Pediatric Thyroid Cancer: Postoperative Classifications and Response to Initial Therapy as Prognostic Factors

Transcription

1 ORIGINAL ARTICLE Pediatric Thyroid Cancer: Postoperative Classifications and Response to Initial Therapy as Prognostic Factors Liora Lazar, Yael Lebenthal, Karl Segal, Adam Steinmetz, Yulia Strenov, Maya Cohen, Isaac Yaniv, Michal Yackobovitch-Gavan, and Moshe Phillip The Jesse Z. and Sara Lea Shafer Institute for Endocrinology and Diabetes, National Center for Childhood Diabetes (L.L., Y.L., M.Y.-G., M.P.), Department of Pediatric Hematology-Oncology (I.Y.), Schneider Children s Medical Center of Israel, Petah Tikva 49202, Israel; Departments of Otolaryngology and Head and Neck Surgery (K.S.), Nuclear Medicine (A.S), and Imaging (M.C.), and Institute of Pathology (Y.S.), Rabin Medical Center, Beilinson Campus, Petah Tikva 49100, Israel; and Sackler Faculty of Medicine (L.L., Y.L., K.S., M.C., I.Y., M.P.) Tel Aviv University, Tel Aviv 69978, Israel Context: Prognostic factors for pediatric differentiated thyroid cancer (DTC) are not well established. Objective: The objective of the study was to retrospectively compare the postoperative risk-stratification systems: American Thyroid Association (ATA) risk categories, Schneider Children s Medical Center of Israel (SCMCI) score, and the response to initial therapy as predictors for disease outcome. Patients and Methods: Fifty-four DTC patients, median age at diagnosis 13.9 years (range y), followed up for a median of 8.8 years (range y) were stratified into prepubertal (n 9), pubertal (n 25), and postpubertal (n 20) groups. All patients underwent total/near-total thyroidectomy; 48 received radioiodine therapy. The extent of DTC was evaluated by applying the ATA risk categories and the novel SCMCI score. Postoperative risk stratifications (low/intermediate/ high) were determined using histopathological, laboratory, and imaging findings. Response to initial therapy (complete/acceptable/incomplete) was based on stimulated thyroglobulin and imaging results during the first 2 years of follow-up. Results: The risk for recurrent/persistent disease, as assessed by the postoperative ATA risk-stratification system and the SCMCI score and by the response to initial therapy, was higher in the prepubertal group (P.001, P.002, and P.02, respectively). Outcome prediction by the risk-stratification systems was applicable: ATA risk categories, P.014, R , predictive ability 80.4%; SCMCI score, P.001, R , predictive ability 86.3%; and response to initial therapy stratification, P.001, R , predictive ability 96.1%. The proportion of variance explained by the ATA risk categories (0.25), SCMCI score (0.44), and response to initial therapy (0.79) indicated that the latter was the most precise predictor and that the SCMCI score reflected the disease outcome better than ATA risk categories. Conclusions: Our data confirm that the postoperative pediatric ATA stratification system and the novel SCMCI score are suitable for predicting the risk of recurrent/persistent disease in this population. The response to initial therapy classification performed 1 2 years after the initial therapy may be more appropriate for guiding surveillance recommendations. (J Clin Endocrinol Metab 101: , 2016) ISSN Print X ISSN Online Printed in USA Copyright 2016 by the Endocrine Society Received November 15, Accepted February 22, First Published Online March 1, 2016 Abbreviations: ATA, American Thyroid Association; CT, computed tomography; DTC, differentiated thyroid carcinoma; FNA, fine-needle aspiration; FVPTC, follicular variant of PTC; LN, lymph node; PTC, papillary thyroid carcinoma; RAI, radioactive iodine; SCMCI, Schneider Children s Medical Center of Israel; Tg, thyroglobulin; US, ultrasound; WBS, wholebody scan press.endocrine.org/journal/jcem J Clin Endocrinol Metab, May 2016, 101(5): doi: /jc

2 doi: /jc press.endocrine.org/journal/jcem 1971 Differentiated thyroid carcinoma (DTC) accounts for 0.5% 3% of all pediatric malignancies, with an annual incidence of case per children (1, 2). Most of the affected patients younger than 18 years are diagnosed during adolescence, but the disease may also occur in childhood (1 11). The behavior of DTC in the pediatric population differs by age group; the disease may be more aggressive, with a less favorable prognosis, in prepubertal children (3, 6 13). In an attempt to better assess the long-term prognosis and thereby to improve the management approach, stratification of adults with DTC into risk levels according to postoperative tumor classification has been suggested by the American Thyroid Association (ATA) (14). The several prognostic staging systems validated in adults were found not to be applicable in pediatric patients (15 17). In the recently published ATA guidelines on pediatric thyroid cancer, the American Joint Cancer Committee tumor node metastasis staging system, based on the evidence of regional invasion and metastasis, was recommended as the most appropriate scale for assessing the risk of recurrent/persistent disease in pediatric patients (18). It was further suggested that follow-up findings in the postoperative stage could be important in tailoring decisions regarding the need for radioactive iodine (RAI) ablation and the degree of TSH suppression as well as the frequency and modality of follow-up studies. It is well established that in adult patients with DTC the disease outcome is also associated with the response to initial treatment. Tuttle et al (19) in 2010 reported a risk-stratification scheme that integrated the effect of initial therapy with the ATA risk stratification of the disease. The clinical data obtained during their first 2 years of follow-up, including suppressed and stimulated thyroglobulin (Tg) levels, neck ultrasound (US), and whole-body scan (WBS) after RAI administration, were used to categorize the response to therapy. They found that the response to initial therapy altered the postoperative risk estimates and therefore provided more accurate prognostic information. To the best of our knowledge, the impact of the response to initial therapy on the course and outcome of DTC has never been studied in the pediatric population. In the past 20 years, 54 pediatric patients with DTC have been treated in our single tertiary care center at the Schneider Children s Medical Center of Israel (SCMCI). The objectives of this retrospective study were to evaluate the clinical and pathological characteristics and outcome of DTC in prepubertal, pubertal, and postpubertal patients; to compare the ATA risk-stratification system, the novel SCMCI score developed by our center, and the response to initial therapy stratification as predictors for persistence/recurrence of disease; and to assess which of the risk-stratification systems may best predict the course and outcome. Patients and Methods Patients The 54 DTC pediatric patients were diagnosed in our center between 1994 and The tumor was sporadic in 48 patients, part of a genetic syndrome associated with multiple tumors in three patients (one with APC associated polyposis, two with PTEN hamartoma syndrome), and postexternal irradiation in three patients (for treatment of acute lymphatic leukemia, Wilms tumor, and choroid plexus carcinoma). Throughout the follow-up period, all were treated by a multidisciplinary team (pediatric endocrinologists; ear, nose, and throat surgeons; nuclear medicine physicians; and oncologists). Initial treatment was based on surgery, RAI administration, and thyroid hormone therapy. All patients underwent total/near-total thyroidectomy, in either a one- or two-step procedure (the latter in cases of indeterminate fine needle aspiration [FNA] cytology and DTC on pathology), with cervical central compartment lymph node (LN) dissection as indicated and uni-/bilateral neck dissection when suspicious LNs were present. In 48 patients, postoperative I 131 treatment was given 4 6 weeks after surgery for the ablation of thyroid remnants or distant metastases, with a selected adult dose (tumor limited to the thyroid gland, mci; tumor invading the thyroid capsule, surrounding tissues and/or with metastases in the neck or mediastinal LNs, 150 mci; distant metastases, mci) scaled down according to the pediatric patient s weight, using the following formula: pediatric dose adult dose patient s body weight (kilograms)/70 (kilograms) (9). One prepubertal girl who was lost to follow-up and five adolescents in whom the tumor was less than 1 cm and confined to one thyroid lobe did not receive RAI therapy. All patients received TSH-suppressive (TSH 0.05 miu/l) doses of L-thyroxine. Follow-up included the following: physical examination; testing of thyroid functions; determination of levels of serum Tg and Tg antibodies while on TSH-suppressive and on TSH-stimulated, off-thyroid hormone therapy or after administration of two injections of recombinant human TSH (Thyrogen; Genzyme Corp); neck US; and diagnostic WBS with I 131 (2 4 mci). Intervals between evaluations ranged from 4 to 6 months during the first 2 3 years. In patients with complete remission, repeated evaluation was performed every 1 2 years for the subsequent 4 years and every 5 years thereafter. Patients with suspicion of persistent or recurrent disease underwent extensive evaluation followed by repeated I 131 treatment or additional surgery as required. In assessing the outcome, disease free was considered to be expressed by two consecutive undetectable ( 2 ng/ml) TSH-stimulated Tg levels and negative diagnostic WBS, with no anatomical evidence of disease on neck US; persistent disease by continuously TSH-suppressed Tg levels greater than 1 ng/ml or TSH-stimulated Tg levels greater than 2 ng/ml, positive WBS, and/or anatomical evidence of disease on imaging for more than 2 years after initial therapy; and recurrent disease by evidence of disease detected on physical examination or identified by neck US, by increased Tg levels, and/or by positive WBS in a patient

3 1972 Lazar et al Pediatric DTC: Prognostic Classification Systems J Clin Endocrinol Metab, May 2016, 101(5): who had been considered disease free for at least 12 months after initial therapy. This retrospective study was approved by our institutional review board. Methods The data obtained from the medical file of each patient included the following: demographic data; possible predisposing factors, eg, Hashimoto thyroiditis, familial DTC (first degree relative/s with DTC); genetic syndromes associated with multiple tumors; previous exposure to external irradiation; presenting complaint and clinical findings; FNA cytology results; initial treatment (extent of operation, I 131 dose); initial pathology and WBS findings; laboratory (Tg and Tg antibodies) and imaging (cervical US, chest computed tomography [CT] or magnetic resonance imaging, or WBS) findings from diagnosis throughout follow-up; additional treatments; course; and outcome. Laboratory evaluation included thyroid function tests; antithyroid antibodies (normal reference: antimicrosomal 75 IU/mL and anti-tg 150 IU/mL) initially measured by the hemagglutination method and in recent years by an ELISA (Orgentec Diagnostika GMBH); and Tg levels measured using a chemiluminescence assay (Immulite, 2000; Diagnostic Products Corp). Sonography was performed by the same highly experienced radiologist specialized in neck US, using the high-resolution GE VOLUSON Expert, Philips IU22, or the Philips HDL 5500 US machine (Philips ATL). FNA was performed and interpreted by the same skilled cytopathologist; it was initially reported as nondiagnostic, benign, atypical, or malignant (20) and since 2009 according to the Bethesda classification (21). Pathological findings were recorded from the original pathology reports and included tumor size and type (papillary thyroid carcinoma [PTC], follicular variant of PTC [FVPTC], follicular thyroid carcinoma); evidence of tumor multifocality; tumor extension; lymphocytic infiltration; vascular invasion; cervical LN metastases; and extracapsular LN invasion. Lung metastases were detected by RAI scan and/or by chest CT: RAI scan by focal I 131 uptake for focal pulmonary metastases; bilateral diffuse homogenous I 131 uptake by diffuse lung metastases. Chest CT detected distinct nodules for focal pulmonary metastases or metastases disseminated throughout the pulmonary parenchyma for diffuse lung metastases (22). Postoperative staging system The extent of DTC was evaluated in each patient by the recently proposed ATA risk-stratification system (18) and also by applying a novel numerical SCMCI score developed by us. The SCMCI score included eight parameters, each ranging from 0 to 3 points (Table 1). The numerical SCMCI score was divided into tertiles corresponding to the three risk-groups: low risk (score 0 3), intermediate risk (score 4 7), and high risk (score 8 14). In both classification systems, the risk grade (low risk 1; intermediate risk 2; high risk 3) was used to compare the possibility of recurrent/persistent disease in each group and to determine its association with the course and outcome of the disease (Table 2). Response to initial therapy was assessed between 1 and 2 years from diagnosis using the TSH-stimulated Tg levels, the diagnostic WBS results, and the neck US findings. The stratification scheme was adapted from that for adult DTC patients (19). The medical status of each patient was categorized into three groups: complete response, acceptable response, and incomplete response. The response grade (complete 1; acceptable 2; incomplete 3) was used for comparison of the response to initial therapy in each group and for determining its association with course and outcome of the disease (Table 2). All patients had adequate information (based on histopathological and imaging findings at presentation) for the analysis by the ATA risk-stratification system and the SCMCI score; only 51 patients had sufficient information for evaluation of the response to initial therapy during the first 2 years of follow-up (two patients were lost to follow-up and one was diagnosed only 13 months prior to analysis). Table 1. SCMCI Score: A Novel Postoperative Staging System for Pediatric DTC Tumor foci within the Unifocal Multifocal thyroid gland Tumor involvement One lobe Two lobes Tumor invasion Confined to the thyroid gland with no capsular invasion Confined to the thyroid gland, non-encapsulated Minimal extraglandular microscopic invasion into the perithyroidal soft tissues Extensive extraglandular invasion to sc soft tissues, larynx, trachea, esophagus, or recurrent laryngeal nerve Vascular invasion No tumor spread into blood vessels Tumor spread into one or more blood vessels Cervical LN No Only in the central Ipsilateral Bilateral compartments metastases compartment compartments Extracapsular LN No Yes invasion Lung metastases No Micronodular focal Micronodular diffuse Macronodular Postoperative TSHstimulated Tg levels, ng/ml Undetectable 2 to to

4 doi: /jc press.endocrine.org/journal/jcem 1973 Table 2. Definition of Risk Groups for Recurrent/Persistent DTC According to Postoperative Staging and Response to Initial Therapy Classifications Postoperative Staging Low Risk (Grade 1) Intermediate Risk (Grade 2) High Risk (Grade 3) ATA a Disease confined to the thyroid No local or distant metastases Extensive LN metastases to cervical level VI compartment or minimal invasion to cervical levels I, II, III, Regionally extensive disease or locally invasive disease with or without distant metastasis No invasion of locoregional tissues IV, or V SCMCI score Response to Therapy b Complete (Grade 1) Acceptable (Grade 2) Incomplete (Grade 3) Tg levels Suppressed 1 ng/ml, stimulated 2 ng/ml Suppressed 1 ng/ml, stimulated 2 to 10 ng/ml Neck US No evidence of disease Residual gland stable ( ) LNs 1 cm WBS No I 131 uptake Nonspecific I 131 uptake in the cervical region a American Thyroid Association risk of recurrence classification (18). b Tuttle et al (19). Suppressed 1 ng/ml Stimulated 10 ng/ml or rising values Persistent or newly identified disease ( ) LNs 1 cm ( ) I 131 uptake in the cervical region and/or in the lungs Statistical methods Data analysis was performed using SPSS software, version 22.0 (SPSS, Inc). Differences between the three groups (prepubertal, pubertal, and postpubertal) were tested using a one-way ANOVA for variables with normal distribution, 2 tests for categorical variables, and the independent-samples Kruskal-Wallis test for continuous variables with skewed distribution. Stepwise multiple logistic regression analyses were conducted to predict disease outcome, and a receiver-operating characteristic curve analysis was used to define the optimal cutoff point within the SCMCI score, which may discriminate between disease-free and persistent disease. The proportion of variance explained was estimated using Nagelkerke s method and the R 2 option in PROC LOGISTIC of SAS/SAT (19, 23). A value of P.05 was considered significant. Results Demographic and clinical data, tumor characteristics, and initial therapy (Table 3) Age at diagnosis of the entire cohort ranged from 1.9 to 17 years, with a median of 13.9 years. Of the 54 patients, nine were prepubertal (Tanner stage 1), 25 pubertal (Tanner stage 2 4), and 20 postpubertal (Tanner stage 5). The male to female ratio was 1:3.5. Predisposing factors, found in more than half of the cohort (32 of 54 patients), were present in only one prepubertal patient (P.012). Most pubertal and postpubertal patients presented with an asymptomatic thyroid nodule, whereas in the prepubertal group, diffuse firm goiter associated with hoarseness and stridor was more common (three of nine vs 1 of 25 vs 0 of 20, P.004). Cervical lymphadenopathy was more prevalent in the prepubertal group (eight of nine vs 6of25vs3of20,P.002). The initial FNA results and the pathological subtypes showed a similar distribution in the three groups. The postoperative histopathological findings revealed significantly larger tumors in the pubertal and postpubertal patients (P.014). The diffusely infiltrative form (massive diffused spread of papillary tumor throughout the entire parenchyma) was mostly found in the prepubertal patients (three of nine vs 0 of 25 vs 1 of 20, P.004). The presence of multifocal tumor, bilateral lobe involvement, and background thyroiditis was similar in the three groups. Tumor extension was significantly more prominent in the prepubertal patients: tumor invasion (P.001); blood vessel invasion (P.002); cervical LN metastases (P.003); and lung metastases (P.001). Postoperative Tg levels were higher in the prepubertal group (340 vs 2 ng/ml, P.001). Positive Tg-antibody was detected in 3 of 54 patients (two pubertal, one postpubertal) at the initial postoperative evaluation; during the first two-year surveillance period, all patients became Tg antibody negative. The absolute initial RAI dose ranged from 10 to 180 mci RAI. The prepubertal children with more invasive disease received a significantly higher weight-adjusted dose than the adolescents (P.004). Post-RAI scan showed that the percentage of patients with positive I 131 uptake in the cervical region (only in thyroid bed or in thyroid bed and regional LN) was similar in the three groups, whereas lung metastases were more prevalent in the prepubertal than in the pubertal and postpubertal patients (seven of eight vs 5 of 23 vs 3 of 17, P.001).

5 1974 Lazar et al Pediatric DTC: Prognostic Classification Systems J Clin Endocrinol Metab, May 2016, 101(5): Table 3. Clinical and Morphological Characteristics of DTC in 54 Children and Adolescents All (n 54) Prepubertal (n 9) Pubertal (n 25) Postpubertal (n 20) P Value CA at diagnosis (mean SD) Males 12 (22.2) 4 (44.4) 4 (16.0) 4 (20.0).203 Predisposing factors Hashimoto thyroiditis 15 (27.8) 0 9 (36.0) 6 (30.0).113 Familial DTC 11 (20.4) 1 (11.1) 8 (32.0) 2 (10.0).143 Malignant syndromes 3 (5.6) 0 2 (8.0) 1 (5.0).662 External irradiation 3 (5.6) 0 1 (4.0) 2 (10.0).244 All factors 32 (59.2) 1 (11.1) 20 (80) 11 (55.0).012 Initial clinical symptoms Single nodule 27 (50.0) 5 (55.6) 11 (44.0) 11 (55.0).715 MNG 23 (42.6) 1 (11.1) 13 (52.0) 9 (45.0).101 Diffuse goiter 4 (7.4) 3 (33.3) 1 (4.0) Cervical LN 17 (31.5) 8 (88.9) 6 (24.0) 3 (15.0).002 Hoarseness 4 (7.4) 3 (33.3) 1 (4.0) FNA Benign 2 (3.7) 0 1 (4.0) 1 (5.0) Suspicious/follicular 24 (44.4) 3 (33.3) 14 (56.0) 7 (35.0).538 Malignant 28 (51.9) 6 (66.7) 10 (40.0) 12 (60.0) Operation Total central ND 39 (72.1) 3 (33.3) 20 (80.0) 16 (80.0) Total unilateral ND 11 (20.4) 4 (44.4) 4 (16.0) 3 (15.0).061 Total bilateral ND 4 (7.4) 2 (22.2) 1 (4.0) 1 (5.0) Re-operation 4 (7.4) 0 2 (8.0) 2 (10).659 Tumor pathology Tumor type PTC 30 (55.6) 7 (77.8) 12 (48.0) 11 (55.0) FVPTC 22 (40.7) 2 (22.2) 12 (48.0) 8 (40.0).632 FTC 2 (3.7) 0 1 (4.0) 1 (5.0) Tumor size 1 cm 2 (3.7) 1 (11.1) 0 1 (5.0) 1 cmto 2 cm 13 (24.1) 2 (22.2) 9 (36.0) 2 (10.0) 2 cmto 4 cm 25 (46.3) 1 (11.1) 12 (48.0) 12 (60.0) 4 cm 10 (18.5) 2 (22.2) 4 (16.0) 4 (20.0) Diffuse 4 (7.4) 3 (33.3) 0 1 (5.0).004 Tumor characteristics Multifocality 30 (55.6) 5 (55.6) 12 (48.0) 13 (65.0).522 Two-lobe involvement 19 (35.2) 5 (55.6) 9 (36.0) 5 (25.0).279 Background thyroiditis 12 (22.6) 0 7 (28.0) 5 (25.0).203 Tumor invasion Extracapsular 9 (16.7) 0 5 (20.0) 4 (20.0) Extraglandular, minimal 9 (16.7) 4 (44.4) 4 (16.0) 1 (5.0) Extraglandular, extensive 8 (14.8) 5 (55.6) 2 (8.0) 1 (5.0).001 Vascular invasion 11 (20.4) 4 (44.4) 3 (12.0) 4 (20.0) Cervical LN metastases Central 11 (20.4) 1 (11.1) 5 (20.0) 5 (25.0) Ipsilateral 11 (20.4) 2 (22.2) 5 (20.0) 4 (20.0).002 Bilateral 9 (16.7) 6 (66.6) 2 (8.0) 1 (5.0) Extracapsular LN invasion 10 (18.5) 5 (55.6) 1 (4.0) 4 (20.0).003 Lung metastases 15 (27.8) 7 (77.7) 5 (20.0) 3 (15.0).005 Micronodular focal 11 (20.4) 5 (55.6) 4 (16.0) 2 (10.0) Micronodular diffuse 4 (7.4) 2 (22.2) 1 (4.0) 1 (5.0) Tg levels, ng/ml 3.8 (1.2, 33.7) 340 (30.2, ) 2 (1.0, 6.6) 2 (1.0, 16.9).001 RAI therapy Number of patients Residual gland only 13 (27.1) 0 7 (30.4) 6 (35.3).159 Residual gland and LN 20 (41.7) 1 (12.5) 11 (47.8) 8 (47.1).186 Lung metastases 15 (31.2) 7 (87.5) 5 (21.7) 3 (17.6).001 Number of RAIs 1 27 (50.0) 2 (22.2) 14 (56.0) 11 (55.0) 2 15 (27.8) 2 (22.2) 9 (36.0) 4 (20.0) (11.1) 4 (44.4) 0 2 (10.0) Abbreviations: CA, chronological age; MNG, multinodular goiter; ND, neck dissection. The data are presented as number (percentage) or median (range) unless otherwise indicated..014

6 doi: /jc press.endocrine.org/journal/jcem 1975 The risk for recurrent/persistent disease, as assessed by the postoperative ATA risk-stratification system and SC- MCI score, was higher in the prepubertal group (P.001 and P.002, respectively) (Figure 1). The significantly worse ATA risk-stratification and SCMCI score grades of the prepubertal patients, compared with those of the pubertal and postpubertal patients (three vs one vs one, with P.001 for both variables), corresponded with these findings. It should be noted that although the Tg score was not adjusted for the 3 of 54 patients with postoperative interfering antibodies, the risk category of these patients would have remained unchanged, even after adjustment of Tg score by 1 point (intermediate risk in one pubertal and one postpubertal patient and high risk in one pubertal patient). Figure 1. Distribution of postoperative risk-stratification systems (ATA risk classification and SCMCI score) and response to initial therapy system in 51 pediatric patients with DTC: comparison between prepubertal, pubertal, and postpubertal patients. Response to initial therapy, course, and outcome (Tables 3 and 4) The median duration of follow-up was 8.8 years (range y) and did not differ significantly in the three groups. Two years from the initial therapy, 35 patients (only two prepubertal) showed a complete therapeutic response, 29 after the first treatment and six after a second dose of RAI therapy. In 10 of 11 patients with focal lung metastases, a complete therapeutic response was achieved after one to two RAI treatments, whereas in the four patients with diffuse lung metastases, persistent disease was evidenced. It is of note that the RAI therapy was less aggressive in patients with focal disease than in those with diffuse lung disease (the number of RAI treatments, one to two vs three to four, P.003; the cumulative RAI dose, mci vs mci, P.001). In 16 patients (six prepubertal), the disease remained persistent; in six of these (only one prepubertal), the therapeutic response was acceptable, whereas in 10 (five prepubertal), it was incomplete. The percentage of prepubertal patients with initial persistent disease was significantly higher than that of pubertal and postpubertal patients (75% vs 25% vs 21%, P.02). The risk for recurrent/persistent disease as assessed by the response to initial therapy system was higher in the prepubertal group (P.02) (Figure 1). The significantly worse grade of response to initial therapy of the prepubertal group, as compared with that of the pubertal and postpubertal groups (three vs one vs one, P.031), corresponded with this finding. During follow-up, eight patients initially disease-free (five pubertal, three postpubertal) developed recurrent disease in the thyroid bed and/or cervical LNs. The time elapsed to recurrence ranged from 2.5 to 13.5 years. After the additional treatment (RAI alone in four patients, RAI and surgery in four), all but one (who became partially iodine negative) were disease-free. Six patients with initial persistent disease (two prepubertal, three pubertal, one postpubertal) eventually became disease-free, one after reoperation and five after additional RAI therapy. At the time of the last evaluation, the survival rate was 100%. At this time point, 40 patients (78.4%) were diseasefree. Outcome was similar in all age groups (P.103). In 11 patients there was persistent disease, lasting from 4 to 16.5 years. In six of these (four pubertal, two postpubertal), the tumor became partially iodine negative. The five patients with persistent iodine-avid disease had lung metastases (micronodular diffuse in four and micronodular focal in one patient) at presentation. In these patients a continuous decline in Tg levels was seen over time, with no evidence of progressing metastases, although additional RAI therapy was withheld after the fourth treatment.

7 1976 Lazar et al Pediatric DTC: Prognostic Classification Systems J Clin Endocrinol Metab, May 2016, 101(5): Table 4. Clinical Course and Outcome for Each Classification System (ATA Risk Category, SCMCI Score, and Response to Initial Therapy) in 51 Children and Adolescents With DTC ATA Risk Category SCMCI Score Response to Initial Therapy Low (n 29) Intermediate (n 6) High (n 16) Low (n 25) Intermediate (n 14) High (n 12) Complete (n 35) Acceptable (n 6) Incomplete (n 10) Follow-up Disease free (n 30) 70.0% (20) 50.0% (3) 43.7% (7) 80.0% (20) 42.9% (6) 25.0% (4) 82.9% (29) 16.7% (1) 0% (0) Persistent disease (n 13) 10.3% (3) 16.7% (1) 56.3% (9) 8.0% (2) 21.4% (3) 75.0% (8) 5.7% (2) 16.7% (1) 100% (10) Recurrent disease (n 8) 20.7% (6) 33.3% (2) 0% (0) 12.0% (3) 35.7% (5) 0% (0) 11.4% (4) 66.7% (4) 0% (0) Last visit Disease free (n 40) 89.6% (26) 83.3% (5) 56.2% (9) 92.0% (23) 78.6% (11) 50.0% (6) 97.1% (34) 83.3% (5) 10.0% (1) Persistent disease (n 11) 10.4% (3) 16.7% (1) 43.8% (7) 8.0% (2) 21.4% (3) 50.0% (6) 2.9% (1) 16.7% (1) 90.0% (9) The data are presented as percentage (number). Clinical course and outcome for each classification system (ATA risk category, SCMCI score, and response to initial therapy) are presented in Table 4. During follow-up and at the last visit, disease-free was the most prevalent outcome in patients whose risk estimate was low grade by both the ATA risk-stratification system and the SCMCI score as well as for those with complete response to initial therapy. Persistent disease was more common during initial follow-up among patients with a high-grade postoperative risk estimate and on the last visit in patients with incomplete response to initial therapy. Recurrent disease occurred more frequently in patients with a low- or intermediate-grade postoperative risk estimate and among those with only an acceptable response to initial therapy. Prediction of disease outcome Stepwise multiple logistic regression models to predict disease outcome, which assessed the independent parameters of age at diagnosis, pubertal status, gender, predisposing factors, type of operation, initial RAI dose, and initial postoperative Tg levels as well as the grade yielded by one of the classification systems, revealed that in each model the classification system remained the sole potential explanatory variable for long-term disease outcome (the ATA risk stratification system: P.014, R , predictive ability 80.4%; the numerical SCMCI score: P.001, R , predictive ability 86.3%; the response to initial therapy classification P.001, R , predictive ability 96.1%). The proportion of variance explained associated with each system (0.79 for response to initial therapy classification, 0.44 for SCMCI score, 0.25 for ATA risk stratification system) indicated that the response to initial therapy classification explained a greater proportion of the variance and therefore was a more precise predictor for disease outcome. The predictive ability for disease outcome of the numerical SCMCI score was better than that of the ATA risk-stratification system. A receiver-operating characteristic curve analysis showed that the optimal cutoff for prediction was a score of 5 (area under the curve with 95% confidence interval 71% 99%, sensitivity 90%, and specificity 76%, P.001); hence, patients with a SCMCI score of less than 5 were at low risk for eventual recurrent/persistent disease. Discussion The recently proposed ATA risk of recurrence stratification system aimed to define the likelihood of recurrent or persistent disease after initial surgery in pediatric DTC and to identify the patients who would benefit from RAI treatment. In the present study, we demonstrated that the application of risk-stratification systems according to postoperative findings (both the ATA risk stratification system and the novel SCMCI score) as well as the response to initial therapy, may facilitate predicting the disease course and outcome. The ATA risk-stratification system categorizes pediatric patients into three risk groups according to regional LN and distant metastasis staging by using the tumor node metastasis classification system (18). However, it does not encompass all of the tumor characteristics necessary for accurate assessment of the prognosis after the diagnosis of DTC. The SCMCI score developed by us is a more detailed classification system, comprising information on the histopathological characteristics of the tumor (multifocality, involvement of opposite lobe, extraglandular extension, regional LN metastases levels and laterality with or without extracapsular invasion, and vascular invasion), lung metastases, either focal or diffuse, and postoperative Tg levels, all known to be associated with disease outcome (16, 17, 24 30). The SCMCI scoring system assigned higher scores for bilateral cervical LN metastases and extracapsular LN invasion because these reflect more extensive disease requiring more aggressive treatment. The diffuse pattern of lung metastases was scored higher than the focal pattern because the number of RAI treatments, the total dose of RAI, and the prevalence of persistent disease were higher in our patients with diffuse lung metastases. Tumor size was not included in our scoring system: in adults this correlates with the extent of

8 doi: /jc press.endocrine.org/journal/jcem 1977 disease at presentation and has a role in predicting disease outcome (31), but in young children the relatively small thyroid volume and its changes with age make it an unreliable prognostic criterion (32). Assessment of the patients in our cohort by both the postoperative ATA risk-stratification system and our SC- MCI score showed that the estimated risk was significantly higher in the prepubertal children. Indeed, the clinical complaints, physical and imaging findings, and pathological features indicating a more aggressive disease as well as the higher postoperative Tg levels, a sensitive marker of residual disease, were strikingly more common in the prepubertal children. These data are consistent with previous studies showing that children s age inversely correlates with morphological and clinical aggressiveness of DTC and extends previous reports, including ours, that DTC in prepubertal children is more invasive at diagnosis (5 13, 26, 27, 33, 34). Treatment modality (extent of surgery, RAI therapy, the degree of TSH suppression) and response to initial therapy are all well known to affect the course and outcome of DTC (1, 3, 7 13, 25 29, 32, 34, 35). Although all the patients in our study underwent total/near-total thyroidectomy and most also RAI, the response to initial therapy varied. Assessment of the response to initial therapy based on the risk stratification proposed by Tuttle et al (19) revealed that whereas complete response was observed in most of the pubertal and postpubertal patients during the first 2 years of follow-up, incomplete response was more frequently found in the prepubertal children. Accordingly and similar to previous studies (6, 7, 13), the initial course was characterized by a higher prevalence of persistent disease in the prepubertal children, whereas most pubertal and postpubertal patients demonstrated a disease-free progression. The less satisfactory response to initial therapy in the prepubertal patients was possibly associated with the extent of disease at diagnosis and with the biological characteristics of the tumor (4, 15 17, 24, 25, 29, 33, 34, 36). Nevertheless, the additional therapeutic interventions, administered more frequently to these children, were effective in bringing most of them into remission. Thus, in the last follow-up visit, the rate of disease-free status among the prepubertal patients tended to approach that of the adolescents. The initial therapeutic approach for children and adolescents is a controversial issue. In the past a routine of intensive management consisting of total thyroidectomy followed by RAI administration was the standard care for most pediatric DTC patients. Unfortunately, recent studies have revealed an increase in all-cause mortality for survivors of childhood DTC, predominantly due to secondary malignancies in children treated with radiation (10). The therapeutic challenge has been to achieve a decrease of disease morbidity and recurrence while avoiding the potential complications of therapy. The extent of disease at diagnosis as assessed by the postoperative ATA risk-stratification analysis and the SC- MCI score may serve as prognostic factors for the course of the disease and eventual outcome in the pediatric age group. These two systems were able to identify low- and high-risk patients and thereby may potentially define those cases for which meticulous surveillance and more aggressive therapy are indicated while avoiding overtreatment in those with less extensive disease. The SCMCI score, which accounted for a greater portion of variance, proved to be superior to the ATA risk-stratification system in predicting disease outcome. Yet, as in adults (19), the best predictive system for disease outcome was found to be the response to initial therapy, which by incorporating the effect of therapy variables allowed for modification of risk during follow-up (lowering of risk level in patients initially stratified as high risk but exhibiting a complete response, or higher in patients initially stratified as low risk but having an incomplete response). The modified risk categories seem to be more appropriate than the initial postoperative risk levels when planning the continuing surveillance of pediatric DTC patients. It should be emphasized that the modality and intensity of management in the modified risk level should be in accordance with the ATA guidelines. Strength and limitations The major strength of our study is that the entire cohort was followed up at a single center by the same expert multidisciplinary team from diagnosis to the last visit. Thus, all aspects of initial management were quite uniform, including extent of surgery, RAI treatment (adherence to a low iodine diet before ablation, adequacy of TSH stimulation for ablation, dose of I 131, and interpretation of diagnostic and follow-up WBS), the degree of TSH suppression, and the rigorous surveillance protocol. The main limitations of our study are the relatively small cohort and relatively short median follow-up period of 9 years, making it imperative that we continue to exercise caution and avoid generalizations when dealing with each individual case. Also, to be noted as a limitation is the fact that the Tg score was not adjusted for the 3 of 54 patients with postoperative interfering antibodies. Conclusions This study confirms that the recently proposed ATA risk classification system and the novel SCMCI score developed by us can serve as good initial predictors of recurrent/persistent disease in children and adolescents with DTC. Applic-

9 1978 Lazar et al Pediatric DTC: Prognostic Classification Systems J Clin Endocrinol Metab, May 2016, 101(5): ation of the system of response to initial therapy in addition to these risk-stratification systems provides a better estimate of the long-term disease course and outcome. Incorporation of these systems as part of routine evaluation may well lead to improved management strategies and surveillance decisions for pediatric DTC patients, not only for the first 2 years from diagnosis but throughout follow-up, thus serving to decrease the morbidity caused by recurrent/persistent disease or by unnecessary aggressive treatment. Continued surveillance of our pediatric DTC cohort and larger, long-term prospective studies are needed to validate these risk-stratification systems. Acknowledgments Address all correspondence and requests for reprints to: Professor Liora Lazar, MD, The Jesse Z. and Sara Lea Shafer Institute for Endocrinology and Diabetes, National Center for Childhood Diabetes, Schneider Children s Medical Center of Israel, 14 Kaplan Street, Petah Tikva 49202, Israel. liora.lazar@gmail.com. Disclosure Summary: The authors have nothing to disclose. References 1. Hogan AR, Zhuge Y, Perez EA, Koniaris LG, Lew JI, Sola JE. Pediatric thyroid carcinoma: incidence and outcomes in 1753 patients. J Surg Res. 2009;156: Vergamini LB, Frazier AL, Abrantes FL, Ribeiro KB, Rodriguez- Galindo C. Increase in the incidence of differentiated thyroid carcinoma in children, adolescents, and young adults: a population-based study. J Pediatr. 2014;164: Schlumberger M, De Vathaire F, Travagli JP, et al. Differentiated thyroid carcinoma in childhood: long term follow-up of 72 patients. J Clin Endocrinol Metab. 1987;65: Zimmerman D, Hay ID, Gough IR, et al. Papillary thyroid carcinoma in children and adults: long-term follow-up of 1,039 patients conservatively treated at one institution during three decades. Surgery. 1988;104: Samuel AM, Sharma SM. Differentiated thyroid cancer in children and adolescents. Cancer. 1991;67: Alessandri AJ, Goddard KJ, Blair GK, Fryer CJ, Schultz KR. Age is the major determinant of recurrence in pediatric differentiated thyroid carcinoma. Med Pediatr Oncol. 2000;35: Landau D, Vini L, A Hern R, Harmer C. Thyroid cancer in children: the Royal Marsden Hospital experience. Eur J Cancer. 2000;36: Hung W, Sarlis NJ. Current controversies in the management of pediatric patients with well-differentiated nonmedullary thyroid cancer: a review. Thyroid. 2002;12: Lazar L, Lebenthal Y, Steinmetz A, Yackobovitch-Gavan M, Phillip M. Differentiated thyroid carcinoma in pediatric patients: comparison of presentation and course between pre-pubertal children and adolescents. J Pediatr. 2009;154: Hay ID, Gonzalez-Losada T, Reinalda MS, Honetschlager JA, Richards ML, Thompson GB. Long-term outcome in 215 children and adolescents with papillary thyroid cancer treated during 1940 through World J Surg. 2010;34: O Gorman CS, Hamilton J, Rachmiel M, Gupta A, Ngan BY, Daneman D. Thyroid cancer in childhood: a retrospective review of childhood course. Thyroid. 2010;20: Pawelczak M, David R, Franklin B, Kessler M, Lam L, Shah B. Outcomes of children and adolescents with well-differentiated thyroid carcinoma and pulmonary metastases following 131-I treatment: a systematic review. Thyroid. 2010;20: Jarzab B, Handkiewicz Junak D, et al. Multivariate analysis of prognostic factors for differentiated thyroid carcinoma in children. Eur J Nucl Med. 2000;27: Haugen BR, Alexander EK, Bible KC, et al. American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer. Thyroid. In press. 15. Shayota BJ, Pawar SC, Chamberlain RS. MeSS: a novel prognostic scale specific for pediatric well-differentiated thyroid cancer: a population-based, SEER outcomes study. Surgery. 2013;154: Powers PA, Dinauer CA, Tuttle RM, Robie DK, McClellan DR, Francis GL. Tumor size and extent of disease at diagnosis predict the response to initial therapy for papillary thyroid carcinoma in children and adolescents. J Pediatr Endocrinol Metab. 2003;16: Powers PA, Dinauer CA, Tuttle RM, Francis GL. The MACIS score predicts the clinical course of papillary thyroid carcinoma in children and adolescents. J Pediatr Endocrinol Metab. 2004;17: Francis GL, Waguespack SG, Bauer AJ, et al. Management guidelines for children with thyroid cancer. Thyroid. 2015;25: Tuttle RM, Tala H, Shah J, et al. Estimating risk of recurrence in differentiated thyroid cancer after total thyroidectomy and radioactive iodine remnant ablation: using response to therapy variables to modify the initial risk estimates predicted by the new American Thyroid Association staging system. Thyroid. 2010;20: Cooper DS, Doherty GM, Haugen BR, et al. Management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 2006;16: Cibas ES, Ali SZ. The Bethesda system for reporting thyroid cytopathology. Am J Clin Pathol. 2009;132: Song HJ, Qiu ZL, Shen CT, Wei WJ, Luo QY. Pulmonary metastases in differentiated thyroid cancer: efficacy of radioiodine therapy and prognostic factors. Eur J Endocrinol. 2015;173: Nagelkerke NJD. A note on a general definition of the coefficient of determination. Biometrika. 1991;78: Welch Dinauer CA, Tuttle RM, Robie DK, et al. Clinical features associated with metastasis and recurrence of differentiated thyroid cancer in children, adolescents and young adults. Clin Endocrinol (Oxf). 1998;49: Jarzab B, Handkiewicz-Junak D, Wloch J. Juvenile differentiated thyroid carcinoma and the role of radioiodine in its treatment: a qualitative review. Endocr Relat Cancer. 2005;12: Mihailovic J, Nikoletic K, Srbovan D. Recurrent disease in juvenile differentiated thyroid carcinoma: prognostic factors, treatments, and outcomes. J Nucl Med. 2014;55: Markovina S, Grigsby PW, Schwarz JK, et al. Treatment approach, surveillance, and outcome of well-differentiated thyroid cancer in childhood and adolescence. Thyroid. 2014;24: Popovtzer A, Shpitzer T, Bahar G, Feinmesser R, Segal K. Thyroid cancer in children: management and outcome experience of a referral center. Otolaryngol Head Neck Surg. 2006;135: Demidchik YE, Demidchik EP, Reiners C, et al. Comprehensive clinical assessment of 740 cases of surgically treated thyroid cancer in children of Belarus. Ann Surg. 2006;243: Piccardo A, Arecco F, Puntoni M, et al. Focus on high-risk DTC patients: high post-operative serum thyroglobulin level is a strong predictor of disease persistence and is associated to progression-free survival and overall survival. Clin Nucl Med. 2013;38: McHenry CR, Huh ES, Machekano RN. Is nodule size an independent predictor of thyroid malignancy? Surgery. 144: ; discussion Rivkees SA, Mazzaferri EL, Verburg FA, et al. The treatment of

10 doi: /jc press.endocrine.org/journal/jcem 1979 differentiated thyroid cancer in children: emphasis on surgical approach and radioactive iodine therapy. Endocr Rev. 2011;32: LiVolsi VA, Abrosimov AA, Bogdanova T, et al. The Chernobyl thyroid cancer experience: pathology. Clin Oncol (R Coll Radiol). 2011;23: Fridman M, Savva N, Krasko O, et al. Initial presentation and late results of treatment of post-chernobyl papillary thyroid carcinoma in children and adolescents of Belarus. J Clin Endocrinol Metab. 2014;99: Handkiewicz-Junak D, Wloch J, Roskosz J, et al. Total thyroidectomy and adjuvant radioiodine treatment independently decrease locoregional recurrence risk in childhood and adolescent differentiated thyroid cancer. J Nucl Med. 2007;48: Collini P, Mattavelli F, Pellegrinelli A, Barisella M, Ferrari A, Massimino M. Papillary carcinoma of the thyroid gland of childhood and adolescence: morphologic subtypes, biologic behavior and prognosis: a clinicopathologic study of 42 sporadic cases treated at a single institution during a 30-year period. Am J Surg Pathol. 2006; 30: