Insular thyroid cancer

A population-level analysis of patient characteristics and predictors of survival

Authors

  • Hadiza S. Kazaure BSc,

    1. Department of Surgery, Yale University School of Medicine, New Haven, Connecticut
    Search for more papers by this author
  • Sanziana A. Roman MD,

    1. Department of Surgery, Yale University School of Medicine, New Haven, Connecticut
    Search for more papers by this author
  • Julie A. Sosa MD, MA

    Corresponding author
    1. Department of Surgery, Yale University School of Medicine, New Haven, Connecticut
    • Divisions of Endocrine Surgery and Surgical Oncology, Department of Surgery, Yale University School of Medicine, 330 Cedar Street, Tompkins 208, PO Box 208062, New Haven, CT 06520

    Search for more papers by this author
    • Fax: (203) 785-2499


  • A portion of these data were presented as a poster at the American Thyroid Association meeting, October 26-30, 2011, Indian Wells, CA

Abstract

BACKGROUND:

Insular thyroid cancer (ITC) is an uncommon, poorly differentiated thyroid malignancy. To date, there have been no population-level studies of the characteristics and outcomes of patients with ITC.

METHODS:

The authors used the Surveillance, Epidemiology, and End Results (SEER) database from 1999 to 2007 to compare the characteristics and prognosis of patients who had ITC with those of patients who had well differentiated thyroid cancer (WDTC) and anaplastic thyroid cancer (ATC). Data analyses were performed using chi-square tests, analyses of variance, log-rank tests, and multivariate regression.

RESULTS:

There were 114 patients with ITC, 497 patients with ATC, and 34,021 patients with WDTC. The mean age of patients with ITC was 62.1 years versus 48.1 years for patients with WDTC and 69.5 years for patients with ATC (P < .001). The mean ITC tumor size was 5.9 cm versus 2.0 cm for WDTC and 6.4 cm for ATC (P < .001). Distant metastasis occurred in 31% of patients with ITC versus 4.5% of patients with WDTC and 59.1% of patients with ATC (P < .001). Insular histology was associated independently with compromised survival in the overall study sample (hazard ratio [HR], 2.1; P = .001). The 5-year disease-specific survival rate was 72.6%, 97.2%, and 9.1% for patients with ITC, WDTC, and ATC, respectively (P < .001). After adjustment, radioiodine therapy (HR, 0.15; 95% confidence interval, 0.04-0.5) and distant metastasis (HR, 15.3; 95% confidence interval, 3.7-62.2) were associated independently with ITC survival. The mortality rate was 7.1%, 12%, and 54.3% for patients with localized, regional, and distant stage ITC, respectively (P < .001). For patients who had ITC with distant metastasis, thyroidectomy and radioiodine therapy independently improved survival.

CONCLUSIONS:

ITC is rare and aggressive. The current results indicated that its treatment should include total thyroidectomy and high-dose radioiodine for all patients and neck dissections for patients with lymph node disease. Early diagnosis and close surveillance are essential in the management of patients with ITC. Cancer 2011. © 2011 American Cancer Society.

INTRODUCTION

Insular thyroid cancer (ITC) is a rare thyroid malignancy. Clinical and morphologic characteristics that are intermediate to those of well differentiated thyroid cancer (WDTC) and undifferentiated (anaplastic) thyroid cancers (ATC) have been observed in ITC1-4; as such, ITC is classified among poorly differentiated thyroid cancers.5-7 The estimated incidence of ITC ranges from <1% to 10% of all thyroid cancers.8-11 ITC histology is characterized by solid nests (insulae) of small, uniform carcinoma cells; small follicles containing thyroglobulin; frequent necrotic foci; and variable but consistently present mitotic activity.1, 10, 12 The histogenesis of ITC is unclear; there appears to be an association with long-standing goiter.9, 13, 14

On the basis of results from small series, it has been observed that patients with ITC have a higher frequency of metastasis, recurrence, and mortality than patients with WDTC.1, 3, 4, 12, 15, 16 However, some studies did not report a significant difference in the prognosis for patients with ITC versus the prognosis for patients with papillary and/or follicular thyroid carcinomas.17, 18 The prognosis for patients with ITC appears to be better than that for patients with ATC.2, 9, 19 Given the relative paucity of data, there is no consensus about the prognostic relevance of ITC11, 20; indeed, there is debate about whether an insular histology is associated independently with prognosis.11, 12, 18, 20-22

ITC data are culled largely from small single institution-based case series; to our knowledge, no population-level studies have been published. In addition, understanding of ITC is limited by the finding that most studies did not analyze ITC as a distinct entity among the heterogeneous group of poorly differentiated thyroid cancers,5, 23-29 which also include trabecular and solid variants.28, 30, 31 In the current study, we compare the demographic, pathologic, and clinical characteristics of patients who had ITC with those of patients who had WDTC and ATC and identify the prognostic factors, including treatment interventions, associated with survival among patients with ITC.

MATERIALS AND METHODS

Data Source and Study Participants

We used the Surveillance, Epidemiology, and End Results (SEER) database (1999-2007), which captures data representing 26% of the US population. SEER is the only comprehensive source of population-based cancer information in the United States that includes stage of cancer at the time of diagnosis and patient survival data.32

Patients were identified using the following International Classification of Diseases for Oncology, third Edition (ICD-O-3) histology codes: 8337 (ITC); 8050, 8260, 8330, 8331, and 8332 (WDTC); and 8021 (ATC). WDTC comprises papillary and follicular cell thyroid cancers; tall cell, columnar cell, diffuse sclerosing, and Hurthle cell variants were excluded.

Demographic variables of interest were patient age at diagnosis, sex, and race. Clinical variables included thyroid surgery, radiation therapy, lymph node examination, and survival status as of December 31, 2007. Thyroid surgery was classified into 3 treatment groups: 1) no surgery; 2) lobectomy with or without isthmusectomy; and 3) subtotal, near-total, or total thyroidectomy. In SEER, information on lymph node examination (no vs yes) is derived based on clinical, operative, or pathology assessment. Radiation therapy was categorized into 5 groups: none, radioiodine (RAI), external-beam radiation (EBRT), radioactive implants, and other. Survival was calculated in years from the time of diagnosis to death, the last date the patient was known to be alive, or December 31, 2007, whichever came first. Overall survival and disease-specific survival rates were calculated.

Pathologic information included tumor size, focality (solitary or multifocal), extension (intrathyroid vs extrathyroid), lymph node metastasis, and disease stage at diagnosis. Tumor size was subdivided into 3 groups (≤2.0 cm, 2.1-4.0 cm, and >4.0 cm), which correlate with the tumor classifications T1, T2, and T3, respectively, in the seventh edition of the American Joint Commission on Cancer (AJCC) Cancer Staging Manual for the TNM classification of WDTCs.33 For tumor extension, extrathyroid extension was defined as tumor invasion beyond the thyroid capsule. Lymph node metastasis was treated as a binary variable (none vs ≥1 positive lymph node). Disease stage at diagnosis was based on the staging system used by SEER. According to this system, disease is localized if the tumor is confined to the thyroid and without metastasis; regional if the tumor extends beyond the thyroid into surrounding tissues or metastasizes to regional lymph nodes; and distant if metastasis to extracervical lymph nodes or organs is present. The SEER stages were not recoded into the AJCC-TNM staging system, because it was impossible to distinguish between N1a and N1b levels of lymph node metastasis.

Statistical Analysis

Summary statistics were used to describe baseline characteristics. Chi-square and Student t tests were used to analyze categorical and continuous variables, respectively. The Kaplan-Meier method was used for univariate analyses of survival, and the log-rank test was used to determine the statistical significance of differences in survival. Cox multivariate logistic regression was used to identify factors independently associated with survival. Hazard ratios (HR) and 95% confidence intervals (CIs) were calculated for factors that were associated with survival. Multivariate logistic regression was used to identify factors independently associated with occurrence of distant metastasis. To avoid over-adjustment, 2 stepwise regression models were run: 1 model included a lymph node examination variable treated as a binary variable (yes vs no), and the other included the “positive lymph node variable” categorized as no lymph node examined versus all lymph nodes examined negative versus at least 1 lymph node examined positive. All tests were 2-sided, and a P value < .05 was considered statistically significant.

Only adult patients (age at diagnosis ≥18 years) with active follow-up in the SEER database were included in the analyses; those diagnosed at autopsy were excluded. Data analyses were performed with the SPSS (version 18; SPSS Inc., Chicago, Ill) and SAS (version 9.2; SAS Institute, Inc., Cary, NC) statistical software packages. SEER is a public database with deidentified information; as such, this study was granted an exemption from institutional review board approval at our institution.

RESULTS

In total, 114 patients with ITC, 34,021 patients with WDTC, and 497 patients with ATC were abstracted for analyses; the patients with ITC represented 0.3% of the study sample. The mean follow-up for patients with ITC, WDTC, and ATC were 2.2 years, 3.2 years, and 4.0 months, respectively.

Characteristics

The mean age of patients with ITC was approximately 62 years (Table 1). There was a preponderance of women for all 3 subtypes of thyroid cancers; however, ITC was more frequent among men (47.4% vs 24% WDTC and 37.4% ATC; P < .001). One patient with ITC had a history of follicular thyroid cancer. Approximately 94.8% of patients with ITC underwent thyroid surgery; the rate of this intervention was lower in patients with distant metastasis (88.6%). Greater than 70% of patients with ITC received radiation therapy; 15.8% received EBRT, and 2 patients received both EBRT and RAI. Approximately 7.1%, 10%, and 31.4% of ITC patients with localized, regional, and distant disease stages, respectively, received EBRT. There was no association between receipt of RAI and disease stage (P = .106). Overall, most patients with ITC underwent thyroid surgery and received radiation-based therapy (72.8% vs 48.7% WDTC and 25.6% ATC; P < .001). Only 36.8% of patients with ITC had lymph nodes examined.

Table 1. Baseline Characteristics of Patients With Well Differentiated, Insular, and Anaplastic Thyroid Cancer: Surveillance, Epidemiology, and End Results Database From 1999 to 2007, n=34,632
 Percentage of Patientsa 
CharacteristicWDTC, n = 34,021Insular, n = 114Anaplastic, n = 497P
  • Abbreviations: EBRT, external-beam radiation therapy; RAI, radioiodine; SEM, standard error of the mean; WDTC, well differentiated thyroid cancer.

  • a

    Unless otherwise noted, values for characteristics are shown in percentages; percentages have been rounded and may not add up to 100%.

Age, y   <.001
 Mean±SEM48.1±0.162.1±1.569.5±0.6 
 <4544.419.33 
 45-6439.431.631.8 
 ≥6516.249.165.2 
Gender
 Women7652.662.6<.001
Race   .01
 White82.882.581.3 
 Black4.87.97.8 
 Other12.49.610.9 
Thyroid surgery   <.001
 No surgery5.85.256.3 
 Lobectomy13.613.217.5 
 Thyroidectomy80.681.626.2 
Lymph nodes examined    
 Yes4636.825.9<.001
Radiation therapy   <.001
 None48.123.740.6 
 RAI46.1571 
 EBRT1.915.854.2 
 Radioactive implant1.81.81 
 Other/unknown2.21.83.2 

The mean ITC tumor size was 5.9 cm (Table 2), and only 1 patient had a tumor ≤1 cm. Compared with WDTCs and ATCs, ITCs were more likely to be solitary tumors; the rates of extrathyroid extension were intermediate to those of WDTCs and ATCs. Regional disease without distant metastasis was more common in ITCs compared with WDTCs or ATCs; 31% of patients with ITC had distant metastasis.

Table 2. Pathologic Characteristics of Patients With Well Differentiated, Insular, and Anaplastic Thyroid Cancer: Surveillance, Epidemiology, and End Results Database From 1999 to 2007, n = 34,632
CharacteristicaWDTC, n = 34,021Insular, n = 114Anaplastic, n = 497P
  • Abbreviations: DFS, disease-free survival; IQR, interquartile range; OS, overall survival; SEER, Surveillance, Epidemiology, and End Results Program of the National Cancer Institute; SEM, standard error of the mean; WDTC, well differentiated thyroid cancer.

  • a

    Unless noted otherwise, characteristics are in percentages; percentages have been rounded and may not add up to 100%.

  • b

    Data on tumor size were available for 90.3% of patients with WDTCs, 86% of patients with insular thyroid cancers, and 60.8% of patients with anaplastic thyroid cancers.

  • c

    Only patients who had ≥1 lymph node(s) examined were included in the analysis.

  • d

    Staging was according to SEER: Data were missing for approximately 2.2% of patients with WDTCs, 0.9% of patients with insular thyroid carcinomas, and 6.4% patients with anaplastic thyroid cancers.

Tumor size, cmb    
 Mean±SEM2.0±0.15.9±0.36.4±0.3<.001
 ≤2.0659.17 
 2.1-4.025.928.618.2 
 >4.09.162.374.8 
Focality, multifocal2919.524.7<.001
Extrathyroid extension19.347.382.6<.001
≥1 Positive lymph nodec52.361.980.6<.001
 No. of positive lymph nodes: Median [IQR]3 [1-7]3 [1-7]2 [1-6] 
SEER staged   <.001
 Localized58.424.86.9 
 Regional37.144.234 
 Distant4.53159.1 
Rate
 5-y DFS97.272.69.1<.001
 5-y OS9361.96.3<.001

Survival Analyses

The overall survival rate for patients with ITC, WDTC, and ATC were 76.3%, 94.7%, and 12.1%, respectively (P < .001); and the disease-specific survival rates were 84.2%, 97.8%, and 20.7%, respectively. The 5-year survival rates are listed in Table 2 and illustrated in Figure 1. After adjusting for multiple risk factors, including age, sex, tumor size, focality and histology, extent of tumor and disease, and treatment interventions in the overall study sample, ITC histology was among the factors independently associated with compromised survival (HR, 2.1; 95% CI, 1.3-3.3; P = .001).

Figure 1.

Disease-specific survival curves are shown for well differentiated thyroid cancer (WDTC), insular thyroid cancer (ITC), and anaplastic (ATC) thyroid cancers (based on data from the Surveillance, Epidemiology, and End Results database, 1999-2007).

Tumor size, extrathyroid extension, distant metastasis, and total thyroidectomy were associated with the survival of patients with ITC in univariate analyses (Table 3). After multivariate adjustment, RAI therapy (HR, 0.15; P = .005) and distant metastasis (HR, 15.3; P < .001) were the 2 factors independently associated with survival of patients who had ITC (Fig. 2). The adjusted HR for radiation therapy increased (HR, 0.31; P = .023) when EBRT was grouped with RAI therapy in multivariate analysis of ITC survival.

Table 3. Univariate Analyses of Factors Associated With Disease-Specific Survival in Patients With Insular Thyroid Cancer: Surveillance, Epidemiology, and End Results Database From 1999 to 2007
Risk FactoraUnadjusted HR95% CIP
  • Abbreviations: CI, confidence interval; EBRT, external-beam radiation therapy; HR, hazard ratio; RAI, radioiodine.

  • a

    The referent categories were as follows: Sex, women; age, <45 y; tumor size was analyzed was analyzed a continuous variable; tumor extension, intrathyroid; focality, solitary; extent of disease, localized; thyroid surgery, none; radiation, none; lymph node examination, none.

Men1.40.5-3.5.5
Age ≥45 y2.60.6-11.3.2
Tumor size1.21.1-1.4.004
Extrathyroid extension3.71.3-10.5.01
Multifocality2.60.9-7.2.06
Extent of disease   
 Regional0.80.7-4.8.8
 Distant8.71.9-39.5.005
Thyroid surgery   
 Lobectomy0.50.1-2.7.4
 Thyroidectomy0.20.04-0.7.04
Radiation   
 RAI0.40.1-1.3.1
 EBRT2.10.7-6.5.2
Lymph nodes examined, yes1.10.4-2.8.9
Figure 2.

Cox multivariate regression analysis was used to identify the factors associated independently with survival for insular thyroid cancer (based on data from the Surveillance, Epidemiology, and End Results database, 1999-2007). Stepwise multivariate regression models were adjusted for age (with age <45 years as the referent category), sex (with women as the referent category), thyroid surgery (no surgery [referent] vs lobectomy vs thyroidectomy), radiation therapy (no radiation [referent] vs radioiodine vs external-beam radiation therapy [EBRT] vs other), tumor size (treated as a continuous variable as well as a categorical variable based on American Joint Committee on Cancer tumor size cutoffs for thyroid cancer), focality (with not multifocal as the referent category), tumor extension (with not extrathyroid as the referent category), lymph node assessment (as described in the text; see Materials and Methods), and disease stage (localized [referent] vs regional vs distant disease).

The mortality rate was 7.1% (n = 2), 12% (n = 6), and 54.3% (n = 19) for localized, regional, and distant ITC disease stages (P < .001). The 5-year survival rates for patients with ITC by disease stage differed significantly (P < .001) (Fig. 3). Analyses of the prognostic factors associated with survival for patients who had ITC with localized and regional disease were not performed because of a paucity of fatal events. After adjustment, extrathyroid extension (adjusted odds ratio, 12.8; P < .001) was the only factor that predicted distant metastasis in patients with ITC; multivariate analyses of survival for patients who had ITC with distant metastasis revealed that total thyroidectomy and RAI therapy were the 2 factors independently associated with improved survival (Fig. 2). Age was not a predictor of occurrence of ITC distant metastasis or survival of patients who had ITC with distant metastasis.

Figure 3.

Disease-specific survival curves are shown for insular thyroid cancer according to disease stage(based on data from the Surveillance, Epidemiology, and End Results database, 1999-2007).

DISCUSSION

To our knowledge, this study comprises the largest cohort of ITC patients in the literature. Consistent with results from smaller studies,1, 4, 9, 17, 19, 34, 35 we observed that aggressive clinical and pathologic behavior were common in ITC. The prognosis for patients with ITC was worse than that for those with WDTC despite high rates of thyroid surgery (>90%) and radiation therapy (>70%) for patients with ITC. The survival rate for patients with ITC was significantly higher than that for patients with ATC.

In agreement with results from 1 study21 of 57 patients in which the insular histotype remained the only variable independently predictive of mortality, we demonstrated that ITC histology was independently associated with survival after controlling for several risk factors, including age and tumor size. This result has potential prognostic, therapeutic, and staging implications for patients with WDTC in which there is an insular component. In a review of 2457 papillary and follicular carcinomas, Sasaki et al3 observed that the presence of an insular component was independently associated with poor survival. Likewise, Van den Brekel and colleagues19 identified insular differentiation as a negative prognostic marker for disease-free interval; there was no difference in the outcomes of patients with WDTCs who had a focal insular component compared with those who had a predominantly insular component. In a review of 1230 WDTCs, Decaussin and colleagues15 demonstrated that patients with an insular component were 17 times more likely to have distant metastasis, which is a key prognostic marker of poor survival known to occur in 4% to 23% of patients with WDTC.36, 37

Greater than 75% of patients with ITC in the current study had nonlocalized disease. The steep increase in mortality with the occurrence of distant metastasis highlights the importance of early diagnosis. After adjusting for several risk factors, distant metastasis was the most consistent factor predictive of reduced survival of patients with ITC. Likely because of sample size limitations that result from the overall rarity of ITC, there is a paucity of studies in which multivariate analyses have been performed to identify prognostic factors associated with survival of ITC patients.20, 24 Our results parallel those of Jung and colleagues,24 who associated distant metastasis with survival in a multivariate analysis of 49 patients with poorly differentiated thyroid cancer, including 17 patients with ITC.

RAI, not EBRT, was associated with improved survival for ITC; the benefit of RAI was observed even in patients with distant metastasis. This result is noteworthy considering the higher rate of EBRT use in patients with distant metastasis compared with those assigned a lower disease stage. Although RAI uptake in ITC has been reported, there are conflicting data regarding the impact of radiation-based therapy on the outcomes of patients with ITC.14, 19, 21, 22, 24, 38 Similar to our results, Rufini et al22 demonstrated improved outcomes for patients with ITC who received RAI treatment. In contrast, Pellegriti et al21 observed a clinical benefit from RAI in only 1 patient despite good RAI uptake in all 11 patients who had ITC with metastatic disease in their study. Lai and colleagues38 did not observe a survival benefit from RAI or radiation therapy in their meta-analysis of 82 patients with ITC. On the basis of expert opinion (Category C recommendation), the American Thyroid Association encourages the use of high-dose RAI (100-200 mCi) for patients with aggressive tumor histology, including the insular subtype.39

Appropriate staging of cancer patients is essential for optimal management. The current AJCC-TNM33 system does not provide staging specifically for ITC. In the current study, >60% of patients with ITC had at least pathologic T3 tumors according to AJCC staging. Survival was associated with disease stage, which takes into consideration tumor extent as well as lymph node status and distant metastasis. In addition, we observed that age was not associated with ITC survival, or the risk of distant metastasis, or prognosis. It is unclear whether the age stratification used for staging WDTC is pertinent for staging ITC. In 1 study,21 compromised survival was observed among patients who had ITC compared with patients of similar age and tumor size who had papillary and follicular thyroid carcinoma; a similar finding was observed in another study22 that compared the survival of patients with ITC versus patients with WDTC who were matched by age, tumor size, and extent of differentiation. In contrast, age >45 years was associated with worse survival in a meta-analysis of 82 patients with ITC.38

Our current study has limitations inherent to the use of the SEER database, such as coding errors and misclassification. SEER performs annual audits, and quality-control studies have indicated the accuracy of the database40, 41; however, a centralized review of pathologic data by an experienced thyroid pathologist is not performed. Data related to patient comorbidities, iodine deficiency, and history of radiation exposure; relevant biochemical information, such as thyroglobulin, thyroid-stimulating hormone, and thyroxine levels; adjuvant RAI dose, reoperations, completion thyroidectomies, or recurrence are not captured in the SEER database; and information about the presence of concurrent thyroid cancers, such as ITC and WDTC in the same patient, is not provided. The prognostic role of lymphadenectomy is unclear given the low rates of lymph node examination in this study. Moreover, the SEER database does not provide details about surgical margins or the extent of lymph node dissection; and information regarding ultrasound evaluation of lymph nodes is not collected. Sample size limitations precluded extensive analyses of disease-specific survival as well as more disease stage-specific analyses. Considering that almost 95% of patients with ITC underwent thyroid surgery, only a small (control) group of patients managed nonsurgically were available for comparison after adjusting for other risk factors; a type II error may explain the nonsignificant role of thyroid surgery on ITC survival in multivariate analysis. In contrast, the impact of thyroid surgery on the survival of patients with distant metastasis was demonstrated, because the lower thyroid surgery rates for this subgroup allowed valid comparison. In this study, only 2 patients received EBRT in addition to RAI; thus, the role of RAI together with EBRT therapy in the outcome of patients with ITC is unclear.

ITC is a distinct clinical entity for which data are scarce. The propensity of ITC to metastasize and its relatively poor prognosis highlight the need for early detection, aggressive intervention, and close surveillance of affected patients. Given the demonstrated survival benefit of tumor debulking, patients with ITC (including those with distant metastasis) are candidates for aggressive treatment involving total thyroidectomy, high-dose RAI, and appropriate lymph node dissections for those with lymph node metastasis. Although the collective analysis of poorly differentiated thyroid cancers has improved understanding of this group of thyroid cancers, analyses of subtypes should be performed whenever possible. Considering the rarity of ITC and other poorly differentiated thyroid cancers, a multi-institutional, international registry to achieve adequate statistical power for answering important clinical questions could be considered. More research is needed to investigate the clinical impact of concurrent ITC and WDTC. Exploration of the molecular changes in ITC may be helpful to tailor therapeutic interventions, and patient enrollment in clinical trials should be encouraged when systemic therapy for disseminated disease is required. Thus, as results from future studies become available, a staging system specific for ITC could merit further exploration.

FUNDING SOURCES

No specific funding was disclosed.

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary