Bone Metastases and Skeletal-Related Events in Medullary Thyroid Carcinoma

Abstract

Context:

Bone metastases (BM) can lead to devastating skeletal-related events (SREs) in cancer patients. Data regarding medullary thyroid carcinoma (MTC) with BM are lacking.

Objective:

We evaluated the natural history of BM and SREs in MTC patients identified by a cancer center tumor registry.

Setting:

The study was conducted at a tertiary cancer center.

Patients and Main Outcome Measures:

We retrospectively reviewed the charts of MTC patients with BM who received care from 1991 to 2014 to characterize BM and SREs.

Results:

Of 1008 MTC patients treated, 188 were confirmed to have BM (19%), of whom 89% (168 of 188) had nonosseous distant metastases. Median time from MTC to BM diagnosis was 30.9 months (range 0–533 mo); 25% (45 of 180) had BM identified within 3 months of MTC diagnosis. Median follow-up after detecting BM was 1.6 years (range 0–23.2 y). Most patients (77%) had six or more BM lesions, most often affecting the spine (92%) and pelvis (69%). Many patients (90 of 188, 48%) experienced one or more SREs, most commonly radiotherapy (67 of 90, 74%) followed by pathological fracture (21 of 90, 23%). Only three patients had spinal cord compression. Patients with more than 10 BM lesions were more likely to experience SREs (odds ratio 2.4; P = .007), with no difference in 5-year mortality after MTC diagnosis between patients with (31%) and without SREs (23%) (P = .11).

Conclusions:

In this large retrospective series, BM in MTC was multifocal, primarily involving the spine and pelvis, supporting screening these regions for metastases in at-risk patients. SREs were common but spinal cord compression was rare. Antiresorptive therapies in this population should be investigated further with prospective trials.

We studied the characteristics and clinical behavior of bone metastases in medullary thyroid cancer patients and found they are often multifocal and that SREs are common.

Thyroid cancer is the most common endocrine malignancy with an age-standardized incidence of 9.1 per 100 000 females and 2.9 per 100 000 males in developed countries (1). Medullary thyroid carcinoma (MTC), arising from the parafollicular C cells of the thyroid, accounts for 1%–2% of thyroid cancers. Patients with MTC who present with a palpable thyroid nodule are found to have cervical lymph node metastases in 75% and distant metastases 10%–15% of the time (2). Distant metastases often affect multiple organs including the lungs, liver, and bones. Ten-year survival rate for patients with stage IV MTC is 21% (3).

Bone is a frequent site of cancer metastasis due to high blood flow, adhesion molecules secreted by tumor cells, and local growth factors (4). Bone metastases (BMs) occur in up to 70% of patients with advanced breast or prostate cancers (5). In malignant pheochromocytoma and paraganglioma, which are other neuroendocrine neoplasms, the prevalence of BM is 71% (6). The stimulation of osteoclastogenesis by tumor cells through various paracrine factors contributes to metastasis-induced bone destruction. As a result, BMs can cause skeletal related events (SREs), defined as spinal cord compression, pathological fracture, radiation or surgery to bone, and hypercalcemia of malignancy (HCM), which can significantly decrease the quality of life in cancer patients (7).

Randomized controlled trials have evaluated and demonstrated the efficacy of antiresorptive therapies (ARTs), bisphosphonates, and denosumab (DMab) in preventing SREs in various types of solid tumors (8–10). Pamidronate (PAM) and zoledronic acid (ZA) are the two most commonly used iv bisphosphonates. PAM received approval for treatment of HCM in 1991 and osteolytic BM from breast cancer in 1996. ZA was approved for HCM in 2001 and BM in 2002. DMab, an antibody to receptor activator of nuclear factor-κ-B ligand, was approved for BM in 2010. ARTs have become a standard of care for patients with BM from various solid tumors.

Data regarding BM, SRE, and response to ARTs in thyroid cancer are limited. In a retrospective review of 245 differentiated thyroid cancer (DTC) patients with BM, 78% of the patients presented with or developed at least one SRE (11). Bisphosphonates were effective in reducing the number of SREs experienced in DTC patients with BM in a retrospective study by Orita, et al (12), whereas a subsequent prospective study reported 42% of DTC patients with BM treated with bisphosphonates experienced SREs. However, both studies included small cohorts (13). In patients with MTC, BM was reported in 40% of 50 patients with elevated calcitonin after treatment with surgery, radiation, or chemotherapy (14). However, no detailed characterization of BM or SRE or benefits or risks of ART in patients with MTC has been described to date (15). The use of ARTs is often based on clinical judgment or extrapolation from studies with other solid tumors. In this single-center retrospective study at our tertiary cancer center, we sought to describe the characteristics and distribution of BM in MTC, define the natural history of SREs in MTC patients with BM, and evaluate the effect and risk of ART treatment in these patients.

Patients and Methods

Patients

Institutional review board approval was obtained for this retrospective study at the University of Texas, M. D. Anderson Cancer Center. We reviewed the charts of patients with MTC and BM who received care from January 1, 1991, to December 29,2014, since the first agent, PAM, was approved in 1991 for HCM. A total of 202 patients (20%) were confirmed to have BM among 1008 patients with MTC identified by the institutional tumor registry, of whom 14 were excluded from analysis due to the presence of other active malignancy (n = 8) or insufficient clinical data (n = 6) (Figure 1). BMs were confirmed by review of clinical notes and radiological reports of computed tomography, magnetic resonance imaging, position emission tomography, or bone scan. Serial imaging was reviewed to confirm the sites of involvement, number of bone lesions, and development of new lesions.

Figure 1.

Flow chart describing the selection of patients. a, Excluded from subanalysis on patients who received ARTs.

The types of SREs assessed in the study included spinal cord compression, pathological fracture, radiation or surgery to areas of BMs, or HCM. HCM was defined by a corrected serum calcium level greater than 11.5 mg/dL (correlates with grade 2 Common Terminology Criteria for Adverse Events, version 4.03) excluding oversupplementation with calcium and/or vitamin D/calcitriol (16). The occurrence of any type of SRE at the same site within 30 days was recorded as the same SRE or incident. A subanalysis of patients who received ARTs (PAM, ZA, and DMab) was performed. ART information was obtained from clinical notes and research pharmacy records if ARTs were dispensed at M. D. Anderson Cancer Center. Patients receiving ARTs for osteoporosis (n = 4) or lacking treatment information (n = 7) were excluded from this subanalysis. Hypocalcemia after ART was graded according to Common Terminology Criteria for Adverse Events definition: grade 0, within normal range; grade 1, less than the lower limit, 8.0 mg/dL; grade 2, less than 8.0–7.0 mg/dL; grade 3, less than 7.0–6.0 mg/dL; and grade 4, less than 6.0 mg/dL. Any occurrences of osteonecrosis of the jaw (ONJ) were recorded.

Statistical analysis

Summaries for patient demographics and clinical characteristics were produced for all patients and by ART treatment group. Associations between the ART treatment group and categorical measures were assessed using either a Fisher's exact test or a generalized Fisher's exact test, whereas differences in continuous measures between the ART treatment groups were evaluated using a Wilcoxon rank sum test. In addition, associations between SREs and the number of bone lesions, age, gender, nonosseous distant metastases, and presence of axial bone metastases were determined using logistic regression models. Data were analyzed using SAS software version 9.3 for Windows (copyright 2011 by SAS Institute Inc). All P values were two sided and a value of P < .05 was considered significant.

Results

Demographic and clinical characteristics of patients with bone metastases

A total of 188 MTC patients with confirmed BM were included (112 males and 76 females). The patient demographics and clinical characteristics are summarized in Table 1. The median age at MTC diagnosis was 47 years (range 11.4–84.1 y) and the median follow-up time after MTC diagnosis was 5.5 years (range 0.1–54.0 y). Eighty-nine percent of the patients (168 of 188) had distant metastases involving other sites in addition to bone. Thirty-three percent had received radiation to the neck and 59% received systemic therapy. Of 120 patients who had germline RET testing, 25% were positive for a RET mutation, the same frequency as expected in the general MTC population.

Table 1.

Demographic and Clinical Characteristics

	All Patients (n = 188)
Age at diagnosis of MTC, y
Median	47.0
Minimum, maximum	11.4, 84.1
Gender, n, %
Male	112 (60)
Female	76 (40)
Ethnicity, n, %
White	141 (75)
Hispanic	24 (13)
Black	13 (7)
Other	10 (5)
Treatment, n, %
Total thyroidectomy	170 (90)
Radiation to neck	62 (33)
Systemic therapy (n = 186)	111 (60)
Germline RET mutation, n, %
Yes	30 (25)
RET, unknown codon	7
RET 618	8
RET 634	7
RET 620	3
RET 609, 791, 891, 804, 918	1 each
No	90 (75)
Missing	68
Somatic mutation, n, %
Yes	16 (70)
RET 918	14
HRAS	1
MET	1
No	7 (30)
Missing	165
Nonosseous distant metastases, n, %
Yes	168 (89)
No	20 (11)
Vital status, n, %
Died	111 (59)
Alive	77 (41)
Duration of follow-up from MTC diagnosis, y
Median	5.50
Minimum, maximum	0.12, 53.95

Characteristics and distribution of bone metastases in entire cohort

The median duration between MTC and BM diagnoses was 30.9 months (range 0.0–533.4 mo), whereas 45 of 180 patients (25%) were identified to have BM within 3 months of MTC diagnosis. Of note, eight patients were excluded from this particular analysis because the dates of bone metastasis diagnosis were unknown. In addition, of the 168 patients with nonosseous distant metastases, 159 patients with available dates of diagnoses for BM and distant metastases were reviewed to determine the timing of the BM diagnosis relative to the diagnosis of nonosseous metastases. Twenty-three of 159 patients (14%) were diagnosed with BM before evidence of nonosseous distant metastases, with a median duration between the BM diagnosis and nonosseous distant metastases of 5.3 months (range 0.0–74.9 mo) in this group. Thirty-eight of 159 patients (24%) had concomitant BM identified at the time of the distant metastases diagnosis, and 98 of 159 (62%) had BM diagnosed after nonosseous the distant metastases diagnosis. The median duration of follow-up after the detection of the first BM was 1.6 years (range 0–23.2 y).

Sixty-five percent of patients had more than 10 bone lesions, 12% had 6–10 lesions, 16% had two to five lesions, and only 6% had a single lesion. The median duration of follow-up after the BM diagnosis was 8.1 month (range 0.3–44.7 mo) in patients with a single BM lesion and 20.1 months (range 0.0–278.1 mo) in patients with more than one BM lesion (P = .13). Most common sites affected were the spine in 92% and the pelvis in 69% of patients, followed by the ribs (53%), the extremities (38%), and the skull (32%) (Figure 2). Of the patients with BM at sites other than spine and pelvis, the majority (93%) had concomitant spine or pelvis involved.

Figure 2.

Location of BMs.

Description of skeletal-related events in entire cohort

Forty-eight percent of our cohort (90 of 188) developed at least one SRE. The first SRE occurred at a median duration of 2.7 months (range 0–104.4 mo) after the first BM was identified (Table 2). For 14 patients who had only a pathological fracture or HCM as their first SRE, the median duration of time from BM diagnosis to the first SRE was 3.6 months (range 0–38.3 mo). Thirteen of 87 patients (15%) experienced their first SRE at the same time BM was identified, of whom four had symptomatic fracture leading to the diagnosis. The most common type of first SRE was radiotherapy (74%), followed by pathological fracture (23%), surgery (13%), HCM (7%), and spinal cord compression (3%). The first SRE most often affected the spine (58%), followed by the pelvis (17%) and the extremities (11%). Thirty-six of 188 patients (19%) developed at least two SREs. Similar to the first SRE, most of the second SREs were radiotherapy (69%) or pathological fracture (31%). The median duration between the first and second SREs was 8.1 month (range 0.3–78.7 mo).

Table 2.

Features of Skeletal-Related Events

	All Patients (n = 188)
Patients with one or more SREs, n, %
Yes	90 (48)
No	98 (52)
Number of SREs per patient, n, %a
1	54 (60)
2	26 (29)
3	8 (9)
4	1 (1)
10	1 (1)
First SRE presented concurrent with first BM identification, n, %a
Yes	13 (15)
No	74 (85)
Type of first SRE, n, %a
Radiation therapy to bone	67 (74)
Pathological fracture	21 (23)
Surgery to bone	12 (13)
HCM	6 (7)
Spinal cord compression	3 (3)
Location of first SRE, n, %a
Spine	52 (58)
Pelvis	15 (17)
Extremity	10 (11)
Rib	8 (9)
Scapula	3 (3)
Hip	3 (3)
Skull	3 (3)
Clavicle	2 (2)
Shoulder	1 (1)
Type of second SRE, n, %b
Radiation therapy to bone	25 (69)
Pathological fracture	11 (31)
Surgery to bone	7 (19)
Spinal cord compression	6 (17)
Duration between first and second SRE, mo
Median	8.1
Range	0.3–78.7

^aPercentage calculated of 90 patients with one or more SREs.

^bPercentage calculated of 36 patient with two or more SREs.

Associations between the incidence of SREs and demographic and clinical measures of interest were evaluated. Patients with more than 10 bone lesions were more likely to experience an SRE compared with those with 10 or fewer bone lesions (odds ratio [OR] 2.4; P = .007). Conversely, there was no significant difference in age, gender, or presence of axial bone involvement between patients who developed an SRE and SRE-free patients (P > .05). The 5-year mortality rate was not different between patients who experienced SREs and SRE-free patients, in which 42 of the 90 patients with SREs (47%) died within 5 years after the MTC diagnosis compared with 57 of the 98 SRE-free patients (58%) (P = .11).

Subanalysis of antiresorptive therapy in patients with BMs

After excluding four patients receiving ARTs for osteoporosis and seven who lacked treatment information, we performed a subanalysis of the remaining 177 patients regarding the use of ART in our population (Figure 1). Eighty-four of 177 patients (47%) received ARTs for BM (ART+ group) and 93 of 177 patients (53%) were not treated with ARTs (ART− group). In the ART+ group, 68% received ARTs from the pharmacy of M. D. Anderson Cancer Center. ARTs were initiated before developing SREs in 56 of 84 patients (67%). Thirteen patients received multiple agents during the assessment period. ZA was the most frequently used agent (64 patients), whereas 16 patients received PAM and 12 patients were given DMab. The median duration of ART treatment was 3.9 months (range 0–117 mo) with a median total number of doses of two (range 1–64). On average, ARTs were administered on a monthly basis (range 0.25–12 mo).

Patients in the ART+ and ART− groups were comparable in terms of age at diagnosis, gender, ethnicity, germline and somatic mutations, and nonosseous distant metastases (data not shown, all P > .05). However, compared with patients in the ART− group, a higher percentage of patients in ART+ group had more bone lesions (P = .006) and more involved sites (P = .026) (Table 3). The median duration of follow-up from MTC diagnosis was 5.6 years in the ART+ group. During that period, 56 of 84 patients (67%) died. For the patients in the ART− group, the median duration of follow-up from MTC diagnosis was 5.5 years, of which 49 of 93 patients died (53%) (P = .067). When a multivariable logistic regression model adjusting for the number of bone lesions and involved sites is performed, the P value comparing ART+ patients with ART− patients increases from P = .06 (OR [95% confidence interval [CI]] 1.80 [0.98, 3.30]) to P = .22 (OR [95% CI] 1.51 [0.78, 2.92]), indicating there is no difference in the incidence of death. No deaths were related to an SRE.

Table 3.

Comparison of BM Between ART+ and ART− Groups

Groups	ART+ Patients (n = 84)	ART− Patients (n = 93	P Value
Number of bone lesions, n, %			.006a
1–10	19 (24)	40 (45)
>10	61 (76)	49 (55)
Missing	4	4
Total number of skeletal sites involved, n, %			.026b
1	8 (10)	23 (25)
2	13 (16)	20 (22)
3	20 (24)	17 (18)
4	13 (16)	15 (16)
5	13 (16)	10 (11)
6–9	16 (19)	7 (8)
Missing	1	1

^aFisher's exact test.

^bGeneralized Fisher's exact test.

Four patients in the ART+ group, compared with none in the ART− group, developed ONJ. Three of the four patients were diagnosed with ONJ while on concurrent ARTs (ZA, two patients; DMab, one patient) and antiangiogenic tyrosine kinase inhibitors, of whom one also had prior radiotherapy to the neck and brain. The fourth patient, who had a prior history of radiation for brain metastasis, developed ONJ 6 years after stopping PAM and 6 months after holding sorafenib for a jaw fracture. One patient required surgery for ONJ and the other three received conservative, noninvasive treatment. Within the ART+ group, pretreatment hypoparathyroidism was noted in 9 of 80 patients (11%) and vitamin D deficiency in 32 of 56 patients (57%). Prior to starting ARTs, of 59 patients with available calcium measurements, hypocalcemia was noted as follows: grade 1 in 75%, grade 2 in 14%, grade 3 in 3%, and grade 4 in 0%. Worsening hypocalcemia after ART treatment was observed in 18 patients, most to a grade 1 or 2 level (grade 0 to 1 [n = 3], grade 1 to 2 [n = 10], grade 1 to 3 [n = 2], grade 2 to 3 [n = 2], and grade 3 to 4 [n = 1]). Creatinine levels in the patients remained stable after the AR treatment.

Incidences of SREs identified in each ART group were previously presented in abstract form, but due to the retrospective nature of this study and the heterogeneity of the two populations, these data are not included in this report (17).

Discussion

To our knowledge, this report of 188 medullary thyroid cancer patients with bone metastases is the largest in the literature describing the extent and natural history of BM in this rare cancer. Bone metastasis occurs commonly in MTC, particularly in association with other sites of metastatic disease. Our prevalence of 19%, coupled with a previous report by Giraudet et al (14) of 40%, highlights the importance of bone metastasis in this group of patients and confirms recommendations to screen for the presence of bone metastasis in patients with significantly elevated serum calcitonin or carcinoembryonic antigen levels.

Our study revealed that BMs in MTC were almost uniformly multifocal, with 77% of patients having at least six documented lesions. The spine and pelvis were the most common sites of bone metastases. In the patients who had BMs found outside the spine or pelvis, only 7% did not have metastatic disease in the spine or pelvis. These findings support the American Thyroid Association's published guidelines on MTC recommending a magnetic resonance imaging of the axial skeleton and pelvis and bone scintigraphy to identify central and peripheral BM, respectively (18).

The incidence of SREs was high (48%) in patients with identified bone metastasis; however, it is lower than the 78% reported in patients with DTC (11). Consistent with other solid tumors and DTC, the most common type of SRE was radiotherapy, which is usually indicated for painful lesions, existing or impending fracture, or pending spinal cord compression. The most frequent site of SRE was in the spine, the location with greatest risk of developing a critical consequence, namely spinal cord compression. However, spinal cord compression was the least common SRE (3%) in our cohort. In contrast, in an analysis of 202 DTC patients with spinal metastases, 72% of patients with follicular thyroid cancer and 36% of patients with papillary thyroid cancer developed neural structure compression (19). The difference in the prevalence of spinal cord compression seen between MTC and DTC patients may be related to the more indolent pace of MTC or may reflect the different biology of these two tumor types. SREs are prognostic indicators of poor survival in prostate and breast cancer patients (20, 21); however, no difference in the mortality rate was seen between the patients with and without SREs in our cohort.

ARTs have an established benefit in prolonging SRE-free survival and possible overall survival in prostate, breast, colorectal, and lung cancer (8–10, 22, 23) and perhaps also in malignant pheochromocytoma and paraganglioma (6). Additionally, studies have demonstrated the possible antitumor and antimetastatic effect of ARTs (24, 25). Because the patients in our study represented a heterogeneous group who were not randomized to ART in a prospective fashion based on predetermined criteria and did not receive standardized regimens of ART, we cannot make any definitive conclusions regarding the efficacy of ART in our MTC patients with BMs.

Patients with metastatic thyroid cancer have predisposing factors for developing ONJ or worsening hypocalcemia from ART therapy, including postoperative hypoparathyroidism, radiation exposure to the jaw during treatment to the neck, and use of antiangiogenic systemic chemotherapy. The reported incidence of ONJ in patients receiving iv bisphosphonates is 0.6%–9%, with an increased risk in patients receiving high cumulative doses of bisphosphonates (26). In our study, 4 of 84 patients (4.8%) who received ART developed ONJ. Although worsening of hypocalcemia was seen after ART treatment, only 5 of 84 (6%) had grade 3 or 4 hypocalcemia. These findings raise the importance of educating patients and providers about the potential risks with ART treatment and to implement strategies to prevent these side effects, such as maintaining good dental hygiene, smoking cessation, avoiding invasive dental procedures, and correcting hypocalcemia and vitamin D deficiency before starting an ART.

We recognize that there are limitations with any retrospective study including incomplete documentation, problematic verification of information, and differences in patient population. Patients in our study had varying radiology tests, sites of imaging, and intervals of monitoring, which can underestimate the true prevalence and distribution of bone metastases. Radiation therapy is a surrogate for cancer-induced, skeletal-related events of pain, impending fracture, or risk of spinal cord compression; however, in clinical practice, clinicians sometimes recommend radiotherapy based on individual clinical judgment. Thus, it is difficult to know what the outcomes of BM in this population would be without radiotherapy. Symptomatic SREs defined by symptomatic fracture rather than fracture incidentally identified on imaging, have been used in recent clinical trials and likely are more clinically relevant (10). However, data regarding symptomatic, pathological fractures in our study were not routinely documented and could not be analyzed. Calcitonin and carcinoembryonic antigen levels at the time of BM detection and occurrence of SREs were not uniformly available, preventing us from correlating BM and SREs with tumor markers. Despite these limitations, the findings from our study are of clinical relevance with respect to more thoroughly describing the prevalence, clinical characteristics, and natural history of BM in MTC patients and emphasizing that bone metastases leading to SREs are highly prevalent in this population.

Conclusion

This study describes the prevalence, clinical characteristics, and natural history of bone metastases and skeletal-related events in the largest cohort of medullary thyroid carcinoma patients reported to date. The spine and pelvis are the sites most often involved, and SREs (primarily radiotherapy and pathological fracture) are common. Reassuringly, the clinically devastating SRE of spinal cord compression is rare in patients with MTC. The questions regarding optimal dosing regimen and efficacy of antiresorptive treatment for bone metastases in MTC patients in reducing subsequent SREs or its antitumoral role in MTC should be investigated by a randomized, prospective study.