A, Kaplan-Meier curve showing overall 5-year survival by race/ethnicity. B, Kaplan-Meier curve showing disease-specific 5-year survival by race/ethnicity. See the Predictor Variables subsection in the Methods section for a description of “Other.”
Survival analysis based on surgical intervention. LN indicates lymph node examination.
eTable. Multivariable Analyses of Predictors of Treatment
eFigure. Five-Year Survival by Year of Diagnosis
Customize your JAMA Network experience by selecting one or more topics from the list below.
Roche AM, Fedewa SA, Chen AY. Association of Socioeconomic Status and Race/Ethnicity With Treatment and Survival in Patients With Medullary Thyroid Cancer. JAMA Otolaryngol Head Neck Surg. 2016;142(8):763–771. doi:10.1001/jamaoto.2016.1051
Medullary thyroid cancer (MTC) is a relatively rare neoplasm of the thyroid but accounts for 14% of thyroid cancer–related deaths. Female sex, young age, and stage at presentation have been found to predict survival and treatment. However, patterns of survival and treatment by socioeconomic status and race/ethnicity have not been fully described.
To determine whether socioeconomic status and race/ethnicity are associated with survival and treatment in patients with MTC.
Design, Setting, and Participants
Data for 1647 patients with MTC from January 1, 1998, to December 31, 2011, in the Surveillance, Epidemiology, and End Results (SEER) Program registry were examined. Data analysis was conducted from June 1, 2013, to July 31, 2014.
Main Outcomes and Measures
Differences in receipt of thyroidectomy and lymph node examination by race/ethnicity were examined using logistic regression models. Overall and disease-specific survival were examined by race/ethnicity using Kaplan-Meier survival curves and adjusted Cox proportional hazards regression models.
Of the 1647 patients with MTC were 1192 white (72.4%), 139 black (8.4%), 222 Hispanic (13.5%), and 94 other races/ethnicities (5.7%). Of these, 1539 (93.4%) underwent surgical treatment. There were no differences in receipt of thyroidectomy by race/ethnicity; however, black patients (adjusted odds ratio, 0.61; 95% CI, 0.39-0.93) and female patients (adjusted odds ratio, 0.76; 95% CI, 0.59-0.99) were less likely to undergo lymph node examination compared with non-Hispanic white and male patients. Black patients had lower overall (adjusted hazard ratio, 2.40; 95% CI, 1.45-3.98) and disease-specific survival (adjusted hazard ratio, 2.9; 95% CI, 1.64-5.14) compared with non-Hispanic white patients.
Conclusions and Revelance
In this population-based study of patients with MTC, black patients were less likely to have lymph node examination following surgery. Furthermore, Hispanic and black patients had poorer overall and disease-specific survival compared with non-Hispanic white patients after accounting for clinical factors. Racial/ethnic disparities exist in the type of treatment as well as outcomes in patients with MTC.
Medullary thyroid cancer (MTC) is a neoplasm of the thyroid neuroendocrine parafollicular or calcitonin cells and can occur sporadically or may be hereditary. Medullary thyroid cancer accounts for approximately 3% of all thyroid cancer, but it is responsible for 14% of thyroid cancer–related deaths.1,2 Patients with MTC have a slightly lower 10-year relative survival risk compared with patients with papillary and follicular carcinoma (0.80 vs 0.98 and 0.92, respectively),2 and presence of cervical lymph node involvement confers a poorer prognosis for patients with MTC.3 Based on observational studies and cooperative group consensus, the treatment for MTC is complete surgical extirpation, which includes at minimum a total thyroidectomy and central compartment neck dissection4-6 and may include varying levels of lateral neck dissection, depending on the clinical presentation.7,8 Radioactive iodine I-131 is not used to treat MTC because parafollicular calcitonin cells do not actively take up iodine. External beam radiotherapy has been shown to have no improvement in overall survival,9 and has been shown to be an independent predictor of decreased survival.10 Existing literature has identified several predictors of survival for MTC, including female sex, young age at the time of diagnosis, localized stage, well- or moderately differentiated histologic features, and absence of cervical lymph node involvement.3,10,11 Medullary thyroid cancer is treated surgically, and research has shown that more extensive thyroid surgery (total thyroidectomy vs lobectomy) contributes to better overall survival.10 Robust literature exists about the prognostic role of histologic and staging characteristics of patients; however, there is a dearth of information on the role that race/ethnicity and socioeconomic status play in treatment and survival outcomes. Our objective is to determine the effect that race/ethnicity and socioeconomic status have on treatment and survival outcomes by using the population-based Surveillance, Epidemiology, and End Results (SEER) Program registry. We hypothesize that treatment and survival outcomes are influenced by race/ethnicity and socioeconomic status.
Question What are the effects of race/ethnicity and socioeconomic status on the treatment of and survival in patients with medullary thyroid cancer?
Findings In this retrospective survival analysis, black and female patients were less likely to undergo lymph node examination in addition to thyroidectomy. Black patients had lower overall and disease-specific survival compared with non-Hispanic white patients.
Meaning Racial/ethnic disparities exist in the type of treatment for and outcomes in patients with medullary thyroid cancer.
Data from the SEER-18 population-based cancer registry covering approximately 28% of the United States population were used in this study. The 18 SEER regions included Alaska natives; metropolitan Atlanta, Georgia; California (excluding San Francisco and Los Angeles); Connecticut; Detroit, Michigan; greater Georgia; Hawaii; Iowa; Kentucky; Los Angeles, California; Louisiana; New Jersey; New Mexico; rural Georgia; San Francisco and Oakland, California; metropolitan San Jose and metropolitan Monterey, California; Seattle, Washington; and Utah. The SEER database is publicly available; therefore, the Emory Institutional Review Board waived approval for this study. All patient data within the SEER database are deidentified. A total of 1754 adults (aged 18 years and older) who were diagnosed with their first primary malignant MTC between January 1, 1998, and December 31, 2011, were selected from the SEER database. Only patients diagnosed after 1998 were selected based on the availability of data on primary treatment and lymph node dissection. Medullary carcinomas were included using the following codes based on the International Agency for Cancer Research classification: 8510 to 8513 and 8345. Cases missing data on race/ethnicity (n = 27) were excluded, as were cases with missing data on surgery (n = 16), radiotherapy (n = 33), and lymph node dissection (n = 31). The total analytic cohort contained 1647 patients with MTC.
The primary outcome was first primary treatment for MTC, which included surgery, radiotherapy, and lymph node dissection. Surgical treatment was categorized as thyroidectomy, which included subtotal thyroidectomy. We also studied rates of thyroidectomy with examination of lymph nodes and thyroidectomy with radiotherapy. Our secondary outcome was overall and disease-specific survival of patients with MTC.
Race/ethnicity was classified as non-Hispanic white, Hispanic, black, and other (Asian or Pacific Islander, Native Americans, Native Alaskans, and other race/ethnicity as specified by the patient). Asian or Pacific Islander was included in the category of other race/ethnicity owing to the small numbers of these patients. Age was considered as a continuous variable. Sex was also included. County-level socioeconomic status, as measured by the percentage of adults living in households below the federal poverty level according to the 2000 Census, was also measured and dichotomized into low and high poverty using the median cutoff point (low poverty being ≥15% of adults living below the federal poverty level) in this study population. Time was divided into 3 periods: 1998-2002, 2003-2007, and 2008-2011. Marital status was also considered and grouped as married and nonmarried, which included single, widowed, and separated. Clinical characteristics, including tumor size (<1.0, 1.0-1.9, 2.0-2.9, 3.0-3.9, and ≥4.0 cm) and nodal spread (localized, regional, and distant) were considered. Beginning in 2004, SEER collected data on whether tumors were unifocal or multifocal as part of the collaborative stage system12; and sensitivity analyses limited to patients diagnosed in 2004 and after were also conducted.
Analyses were performed from June 1, 2013, to July 31, 2014, with SAS, version 9.2 (SAS Institute Inc). χ2 Statistics were calculated to analyze the association between race/ethnicity and categorical covariates, with P = .05 considered significant. t Tests and F tests were used to examine variations in continuous variables (age at diagnosis) by race/ethnicity. Adjusted logistic models were used to estimate odds ratios (aORs), and 95% CIs were used to assess characteristics associated with treatment. Three models examining treatment were conducted: the first examined receipt of thyroidectomy vs no surgery, the second compared thyroidectomy with lymph node dissection vs thyroidectomy with no lymph node dissection, and the third compared thyroidectomy plus radiotherapy vs thyroidectomy alone. Kaplan-Meier survival curves were used to calculate unadjusted cumulative 5-year disease-specific and overall survival estimates and statistical significance was determined by the log-rank test. Adjusted Cox proportional hazards regression models were used to estimate 5-year disease-specific and 5-year overall survival hazard ratios (aHRs) and 95% CIs. For the Cox proportional hazards regression models, we tested significance of the proportional hazards assumptions using time and covariate interaction terms. All models were restricted to cases with complete information on all covariates included in the model.
A total of 1647 patients were included in this analysis (Table 1). Most patients were non-Hispanic white (1192 [72.4%]), while 222 [13.5%] were Hispanic, 139 [8.4%] were black, and 94 [5.7%] were other race/ethnicity. The mean (SD) age at diagnosis was 52.8 (15.6) years. Non-Hispanic white patients were older (54.0 [15.5] years) compared with Hispanic (49.1 [15.2] years), black (49.7 [15.2] years), and other race/ethnicities (50.0 [15.5] years). Most patients were women (981 [59.6%]), and sex did not vary significantly across race/ethnicity (P = .07). Regarding socioeconomic factors, 108 non-Hispanic white patients (9.1%) diagnosed with MTC live in counties with 15% or more of the adults living below the federal poverty level, compared with 32 Hispanic patients (14.4%) and 18 black patients (12.9%) (P = .006). Tumor size at presentation varied by race/ethnicity (P = .003). Hispanic and black patients tended to present with tumor size of 4 cm or greater compared with non-Hispanic white patients (42 [30.2%] and 55 [24.8%] vs 247 [20.7%], respectively), and non-Hispanic white patients more frequently presented with tumor sizes between 1 and 2 cm compared with Hispanic and black patients (295 [24.8%] vs 33 [14.9%] and 26 [18.7%], respectively).
Most patients in our study (1093 [66.4%]) underwent combined thyroidectomy and lymph node examination (Table 1). Thyroidectomy alone was performed in 446 patients (27.1%), and 108 patients (6.6%) did not undergo surgery. Patients 65 years or older were less likely to undergo thyroidectomy (353 of 395 [89.4%]) compared with patients younger than 65 years (1186 of 1252 [94.7%]) and had lower adjusted odds of undergoing surgery (aOR, 0.38; 95% CI, 0.21-0.70). Among patients undergoing surgery, those diagnosed between 1998 and 2002 (152 of 387 [39.3%]; aOR, 0.36; 95% CI, 0.26-0.49) and 2003 and 2007 (151 of 548 [27.5%]; aOR, 0.72; 95% CI, 0.54-0.96) were less likely to receive lymph node examination compared with patients diagnosed between 2008 and 2011 (143 of 604 [23.7%]) (P < .001). Type of surgical intervention varied by tumor size, as patients with larger tumors (≥4 cm, 259 of 362 [71.6%]; 1-2 cm, 254 of 370 [68.7%]; 2-3 cm, 221 of 295 [74.9%]; 3-4 cm, 147 of 215 [68.4%]) were generally more likely to undergo thyroidectomy with lymph node examination compared with patients with smaller tumors (<1 cm, 161 of 273 [59.0%]) in unadjusted and adjusted analyses (eTable in the Supplement). Radiotherapy was administered in 274 patients (16.6%). Of the 108 patients who did not undergo surgery (6.6%), 32 (29.6%) underwent radiotherapy. Of the 446 patients (27.1%) who underwent thyroidectomy without lymph node examination, 41 (9.2%) underwent radiotherapy. Among 1093 patients who underwent surgery and had lymph nodes removed, 1201 (18.4%) received radiation therapy, which did not vary by race/ethnicity.
In adjusted analyses, factors that predicted thyroidectomy vs no surgery included tumor size of 4 cm or more (aOR, 0.36; 95% CI, 0.13-0.97), presence of distant nodes (aOR, 0.02; 95% CI, 0.003-0.18), and age at diagnosis (aOR, 0.97; 95% CI, 0.95-0.98). Patients who were married were more than 2 times more likely to undergo thyroidectomy compared with patients who were widowed, divorced, or single (aOR, 2.80; 95% CI, 1.54-5.12) (Table 2).
Factors that predicted lymph node examination vs no lymph node examination among patients who underwent thyroidectomy included race/ethnicity, sex, tumor size, age at diagnosis, year of diagnosis, and poverty level (eTable in the Supplement). Black patients were less likely to undergo lymph node examination (aOR, 0.61; 95% CI, 0.39-0.93) compared with non-Hispanic white patients, and female patients were less likely to undergo lymph node examination (aOR, 0.76; 95% CI, 0.59-0.99) compared with male patients. In addition, patients living in counties with 15% or more of the adults living below the federal poverty level were less likely to undergo lymph node examination (aOR, 0.58; 95% CI, 0.39-0.86). Patients with tumors larger than 1 cm were more likely to undergo lymph node examination (1-2 cm: aOR, 1.92; 95% CI, 1.34-2.74; 2-3 cm: aOR, 2.42; 95% CI, 1.64-3.56; 3-4 cm: aOR, 1.95; 95% CI, 1.29-2.95; and ≥4 cm: aOR, 2.81; 95% CI, 1.92-4.12). Patients diagnosed a longer amount of time before surgery were less likely to undergo lymph node examination (aOR, 0.72; 95% CI, 0.54-0.96 in 2003-2007; and aOR, 0.36; 95% CI, 0.26-0.49 in 1998-2002) when compared with patients diagnosed after 2008.
Factors that predicted whether patients who underwent thyroidectomy received radiotherapy included tumor size and presence of involved regional lymph nodes (eTable in the Supplement). Patients whose tumors were between 1 and 2 cm were twice as likely to receive radiotherapy compared with those whose tumors were less than 1 cm (aOR, 2.06; 95% CI,1.03-4.08). Patients whose tumors were between 3 and 4 cm were more than 3 times as likely to receive radiotherapy (aOR, 3.64; 95% CI, 1.81-7.33) and those whose tumors were 4 cm or larger were 4 times more likely to receive radiotherapy (aOR, 4.01; 95% CI, 2.08-7.72) compared with patients whose tumors were smaller than 1 cm. Patients with involvement of regional lymph nodes were 4 times more likely to receive radiotherapy (aOR, 4.32; 95% CI, 3.07-6.07).
The Kaplan-Meier curve for 5-year overall survival, analyzed by race/ethnicity, shows that black patients have a lower overall survival (74.8%) than non-Hispanic white (83.4%) and Hispanic patients (79.1%), although this finding was not statistically significant (P = .06) (Figure 1A). Disease-specific 5-year survival was significantly lower for black patients (79.6%) when compared with non-Hispanic white (88.0%) and Hispanic patients (86.1%) (P = .046) (Figure 1B). In a survival analysis based on surgical intervention, disease-specific 5-year survival was notably lower in patients who did not undergo any type of surgical intervention (Figure 2).
When analyzed by year of diagnosis, overall survival was lower for patients diagnosed in 1998-2002 (78.7%) compared with those diagnosed in 2003-2007 (82.9%) and 2008-2011 (84.5%), although this finding was not statistically significant (P = .09) (eFigure, A, in the Supplement). Disease-specific survival was also significantly lower in 1998-2002 (83.0%) when compared with later years of diagnosis (89.3% in 2003-2007; 9.1% in 2008-2011) (P = .009) (eFigure, B, in the Supplement).
Multivariable Cox proportional hazards regression models examining 5-year disease-specific and overall survival are presented in Table 3. Other race/ethnicity could not be determined in this model owing to unstable estimates for the Cox proportional hazards regression models and were thus not included. Black patients (aHR, 2.40; 95% CI, 1.45-3.98) and those who were older at diagnosis (aHR, 1.07; 95% CI, 1.05-1.08) had higher overall hazards of death compared with non-Hispanic white and younger patients, respectively. Married patients had an increased overall survival compared with those who were widowed, divorced, or single (aHR, 0.71; 95% CI, 0.50-0.99). Presence of regional lymph node involvement (aHR, 2.72; 95% CI, 1.88-3.93) was associated with an increased risk of all causes of death. When examining the role that radiotherapy had on outcomes, patients who underwent radiotherapy alone were nearly 4 times more likely to die (aHR, 3.81; 95% CI, 1.69-8.59), while patients who underwent surgery and radiotherapy (aHR, 0.39; 95% CI, 0.21-0.72) and surgery alone (aHR, 0.23; 95% CI, 0.13-0.42) were much less likely to die. Increasing tumor size was also associated with increased risk of death. Receipt of thyroidectomy, regardless of lymph node dissection, was associated with a decrease in the risk of death compared with no surgery. Patients who were diagnosed most recently appeared to have a slightly decreased risk of death (aHR, 0.92; 95% CI, 0.87-0.96). Patients who underwent radiotherapy had poorer overall survival compared with patients who underwent surgery alone (aHR, 1.54; 95% CI, 1.04-2.28) even after controlling for the available clinical information (tumor size and lymph node spread).
Estimates of 5-year disease-specific survival with a Cox proportional hazards regression model are presented in Table 3. Disease-specific results were similar to overall survival results: black patients had a greater than 2-fold increase in the risk of death owing to MTC (aHR, 2.85; 95% CI, 1.61-5.01), while regional lymph node involvement increased the chance of death 4-fold (aHR, 4.34; 95% CI, 2.70-6.93). Tumor size of 4 cm or larger (aHR, 1.77; 95% CI, 1.10-2.84 was also associated with increased risk of death. Diagnosis in earlier years increased the risk of death from MTC (aHR, 0.90; 95% CI, 0.85-0.95).
Receipt of thyroidectomy regardless of lymph node dissection was associated with a decreased risk of disease-specific death (thyroidectomy alone: aHR, 0.26; 95% CI, 0.11-0.58; thyroidectomy with lymph node examination: aHR, 0.14; 95% CI, 0.08-0.24). As in overall survival, patients diagnosed more recently were less likely to die from causes related to MTC (aHR, 0.90; 95% CI, 0.85-0.95) and patients who underwent radiotherapy alone had worse disease-specific survival (aHR, 0.17; 95% CI, 0.09-0.34) compared with patients who underwent surgery and radiotherapy (aHR, 0.36; 95% CI, 0.18-0.71).
In our population-based study of the largest analyzed cohort of patients diagnosed with MTC, nonclinical factors, specifically black race, were independently associated with treatment, overall survival, and disease-specific survival. Although black race was not associated with undergoing thyroidectomy vs no thyroidectomy, it was associated with whether lymph nodes were examined among those receiving surgical treatment. Additional nonclinical factors that were associated with treatment include age at diagnosis, marital status, and socioeconomic status; older patients were less likely to undergo surgery, married patients were more likely to undergo surgery, and patients who lived in areas of high poverty were less likely to undergo lymph node examination. Among the patients who underwent surgery, both clinical and demographic factors predicted whether patients underwent lymph node examination: predictors of patients who did not undergo lymph node examination included black race, female sex, larger tumor size, high county poverty level, and remote (1998-2007) year of diagnosis.
Risk of death for overall survival was predicted by black race, older age at presentation of MTC, large tumor size, and regional lymph node involvement. Risk of death for disease-specific survival was predicted by black race, tumor size, regional lymph node involvement, and year of diagnosis. Our findings regarding the effect of race/ethnicity on survival substantiate those of Gilliland et al13 in their analysis of thyroid cancer; however, their study included all histologic subtypes of thyroid cancer and did not control for treatment.
Our results support previous findings that patients who undergo surgery for MTC live longer than patients who do not undergo surgery10; patients in our study who did not undergo surgery had lower overall and disease-specific 5-year survival. Our findings also corroborate previous literature that identifies advanced age at diagnosis, larger tumor size, and lymph node involvement as predictors of poor prognosis.1,3,10,14
We found that patients who underwent radiotherapy had poorer overall survival and disease-specific survival; radiotherapy was also an independent predictor of increased risk of death, a finding that has been previously reported.10 In patients who underwent only radiotherapy and who had overall poor outcomes, it is likely that these patients underwent radiotherapy as palliative care. We also found that marital status has a protective effect on both treatment and overall 5-year survival, but not disease-specific survival, a finding that was not previously demonstrated in the subgroup analysis by Gilliland et al13 of patients with MTC. In an analysis of socioeconomic status in Asian and Pacific Islanders with head and neck cancer, Chu et al15 demonstrated that nonmarried status was associated with a significantly higher risk of death in patients with upper aerodigestive tract cancer; however, this study did not include patients with thyroid cancer.
In our survival models, we did not find that high poverty affected overall 5-year survival or disease-specific survival. Chu et al15 found that lower socioeconomic status was significantly associated with disease-specific survival in patients with head and neck cancer after adjusting for numerous clinical and demographic factors; however, their analysis was limited to patients with aerodigestive malignant neoplasms. We did find that black and Hispanic patients were more likely to live in areas of high poverty compared with non-Hispanic white patients, as demonstrated by Chu et al.15
The role of race in treatment and outcomes is complex and may be associated not only with differences in access to care but also with perceived discrimination by patients and racial bias inherent in institutional mechanisms. In their discussion of racial bias in health care and health, Williams and Wyatt16 state that segregation continues to isolate black patients to areas of urban poverty, with limited access to education and employment opportunities. Acknowledging the existence of racial disparities raises awareness among physicians and represents a step in the process of minimizing the effect that race has on the delivery of health care.
Like any population-based registry, the SEER database has inherent limitations. Data regarding the extent of lymph node dissection and examination are not included in the database throughout the entire time period, nor is a description of the circumstances under which the lymph nodes were examined (ie, elective vs intraoperative). In using socioeconomic data in our analysis, we acknowledge that information contained in the SEER database does not include individual-level socioeconomic status, but rather county-level socioeconomic status. These measures are likely correlated, although they may each represent independent risk factors in health outcomes.17 Last, the SEER database did not include data on chemotherapeutic agents or targeted therapy, and we were therefore unable to account for the effect that this treatment modality may have had on outcomes related to MTC.
We found that, in the largest cohort of patients diagnosed with MTC, to our knowledge, black race significantly affects treatment and overall survival in those diagnosed with MTC. In most patients, MTC continues to be treated largely with thyroidectomy with or without lymph node examination. There remains a small percentage of patients who are treated with radiotherapy only, despite national guidelines recommending surgery; however, this group may represent patients who underwent palliative radiotherapy. There is a trend toward more extensive surgery over time, as patients more recently diagnosed are more likely to undergo lymph node examination in addition to thyroidectomy. Overall and disease-specific 5-year survival rates were increased when surgery was performed and in patients diagnosed more recently. Differences in treatment and survival outcomes in black patients may be due in part to disparities in access to health care for patients living in lower socioeconomic regions. Societal racial bias may also contribute to disparities in treatment and outcomes. Public health efforts should be made to improve not only access to quality health care for all patients but also to raise awareness of potential societal racial discrimination within the health care system.
Accepted for Publication: April 2, 2016.
Corresponding Author: Ansley M. Roche MD, Department of Otolaryngology–Head and Neck Surgery, Emory University, 550 Peachtree St, Ste 1135, Atlanta, GA 30308 (email@example.com).
Published Online: June 2, 2016. doi:10.1001/jamaoto.2016.1051.
Author Contributions: Ms Fedewa had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Roche, Chen.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Roche, Chen.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: All authors.
Adminstrative, technical, or material support: Roche, Chen.
Study supervision: Chen.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Previous Presentations: This study was presented as a poster at the 15th International Thyroid Congress and 85th Annual Meeting of the American Thyroid Association; October 22, 2015; Orlando, Florida.
Create a personal account or sign in to: