Predictive risk factors for distant metastasis in pediatric differentiated thyroid cancer from Saudi Arabia

Background Despite their excellent prognosis, children and young adults (CAYA) with differentiated thyroid cancer (DTC) tend to have more frequent occurrence of distant metastasis (DM) compared to adult DTC. Data about DM in CAYA from Middle Eastern ethnicity is limited. Methods Medical records of 170 patients with DTC ≤18 years were retrospectively reviewed. Clinico-pathological factors associated with lung metastasis in CAYA, their clinical presentation and outcome were analyzed. Rick factors related to distant metastasis-free survival (DMFS) for the whole cohort were evaluated. Results DM was observed in 27 patients and all were lung metastasis. Lung metastasis was significantly associated with younger age (≤15 years), extrathyroidal extension (ETE), multifocal tumors, bilaterality, presence of lymph node (LN) disease and high post-operative stimulated thyroglobulin (sTg). Highest negative predictive values were seen with low post-operative sTg (97.9%), absence of LN disease (93.8%), absence of ETE (92.2%) and age older than 15 years (92.9%). Post-therapy whole body scan (WBS) identified most of the lung metastasis (21 of 27; 77.8%). Upon evaluating patients response according to ATA guidelines, excellent response was seen in only one patient, while biochemical persistence and structural persistence were seen in 11.1% (3/27) and 77.8% (21/27), respectively. Elevated post-operative sTg (>10ng/ml) was the only risk factor found to be significantly associated with both biochemical persistence (with or without structural persistence (p = 0.0143)) and structural persistence (p = 0.0433). Cox regression analysis identified age and post-operative sTg as independent risk factors related to DMFS. Based on these two risk factors for DMFS, patients were divided into 3 groups: low risk (no risk factors), intermediate risk (1 risk factor) and high risk (both risk factors). 20-year DMFS rates in the low-, intermediate- and high-risk groups were 100.0%, 81.3% and 23.7% respectively (p < 0.0001). Conclusion Higher suspicion for metastatic pediatric DTC should be considered in patients who are young, have LN disease, extrathyroidal extension and elevated post-operative sTg. Persistent disease, despite therapy, is very common and it appears to be related to post-operative sTg level. Hence, risk adaptive management is desirable in CAYA with DTC.


Introduction
Differentiated thyroid cancer (DTC) is a common solid tumor in children and young adults (CAYA) (1,2).The incidence of CAYA DTC has been increasing lately (3).Despite similarity in pathophysiology, pediatric and adult DTC have different clinical behavior.DTC in CAYA tend to present with more aggressive clinical features at diagnosis such as large tumor size and high proportion of bilateral and multifocal tumors (4,5).Furthermore, children and adolescents with DTC are more likely to present with advanced disease with higher rate of recurrence, persistence, regional and distant metastasis (6,7).
Even so, the reported mortality rate in very low compared to adult, with a disease specific survival of 99% in affected children (6,8).The optimal approach to evaluation and treatment of CAYA with distant metastasis and particularly lung metastasis, remains unresolved (9).According to the American Thyroid Association (ATA) treatment guidelines (10), total thyroidectomy is recommended for all pediatric DTC patients.In clinical practice, pediatric DTC patients are generally treated with surgery followed by possible radioactive iodine (RAI) therapy, which is usually given in the context of post-operative staging based on diagnostic thyroid scan and stimulated thyroglobulin (sTg) level (10,11), especially in cases of known distant metastatic pediatric DTC.The lungs are the primary site of DM, detected in up to 20% of childhood DTC (6,7,12).Despite the fact that most pediatric patients with DM do not reach remission, they do have an extremely low disease specific mortality (7,9,(12)(13)(14).
The management of pediatric DTC with DM is challenging and recent reports have highlighted the importance of personalized therapy, based on the risk of poor outcome, to maintain the balance between treating aggressive disease and minimizing the potential long-term sequelae of over-zealous therapy in children with DTC (15)(16)(17).Therefore, it is extremely important to identify the subset of pediatric DTC who are at risk of poor clinical outcome, especially those with distant metastasis, where data regarding CAYA DTC from Middle Eastern ethnicity is limited.
In the present study, we retrospectively analyzed a large cohort of children and young adults with distant metastatic DTC, treated at a single tertiary care hospital in Saudi Arabia.Prevalence, predictive markers and response to therapy of CAYA with distant metastatic DTC were evaluated.Risk factors for DM were evaluated and risk stratification was attempted.

Patient selection
One hundred and seventy CAYA (≤ 18 years) DTC patients diagnosed between 1988 and 2018 at King Faisal Specialist Hospital and Research Centre (Riyadh, Saudi Arabia) were included in the study.Cases were identified based on clinical history followed by fine needle aspiration biopsy for confirmation.The Institutional Review Board of the hospital approved this study and since only retrospective patient data were used, the Research Advisory Council (RAC) provided waiver of consent under project RAC # 221 1168 and # 2110 031.The study was conducted in accordance with the Declaration of Helsinki.

Clinico-pathological and follow-up data
Baseline clinico-pathological data were collected from case records and have been summarized in Table 1.Staging of DTC was performed using the eighth edition of American Joint Committee on Cancer (AJCC) staging system (18).The patients were seen 6 to 8 weeks after surgery, at which time, a diagnostic radioactive iodine (I-123) whole body scan (DxWBS) and neck ultrasonography were performed, and stimulated thyroglobulin (sTg), anti-Tg antibodies, TSH and free T4 were measured.Radioactive iodine (I-131) was administered at activities that averaged 150 to 200 mCi for a 70-kg patient, adjusted for body weight, with consideration of the extent of lung involvement found on the DxWBS.The initial I-131 therapies were administered every 6 to 12 months.Monitoring of these patients was achieved by periodic (every 3 to 9 months) measurement of TSH, FT4, Tg, and anti-Tg antibodies, neck ultrasonography, DxWBS, and CT scan of the lungs.
Response to therapy was defined as ''excellent,'' with basal (unstimulated) Tg <0.2 ng/mL or thyrotropin-stimulated Tg (sTg) <1 ng/mL (with normal anti-Tg titers) and no structural disease on imaging; biochemical persistence, with abnormal biochemistry and no structural disease; structural persistence, with disease seen on imaging (regardless of Tg and anti-Tg levels); or indeterminate, wherein anti-Tg was elevated or data were missing (19).

DNA isolation
DNAs were extracted from PTC formalin-fixed and paraffinembedded (FFPE) tumor tissues utilizing Gentra DNA isolation kit (Gentra, Minneapolis, MN, USA) according to manufacturer's protocols as elaborated in the previous studies (20).

Sanger sequencing analysis
PCR and Sanger sequencing analysis of hotspot mutations in BRAF, KRAS, HRAS, NRAS and promoter region in TERT gene were carried out as described previously (21).Primer 3 online software was utilized to design the primers (available upon request).Reference sequences were downloaded from NCBI GenBank.Sequencing results were compared with the reference sequence by Mutation Surveyor V4.04 (Soft Genetics, LLC, State College, PA).

Fluorescence in situ hybridization analysis
FISH assay for NTRK3 gene fusions was performed as described previously on a tissue microarray format following the manufacturer's instructions (22).NTRK3 dual-color break-apart probe (Empire Genomics, Williamsville, NY, USA) were utilized.NTRK3 fusion was termed as positive if >20% of tumor cells showed split red and green signals or a single signal in addition to a single fused signal; otherwise, the specimen was classified as fusion negative (22).

Statistical analysis
The associations between clinico-pathological variables and distant metastases was performed using contingency table analysis and Chi square/Fisher exact tests.Unpaired t tests were used for comparisons of continuous variables.Distant metastasis-free survival (DMFS) was the primary endpoint and progression-free survival (PFS) was the secondary endpoint.DMFS and PFS rates were calculated by the Kaplan-Meier method.Cox proportional hazards model was used for analyzing the impact of prognostic factors on DMFS and PFS in univariate and multivariate manner.
Risk stratification was performed according to the factors related to survival.Two-sided tests were used for statistical analyses with a limit of significance defined as p value < 0.05.Data analyses were performed using the JMP14.0(SAS Institute, Inc., Cary, NC) software package.

Patient and tumor characteristics
Mean age of the entire cohort was 14.9 years, with a male: female ratio of 1:3.Majority of the tumors were PTC (92.3%; 157/ 170).Regional lymph node metastasis (LNM) was noted in 63.5% (108/170) of cases and distant metastasis was present in 15.9% (27/ 170) (Table 1), with all of them being lung metastasis.Of the 27 cases with lung metastasis, 20 (74.1%) were identified within 6 months of diagnosis.Pattern of lung metastases were as follows: micronodular (≤ 1cm) in 22 patients, macronodular (> 1cm) in two patients and no apparent nodule (found only on DxWBS) in three patients (Table 2, Figure 1).Median follow-up duration was 8.1 years (range 0.8 -32.9 years).
Although no mutations were detected in the mitogen activated protein kinase and TERT genes among the 16 lung metastatic cases tested, NTRK3 fusions were noted in 30.8% (4/13) and were significantly associated with lung metastasis (Table 1).
We further analyzed the clinico-pathological associations for BRAF mutation and NTRK3 fusion in the entire cohort.BRAF mutation was significantly associated with older age (>15 years, p = 0.0058) and inversely associated with extrathyroidal extension (p = 0.0390), T stage (p = 0.0054) and lung metastasis (p = 0.0062) (Supplementary Table 2).NTRK3 fusions were found to be significantly associated with only lung metastasis (p = 0.0379) (Supplementary Table 3).

Detection of lung metastasis, treatment and response to therapy
74.1% (20/27) lung metastases were diagnosed within 6 months of thyroidectomy.Lung metastasis were diagnosed by preoperative CT in 17/27 (63%) patients, of whom 16 patients were also positive by pre-ablation 123 I WBS.One patient had missing information about pre-ablation 123 I WBS.Three patients had disease detected only by pre-ablation WBS.Six patients had a normal DxWBS but developed lung metastasis after a period ranging from 1-10 years, identified by surveillance 123 I WBS, despite negative prior scans.
All patients with lung metastases were treated by RAI therapy.The median cumulative RAI activity was 380 mCi (range 75-835), and the median total number of RAI courses was 3 (range 1-5).At approximately one year after treatment, only one patient (3.7%) achieved an excellent response.Biochemical persistence was noted in 3 of 27 (11.1%)patients, structural persistence in 21 of 27 (77.7%)patients (including one death) and indeterminate in 2 of 27 (7.4%)patients (Table 2).Elevated post-operative sTg was the only risk factor found to be significantly associated with both biochemical persistence (with or without structural persistence (p = 0.0143)) and structural persistence (p = 0.0433).Neither RAI cumulative activity nor lung metastasis pattern had an impact on therapy response in this cohort (Table 3).
We also analyzed the PFS with regards to predictive factors and risk stratification.Similar to DMFS, we found that age and high post-operative sTg were independent predictors of PFS (Supplementary Table 1) and risk stratification showed that 20-year PFS rates in the low-, intermediate-and high-risk groups were 84.1%, 73.7% and 19.2% respectively (p = 0.0005) (Supplementary Figure 1).

Discussion
The management of CAYA DTC with DM is clinically challenging, given the favorable prognosis and protracted clinical course.In addition, there are concerns regarding the potential long-term complications of multiple treatments for CAYA DTC (15-17, 23, 24).In this study, we present a relatively large cohort of CAYA DTC from Middle Eastern ethnicity.DM was observed in 27 patients, and all were lung metastases.The prevalence of lung metastasis in our CAYA cohort is ~16% and is comparable with previous report conducted on series of pediatric DTC, where the prevalence of DM ranged from 15-25%, depending on the population studied and the cohort size (7,12,(25)(26)(27).
We found that lung metastasis was significantly associated with age younger than 15 years, presence of LNM, bilateral tumors, multifocal tumors, extrathyroidal extension and post-operative sTg.There are few reports that investigated risk factors associated with DM in CAYA DTC, where they identified additional factors such as tumor size and T-stage as independent risk factors (12,28,29), whereas these were not significant factors in the present study.Interestingly, age older than 15 years, absence of LNM and low post-operative sTg levels had negative predictive value of more than 90% for lung metastasis.Post-operative sTg levels and age younger than 15 years were significant independent predictors for lung metastasis and both were shown to be an accurate predictor of lung metastasis by ROC curve analysis.Indeed, in concordance with our findings, a recent study showed that sTg was an independent predictor of 1-year treatment response and final outcome in pediatric DTC (30).
Most lung metastasis 20/27 (74%) were identified within 6 months from thyroidectomy.Pattern of lung metastasis were mostly micronodular (22/27; 81.5%), with macronodular pattern seen in 2 cases and no apparent nodule (negative by CT Scan and detected by pre-ablation WBS) in 3 patients.
Genetic analysis for mitogen activated protein kinase and TERT mutations were assessed in 117/170 and 106/170 patients, respectively.None of the DM cases were positive for these mutations.NTRK3 fusion was assessed successfully in 68/170 cases.30.8% (4/13) of DM patients tested were found positive for NTRK3 fusion, compared to 7.3% (4/55) of non-metastatic cases.RET fusions were not analyzed due to technical reasons.Despite the small number of cases tested for NTRK3 fusion, the significant association between NTRK3 fusion and lung metastasis further highlights our previous finding (22) that NTRK3 plays an important role in aggressive pediatric DTC and confirms the notion that pediatric tumors are commonly driven by oncogenic fusions involving NTRK3 and less commonly driven by BRAF and TERT mutations.Indeed, this is in concordance with recent studies that have shown NTRK fusions to be enriched in pediatric lung metastasis cases, suggesting that NTRK fusions may be associated with tumor progression in pediatric DTC (31,32).
Lung metastases were diagnosed by preoperative CT in 17/27 (63%) patients, of whom 16 patients were also positive by preablation 123 I WBS.One patient had missing information about pre-ablation 123 I WBS.Three patients had disease detected only by Computed tomography (CT) scan images of lung metastasis in pediatric differentiated thyroid carcinoma patients showing macronodular (A) and micronodular (B) pattern.
pre-ablation WBS.All of these three patient were classified as high risk according to ATA risk stratification.Six patients had normal diagnostic WBS but developed lung metastasis after 1-10 years, identified by surveillance 123 I WBS, despite negative prior scans.These finding highlight the importance of multimodality imaging, especially in patients at high risk for recurrence.
All patients with DM were treated by RAI therapy.The overall response was evaluated using the ATA guidelines (19).After a median follow-up of 12.1 years, only one patient (3.7%) achieved an excellent response.Similarly low excellent response rates were reported previously in patients under 20 years (9,12,33).Despite a high rate of structural persistent disease, overall survival rate was excellent (96.3%).We did not identify any association between cumulative RAI activity (median = 380 mCi), number of RAI courses (median = 3) or pattern of lung metastasis and structural outcome.This could suggest that higher RAI doses might not improve the outcome of these patients and different systemic therapy might have to be considered for those with progressive DM.
In this study, we also estimated the DMFS for all patients.The 5-, 10-and 20-year DMFS rates were 89.3%, 83.2% and 63.8%, respectively.Multivariate analysis showed that age younger than 15 years and high post-operative sTg are the only predictive markers of poor DMFS in this cohort.Furthermore, based on the number of risk factors related to poor DMFS, we divided the patients into low-risk, intermediate-risk and high-risk groups.Interestingly, DMFS was significantly different among the three risk groups.The 20-year DMFS group rate in low-, intermediate-, and high-risk groups were 100.0%,81.3% and 23.7%, respectively.Interestingly, all three patients with DM detected only by 131 I WBS were in the high-risk group according to our proposed risk stratification.The low number of patients might limit the applicability of our risk stratification but should encourage a risk-adapted management for CAYA with DTC.
This study had certain limitations.Firstly, the retrospective nature of this study from a single institution is prone to bias.Secondly, the molecular alterations were not analyzed on the whole cohort samples and hence the results should be interpreted with caution.Thirdly, risk stratification was done in a limited number of patients from specific ethnicity, which precludes generalizability.Larger studies from other populations should be encouraged before applying the stratification.
In conclusion, in this cohort of CAYA DTC, lung metastases were associated with younger age, extrathyroidal extension, bilateral tumors, LN disease, a high post-operative sTg and NTRK3 fusions.Excellent response is rare and multiple RAI courses do not affect the outcome in these patients.Younger age and post-operative sTg have a potential role in stratifying CAYA DTC with worse DMFS.

TABLE 1
Clinico-pathological characteristics and associations of CAYA DTC with and without lung metastasis.

TABLE 2
Characteristics of lung metastasis: pattern, treatment and outcome.

TABLE 3
Clinico-pathological characteristics associated with structural and biochemical persistence in patients with lung metastasis.

TABLE 4
Clinico-pathological factors associated with distant-metastasis-free survival.