Thyroid cancer bone metastasis: Survival and genomic characteristics of a large tertiary care cohort.

Abstract

Purpose:

Bone metastasis (BM) in differentiated thyroid cancer (DTC) is the second most common site of metastasis after lung. BM are associated with worse prognosis in DTC. In this study, we examined risk factors for overall survival in patients with BM and for the first time explore the pattern of genomic alterations in DTC BM. Materials and Methods: A Health Insurance Portability and Accountability Act (HIPAA) compliant, institutional review board (IRB) approved retrospective evaluation of the medical record was performed for all patients treated at a single institution for thyroid cancer over a 16-year period. 74 patients met inclusion criteria. Multiple prognostic factors including age, gender, genes, radioactive iodine, and radiation or kinase inhibitor therapies were analyzed. Univariate and multivariate analyses were performed.

Results:

Treatment with external beam radiation (EBRT) was found to significantly increase survival (p=0.03). The five-year survival rate was 59% and median survival was 92 months. Patients who developed bone metastasis earlier tend to live longer (p=0.06). The presence of TERT and BRAF mutations did not significantly worsen the prognosis (p=0.10).

Conclusion:

DTC patients can benefit from early treatment with EBRT, especially those who develop bone metastasis within 3 years of primary TC diagnosis. Kinase inhibitor treatment tended to prolong survival but not in a statistically significant manor. Gender, age, and TERT or BRAF genetic mutations did not significantly affect the prognosis.

Introduction

Incidence of thyroid cancer (TC) has been rising in the past three decades [1]. Systemic metastatic disease is a common cause of morbidity and mortality in differentiated thyroid cancer (DTC). Lung and bone are the two most common sites for the spread of distal metastasis. Bone metastasis (BM) are typically associated with an overall worse prognosis than the lung metastasis [2–5]. The incidence of bone metastasis is reported to be between 1%−20% for DTC [6]. Approximately 44% of patients with metastatic DTC have disease that has spread to the bone [7].

Conventional treatment methods for BM include radioactive iodine (RAI), surgical resection, external beam radiation (EBRT), and chemotherapy. The use of kinase inhibitors has been increasing as a potential therapy for DTC. Some studies have reported that multi-kinase inhibitors, such as sorafenib or lenvatinib, have significantly increased median progression-free survival in radioiodine-refractory metastatic thyroid cancer [8, 9]. However, a few studies have reported that kinase inhibitors have minimal benefit in treating BM [10, 11].

The clinical use of genetic and molecular therapies in the treatment of metastatic DTC has increased. An evaluation of the most common genetic biomarkers to identify their implication on outcomes would be an invaluable addition to the current available literature. Studies have reported prognostic factors for BM in TC, however, no prior genetic profiling studies have been performed on DTC BM [2, 12]. This retrospective cohort study aims to evaluate how common genetic mutations affect the clinical course and survival in patients with DTC and BM, examining these potential prognostic factors and clinical course of these patients at a large tertiary clinical cancer center.

Materials and Methods

A Health Insurance Portability and Accountability Act (HIPAA) compliant, institutional review board (IRB) approved retrospective evaluation of the medical record was performed at our institution for all patients treated for TC from January 2000 to November 2016 by searching hospital billing codes in line with the international classification of diseases. A total of 14,411 patients were treated. We excluded patients who did not have BM, patients with a histopathologic diagnosis of aggressive anaplastic or medullary thyroid cancer, patients with synchronous cancers, and those who did not receive pathology that included an evaluation of the underlying genetic panel for the tumor. At our institution a custom DNA probe is applied, which was designed to target sequences of all exons and selected introns of 341 oncogenes, tumor suppressor genes, and members of pathways deemed actionable by targeted therapies, named “IMPACT.” IMPACT was performed on samples from thyroidectomy, neck dissection, or needle biopsy. A total of 74 patients met the criteria for inclusion in our evaluation (Table 1).

Table 1.

Retrospective review of the medical record from January 2000 to November 2016 Inclusion criteria

The specific diagnosis for the type of TC was confirmed by either biopsy or thyroidectomy with histopathologic diagnosis performed by pathology. Primary thyroid tumors were classified according to the World Health Organization (WHO) criteria. Poorly differentiated thyroid cancer was defined by proliferative grading features: ≥ 5 mitoses/10 high-power fields and/or tumor necrosis regardless of architectural pattern. This definition differs from the most recent Turin proposal, which requires the presence of a solid/trabecular/insular growth pattern in addition to proliferative grading [13]. However, in a recent study, both classifications were found to be comparable in predicting intermediate prognosis of thyroid cancer [14]. The staging was based on most updated 8th edition guidelines from American Joint Committee in 2017.

Presence of BM was confirmed by imaging; these included technetium-99mmethylene diphosphonate (MDP) bone scintigraphy, computed tomography (CT), magnetic resonance imaging (MRI), or Fludeoxyglucose positron emission tomography (PET/CT) (Figure 1).

Figure 1.

(A) Four days post-I-131 treatment (290 mCi), focal uptake in the spine (L2 and L5) and right skull with hazy uptake in the lungs is seen (Red arrows). (B) T1 weighted MRI of the spine with focal loss of signal within the L2 (upper) and L5 (lower) vertebral bodies consistent with metastases (White arrows). (C) Fused FDG PET/CT images of hypermetabolism within the L2 (upper) and L5 (lower) metastases. (D) CT images of lytic osseous metastases within L2 (upper) and L5 (lower) vertebral bodies (White arrows).

Treatment of the BM was at the discretion of the primary oncologist, based on the available treatment options including RAI, surgical resection, EBRT, chemotherapy, and various multi-kinase or single kinase inhibitors. These treatments could be alone or in combination. For example, with EBRT patients usually had identified iodine uptake and received RAI.

Statistical analysis methods

Survival analysis was performed using the Kaplan-Meier method. Survival time was defined as time from the date of BM diagnosis to date of death, or to date of last follow up for censored patients. The log-rank test was used in univariate analysis to identify covariates associated with survival. Continuous covariates were categorized into groups for the log-rank test. P values ≤ 0.05 were considered statistically significant. Statistical analysis was performed using SAS version 9.4 (Cary, NC).

Results Clinicopathologic characteristics

The clinical details of the 74 patients are presented in Table 2. Median age at diagnosis of BM was 62 years old (33–87). 58% male (n=43) and 42% female (n=31). Mean time from DTC diagnosis to development of BM was approximately 1.6 years. 32 patients (43%) had distal metastases at the time of diagnosis.

Table 2.

Clinical characteristics of patients with bone metastasis from Thyroid Cancer for inclusion (N = 74)

Characteristic	No. (%)
Age at diagnosis of BM (years) – median (range)	62.0 (33, 87)
Time from diagnosis of TC to BM (years) – median (range)	1.6 (0, 22.9)
Sex
Male  Female	43 (58.1) 31 (41.9)
Histology
Poorly differentiated	34 (46.0)
Papillary	22 (29.7)
Hurthle cell	14 (18.9)
Follicular	4 (5.4)
TC stage at diagnosis*
I	13 (18.6)
II	28 (40.0)
III	7 (10.0)
IV	22 (31.4)
Site of metastasis
Bone only	19 (25.7)
Bone and other organ	55 (74.3)
Extrathyroidal extension†
None	25 (36.8)
Minimal	24 (35.3)
Gross	19 (27.9)
Vascular invasion present‡	52 (80.0)
Number of BM sites
1	29 (39.2)
2	13 (17.6)
3	4 (5.4)
>3	28 (37.8)
EBRT after BM§	51 (69.9)
Kinase inhibitor given	45 (60.8)
RAI treatment given	63 (85.1)
RAI avidity‖	27 (65.9)

Abbreviations: BM, bone metastasis; TC, thyroid cancer, RAI, radioactive iodine; EBRT, external beam radiation therapy.

Values reported as frequency (percent), unless otherwise noted. Values from two-level variables are presented for only one level.

^*Unknown for 4 patients

^†Unknown for 6 patients

^‡Unknown for 9 patients

^§Unknown for 1 patient

^‖Unknown for 33 patients

In our cohort, poorly-differentiated papillary cancer consisted of 46% of the pathologic diagnosis, 30% papillary, 19% hurthle cell, and 5% follicular cancer. By stage; 19% were classified as stage I, 40% as stage II, 10% as stage III, and 31% as stage IV. Fifty-one patients (70%) received EBRT and 45 (61%) received kinase inhibitor treatment. Sixty-three patients (85%) received RAI treatment. Among them, 41 patients had records of post-I-131 treatment scans, of which 27 (66%) exhibited iodine avidity within the BM. These scans were performed at various times after treatment. Figure 2 is an example of a whole body I-131 scan at four days post treatment with iodine avid metastasis.

Figure 2.

Anterior (left) and posterior (right) whole body MDP bone scan images with radiotracer uptake within the right proximal humerus (Black arrows). (B) Radiograph showing expansile lytic lesion within the right proximal humerus (Black arrows). (C) Soft tissue and (D) bone windowing of CT images of right humeral lesion displaying large soft tissue component of right proximal humerus osseous metastasis (White arrows).

IMPACT Genetic Analysis

Presented in Table 3; 53 patients (72%) were TERT-positive, 15 (20%) were BRAF-positive, and 31 (42%) RAS-positive (23 NRAS; 2 KRAS; 6 HRAS). 13 were TERT and BRAF double-positive, 26 were TERT and RAS double-positive.

Table 3.

Univariate analysis of predictors associated with survival in patients with bone metastasis from thyroid cancer

Variable	N	3-year Survival Probability (95% CI)	P
Age (years)
<55	18	74.3 (44.1, 89.8)	0.37
≥55	56	71.7 (56.5, 82.3)
Sex
Male	43	69.6 (52.5, 81.6)	0.48
Female	31	76.4 (53.8, 89.0)
Time interval (years)
<3	46	80.3 (64.1, 89.8)	0.06
≥3	28	57.3 (33.6, 75.2)
Histology
Papillary	22	61.3 (36.8, 78.6)	0.17
Other	52	76.9 (60.9, 87.0)
Site of metastasis
Bone, only	19	81.7 (53.1, 93.8)	0.052
Bone and other organ	55	68.8 (53.0, 80.2)
Number of BM sites
<3	42	72.1 (53.8, 84.1)	0.93
≥3	32	72.9 (52.7, 85.6)
EBRT after BM
No	22	56.1 (32.0, 74.6)	0.03
Yes	51	78.5 (62.4, 88.4)
Kinase inhibitor after BM
No	29	65.9 (43.9, 80.9)	0.13
Yes	45	75.9 (58.3, 86.8)
RAI avidity
No	14	53.9 (24.3, 76.3)	0.01
Yes	27	95.5 (71.9, 99.3)
BRAF mutation
No	59	76.4 (61.9, 86.0)	0.11
Yes	15	56.0 (26.6, 77.6)
TERT mutation
No	21	70.5 (45.6, 85.6)	0.41
Yes	53	72.6 (56.1, 83.7)
TERT+BRAF mutations
No	61	75.6 (61.4, 85.2)	0.10
Yes	13	56.4 (24.4, 79.3)
RAS mutation
No	43	47.1 (23.9, 67.4)	0.18
Yes	31	50.5 (21.7, 73.6)

Overall survival was 59% (95% CI: 43, 71) at 3 years. Median survival time after BM was 7.7 years.

Abbreviations: BM, bone metastasis; EBRT, external beam radiotherapy, RAI, radioactive iodine.

Survival analysis result

25 died (34%), causes of death included thyroid cancer (n=17) and miscellaneous other reasons (n=8). The median survival time after diagnosis of BM was 92 months. Survival was 83% (95% CI: 72, 90) at 1-year and 59% (95% CI: 43, 71) at 3years (Figure 3).

Figure 3.

Kaplan-Meier for our included patient population evaluating survival during follow-up.

Two covariates were associated with a significantly longer survival (Table 3), treatment with EBRT and tumor avidity post RAI treatment. Patients with EBRT had a better survival outcome (3-year survival probability: 79%; 95% CI: 62, 88) compared to patients with no EBRT (3-year survival probability: 56%; 95% CI: 32, 75; p=0.03). RAI avidity was known for 41 patients in the cohort. In this subset, patients with RAI avidity (n=27, 66%) had better survival outcome (3-year survival probability: 96%; 95% CI: 72, 99) compared to patients with no RAI avidity (3-year survival probability: 54%; 95% CI: 24, 76; p=0.01) Figure 4.

Figure 4.

Kaplan-Meier graph of overall survival after the diagnosis of bone metastasis after a number of factors. (A) Treatment with external beam radiation; (B) Avidity of the metastasis on whole body iodine imaging.

Patients with bone-only metastasis exhibited borderline significant improved survival (three-year survival probability: 82%; 95% CI: 53, 94) compared to patients with bone plus other metastatic sites (three-year survival probability: 69%; 95% CI: 53, 80; p = 0.052). Additionally, patients with development of BM less than three years from DTC diagnosis had better survival (three-year survival probability: 80%; 95% CI: 64, 90) when compared with development of BM three years and beyond, albeit borderline significant (three-year survival probability: 57%; 95% CI: 34, 75; p = 0.06).

Discussion

BM is the second most common distal metastases in DTC and associated with poor prognosis [3]. Previous studies found that older age, follicular cell type, early BM diagnosis (within 3 years of primary diagnosis), RAI treatment, or no avidity in metastasis on the post RAI scan are poor prognostic factors [2–4]. In keeping with the literature, our study found that RAI avid BM were associated with better survival outcome. This result confirmed previous finding about the implacable role of RAI treatment [15]. 2015 American Thyroid Association revised guideline strongly recommend RAI treatment because the therapy is associated with improved survival, although RAI is rarely curative [16].

Our results also demonstrate EBRT significantly increased survival. This finding has not been recognized in previous study. EBRT has typically been utilized as a palliative therapy for BM, specifically in the events of pain, fracture risk, and neurological symptoms from vertebral compression [7]. Our data suggest that EBRT may potentially have a more prominent role in the treatment of BM and further study is needed to confirm our novel finding.

The overall survival in our cohort is similar to previous studies, with an overall 5year survival rate of 58.7% and 7-year survival rate of 54.2% [4, 7]. Surprisingly, patients with early BM diagnosis (i.e., within 3 years of diagnosis of DTC) have better survival than those who develop BM beyond 3 years, although not significant (p=0.06). This difference may be related to further mutations that develop resulting in treatment resistance.

Patients with only BM had significantly better survival outcome (p = 0.053) than those with BM plus other sites (mainly lung). This difference is likely related to increased tumor burden with multiple metastatic sites. A prior study in 283 patients failed to find metastatic site as a significant prognostic factor [17]. Our results indicate that age, gender, or primary thyroid histology are not significant factors that affect the overall prognosis for patients with BM.

A recent paper by Lin et al. divided patients into subgroups depending on the BM diagnosis within or beyond 6 months of RAI treatment [4]. This study emphasized the TNM staging at RAI treatment, which usually happened right after thyroidectomy. More advanced TNM staging was associated with early diagnosis of BM. Our study, on the other hand, looked at the “tumor burden”, which is assumed to increase in the time course of DTC progression.

PET imaging is particularly helpful for poorly-differentiated TC, as this tumor type can be negative on whole body iodine scans due to mutations that eliminate uptake of iodide. FDG avidity was not assessed as a prognostic factor because almost all patients (69/70 patients) had at least one site of FDG avid BM. In addition, the presence of FDG avidity of BM has already been proven to be associated with cancer aggressiveness and poor prognosis [18]. In Figure 5 we see a bone metastasis with FDG avidity after aggressive recurrence.

Figure 5.

T1 weighted and (B) T1 post-contrast MRI images of an enhancing femoral osseous metastasis with extra osseous soft-tissue component. (C) FDG PET/CT images showing radiotracer uptake/hypermetabolism within the soft tissue component of the femoral osseous metastasis (White arrows).

Our cohort study includes almost 50% of poorly-differentiated cancer, in line with another Memorial Sloan Kettering Cancer Center (MSKCC) study, published in 2000, where 68% of 79 BM patients had poorly differentiated or undifferentiated features in the primary and/or metastatic TC [19]. In contrast, there are surprisingly low incidence of follicular TC, which has a propensity for BM [4, 12, 15]. Selection bias may play a role in the different histology group distribution as MSKCC is a tertiary referral center for patients with cancer.

Unlike previous studies, we evaluate genetic changes in advanced DTC. Several important genes have been linked to poor survival and increased tendency of distal metastasis in the DTC; these include BRAF V600E, TERT, and various mutations in the RAS family [19].

TERT is an important gene associated with the advancement and aggressiveness of many cancer types [20]. In DTC, TERT is related to poor survival, higher prevalence of distal metastasis, and recurrence after RAI [21–23]. The high incidence of TERT within our study is unusual, as the overall incidence of somatic TERT mutation was thought to be rare in DTC, with a previously reported incidence of around 9% [20, 24–26]. Our genetic evaluation reveals the presences of the TERT promoter mutations in 53/74, or 72% of patients with BM. In our cohort the TERT gene mutation alone was not related to any change in the survival outcome. However, we observe that the combination of TERT plus the BRAF mutation may influence outcomes. In our study, the BRAF mutation was seen in 15/74 patients. Among them, 13/74 were both TERT and BRAF positive. This hypothesis is supported by the available literature as the presence of BRAF mutation has been shown to promote the TERT promoter mutation and TERT mRNA expression in a statistically significant manner [20, 21, 26]. The deleterious effects of the TERT plus BRAF mutations considerably shortened survival-free time compared to the BRAF alone [27, 28].

RAS family mutations are seen in both benign and malignant TC with NRAS as the predominant mutation. The incidence of RAS is higher in follicular TC, poorly-differentiated TC, and anaplastic TC [29]. Our data suggests similar findings; NRAS mutation was foremost in poorly-differentiated TC (23/25 or 92%).

Kinase inhibitors have been increasingly used in the treatment of advanced TC. However, the benefit to overall survival and clinical outcome remains undefined [17]. Multi-kinase inhibitors, such as sorafenib and lenvatinib are currently US Food and Drug Administration (FDA)-approved medicines for radioiodine-refractory metastatic thyroid cancer based on phase 3 trials [8, 9]. They have been reported to significantly increase median progression-free survival (sorafenib 10.8 months vs. placebo 5.8 months; lenvatinib 18.3 months vs. placebo 3.6 months). Other kinases inhibitors seem to be promising as well. For example, in a clinical trial from MSKCC, an MEK inhibitor, selumetinib, reversed radioiodine refractoriness in patients with metastatic thyroid cancer including the BM [30]. Not all studies have been positive and have reported a limited effect of kinase inhibitor on the DTC BM.

Sorafenib demonstrates efficacy in lung metastasis, but appears less effective for treating BM [10, 11]. 61% of our patients received a kinase inhibitor; these included sorafenib, lapatinib (tyrosine kinase inhibitor), dabrafenib (RAS inhibitor), or selumetinib (MEK inhibitor). Our study failed to establish that kinase inhibitor had a significant increase in the survival outcome of patients with BM (p=0.13). A limitation to this conclusion was our consolidating of all kinase inhibitors, including multi-kinase inhibitors into a single group. However, we felt given the smaller sample size, a clinically relevant conclusion could not be reached with subgroups due to lack of power.

The retrospective nature and small sample size of our study are other limitations. In the future study, we would like to test the genetic profiling of both the original cancer and the metastatic sites by IMPACT. In this way, we may better understand the possible evolution and increased aggressiveness of the metastatic thyroid cancer.

Conclusion

This retrospective study provides some statistical evidence that a small number of clinical features, including EBRT and radioiodine avidity of BM, are independently associated with survival outcome. Age, gender, histology, number of BM, and kinase inhibitor therapy, all had no detectable effect on survival. High incidence of TERT gene mutations were observed without a significant impact on survival, as one might expect in advanced-stage metastatic thyroid cancer. As genomic testing becomes an integrated component of treatment algorithms, these data may be used to personalize treatment.