Pediatric differentiated thyroid carcinoma in stage I: risk factor analysis for disease free survival

Background To examine the outcomes and risk factors in pediatric differentiated thyroid carcinoma (DTC) patients who were defined as TNM stage I because some patients develop disease recurrence but treatment strategy for such stage I pediatric patients is still controversial. Methods We reviewed 57 consecutive TNM stage I patients (15 years or less) with DTC (46 papillary and 11 follicular) who underwent initial treatment at Ito Hospital between 1962 and 2004 (7 males and 50 females; mean age: 13.1 years; mean follow-up: 17.4 years). Clinicopathological results were evaluated in all patients. Multivariate analysis was performed to reveal the risk factors for disease-free survival (DFS) in these 57 patients. Results Extrathyroid extension and clinical lymphadenopathy at diagnosis were found in 7 and 12 patients, respectively. Subtotal/total thyroidectomy was performed in 23 patients, modified neck dissection in 38, and radioactive iodine therapy in 10. Pathological node metastasis was confirmed in 37 patients (64.9%). Fifteen patients (26.3%) exhibited local recurrence and 3 of them also developed metachronous lung metastasis. Ten of these 15 achieved disease-free after further treatments and no patients died of disease. In multivariate analysis, male gender (p = 0.017), advanced tumor (T3, 4a) stage (p = 0.029), and clinical lymphadenopathy (p = 0.006) were risk factors for DFS in stage I pediatric patients. Conclusion Male gender, tumor stage, and lymphadenopathy are risk factors for DFS in stage I pediatric DTC patients. Aggressive treatment (total thyroidectomy, node dissection, and RI therapy) is considered appropriate for patients with risk factors, whereas conservative or stepwise approach may be acceptable for other patients.


Patient
Between 1962 and 2004, 68 pediatric patients (≤ 15 years) with DTC were initially treated at Ito Hospital, which has a special interest in thyroid disease. Among the 68 consecutive patients, 11 patients were defined as TNM stage II because of the presence of synchronous lung metastasis. The remaining 57 TNM stage I patients (7 males and 50 females) were reviewed in this study. The mean age (mean ± standard deviation [SD]) was 13.1 ± 2.3 (range: 7-15) years. The mean follow-up period was 17.4 ± 10.9 (range: 0.7-45.0) years. In 42 patients (73.7%), the duration of follow-up was over 10 years. Their mean height and body weight were 152.7 ± 12.4 (range: 118.3-172.0) cm and 45.4 ± 12.2 (range: 18.0-73.0) kg, respectively. All 57 patients in our series underwent initial treatment at our institution and none of the patients had undergone any previous treatment for DTC elsewhere. Thus, patients with any prior management before diagnosis and therapy in our institution were excluded from the present analysis.

Histology
There were 46 PTCs including 2 follicular variant types and 11 FTCs. Overall, the primary tumor size was 4.3 ± 2.0 cm. Specific variants, such as follicular variant of papillary carcinoma, solid and trabecular pattern, squamous metaplasia or marked psammoma bodies, were found in 15 patients. Hürthle cell carcinoma or insular carcinoma was not found in our series. Medullary thyroid carcinoma or anaplastic thyroid carcinoma were excluded.

Clinical manifestation at diagnosis
The clinical manifestations at diagnosis were thyroid mass in 53 (93.0%), extrathyroid extension in 7 (12.3%), clinical lymphadenopathy in 12 (21.1%), and hoarseness due to RLN palsy in 2 (3.5%) of all 57 patients. No patients presented with synchronous lung metastasis at diagnosis but three patients developed metachronous lung metastasis in the follow up period. In only one patient, PTC was incidentally found by ultrasound examination due to screening during the treatment for Graves' disease.

Family or radiation history
Eleven patients had a family history of thyroid disease (PTC in one, benign thyroid tumor such as adenomatous goiter or follicular adenoma in 6, and Graves's disease in 4). There were no patients who had previous radiation exposure to the head and neck.

Preoperative diagnostic modality
Fine needle aspiration biopsy cytology (FNABC) has been performed since 1974 and ultrasound (US) examination has been performed as a routine preoperative practice to study primary thyroid tumor and regional lymph nodes since 1980 at Ito Hospital. US-guided FNABC has also been performed as a novel diagnostic tool since 1998. Overall, FNABC was performed preoperatively in 36 patients. Cytological diagnoses were PTC, strongly suspected PTC, suspected FTC, and follicular adenoma in 25, 2, 3, and 6 of the 36 patients, respectively. Consequently, conventional or US-guided FNABC contributed to the diagnosis of thyroid carcinoma in 30 of 36 (83.3%) patients before surgery.
Chest X-ray was preoperatively performed as one of the routine examinations for total anesthesia and as a screening purpose of apparent lung metastasis in all patients.
Computed tomography (CT) scan was also performed before surgery to evaluate local disease and the presence or absence of synchronous lung metastasis since 1987.
The level of serum thyroglobulin (Tg) was measured since 1981 and the presence of Tg antibody was evaluated since 1994. In 7 patients, the information of serum Tg was not available. Finally, the level of serum Tg was considered reliable in 15 patients without Tg antibody at the end of follow-up. Thus, the serum Tg was valuable to diagnose disease recurrence or progression only in selected patients in our series. Therefore, the level of serum Tg was not evaluated statistically in this study.

Surgical treatment
Lobectomy and subtotal or total thyroidectomy were performed in 25 and 21 of 46 PTC patients, and in 9 and 2 of 11 FTC patients, respectively. Of the 46 PTC patients, 10, 1, 31, and 4 patients underwent no dissection, only central node dissection, ipsilateral modified neck dissection (MND), and bilateral MND, respectively. Of the 11 FTC patients, 6, 2, and 3 patients underwent no dissection, only central node dissection, and ipsilateral MND, respectively. MND was defined as systematic node dissection in the central compartment (level VI) and ipsilateral lateral compartment (level II to V). Fundamentally, lobectomy was performed in patients with unifocal and unilateral tumor, no clinical lymphadenopathy, and no distant metastasis at diagnosis. Total thyroidectomy was performed in patients with multifocal or invasive tumor, clinical lymphadenopathy, or distant metastasis at diagnosis. Subtotal thyroidectomy was performed to avoid surgical complications, such as permanent hypoparathyroidism or RLN palsy, in selected patients who were basically candidates for total thyroidectomy. For example, patients presenting with ipsilateral RLN palsy underwent less total thyroidectomy to prevent bilateral RLN palsy. MND was usually performed in patients with advanced tumor stage or lymphadenopathy, whereas no or only central node dissection was performed in those with lower tumor stage or no lymphadenopathy.

Pathological diagnosis and TNM staging
Primary tumors and regional lymph nodes were removed in all patients and those specimens were pathologically confirmed to be PTCs or FTCs and nodal metastases.
The sixth edition of the AJCC/UICC TNM classification was used to determine the tumor (T) stage, nodal (N) stage, distant metastasis (M) stage, and TNM staging. Among the 68 pediatric patients, 57 patients without synchronous distant metastasis were defined as TNM stage I and were analyzed statistically in this study. Eleven patients with synchronous lung metastasis were defined as TNM stage II and were excluded from the present analysis.

Radiotherapy
Therapeutic RI treatment has principally been performed when refractory nodal recurrence was observed or when distant metastasis was detected visually by chest X-ray or CT scan. Overall, 6 patients underwent RI ablation and 4 patients received therapeutic dose of RI because 3 patients developed both nodal recurrence and metachronous lung metastasis and one patient developed nodal recurrence.
External beam radiation therapy (EBRT) was performed for only one patient.
TSH suppression therapy TSH suppression therapy was administered to all patients but it was adequately maintained in 48 of 57 patients. In the remaining 9 patients, TSH suppression was inadequate due to patient compliance, or was not maintained because of the curability of the disease (unilateral tumor and no pathological node metastasis). These 9 patients did not develop any recurrent disease at the time of the present analysis.

Therapeutic complications
Overall, permanent surgical complications considered iatrogenic were found in 9 (15.8%) of 57 patients. Permanent hypoparathyroidism was observed in 6 (10.5%) patients (subtotal thyroidectomy in one and total thyroidectomy in 5). RLN palsy considered due to the nerve injury not due to the potential risk of primary disease was observed in 3 (5.3%) patients (subtotal thyroidectomy in one and total thyroidectomy in 2). None of the 10 patients (3 males and 7 females) who received RI ablation or therapy developed any other secondary malignant disease. Among the 7 female patients, 2 patients experienced 3 healthy deliveries.

Follow-up and definition of disease recurrence
Patient follow-up was generally performed every 6 months after a relatively stable condition was obtained. Recurrent disease was defined as local recurrence (remnant thyroid tissue or regional lymph node) or metachronous distant metastasis that developed as new disease during the follow-up period, at least 6 months after the initial surgery. Recurrence was detected by physical and radiological examinations, including US, chest X-ray, CT scan, and/or scintigraphy, and blood test to screen for elevation of the serum Tg in patients treated with total thyroidectomy.

Evaluation of outcomes
Clinicopathological features, treatments, and outcomes, such as age stratification (≤ 13 years old), male gender, histology (papillary carcinoma), specific variant, tumor size (>3.0 cm), extrathyroid extension, multifocality, tumor stage (T3, 4a), clinical lymphadenopathy (cN), pathological node metastasis (pN), the extent of thyroid resection and node dissection, surgical complications, and the rates of disease recurrence and mortality were evaluated. These factors were compared between patients who developed recurrence and those who did not. Disease-free survival (DFS) was evaluated in all and univariate and multivariate analyses were then performed to reveal risk factors for DFS. Disease-specific survival was not evaluated because no patients died of disease. This study was approved by the Ethical boards of our institution. Informed consent was obtained from patients and their parents.

Statistical analysis
The results are expressed as the mean ± standard deviation (SD). Frequencies were compared with the chi-square test and Fisher's exact probability test. DFS was assessed with the Kaplan-Meier method and compared with the logrank test. The effects of each risk factor on DFS were evaluated by multivariate analysis using the Cox proportional hazards models [Hazard ratio (HR), 95% confidence interval (CI)]. Differences were considered statistically significant when p-values were less than 0.05. Statistical analyses were performed with SPSS, version 11.0. Table 1 summarizes the clinicopathological features and outcomes. 26.3% (15 patients) exhibited recurrent disease (local in 12 and both local and lung metastasis in 3). In the 15 patients (regional node recurrence in 13 and both remnant thyroid and regional node recurrence in 2), the clearance of local recurrences was achieved successfully. After the treatment for recurrence, 10 of the 15 patients are currently free of disease. Four patients with nodal recurrence and one patient with both nodal recurrence and lung metastasis are alive with disease.
Optimal treatment strategy and significant risk factors for stage I pediatric DTC patients are not well established. Therefore, we analyzed the risk factors in TNM stage I patients. Total thyroidectomy, node dissection, RI therapy, and TSH suppression therapy are considered superior treatment options for pediatric DTC patients in many centers. However, less aggressive treatments are also recommended in other centers. Newman et al. [21] suggested that the selection of the extent of thyroidectomy is warranted, i.e., lobectomy for localized tumor and subtotal or total thyroidectomy for extensive tumor. We consider that the appropriate treatment strategy should be selected to handle primary or recurrent disease and to avoid surgical complications according to risk factors. Such considerations would improve the DFS and the patient's quality of life in pediatric age population. Our results suggest that male gender, advanced tumor stage, and clinical lymphadenopathy are risk factors for DFS in TNM stage I pediatric DTC patients. In the literature, advanced tumor stage, multifocality, lymphadenopathy, distant metastasis, and lower patient age are considered major risk factors for worse prognosis [6,11,[14][15][16][17]20,21]. Male gender is also a risk factor in pediatric patients [2,11,14]. Thus, TNM stage I pediatric patients may be categorized separately into different substage groups according to risk factors, for example stage IA or stage IB, and a conservative or stepwise approach may be acceptable in pediatric patients without risk factors.
Finally, long-term follow-up is considered essential to evaluate outcomes in pediatric patients [6,35]. A shorter follow-up period may result in inappropriate decision making of the treatment strategy. In the literature, the durations of follow-up period range from 3.8 to 27.6 years [1][2][3][4][5][6][31][32][33][34][35]. The mean follow-up period in our study was 17.4 years, which is considered sufficient for evaluating outcomes in pediatric patients.

Conclusion
Male gender, advanced tumor stage, and clinical lymphadenopathy are risk factors for DFS in TNM stage I pediatric DTC patients, although many patients with recurrence can achieve disease-free after additional treatments. Aggressive treatment (total thyroidectomy with node dissection followed by RI therapy) is appropriate for stage I pediatric patients with risk factors, whereas conservative approach may be acceptable for those without risk factors.