| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pediatric Imaging |
1 From the Laboratory of Thyroidology, Clinical Research Institute for Radiation Medicine and Endocrinology, Minsk, Belarus (A.L., V.D.); Department of Oncology, Belarusian State Medical University, Minsk, Belarus (Y.D.); and Clinic and Policlinic for Nuclear Medicine, University of Wuerzburg, Wuerzburg, Germany (C.R.). Received December 7, 2003; revision requested February 10, 2004; final revision received June 24; accepted August 4. Supported in part by ISTC project 517-B. Address correspondence to A.L., Department of Nuclear Medicine and Diagnostic Imaging, Kyoto University Graduate School of Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, 606-8507 Kyoto, Japan (e-mail: lyshchik@kuhp.kyoto-u.ac.jp).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The study protocol was approved by the institutional review board, and patient’s parents gave full informed consent. One hundred three consecutive pediatric patients with solid thyroid nodules were included in the study. Thirty-five patients had thyroid cancer (mean age, 14.6 years ± 2.6 [standard deviation]; range, 10–18 years), and 68 patients had benign thyroid nodules (mean age, 14.2 years ± 2.9; range 9–18 years). Three-dimensional US was used to determine the volume of thyroid gland and thyroid nodules. Results of nodule cytologic and histologic examination and long-term clinical and US follow-up were used as a proof of final diagnosis. The following US characteristics were evaluated: location, echogenicity, echotexture, outline, presence of a halo, microcalcifications, and type of vascularization. Multivariate logistic regression analysis was used to evaluate the accuracy of US criteria for thyroid cancer in lesions with diameter of 15 mm and smaller and lesions with diameter larger than 15 mm. Qualitative variables were compared by using the 
2 test and quantitative variables were compared by using the Student t test. Significance was defined at P < .05.
RESULTS: In thyroid nodules with diameter of 15 mm and smaller, the most reliable diagnostic criteria for malignancy were an irregular outline (sensitivity, 69.6%; specificity, 86.4%; P < .001), subcapsular location (sensitivity, 65.2%; specificity, 86.4%; P < .001), and increased intranodular vascularization (sensitivity, 69.6%; specificity, 87.9%; P < .01). For thyroid nodules larger than 15 mm in diameter, the accuracy of US diagnosis was much lower than that for smaller nodules. The only reliable criterion for cancer in this group was hypoechogenicity (sensitivity, 60.0%; specificity, 84.0%; P < .01).
CONCLUSION: Study findings indicate that US is most helpful in diagnosis of thyroid malignancy in thyroid nodules with diameter of 15 mm and smaller, with detection of irregular tumor outline, subcapsular location, and increased intranodular vascularization.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Several study findings have shown that thyroid carcinoma in pediatric patients differs from that in adults with respect to its presentation and outcome. In children, it tends to present at a more advanced stage than in adults, with a higher frequency of lymph node and pulmonary metastases (9). Therefore, cancer diagnosis at an early stage is extremely important, especially in a high-risk population, because capsular invasion followed by distant spread or in some cases dedifferentiation can develop during the course of the disease (10).
Systematic screening with ultrasonography (US) of the thyroid gland is used for early detection of clinically undetectable nodules in high-risk groups or in persons in an attempt to diagnose and treat thyroid cancer early in its development (11). However, the diagnostic management of thyroid nodules in children and the US criteria used to identify and characterize suspicious nodules in children are controversial and not well defined (12–15). Moreover, the diagnostic value of single and combined gray-scale and power Doppler US features in children with solid thyroid nodules has not been previously evaluated in detail. Thus, the purpose of our study was to prospectively analyze the accuracy of various diagnostic criteria for cancer in solid thyroid nodules in children on the basis of gray-scale and power Doppler US findings.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Patients were divided into two groups. The first group of patients included 35 children with papillary and follicular thyroid cancer. The final diagnosis was histologically confirmed in all patients. The mean age of the patients in this group was 14.6 years ± 2.6 (standard deviation) (range, 10–18 years), and the male-to-female ratio was 11:24. Patients with thyroid malignancy resided in the area of Belarus with moderate iodine deficiency. In addition, this area was contaminated by radionuclides as a result of the Chernobyl nuclear power plant accident in April 1986. All patients had a confirmed history of radiation exposure.
The second age-matched group of patients included 68 children with solid benign thyroid nodules. Patients with benign thyroid disease resided in the same areas of Belarus (with respect to iodine deficiency and radionuclide contamination) as the patients with thyroid cancer. The mean age of the patients in this group was 14.2 years ± 2.9 (range, 9–18 years), and the male-to-female ratio was 27:41. A final diagnosis of benign thyroid nodular disease was histologically confirmed in 39 patients (57.4%). In the other 29 patients (42.6%) who did not require surgical treatment, the final diagnosis was established with cytologic evaluation and annual clinical and US follow-up. There were no significant differences in age (t test, P = .5) and sex (2 test, P = .4) between the patients with malignant and those with benign thyroid tumors.
Three patients with malignant tumors (one boy and two girls) and 18 patients with benign tumors (six boys and 12 girls) had multinodular disease (8.6% and 26.5%, respectively; 2 test, P < .05). Because all lesions in the patients with multinodular disease differ in their US characteristics, they were evaluated separately. The total number of thyroid nodules analyzed in this study was 129. Histologic and/or cytologic evaluations revealed papillary carcinoma in 36 (27.9%) nodules, follicular carcinoma in two (1.5%) nodules, follicular adenoma in 18 (14.0%) nodules, and adenomatous goiter in 73 (56.6%) nodules.
Radiation History
The mean age of the patients with thyroid cancer at the time of radiation exposure was 2.3 years ± 2.5 and that of patients with benign thyroid tumors was 2.2 years ± 2.0 (t test, P = .8). Thirty-four (97.1%) patients with thyroid cancer and 56 (82.4%) patients with benign thyroid tumors (2 test, P < .05) were exposed to radiation during the first 5 years of life. Among those patients, 10 with thyroid cancer (28.6%) and 13 (19.1%) with benign thyroid tumors (2 test, P = .2) were exposed to radiation during the 1st year of life. All patients with multinodular disease were exposed to radiation at 1–5 years of age.
US Examinations and Interpretation
Gray-scale, power Doppler, and three-dimensional US examinations were performed in all patients by the same investigator (A.L.) using a scanner (HP ImagePoint; Hewlett-Packard, Andover, Mass) with a 7.5-MHz linear probe. Nodule location was reported as subcapsular if no intervening parenchyma between the nodule and the thyroid capsule was identified (Fig 1). Nodule echogenicity was assessed with respect to the surrounding thyroid tissue and was classified as hypoechoic, isoechoic, hyperechoic, or mixed. The outline of the nodule margin was assessed with respect to the smoothness and definability of the nodule contour and was classified as regular or irregular. The nodule echotexture was classified as homogeneous or heterogeneous. The presence of a halo sign (anechoic or hypoechoic rim surrounding the nodule) and of microcalcifications was also recorded. Examples of the nodules with different US characteristics are presented in Figure 2.
|
|
|
|
|
|
|
|
In all patients, all gray-scale and power Doppler images were reviewed and interpreted by two radiologists (A.L. and V.D., with more than 3 and 15 years of experience, respectively, in thyroid US and thyroid biopsies), who were blinded to the patients’ diagnosis. Final decisions regarding the findings were reached by consensus.
Fine-Needle Aspiration Biopsy
After US examination, a US-guided fine-needle aspiration biopsy was performed for all nodules by one of two investigators (A.L. or V.D.). The aspirated material was stained with May-Grunwald-Giemsa stain and evaluated by a cytopathologist, who had more that 5 years of experience in thyroid cytology and was blinded to US findings. Cytologic results were classified as benign, suspicious, or malignant. Patients with a benign cytodiagnosis had no features suggestive of or diagnostic for malignancy. Patients with a suspicious cytodiagnosis had specimens showing hypercellularity and a pattern suggestive of follicular neoplasms or atypical features suggestive of, but not diagnostic for, malignancy. Patients with a malignant cytodiagnosis had cytologic findings that indicated the presence of malignant cells consistent with thyroid carcinoma (19).
Surgery
All patients with suspicious or malignant cytologic evaluation results underwent surgery (total thyroidectomy with lateral neck lymph node dissection). All neoplastic lesions were staged according to the TNM classification (20). Thirty-nine patients with benign cytologic findings and clinical indications such as comparatively large or quickly growing nodules or whose parents preferred surgical treatment despite benign cytologic findings underwent hemithyroidectomy. Histologic evaluations were performed by a pathologist, who had more that 15 years of experience in thyroid histologic examination and who was blinded to US and cytologic findings. In all 74 patients (35 with malignant and 39 with benign disease) who underwent surgery, papillary thyroid cancer was diagnosed in 33 (44.6%), follicular thyroid cancer was diagnosed in two (2.7%), follicular adenoma was diagnosed in 16 (21.6%), and nodular goiter was diagnosed 23 (31.1%) patients.
Follow-up Examinations
Twenty-nine patients with benign cytologic findings who did not receive surgical treatment were followed up with the administration of suppressive doses of levothyroxine. Once a year they underwent repeated clinical and US examinations, US-guided fine-needle aspiration biopsy, and thyroid-stimulating hormone and thyroid hormone tests to control nodule progression and treatment response. The mean follow-up period was 2.1 years ± 0.7, and the mean interval between repeated examinations was 11.2 months ± 1.3. During the observation period, none of these patients had any findings suspicious for malignancy.
Statistical Analysis
Comparison of qualitative variables (clinical, pathologic, and US) was performed by using the 2 test. Quantitative variables (patient’s age and thyroid volumes) were compared by using the paired Student t test. US characteristics of the thyroid gland and of each thyroid nodule were registered separately and were processed blindly for statistical evaluation. The unit of analysis was each nodule rather than each patient. For each gray-scale and power Doppler criteria, the sensitivity, specificity, and overall accuracy of differentiation between benign and malignant lesions were calculated by using standard procedures (21). Correlations between the frequency of visualization of the main US characteristics and nodule size were assessed with the Spearman rank-order correlation coefficient (R). To determine the size dependency of US criteria for thyroid cancer, different cutoff values of the nodule diameter (8, 10, 12, 15, 17, and 20 mm) were tested to find the one that can divide examined lesions into two groups with a statistically significant difference in their US characteristics (2 test). After the cutoff value of the nodular size was established, all US characteristics and their diagnostic accuracy were analyzed for nodules smaller and larger in diameter than the selected threshold value.
A forward stepwise multivariate logistic regression analysis was performed to select independent variables out of examined diagnostic criteria associated with the dependent variable, that is, the tumor type (malignant or benign). The independent variables observed in gray-scale and power Doppler modes were the size of the nodule, hypoechogenicity, heterogeneity, outline irregularity, absence of halo sign, subcapsular location, presence of microcalcifications, and type III vascularization. Each qualitative variable had a dichotomous value (observed or not observed). Only independent variables that showed statistical significance (P < .2) were included in the final multiple logistic regression model. This analysis was performed for all thyroid nodules included in our study, as well as for lesions smaller and larger in diameter than the selected threshold value of the tumor size. Quantitative data are presented as mean ± standard deviation. Significance was defined at P < .05. Statistical analyses were performed by using a statistical software package (StatView, version 5.0; SAS Institute, Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
2 test, P < .01). However, no patient had significant alteration of the surrounding thyroid tissue echogenicity or vascularity by goiter or other concomitant thyroid diseases. US characteristics of all thyroid nodules.—The size of malignant and benign thyroid nodules did not differ significantly. The mean maximal diameter of malignant nodules was 13.8 mm ± 8.8 (range, 5–28 mm), which was comparable with the mean maximal diameter of benign nodules (14.8 mm ± 10.7; range, 5–51 mm; t test, P = .6).
Among the gray-scale characteristics evaluated for all nodules, nodule echogenicity, border outline, and location were statistically significant indicators of thyroid malignancy (Table 1). Hypoechogenicity was found in 55.3% of malignant lesions and in 35.2% of benign lesions (2 test, P < .05). A statistically significant difference was also found in the number of nodules with an irregular outline. This feature was more often diagnosed in malignant nodules (71.1%) than in benign nodules (19.8%; 2 test, P < .01). Furthermore, our analysis revealed one more characteristic that was associated with thyroid cancer in children: a subcapsular location of thyroid nodules (Fig 1a). This location was found in 76.3% of malignant lesions but only in 36.3% of benign tumors (2 test, P < .01). Another 63.7% of the benign lesions were located in the central regions of the thyroid lobes or were separated from the capsule by intervening thyroid parenchyma (Fig 1b).
|
2 test, P = .2). The same findings were observed for the halo sign, which was typically absent in both malignant (78.9%) and benign (61.5%) nodules (2 test, P = .07), and for the presence of microcalcifications (5.3% and 3.3%, respectively; 2 test, P = .6).
Power Doppler US examinations revealed statistically significant differences between malignant and benign lesions (Table 1). Type III nodular vascularization was detected in 72.2% of thyroid cancers (Fig 3), whereas this feature was registered in only 22.0% of benign nodules (2 test, P < .01).
US characteristics of thyroid nodules of different size.—The correlation analysis showed that the frequency of visualization of benign thyroid nodules with certain US characteristics significantly depends on nodule size (Table 2). In benign thyroid lesions, we revealed a significant decrease in the number of hypoechoic nodules (R = –0.23, P < .05) and a significant increase in the number of heterogeneous nodules (R = 0.37, P < .01), with an increase in nodule size. One more important finding was the significant increase in the number of benign nodules with type III vascularization, with increase in their size (R = 0.35, P < .01). In contrast, US characteristics of malignant nodules did not depend significantly on nodular diameter.
|
2 test, P < .01). They were less often heterogeneous (53.0%) and had an irregular outline (13.6%) than did larger nodules (80.0% and 36.0%, respectively; 2 test, P < .01). As for power Doppler features, only 12.1% of benign nodules with maximal diameter of 15 mm and smaller showed type III vascularization. In larger nodules, this feature was registered in 48.0% of lesions (2 test, P < .01). In contrast to benign nodules, US characteristics of malignant nodules smaller and larger than 15 mm did not differ significantly (Fig 5).
|
|
2 test, P = .4). On the other hand, absence of the halo sign became a significant indicator of malignancy in nodules 15 mm and smaller and was registered in 82.6% of malignant and in 59.1% of benign lesions (2 test, P < .05). Other criteria, such as irregular border outline, subcapsular nodule location, and type III vascularization at power Doppler US retained their significant association with malignant lesions (P < .01).
|
2 test, P < .01). Irregularity of the nodule outline remains the primary appearance in malignant lesions larger than 15 mm (73.3% vs 36.6% for lesions 
15 mm; 2 test, P < .05). Unlike in lesions with maximal diameter of 15 mm and smaller, subcapsular location, absence of the halo sign, and type III vascularization were not associated with malignancy in nodules larger than 15 mm.
Diagnostic Accuracy of US Criteria for Thyroid Cancer
Diagnostic accuracy of the US criteria for thyroid cancer among all nodules.—Diagnostic value of US in differentiation of benign and malignant thyroid lesions regardless of their size was moderate (Table 4). Only two diagnostic criteria had an overall accuracy that exceeded 75% and can be recommended for thyroid cancer diagnosis. These criteria were irregularity of nodule outline (71.1% sensitivity and 80.2% specificity) and type III vascularization (71.1% sensitivity and 78.0% specificity). Hypoechogenicity and subcapsular location were less accurate criteria and had 55.3% and 76.3% sensitivity and 64.8% and 63.7% specificity, respectively. Both nodule heterogeneity and absence of the halo sign were not appropriate for diagnosis. They had comparatively high sensitivity (73.7% and 78.9%, respectively) but a very low specificity (39.6% and 38.5%, respectively). In contrast, presence of microcalcifications was very specific (96.7%) but not sensitive (5.3%).
|
Diagnostic accuracy of the US criteria for thyroid cancer among nodules of different size.—After the cutoff value of 15 mm was established for nodular size and a significant difference in the US appearance of the small versus large nodules was found, we estimated the diagnostic sensitivity, specificity, and overall accuracy of various US criteria separately for nodules smaller and larger than 15 mm in maximal diameter (Table 4).
Among nodules with maximal diameter of 15 mm and smaller, the diagnostic value of US criteria for thyroid cancer was higher than that among all lesions, with three criteria demonstrating an overall accuracy of more than 80.0%. These criteria were irregular outline (69.6% sensitivity and 86.4% specificity), subcapsular location (65.2% sensitivity and 86.4% specificity), and type III nodule vascularization at power Doppler US (69.6% sensitivity and 87.9% specificity). A combination of outline irregularity or type III vascularization with subcapsular location of thyroid nodules can increase the specificity up to 97.0% and the overall accuracy up to 85.4%, with some decrease in the sensitivity (52.2%). On the other hand, hypoechogenicity (52.2% sensitivity, 57.6% specificity) and heterogeneity (69.6% sensitivity, 47.0% specificity) were not appropriate for cancer diagnosis. Absence of the halo sign had good sensitivity (82.6%) but low specificity (40.9%). Presence of microcalcifications in this group was extremely specific (98.5%) but not sensitive (8.7%).
The diagnostic value of US criteria for thyroid cancer was much lower for nodules larger than 15 mm in diameter than for smaller nodules. The only reliable criterion was hypoechogenicity, with a 60.0% sensitivity, 84.0% specificity, and 75.0% overall accuracy. Other diagnostic criteria useful for nodules with maximal diameter of 15 mm and smaller, such as irregular outline and type III vascularization, in large lesions showed high sensitivity (73.3%) but low specificity (64.0% and 52.0%, respectively), which decreased their overall accuracy to less than 70%. Subcapsular location, heterogeneity, and absence of the halo sign were also sensitive (93.3%, 80.0% and 73.3%, respectively), but the specificity was extremely low (4.0%, 20.0%, and 32.2%, respectively). Unfortunately, the combination of gray-scale and power Doppler US characteristics did not bring significant improvements to thyroid cancer diagnosis in nodules with a diameter larger than 15 mm.
Multivariate logistic regression analysis.—A forward stepwise multivariate logistic regression analysis was performed to determine independent US predictors for thyroid malignancy (Table 5). When all lesions were evaluated, five criteria (nodule size, subcapsular location, hypoechoic appearance, irregular outline, and type III vascularization) showed significant association with thyroid cancer (P < .05). Same analysis was performed for lesions with maximal diameter smaller and larger than 15 mm. As a result, for small lesions, only four criteria (nodule size, subcapsular location, irregular outline, and type III vascularization) remained to be statistically significant independent predictors for thyroid malignancy (P < .05). In contrast to these findings, among lesions larger than 15 mm in diameter, only one criterion, hypoechoic lesion appearance, was significantly associated with thyroid cancer (P < .01).
|
2 test, P = .2). In patients with nodules smaller than 15 mm, the diagnosis was usually made at less advanced stage. Extrathyroidal cancer spread (stage, pT4) was evident in only 13.6% of the patients and regional lymph node involvement (stages, pN1a and pN1b) was evident in 54.6% of patients. Patients with nodules larger than 15 mm in diameter usually presented at more advanced stage (pT4, 30.8%; pN1a and pN1b, 84.6%). However, this tendency was statistically insignificant (2 test, P > .05).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In a situation when up to 4% of children can have thyroid nodules, reliable criteria are required to determine which nodules can be followed up and which nodules should be aspirated. As previously described, US examination can be successfully used for screening and early detection of thyroid nodules in children (25,26). Unfortunately, the diagnostic accuracy of US criteria in differentiating between benign and malignant thyroid nodules in children is still controversial, and the issue of whether clinically unapparent thyroid nodules should undergo cytologic and histologic evaluation remains unresolved (27).
In the group enrolled in our study, thyroid cancer was diagnosed in 34.0% of the patients. This high percentage can be explained by the methods of patient selection when initial diagnosis of thyroid nodular disease was made with mass US screenings in radiation-contaminated regions.
The volume of the thyroid gland in patients with benign thyroid nodules was significantly higher than that in patients with thyroid cancer. These findings can be explained by the influence of endemic iodine deficiency on the development of benign thyroid nodular disease. The size of malignant and benign nodules did not differ significantly, and nodule dimension was not related to malignancy in our patients.
To evaluate the differences in diagnostic accuracy of US criteria for cancer in solid thyroid nodules of different size, we established a 15-mm cutoff value for the nodule diameter. That cutoff value allowed divisions of all lesions into two groups with a statistically significant difference in their US characteristics. Our results showed a significant difference in the diagnostic accuracy of US in nodules of different size. Among nodules with maximal diameter of 15 mm and smaller, the most reliable criteria for thyroid cancer diagnosis were irregular outline, subcapsular location, and detection of type III vascularization at power Doppler US. However, some criteria widely used for thyroid cancer diagnosis in adults such as nodule hypoechogenicity and heterogeneity showed very low overall accuracy (56.2% and 52.8%, respectively) and cannot be recommended for the differentiation of small thyroid nodules in children. The same results were found for the presence of microcalcifications, which was one of the most accurate diagnostic criteria for thyroid malignancy in adults (15). In our study, presence of microcalcifications had very high specificity (98.5%) but extremely low sensitivity (8.7%), which makes this criterion inappropriate in pediatric patients.
The accuracy of US differentiation between benign and malignant nodules among lesions larger than 15 mm in diameter was lower than that among smaller lesions. The only reliable criterion for cancer in nodules larger than 15 mm was hypoechogenicity. Other diagnostic criteria that were highly valuable in small thyroid nodules were much less sensitive among larger lesions.
Results of multivariate logistic regression analysis showed that in nodules smaller and larger than 15 mm in maximal diameter, different diagnostic criteria should be used for differential diagnosis of thyroid cancer. In smaller lesions, subcapsular location, irregular outline, and type III vascularization enabled a highly accurate differentiation between malignant and benign lesions. However, in larger nodules, only one criterion, hypoechoic tumor appearance, can be successfully used for thyroid cancer diagnosis.
Our results may be in some disagreement with previously published reports (28,29) in which the accuracy of US features for thyroid cancer diagnosis in adults was evaluated. For example, in our study only two patients with thyroid cancer had microcalcifications, which were recognized by other authors as the most common finding in patients with thyroid malignancy (30). Moreover, authors of some studies (31,32) described hypoechogenicity as highly predictive for malignancy, although in our study this was true only for nodules larger than 15 mm in diameter. We also showed that an irregular outline and type III nodular vascularization were specific for thyroid malignancy only among small nodules, and they loose their diagnostic potential in larger nodules because of the high frequency of false-positive results among nodules larger than 15 mm.
On the other hand, in some studies (33,34), the estimated accuracy of US diagnosis for nonpalpable thyroid nodules in adults is highly comparable with our findings. Papini et al (34) showed that irregular outline and type III vascularization can be used as independent risk factors for thyroid malignancy in nonpalpable nodules smaller that 15 mm, which is similar to our results. They also showed that a hypoechoic pattern cannot be used as a reliable predictor for thyroid cancer malignancy; although 87.1% of thyroid cancers were hypoechoic, the majority (56.6%) of benign nodules in their study were also hypoechoic, which is comparable with 42.4% of hypoechoic benign nodules with maximal diameter of 15 mm and smaller in our study.
We should note that the cutoff value of the nodular size used in our study has no diagnostic potential by itself and cannot be used as a prognostic indicator. We did not find statistically significant differences in cancer staging between children with nodules smaller or larger than 15 mm in diameter. Extrathyroidal cancer metastasis, recognized as one of the most substantial risk factors of cancer recurrence, was diagnosed in 13.6% of patients with malignant nodules smaller than 15 mm and in 30.8% of patients with malignant nodules larger than 15 mm (2 test, P = .2). The same findings were observed with the presence of lymph node involvement in patients with small versus patients with large malignant nodules (54.6% vs 84.6%; 2 test, P = .2). These findings are important because some findings of thyroid malignancy in children showed that when the initial disease was confined to the thyroid gland, none of the patients developed recurrent disease after adequate treatment, compared with a 50% recurrence rate in patients with lymph node involvement at the time of diagnosis (9).
Although our results showed a high diagnostic value of US examination in small solid thyroid nodules, some limitations need to be addressed. In the study we did not include solid nodules smaller than 5 mm in diameter, some of which may have been malignant. These patients may require thorough follow-up until the suspected lesion becomes assessable with US-guided fine-needle aspiration biopsy. Finally, the youngest patient in our group was 9 years old; therefore, our study does not address findings in younger children and infants. Further multi-institutional studies will be needed in a larger population to determine the value of US criteria for thyroid cancer and their size dependency in younger children and to support confidence in the outcome results within appropriate limits.
In conclusion, our findings showed the usefulness of gray-scale and power Doppler US in helping to differentiate malignant and benign solid thyroid nodules, especially for lesions with maximal diameter of 15 mm and smaller. Among these nodules, the most reliable criteria for thyroid cancer diagnosis in children were irregular outline, subcapsular nodule location, and type III vascularization at power Doppler US. However, some reliable diagnostic criteria used in adults such as nodule hypoechogenicity, heterogeneity, and presence of microcalcifications showed very low accuracy in our study and may not be applicable for differentiation of small thyroid nodules in children. Among nodules larger than 15 mm in diameter, diagnostic accuracy of gray-scale and power Doppler criteria was low, which can be explained by the significant amount of false-positive results in this group. That is why we conclude that in nodules with maximal diameter of 15 mm and smaller, a US examination can be useful in helping to determine which nodules can be aspirated rather than which should be followed up; however, nodules larger than 15 mm are not reliably differentiated with the US criteria examined and may require cytologic or histologic examination for definite classification. Further well-controlled prospective studies of a larger population may be useful for a more detailed evaluation of the effect of US examination on thyroid cancer diagnosis in children.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, all authors; study concepts and design, all authors; literature research, A.L.; clinical studies, A.L., V.D., Y.D.; data acquisition, A.L., V.D., Y.D.; data analysis/interpretation, all authors; statistical analysis, A.L., V.D.; manuscript preparation, A.L.; manuscript definition of intellectual content, all authors; manuscript editing, revision/review, and final version approval, V.D., C.R.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. O'Kane, E. Shelkovoy, R. J. McConnell, V. Shpak, L. Parker, T. I. Bogdanova, A. Brenner, Y. Naida, A. Frangos, L. Zablotska, et al. Differences in Sonographic Conspicuity According to Papillary Thyroid Cancer Subtype: Results of the Ukrainian-American Cohort Study After the Chornobyl Accident Am. J. Roentgenol., December 1, 2008; 191(6): W293 - W298. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Lyshchik, R. Moses, S. L. Barnes, T. Higashi, R. Asato, M. I. Miga, J. C. Gore, and A. C. Fleischer Quantitative Analysis of Tumor Vascularity in Benign and Malignant Solid Thyroid Nodules J. Ultrasound Med., June 1, 2007; 26(6): 837 - 846. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Lyshchik, T. Higashi, R. Asato, S. Tanaka, J. Ito, M. Hiraoka, M. F. Insana, A. B. Brill, T. Saga, and K. Togashi Cervical Lymph Node Metastases: Diagnosis at Sonoelastography--Initial Experience Radiology, April 1, 2007; 243(1): 258 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Niedziela Pathogenesis, diagnosis and management of thyroid nodules in children. Endocr. Relat. Cancer, June 1, 2006; 13(2): 427 - 453. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||
