All authors 3 authors contributed to 1) conception and design or analysis and interpretation of data, 2) drafting the article or revising it critically for important intellectual content, and 3) final approval of the version to be published.
Hurthle cell carcinoma (HCC) is an uncommon and more aggressive thyroid cancer. To date, there is a paucity of data at a population level. In this study, demographic, clinical, and pathologic characteristics of HCC were investigated and compared with other types of differentiated thyroid cancers (ODTCs). The authors also evaluated disease-specific survival and compliance with American Thyroid Association (ATA) management guidelines from 2009.
The Surveillance, Epidemiology, and End Results (SEER) database from 1988 to 2009 was used to obtain data on patients with thyroid cancer. Data analyses were performed using chi-square tests, analysis of variance, Kaplan-Meier analysis, binary logistic regression, and Cox proportional hazards regression.
In total, 3311 patients with HCC and 59,585 patients with ODTC were identified. Compared with ODTC, HCC was more common among men (31.1% vs 23.0% for ODTC; P < .001) and among older patients (mean age, 57.6 years vs 48.9 years for ODTC; P < .001). Patients with HCC presented with higher SEER disease stage (P < .001), and their tumors were larger (36.1 mm vs 20.2 mm for ODTC; P < .001). Fewer patients underwent total thyroidectomy (P = .028). Both overall and disease-specific survival were lower for patients with HCC (P < .001), and neither improved over the last 2 decades (P = .689). After adjustment, age ≥45 years, not undergoing surgery, and metastatic disease were strongly associated with a worse prognosis (hazard ratio >3.0). Compliance with recommended surgical treatment according to ATA guidelines was lower among patients with HCC aged ≥65 years (odds ratio [OR], 1.43; P = .002) and among unmarried patients (OR, 1.29; P = .004). Predictors of noncompliance with ATA guidelines for treatment with radioactive implants or radioisotopes were age ≥65 years (OR, 1.31; P = .017), diagnosis between 1988 and 1997, no surgery, and partial thyroidectomy (OR, 1.81, 19.48, and 4.02, respectively; P < .001).
Hurthle cells, or oxyphilic cells, are large polygonal cells with distinct borders derived from thyroid follicular epithelium. They contain abundant granular cytoplasm because of an excess of mitochondria and a large nucleus with a prominent nucleolus.1 Hurthle cell carcinoma (HCC), also known as oncocytic or oxyphilic carcinoma, is defined as a thyroid malignancy consisting of at least 75% Hurthle cells and differs from its benign counterpart (Hurthle cell adenoma) by the presence of capsular and/or vascular invasion.2
HCC has been classified by the World Health Organization as a variant of follicular cancer and is treated as such according to the American Thyroid Association (ATA) management guidelines for patients with thyroid nodules and differentiated thyroid cancer.3, 4 It has been noted that patients with HCC have a high propensity for lymph node and distant metastases, late recurrence, and a decreased avidity for I131, making it a more aggressive tumor than other differentiated thyroid cancers (ODTCs).2, 5, 6 HCC represents from 3% to 7% of all differentiated thyroid cancers.7 Most experience with HCC in the literature comes from a small number of single-institutional series; to our knowledge, the only large-scale studies on HCC are based on data that are more than a decade old.8, 9
The objectives of our current study were to compare the outcomes between patients with HCC and patients with ODTC and to evaluate the prognostic factors associated with HCC survival at a population level. We hypothesized that HCC has a more aggressive course than ODTC and that it may warrant consideration for separate management guidelines.
MATERIALS AND METHODS
Data Sources and Study Patients
The Surveillance, Epidemiology, and End Results (SEER) database was used to identify patients who had a diagnosis of HCC and ODTC; ODTC was defined as papillary and follicular thyroid cancers.10 This database uses the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) as the reference for histology coding (HCC, code 8290; papillary thyroid cancer, codes 8050, 8340-8344, and 8350; follicular thyroid cancer, codes 8330-8332, 8335, and 8337).11 Data were collected from 18 registries (San Francisco-Oakland, Connecticut, metropolitan Detroit, Hawaii, Iowa, New Mexico, Utah [from 1973], Seattle-Puget Sound [from 1974], metropolitan Atlanta [from 1975], Alaska, San Jose-Monterey, Los Angeles, rural Georgia [from 1992], greater California [excluding San Francisco, Los Angeles, and San Jose]; Kentucky, Louisiana, New Jersey, and greater Georgia [excluding Atlanta and Rural Georgia; from 2000]); collectively, these regions represent nearly 28% of the US population. The SEER population is comparable to that of the general United States with regard to poverty and education level, but it tends to be more urban and has a higher proportion of foreign-born individuals.12 Patients who were diagnosed with thyroid cancer between 1973 and 1987 were excluded because of a lack of data concerning surgical procedures and the use of radioactive implants or radioisotopes (RAI) use before 1988.
Demographic variables included patient sex, age at diagnosis, race, marital status, year of diagnosis, and survival status as of December 31, 2009. Age at diagnosis was classified into 3 groups: ages 18 to 44 years, 45 to 64 years, and ≥65 years. Race was recoded into white, black, and other (American Indian, Alaska Native, Asian, Pacific Islander, and other unspecified); marital status was coded as married and domestic partner, single, divorced, widowed or separated, and unknown. Year of diagnosis was treated as a categorical variable divided into the following intervals: 1988 to 1997, 1998 to 2006, and 2007 to 2009; these intervals were chosen because of SEER coding changes between 1997 and 1998 and the first release of ATA guidelines for the management of thyroid nodules and differentiated thyroid cancer in 2006.
Clinical variables of interest were the number of primary tumors, thyroid surgery, and radiation therapy. The number of primary tumors was grouped into 1, 2, and ≥3. Surgery was divided into 3 subcategories: no surgery, partial thyroidectomy (SEER codes 10-20 [1988-1997] and 20-30 [1998-2009]), and total thyroidectomy (SEER codes 30-70 [1998-1997] and 40-80 [1998-2009]). Radiation therapy was categorized into 5 groups: none, external-beam radiation therapy, RAI, a combination of external-beam radiation therapy and radioactive implants or radioisotopes (both), and other/unknown. RAI administration was binary: no (none, refused by patient, or external-beam radiation therapy only) and yes (including radioisotopes, radioactive implants, or both for survival analysis and “radiation recommended but unknown if administered” for compliance analysis).
Pathologic information included SEER stage, tumor size, extent of tumor, and lymph nodes examined. The SEER staging system is based on 3 levels of invasion: The tumor is described as local if it is confined entirely to the thyroid, regional if it extends beyond the thyroid into surrounding tissues or has metastasized to regional lymph nodes, and distant if metastases to extracervical lymph nodes or organs are present. According to the American Joint Commission on Cancer (AJCC), tumor size was subdivided into 3 groups (≤2.0 cm, 2.1-4.0 cm, ≥4.1 cm); tumors >15.0 cm were excluded because of the possibility of coding error.12 Tumor extent was classified into intrathyroidal, extrathyroidal, and unknown. The number of lymph nodes examined was categorized as none (no lymph nodes were examined), negative (lymph nodes examined were all negative), or positive.
American Thyroid Association Management Guidelines Recommendations
According to ATA management guidelines recommendation 26, all patients with differentiated thyroid cancer should undergo total or near total thyroidectomy unless there are contraindications to surgery.5 Lobectomy alone may be sufficient for small (<1 cm), low-risk, unifocal, and/or intrathyroid papillary thyroid carcinomas. Recommendation 32 states that expert opinion supports the use of postoperative RAI for higher risk tumors ≥1 cm, such as HCC.5
Simple summary statistics were used to describe demographic, clinical, and pathologic characteristics of the study population. Chi-square tests and analyses of variance were used to analyze categorical and continuous variables, respectively; binary logistic regression was performed to identify independent predictors of noncompliance with ATA management guidelines recommendations 26 and 32.
For univariate analysis of survival, the Kaplan-Meier method was used to determine prognostic factors, and the log-rank test was used to calculate statistical significance. For multivariate analyses, Cox proportional hazard regression modeling was used to analyze the factors associated with the outcomes of interest based on their statistical significance in univariate analysis. Hazard ratios and 95% confidence intervals were calculated for the strength of association between each variable and survival.
Data analysis was performed using Statistical Package for the Social Sciences (SPSS) software (version 19.0; SPSS Inc., Chicago, Ill). All tests were 2-sided, and statistical significance was set at a P value < .05. SEER data are publicly available, and all patient information is deidentified; therefore, this study was deemed exempt from institutional review board approval.
In total, 3311 patients aged ≥18 years with a diagnosis of HCC and 59,585 with a histologically confirmed diagnosis of ODTC were identified between 1988 and 2009.
HCC was less common than ODTC among women (68.9% vs 77%, respectively; P < .001) (Table 1). Patients who were diagnosed with HCC were older (aged 57.6 years vs 48.9 years for ODTC; P < .001), and fewer patients with HCC were alive at the end of the follow-up (82.1% vs 89.2% for ODTC; P < .001). Whites were diagnosed with HCC more often than blacks (85.9% vs 82.8% for ODTC; P < .001), as were single patients (34.8% vs 32% of married patients; P = .004). The incidence of HCC among all differentiated thyroid cancers rose during the late 1990s from 3.8% to 5.7% in 2006 (P < .001) and appeared to stabilize at 5.6% between 2007 and 2009.
Table 1. Demographic, Clinical, and Pathologic Characteristics of Patients With Histologically Confirmed Hurthle Cell Thyroid Cancer Versus Patients With Histologically Confirmed ODTC (Papillary or Follicular Thyroid Cancers): Surveillance, Epidemiology, and End Results, 1988 to 2009 (Hurthle Cell Carcinoma, n = 3311; Other Differentiated Thyroid Cancers, n = 59,585)
Abbreviations: EBRT, external-beam radiation therapy; HCC, Hurthle cell carcinoma; ODTC, other differentiated thyroid cancers; RAI, radioactive implants or radioisotopes; SD, standard deviation; SEER, Surveillance, Epidemiology, and End Results; SEM, standard error of the mean.
aPercentages have been rounded and may not add up to 100%.
Age at diagnosis: Mean±SEM, y
n = 3290
n = 59,130
Year of diagnosis
No. of primary tumors/patients
n = 3261
n = 59,235
n = 3288
n = 59,578
Tumor size: Mean±SD, mm
Extent of tumor
n = 3306
n = 59,480
Lymph nodes examined
n = 3244
n = 58,794
Compared with patients who had ODTC, more patients who had HCC had a second primary malignancy (12.6% vs 18.8%, respectively; P < .001). Fewer patients with HCC underwent total thyroidectomy (76.8% vs 74.9%; P = .028) and received no radiation therapy (51.7% vs 47.6%; P < .001).
Patients with HCC were diagnosed at a higher SEER stage: 51.4% of patients with HCC had localized disease versus 64.3% of patients with ODTC (P < .001). HCC mean tumor size was greater (36.1 mm vs 20.2 mm for ODTC; P < .001), but there were no differences in extrathyroidal spread. More patients with HCC had no lymph nodes examined (73.9% vs 62.5% for ODTC; P < .001); of those who did have lymph nodes examined, fewer patients who had HCC had positive lymph nodes compared with patients who had ODTC (5.3% vs 15.8%, respectively; P < .001).
The overall survival rate was 82.1% for patients with HCC and 89.2% for patients with ODTC (P < .001). Disease-specific survival also was lower in patients with HCC (Fig. 1): Disease-specific mortality occurred in 5.9% of patients with HCC (n = 194) and in 2.7% of patients with ODTC (n = 1584; P < .001). Time-trend analysis for disease-specific survival among patients with HCC indicated that there has been no improvement over the last 2 decades (P = .689) (Fig. 2), whereas disease-specific survival increased over the same period for patients with ODTC. Controlling for disease stage, patients with HCC who were treated in accordance with ATA guidelines had reduced disease-specific survival (P < .001) compared with those who had ODTC and were treated according to guidelines.
In univariate analysis, several factors were associated with compromised disease-specific survival in patients with HCC, including being a man, older age at diagnosis (most pronounced for those aged ≥65 years), and being single (Table 2). Pathologic characteristics that were associated with improved survival included lower SEER stage, tumor confined within the thyroid capsule, tumor size <4 cm, and no lymph node involvement. Patients who did not undergo surgery had an 8.33 times greater risk of dying (P < .001). The administration of postoperative RAI was associated with reduced mortality (hazard ratio, 0.66; P = .005).
Table 2. Univariate Analysis of Disease-Specific Mortality
HR (95 % CI)
Abbreviations: CI, confidence interval; HR, hazard ratio; NS, statistically nonsignificant; RAI, radioactive implants or radioisotopes; SEER, Surveillance, Epidemiology, and End Results.
Age at diagnosis, y
Extent of tumor
Tumor size, cm
Lymph nodes examined
On multivariate analysis, factors that were independently associated with lower disease-specific survival in patients who were diagnosed with HCC included not being married, not undergoing any surgery, extrathyroidal invasion, and tumor size ≥4 cm. Disease-specific survival was strongly associated with disease stage and patient age ≥45 years (Table 3).
Table 3. Cox Multivariate Regression Analysis of Factors Associated Independently With Disease-Specific Mortality (Total, n = 2670; No. of events = 125)
Abbreviations: CI, confidence interval; HR, hazard ratio; SEER, Surveillance, Epidemiology, and End Results.
Reference groups: ages 18-44 years; marital status, married; surgery, total thyroidectomy; SEER stage, localized; extent of tumor, intrathyroidal; tumor size, 1-20 mm.
Ages 45-64 y
Age ≥65 y
Extrathyroidal tumor extent
Tumor size ≥4 cm
American Thyroid Association Management Guidelines Recommendation 26 (Surgery) and Recommendation 32 (Radioiodine)
Adherence to ATA management guidelines was 74.9% for surgery and 50.9% for RAI treatment. There was complete information about surgery for 3040 patients with HCC in the SEER database. In univariate analysis, older patient age was associated with treatment that was not in accordance with the guidelines (77.9% of patients ages 18-44 years and 77.7% of patients ages 45-64 years underwent total thyroidectomy compared with 70.1% of patients aged ≥65 years; P = .005). Marital status also was associated with care that was in accordance with surgery management guidelines: 77.9% of married patients with HCC were treated in accordance with guidelines versus 71.7% of single patients (P < .001). More patients with localized and regional HCC were treated in accordance with ATA guidelines than patients with distant disease (75.6% and 77.9% vs 68.2%, respectively; P = .020). Patients who had no lymph nodes examined had treatment that was less often in accordance with the ATA guidelines (72.3% vs 84.9% with negative lymph node status and 84.8% with positive lymph node status; P < .001).
In multivariate analysis, factors that were associated with practice that was discordant with ATA management guidelines included patient age ≥65 years and single marital status. Patients with HCC who had a lymph node examination more often underwent total thyroidectomy (P < .001); this was most pronounced for patients who had positive lymph nodes (odds ratio, 0.44 vs 0.51). Time-trend analysis revealed no significant changes in the surgical management of patients with HCC over the last 2 decades (percentage in accordance with guidelines: 75.7% during 1988-1997, 73.9% during 1998-2006, and 76.9% during 2007-2009 [the period after publication of the first set of ATA guidelines in 2006]; P = .224).
To evaluate adherence to ATA management guidelines recommendation 32 (RAI), 2558 patients who had tumors that measured ≥1 cm were included in the analysis. In univariate analysis, treatment in accordance with the guidelines was more common among the group ages 45 to 64 years (54.8%), followed by the group ages 18 to 44 years (53.8%), and the group aged ≥65 years (44.5%; P < .001). RAI was received more often by married patients than by single patients (52.5% vs 48.3%; P = .024). Better adherence to guidelines was observed over time; 39.3% of patients who were diagnosed between 1988 and 1997 received RAI compared with 52.3% during 1998 to 2006 and 54.1% during 2007 to 2009 (the period after publication of the first ATA guidelines; P < .001). RAI was received by 5.8% of patients who did not undergo surgery, by 25.6% of patients who underwent partial thyroidectomy, and by 59.1% of patients who underwent total thyroidectomy. Patients with higher SEER stages received care that was less adherent to guidelines (51.4% of patients with localized disease vs 52.3% of patients with regional disease and 38.9% of patients with distant disease; P = .028). Patients with negative lymph nodes were more likely to undergo the recommended treatment compared with those who had positive lymph nodes or no lymph nodes examined at all (60.9% vs 48.2% and 48.4%, respectively; P < .001).
In multivariate analysis, the factors that had an independent association with receiving treatment not in accordance with recommendation 32 of the ATA guidelines included patient age ≥65 years, year of diagnosis from 1988 to 1997, and no surgery or partial thyroidectomy (Table 4). Having lymph nodes examined that were identified as pathologically negative was associated with better guidelines adherence (odds ratio, 0.73; P = .002).
Table 4. Multivariate Analysis of Predictors of Noncompliance With American Thyroid Association Management Guidelines Recommendations 26 (Surgical Treatment, n = 3040) and 32 (Radioactive Implants or Radioisotopes, n = 2558)
Reference groups: ages 18-44 years; marital status, married; year of diagnosis, 1998-2006; surgery, total thyroidectomy; and no LNs examined. A blank cell indicates that the characteristic was not associated significantly with the respective recommendation.
An OR <1 indicates less divergence.
Age ≥65 y
Year of diagnosis 1988-1997
To our knowledge, this study represents the largest and most updated cohort of patients with histologically confirmed HCC. Two prior studies included the SEER database but were performed more than a decade ago.8 In the current study, we observed that HCC was more common in older men than ODTC, and it had a more aggressive course. Haigh and Urbach observed similar differences in demographic characteristics between HCC and non-Hurthle cell follicular carcinoma in cohorts of 172 patients 673 patients, respectively, between 1988 and 1993.9 However, their study did not identify any clinical or pathologic differences between patients who were treated for HCC and ODTC.
Whether HCC has a more aggressive clinical course than ODTC has been the subject of debate in the literature.1, 4, 7, 9, 14, 15 Although HCC is more frequent in a subset of patients with shorter expected overall survival (ie, older patients and men) than ODTC, our data support the finding that HCC appears to have more aggressive behavior, presenting with more advanced disease and reduced disease-specific survival compared with ODTC. Bhattacharyya identified 555 patients with HCC from the SEER database between 1988 and 1998 and determined that older patients, men, and larger tumors were associated with reduced overall survival.8 Our analysis of disease-specific survival confirmed these findings, and indicating that the factors associated independently with worse survival included older patient age, larger tumor size, extrathyroidal tumor extension, higher SEER stage, and not undergoing surgery. Being a man was not associated independently with compromised survival in our data set.
It has been demonstrated that survival from thyroid cancer has improved over time. Trimboli et al retrospectively analyzed a cohort of 500 patients who were diagnosed at age ≥20 years and reported a reduction in mortality of 4% per year.16 Chow et al observed an improvement in survival between 1960 and 2000 among 1348 patients who were diagnosed with differentiated thyroid carcinoma, they attributed this improvement to the detection of cancer at an earlier stage and more aggressive surgical and RAI treatments in more recent years.17 Although our study confirmed a better survival rate for patients with ODTC over the last 2 decades, we did not observe any change in disease-specific survival for patients with HCC. Roman et al documented similar findings in 1252 patients who had medullary thyroid cancer between 1973 and 2002 from the SEER database, indicating no improvement in the survival of patients over the last 30 years.18
RAI therapy for HCC is considered unlikely to have an impact on patient survival; the ATA guidelines suggest that the main purpose of postoperative RAI is ablation of any residual thyroid tissue after total thyroidectomy to improve early detection of cancer recurrence based on thyroglobulin measurement and/or RAI whole-body scan.4, 19, 20, 21 In the current study, we observed that patients with HCC who received RAI treatment had improved disease-specific survival in univariate analysis; however, after adjusting for various factors, RAI was not associated independently with long-term survival.
Famakinwa et al assessed adherence to ATA guidelines in the United States regarding the treatment of 31,486 patients with differentiated thyroid cancer and reported that surgery in accordance with the guidelines decreased with increasing patient age. Moreover, those authors identified 10,998 patients who were eligible for RAI remnant ablation, and the overall adherence rate was 62%.22 In our, we observed similar rates of discordant treatment for patients with HCC. The first ATA practice guidelines were published in 2006; therefore, we performed a time-trend analysis for HCC treatment. We observed that, since the publication of ATA guidelines, there has been an increasing trend toward more treatment that is in line with the recommendations; still, only 76.9% and 54.1% of patients with HCC undergo treatment in accordance with recommendations 26 and 32, respectively.
Survival also was associated with marital status. This has been reported in the literature for other diseases. Wang et al analyzed a cohort of 127,753 patients with colorectal cancer from the SEER database between 1992 and 2006 and identified marital status as a positive prognostic factor; married patients were more likely to be diagnosed at an early stage and to undergo proper surgical treatment (P < .001).23 Being single was associated with compromised survival. It has been postulated that marriage or partnerships are associated with overall better care and support and, thus, with improved survival.
The current study has several limitations, including those inherent to using a database. Nonetheless, SEER has been well validated and has low rate of coding errors overall. SEER does not record comorbidities or intent of surgery; therefore, these variables could not be included in our analysis. Another limitation of using this database is the lack of centralized review of pathologic data by an experienced thyroid pathologist. Recurrence, persistent disease, and reoperations are not recorded in SEER; therefore, this information could not be analyzed. Finally, because of the impossibility of distinguishing between N1a from N1b levels of lymph node metastasis, the SEER stage was not translated into the American Joint Committee on Cancer (TNM) staging system.13 However, SEER has extensive data-quality profiles, and studies have ensured its overall accuracy.24, 25
In conclusion, to our knowledge, the current study is the largest analysis to date of patients with HCC at a population level. There remain clear differences in behavior between HCC and ODTC; the absence of any improvement in disease-specific survival among patients with HCC over the last 2 decades should alert care providers to the reality that efforts to achieve earlier diagnosis and better treatments for advanced disease need to be explored. Given its relative rarity, multi-institutional and international trials should be pursued. HCC appears to be a unique disease entity that behaves in a more aggressive fashion than ODTC. A different staging system for this variant of differentiated thyroid cancer could be proposed to allow better provision of prognosis, and separate management guidelines could be considered in the future.
We acknowledge Risa Wong; Michael Sullivan, MD; and Hadiza Kazaure, MD, for statistical help.
This research was part of the EXTRA PLUS Project, which was developed by the University of Milano-Bicocca, Italy, and was supported by the Cariplo Foundation with the active participation of Assolombarda (Lombardy Entrepreneurial Association).