufficient regions, the prevalence of thyroid nodules in the general population detected by ultrasonography is as high as 50%. Thyroid nodules are encountered 2 to 3 times more frequently in endemic areas.^2-6 Several autopsy studies have estimated thyroid nodule prevalence to range from 40% to 50%.^28-30

Thyroid nodules found incidentally are increasingly common because of the expanding use of imaging studies.^7,8 Ultrasonography characteristics proved to be predictive of the risk of malignancy.^9,10 Thyroid ultrasonography permits the detection of thyroid nodules as small as a few millimeters. Most of these lesions are benign, but the clinical problem is distinguishing those nodules from benign nodules. Contradictory methods for the management of nonpalpable thyroid nodules have been proposed.^3,12,14 Some authors recommend ultrasonography FNAB in nonpalpable thyroid nodules, whereas others consider a simple follow-up with neck palpation to be sufficient when there is no family history of thyroid cancer or head/neck irradiation.^16,17 Fine-needle aspiration biopsy is not useful for differentiating an MNG with malignant foci from a benign MNG.^23,24 The problem for clinicians and radiologists is selecting nodules for FNAB. We are frequently confronted with this problem, as biopsies of all nodules are not possible.

In our study, ultrasonography FNAB was routinely employed for dominant nodules and also performed for suspected malignant nodular lesions when ultrasonography indicated suspicious findings, including a hypoechoic nodule in association with punctuate calcifications and/or irregular borders. Of 365 patients with MNGs associated with a dominant nodule, only 41 (11.2%) patients had thyroid carcinomas in the dominant nodule. We found that ultrasonography FNAB had a sensitivity of 76% and a positive predictive value of 47%.

In a recent study, FNAB guided by manual palpation was carried out on the dominant nodule and any other nodules with clinical features suggesting malignancy.²⁴ Fine-needle aspiration biopsy positively identified 16.6% of confirmed malignancies; 55% of patients with malignancy had a benign FNAB result. Overall sensitivity and the positive predictive value of FNAB was 17% and 32%, respectively.²⁴ Consequently, FNAB was not useful for differentiating an MNG with malignant foci from a benign MNG. An indeterminate result of FNAB was considered negative. If this result of FNAB was considered positive, the sensitivity for detecting carcinomas increased from 17% to 34%.

In several studies, the improved resolution of the ultrasonography and the use of FNAB, usually under ultrasound guidance, enabled the preoperative diagnosis of extremely small papillary thyroid carcinomas.^31-34 Moreover, a significant number of patients with papillary microcarcinomas had locally advanced disease at the time of surgery.^35,36 This finding suggests that the small size cannot guarantee low risk of thyroid papillary carcinoma found incidentally. The treatment of thyroid microcarcinomas is still a matter of discussion.^37,38 Many microcarcinomas may remain occult and are diagnosed as an incidental finding during surgery for benign thyroid disorders.^37,38 However, some microcarcinomas may result in a negative outcome, including distant metastasis and patient death.^35,36

The wide application of screening ultrasonography for the evaluation of thyroid or carotid artery lesions; the use of FNAB, usually under ultrasonographic guidance; and the refinement of pathologic procedures have led to an increase in the diagnosis of incidental thyroid carcinoma.^7-9,32-36 In these studies, the malignancy rate within thyroid incidentalomas was between 13% and 28.8%, all nodules were incidentally found, and incidental carcinoma was diagnosed following FNAB.^32-36 There was not any information regarding the thyroid volumes in those particular patients.

In conclusion, this is the first study to evaluate the effects of thyroid gland volume and tumor size on the detection of suspected malignant thyroid nodules outside the dominant nodule in MNGs. In patients having MNGs with coexistent dominant nodules, ultrasonography might not be useful in detecting the suspected malignant thyroid nodules when the thyroid gland volume is high.

Correspondence: Yeşim Erbil, MD, Department of General Surgery, Istanbul Medical Faculty, University of Istanbul, 34340 Capa, Istanbul, Turkey (yerbil2003@yahoo.com).

Accepted for Publication: January 21, 2007.

Author Contributions:Study concept and design: Erbil, Mete, and Özarmağan. Acquisition of data: Erbil, Barbaros, Salmaslıoğlu, and Yılmazbayhan. Analysis and interpretation of data: Erbil, İşsever, and Tezelman. Drafting of the manuscript: Erbil, Barbaros, Salmaslıoğlu, Mete, and İşsever. Critical revision of the manuscript for important intellectual content: Erbil, Özarmağan, Yılmazbayhan, and Tezelman. Statistical analysis: Erbil and İşsever. Obtained funding: Erbil, Barbaros, Mete, and Yılmazbayhan. Administrative, technical, and material support: Erbil and Yılmazbayhan. Study supervision: Erbil, Özarmağan, and Tezelman.

Financial Disclosure: None reported.