| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Division of Pediatric Endocrinology, Regina Margherita Children Hospital (A.C., S.E., M.A., L.d.S.), 10100 Turin, Italy; Pediatric Unit, Department of Medical Sciences, University of Piemonte Orientale (E.C., G.B.), 28100 Novara, Italy; Endocrine Unit, Department of Pediatrics, University of Milan (G.W., M.V.), 20100 Milan, Italy; Unit of Autoimmune Endocrine Diseases, Bambino Gesu Hospital Research Institute (A.C., S.S.), 00100 Rome, Italy; Department of Pediatrics, University of Pisa (G.C.), 56100 Pisa, Italy; Institute of Pediatrics, University of Messina (M.F.M., F.d.L.), 98100 Messina, Italy; Department of Pediatric University La Sapienza (M.S., A.M.P.), Rome, Italy; Department of Pediatrics, Caldarelli Hospital (M.C.), 86100 Campobasso, Italy; and Department of Pediatrics, University of Turin (G.C.M.), 10100 Turin, Italy
Address all correspondence and requests for reprints to: Dr. A. Corrias, Division of Pediatric Endocrinology, Regina Margherita Children Hospital, 10126 Turin, Italy.
Abstract
In childhood the traditional diagnostic approach to thyroid nodules consists of clinical, laboratory, and imaging evaluations. A safe and accurate procedure is needed to promptly identify patients who require surgery.
In regard to the usefulness of fine needle aspiration biopsy, the data in the literature concerning children and adolescents are scanty. The aim of this study was to evaluate and compare the diagnostic accuracies of clinical, laboratory, and imaging data collected retrospectively in a group of pediatric patients with thyroid nodules submitted to fine needle aspiration biopsy.
Forty-two patients who underwent surgery for thyroid nodules, recruited in 9 Italian pediatric endocrine units, were retrospectively studied. According to histological diagnosis, they were divided into 2 groups, 22 patients with benign lesions and 20 patients with malignant lesions. From clinical records we obtained data about 1) symptoms of neck compression; 2) cervical adenopathy; 3) thyroid function, calcitonin level, and antithyroid antibody titers; 4) ultrasonography; 5) 99mTc scintiscanning; and 6) cytology obtained with fine needle aspiration biopsy. Patients and nodule characteristics were analyzed statistically for associations with the presence of thyroid cancer.
Among clinical findings, only adenopathy was significantly higher in the group with cancer (8 of 22 benign lesions vs. 16 of 20 malignant lesions; P = 0.006). Thyroid function and antibody titers were similar in the 2 groups, whereas the serum calcitonin level was elevated only in 1 patient with malignant lesions. Among ultrasonography findings, no significant statistical difference was found between the 2 groups with regard to number, dimensions, growth progression, or hypoechogenic pattern of the nodules. Regarding scintigraphic findings, no significant difference was found between the 2 groups. However, a positive correlation (r = 0.90; P < 0.0001) was found between fine needle aspiration biopsy cytological findings and histological diagnoses. The sensitivity, specificity, and accuracy of fine needle aspiration biopsy were 95%, 86.3%, and 90.4%, respectively. A multiple regression analysis showed that only fine needle aspiration biopsy (ß coefficient = 0.963; P < 0.0001) significantly contributed to detecting malignancy (multiple r = 0.973; P < 0.0001).
This study provides strong evidence that fine needle aspiration biopsy is a safe technique even in childhood and adolescence, offering the best sensitivity, specificity, and accuracy in detecting malignancy compared with conventional approaches.
THYROID NODULES ARE uncommon in childhood and adolescence; their prevalence ranges between 0.2–1.4% and is 5–10 times lower than in adults (1, 2, 3, 4). However, the frequency of malignancy among thyroid nodules is higher than in adults; different studies have reported a 15–20% incidence, with up to 50% in some series (5, 6, 7, 8, 9). Furthermore, in the U.S.A. and other western countries (10), thyroid cancer is the third most common solid tumor in childhood and adolescence, with an incidence of 1.75 in 100,000. Although it is usually a slow-growing tumor, a prompt diagnosis is recommended, because greater tumor size, distant spread, and greater atypia are associated with increased cumulative mortality.
The traditional diagnostic approach to thyroid nodules consists in clinical, laboratory, and imaging examinations; in particular, a palpable hypoechogenic and scintigraphic cold nodule might suggest thyroid malignancy, especially in the presence of cervical adenopathy. In adults, fine needle aspiration biopsy (FNAB) shows a high degree of accuracy in detecting malignancy, as recently described by some researchers (11, 12, 13, 14, 15, 16); on the contrary, in childhood and adolescence, the data to be found in the literature are scanty and often in disagreement (5, 8, 9, 17).
The aim of this study was to evaluate and compare the diagnostic accuracies of clinical, laboratory, and imaging data collected retrospectively in a group of pediatric patients with thyroid nodules who underwent surgery after FNAB.
Subjects and Methods
Patients
Eighty-seven pediatric patients affected by thyroid nodules were recruited from nine Italian Pediatric Endocrine Units. Forty-five patients with benign lesion obtained by FNAB were excluded from the study group because they had not undergone surgery; only 42 patients who underwent surgery were retrospectively studied. Informed consent was requested and obtained in all cases.
The study group consisted of 34 females and 8 males. At diagnosis, the median age was 13.1 yr (range, 8.6–17.9 yr); 36 (86%) patients were pubertal, and 6 (14%) were prepubertal. Three patients were previously affected with autoimmune thyroiditis, another with Graves’ disease, 3 with goiter, 4 with acute lymphoblastic leukemia, and 1 with non-Hodgkin lymphoma. When the thyroid nodule was discovered, 6 patients were being treated with L-T4 for thyroiditis or goiter, and thyroid function was normal in all cases.
Materials and methods
According to histological diagnosis of thyroid nodules obtained
after surgery, the patients were divided into 2 groups. Group 1 was
composed of 22 patients with benign lesions (BL), and group 2 was
composed of 20 patients with malignant lesions (ML), as shown in Table 1
.
|
The US reports were analyzed for nodule number and dimension. In addition, the echogenic pattern was classified into four categories: hypoechogenic, isoechogenic, hyperechogenic, and mixed echogenic nodules. Progressive growth of nodules was evaluated in 21 patients.
99mTc scintiscanning results were analyzed according to how the nodules concentrated the radioisotope and were classified into four categories: hypofunctional, isofunctional, hyperfunctional, and undetectable.
In patients with two or more nodules, with all of them being either benign or malignant, we decided to consider only US and scintiscanning data of the largest nodule. On the contrary, if there were both benign and malignant nodules we considered US and scintiscanning data of the malignant nodules.
FNAB reports were analyzed independently by two pathologists for evidence of malignancy and then classified into three categories: 1) positive, when the pathologist was able to determine the tumor histology (papillary or medullary carcinoma); 2) suspicious, when the cytological pattern was judged to consist in a follicular neoplasm (including Hurthle cell neoplasms) and in the case of specimens with some, but not all, of the cytological characteristics of papillary thyroid cancer; and 3) negative, when the specimens contained colloid, macrophages, and aggregates of normal-appearing thyroid cells and in cases diagnosed as benign disease (e.g. goiter, colloid cyst, and thyroiditis). Aspirates were considered insufficient when only cyst fluid or a few scattered thyroid cells were identified. In these cases, FNAB had been repeated, and the last reports were included in the study.
Aspirate reports giving conflicting interpretations were submitted to a third pathologist, and the specimens received the classification given by two of the three pathologists if a consensus had not been reached.
A true positive FNAB was defined as suspicious or malignant cytological result from a nodule determined to be malignant after surgery. A false positive was defined as a suspicious or malignant FNAB diagnosis performed on a nodule determined to be benign at surgery. A true negative FNAB was defined as a benign cytological result from a nodule determined to be benign after surgery. A false negative FNAB was defined as a benign cytological result from a nodule determined to be malignant at the time of surgery.
Statistical analysis
To assess the ability of the most common parameters (cervical adenopathy, hypoechogenic nodule, cold nodule, and cytological results) to detect malignancy, sensitivity, specificity, and diagnostic accuracy were calculated.
Comparisons between groups concerning the number, dimension, and
echographic and scintigraphic pattern of nodules were performed by
2 test, and those concerning clinical signs
(palpable nodule, symptoms of neck compression, and cervical
adenopathy) and echographic progressive growth were performed by
Fisher’s exact test. Correlation between cytological and histological
diagnoses was performed by Spearman’s test. The relationship between
the independent variables (clinical, laboratory, and imaging data) and
the malignancy of thyroid nodules were evaluated by multiple regression
analysis. P < 0.05 was considered significant.
Results
The physical examination findings are reported in Table 2. Neither palpable nodules nor signs of
neck compression were significantly different between the 2 groups. On
the contrary, cervical adenopathy was found in 8 of 22 (36%) of the
group with BL, and in 16 of 20 (80%) of the group with ML
(P = 0.006, by Fisher’s test).
|
The US results are shown in Table 3. No
significant statistical difference was found with regard to the number,
dimensions, and progressive growth of the nodules between the two
groups. Hypoechogenic lesions were not significantly more frequent in
the group with ML.
|
. Hypofunctional
lesions were found in 21 cases (66%), of which 9 had BL and 12 had ML
(P = NS). Furthermore, in 9 patients scintiscanning did
not reveal any nodules.
Regarding the cytological findings obtained with FNAB, a positive
result for malignancy was revealed in 17 of 20 (85%) patients in group
ML and in 1 of 22 (4%) in group BL, whereas a negative result was
found in 1 of 20 (5%) patients in group ML and in 19 of 22 (86%) in
group BL. Of the 4 lesions considered suspicious by FNAB cytology, 2
were diagnosed as ML and 2 as BL (P < 0.0001, by
2 test) after surgical excision.
The sensitivity, specificity, and diagnostic accuracy of the main
parameters are reported in Table 4. The sensitivity of the
cytological specimens obtained with FNAB was 95%, whereas the
specificity was 86.3%, and the diagnostic accuracy was 90.4%.
|
.
|
Without considering FNAB, multiple regression analysis showed that the only parameter significantly related to malignancy in our patients was cervical adenopathy (ß coefficient = 0.666; P = 0.002; multiple r = 0.66; P = 0.02).
Discussion
The aim of this study was to evaluate the diagnostic accuracy in 42 patients with 1 or more thyroid nodules submitted to FNAB and, afterward, to surgery.
In childhood and adolescence, the data concerning thyroid nodules is still scanty and often in disagreement (5, 8, 9, 15). In addition, these reports may be influenced by different factors, such as area of origin, iodine deficiency, eventual iodine implementation, and the type of investigation. Moreover, the first studies were performed by physical examination, and it is common knowledge that this is a subjective method. Later, routine US examinations revealed significantly higher frequencies of thyroid nodules (1). However, at least in areas with sufficient iodine intake, the current prevalence of nodular thyroid disease in children has been estimated to be 5–10 times lower than that in adults. On the other hand, in children the frequency of malignancy among thyroid nodules is higher than in adults. Different reports show different frequencies that vary from 2–50%, and in the majority of cases from 15–20% (2, 9, 18).
Although thyroid cancer in children usually has a very indolent course, even with pulmonary metastases, early diagnosis is important to identify as soon as possible patients who need to undergo surgery. Therefore, we decided to evaluate retrospectively a group of pediatric patients affected by thyroid nodules and submitted to surgery to determine the accuracy of FNAB in detecting malignancy compared with those of conventional clinical, laboratory, and imaging approaches.
Recently, FNAB has become the cornerstone in the evaluation of solitary thyroid nodules, cysts, and dominant nodules within multinodular goiters, at least in adults (4, 14, 16, 19). If the procedure is performed properly, it should have very low false negative and false positive rates, with accuracy ranging between 69% and 93% (13, 15). On the contrary, for children few data are available, and there are still many questions without unanimous or clear answers, such as, is FNAB a safe procedure for children? Some consider this procedure of limited usefulness in children because of its discomfort and the high rate of side-effects such as papillary endothelial hyperplasia, hemorrhage, vascular proliferation, vascular thrombosis, fibrosis, cystic change, infarction, and abscess (6, 7, 8, 15, 20). However, in our series most pediatric patients who underwent FNAB did not develop any complications, which agrees with other data (9, 17). Some researchers advocate surgery for the majority of children with solitary thyroid nodules (7). In our experience, considering only the cases with solitary thyroid nodules, we have seen a histological benign lesion in 13 of 27 submitted to surgery and a cytological benign lesion (obtained by FNAB) in 28. If we also consider the cases with 2 or more nodules, a histological benign lesion was found in 9 of 15 cases submitted to surgery and in 17 submitted only to FNAB. In conclusion, the frequency of malignant lesions was only 23% (20 of 87). Consequently, surgery is not recommended as the initial approach to thyroid nodules.
When is FNAB recommended? Some researchers recommend it only when nodules are palpable (21), because epidemiological reports rarely show thyroid cancer in impalpable nodules. Furthermore, to perform FNAB every time that a nodule is accidentally revealed by US would exaggerate the use of this procedure; in addition, the most frequent thyroid cancer in childhood and adolescence is the papillary type (10), which is usually a slow-growing tumor with a very indolent course even after local and pulmonary metastases have occurred (18). However, other researchers show that sometimes even nodules larger than 2 cm may be revealed only by US (22, 23). In fact, physical examination is relatively effective in detecting the nodules localized in the isthmus or in the anterior surface, but is much less so for the nodules localized in the upper pole of the gland, even if they are large. Moreover, thyroid cancer has been diagnosed even in very small nodules (22, 24, 25); in addition, some report that up to 10% of children with thyroid cancer may die of the disease, which is very worrisome (26, 27). Therefore, different researchers suggest different nodule cut-off dimensions to perform FNAB: more than 1 cm (28), 1.5 cm (14), or 2 cm (21). We think that other factors that derive from clinical history must be considered, such as previous external radiation for other malignancies of the head and the neck, family history of thyroid cancer, rapid nodule growth, ecographic characteristics, and suspicious cervical adenopathy.
In our series FNAB was performed in patients with both palpable and impalpable nodules. In these latter (six patients), nodule dimensions ranged from 0.8–1.5 cm, and we decided to perform FNAB precisely because other risk factors for malignancy were present: external radiation for leukemia in one patient, cervical adenopathy associated with subjective symptoms of neck compression in another, rapid spontaneous nodule growth in two patients, rapid growth of a nodule in one treated with L-T4, and suspicious isolated adenopathy in the last patient. The first four children had ML.
The main questions, however, are whether FNAB is the most useful procedure to detect malignancy, whether surgical treatment is needed, and whether this approach compares favorably with conventional clinical, laboratory, and imaging studies. In our series FNAB showed the highest sensitivity, specificity, and accuracy. On the contrary, some researchers report an accuracy lower than those of US and scintigraphy (17), but these data are probably due to small study groups. Other researchers show a correlation between FNAB reports and a histological diagnosis of 100% in a small cohort of patients (18). Furthermore, in our series no side-effects were recorded, which confirms that FNAB is a safe technique that clearly identifies nodules with malignant or suspicious cytology warranting surgery.
In our patients, signs and symptoms of neck compression were similar in the two groups, indicating that benign lesions (e.g. goiter) may lead to discomfort. On the contrary, the frequency of cervical adenopathy was significantly higher in thyroid cancer (78.9%), as noted by others (18, 29), confirming the importance of an accurate physical examination.
The only patient with a high serum calcitonin level was affected with medullary carcinoma (30). The association between autoimmune thyroiditis and thyroid carcinoma is still a matter of debate. We diagnosed papillary carcinoma in three cases with thyroid autoimmune diseases (one Graves’ disease and two thyroiditis), but this number is too small to determine an association in children (7, 10, 31, 32).
US revealed that most nodules were hypoechogenic lesions, both in the group with ML (58.7%) and in the group with BL (28.6%). We found this of limited value in detecting malignancy, because most hypoechogenic nodules are benign. Moreover, hyperechogenic lesions do not exclude malignancy, so none of US-revealed features are sufficiently specific to guide the selection of patients affected by thyroid cancer (33, 34, 35). Also, evaluation of thyroid nodules with color Doppler sonography has not given any specific patterns for malignancy (36). US also identifies the number, dimensions, and progressive growth of nodules. We found that the number of nodules was not related to malignancy, and in particular, the solitary nodule was not significantly more likely to be malignant, but this remains controversial (3, 4, 5, 18, 37) .
Nodule dimensions were not related to malignancy in our patients; 55% of children with ML had a malignant nodule with a diameter less than 2 cm, as reported by others (22, 25). On the contrary, adult series report a higher frequency of thyroid cancer in nodules larger than 4 cm (38). Progressive growth was present in 57% of our group with BL and in 72% of the group with ML. Although this is usually considered predictive of malignancy, it was not so in our patients, three of whom with BL, treated with L-T4, had progressive growth. These findings confirm previous observations describing poor correlation between progressive growth and nodule histology in both adults and children (6, 39).
Thyroid scintiscanning has long been considered the first examination to perform on a thyroid nodule. However, in recent years it has been less frequently used in the initial routine evaluation because most benign and malignant nodules show reduced concentrations of radioisotope, whereas hyperfunctional nodules are occasionally malignant (6, 11, 14, 16). In our series, scintiscanning revealed only hypofunctional and normally functioning nodules, without any relation to malignancy, and failed to reveal any lesions in more than 20% of patients. Therefore, thyroid scintiscanning is useless in detecting malignancy and should be performed only in thyrotoxic patients with thyroid nodules (16).
Repetition of FNAB and/or excision biopsy must be considered if clinical or anamnestic risk factors for malignancy are present (as in one of our patients with negative cytology, but malignant histology). Among the risk factors are nodule growth (especially in cases treated with L-T4), signs or symptoms of cervical compression, and previous therapy for other malignant diseases. With the exception of these cases, we do not perform excisional biopsy in patients with benign lesions detected with FNAB. However, all follicular lesions must undergo surgical evaluation.
In conclusion, in agreement with previous studies in adults that demonstrate the high accuracy of FNAB in detecting malignancy, this study provides evidence that in childhood and adolescence this procedure also shows the greatest sensitivity, specificity, and accuracy compared with other tests. FNAB significantly contributed to the diagnosis of thyroid cancer in our patients, as shown by multiple regression analysis. In addition, when performed properly, FNAB is a safe technique, and we therefore recommend it as the first diagnostic test in euthyroid pediatric patients with thyroid nodule(s).
Acknowledgments
Footnotes
Abbreviations: BL, Benign lesions; FNAB, fine needle aspiration biopsy; ML, malignant lesions; US, ultrasonography.
Received June 16, 2000.
Accepted July 12, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  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]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
