Should All Papillary Thyroid Microcarcinomas Be Aggressively Treated? An Analysis of 18,445 Cases

PAPERS OF THE 131ST ASA ANNUAL MEETING Should All Papillary Thyroid Microcarcinomas Be Aggressively Treated? An Analysis of 18,445 Cases Xiao-Min Yu, MD, PhD, Yin Wan, MS, Rebecca S. Sippel, MD, FACS, and Herbert Chen, MD, FACS Objective: The purpose of this study was to identify the risk factors that predict papillary thyroid microcarcinoma (PTMC)-related death in a large patient population to determine which patients need aggressive treatment. Background: The management of PTMC is controversial and ranges from observation to total thyroidectomy. The lack of consensus is predominantly due to the general excellent overall prognosis, thereby requiring a large cohort to delineate differences in outcome. Methods: All papillary thyroid cancer patients with tumor size of 1 cm or less in the Surveillance, Epidemiology and End Results Cancer Database from 1988 to 2007 were identified. Outcomes, including overall and diseasespecific survival, were compared, and different risk groups were evaluated by multivariate analysis. Results: A total of 18,445 cases of PTMC with surgery were identified. The 10-year and 15-year overall survivals were 94.6% and 90.7%, respectively, while disease-specific survivals were 99.5% and 99.3%. Age greater than 45 years, male sex, African American or minority race, node metastases, extrathyroidal invasion, and distant metastases were stratified to be significant risk factors for overall survival. There were 49 thyroid cancer-related deaths. Forty-five (92%) of the 49 patients had at least 2 risk factors, and 51% of these 49 patients had 3 or more risk factors (vs 5.7% in the rest of the cohort, P < 0.001). Conclusions: Although PTMC is generally associated with an excellent prognosis, 0.5% patients may die of PTMC. The presence of 2 or more risk factors is strongly associated with cancer-related mortality and can help to identify patients who should be considered for more aggressive management. (Ann Surg 2011;254:1 8) Thyroid cancer is one of the most rapidly increasing malignancies with regard to incidence in the United States. The incidence rate of thyroid cancer in 2007 was 11.99 per 100,000, which almost doubled the rate in 1997. 1 Although the reasons for the rapid increase vary extensively, more and more studies have shown that papillary thyroid microcarcinomas (PTMC), defined as papillary thyroid cancers measuring 1 cm or smaller, account for a significant proportion of the increase. It has been reported that between 1988 and 2002, 49% of the increase consisted of small papillary thyroid cancers measuring less than 1 cm. 2 Furthermore, PTMC increased 441% between 1983 and 2006, while the incidence rate of larger papillary carcinomas measuring more than 5 cm remained almost unchanged. 3 Despite the notable increasing incidence of PTMC, the optimal treatment for these patients remains controversial. 4 Although the guidelines on the treatment of thyroid cancer from both the American From the Department of Surgery, University of Wisconsin Madison, WI. Disclosure: The authors declare that they have nothing to disclose. Reprints: Herbert Chen, MD, FACS, Department of Surgery, University of Wisconsin, H4/722 Clinical Science Center, 600 Highland Ave, Madison, WI 53792. E-mail: chen@surgery.wisc.edu. Copyright C 2011 by Lippincott Williams & Wilkins ISSN: 0003-4932/11/25404-0001 DOI: 10.1097/SLA.0b013e318230036d Thyroid Association and the European Thyroid Association suggest total or near-total thyroidectomy as the standard care for papillary thyroid cancer, and thyroid lobectomy only for unifocal PTMC found incidentally, 5,6 some studies have found that close observation alone can result in satisfactory patient outcomes. 7 Even when thyroidectomy is considered, the extent of surgery is another debated issue due to the high frequency of lymph node metastases of PTMC. Wada et al 8 reported that as high as 64% of all PTMCs had central node involvement after they followed a group of 259 patients, all of whom underwent lymph node dissection at the time of thyroidectomy. Furthermore, in another study, lymph node metastases at presentation was found as one of the factors associated with persistent or recurrent disease. 9 Both of these studies advocated routine therapeutic lymph node dissection in PTMC patients with palpable lymph nodes but showed great caution for prophylactic lymph node dissection. On the contrary, in more recent study, prophylactic central lymph node dissection was shown to successfully manage the subclinical central lymph node metastases in PTMC with 0% recurrent rate during 3-year follow-up. 10 The lack of consensus is predominantly due to the general excellent overall prognosis. The overall survival for well-differentiated thyroid carcinoma, including papillary carcinoma, has exceeded 90% 11 and is even better for PTMC. However, there is no doubt that PTMC is not homogeneous, and better prognosis stratification, especially for high-risk patients, will optimize surgical care. The American Joint Committee on Cancer (AJCC)/pathological Tumor Node Metastasis (ptnm) staging algorithm, 12 the most commonly used cancer staging system, is not specific for PTMC and can overestimate the risk for certain cohorts. A recent study divided PTMC patients into 3 different categories based on 2 key factors: presentation with symptoms and the presence of lymph node metastases. 13 Important as these 2 factors may be, it is also possible that other variables may contribute to developing a more accurate prognosis for these patients. Because of the nature of PTMC, the identification of high-risk patients within this group requires larger patient populations with much longer-term follow-up times. Some studies have previously tried to determine the possible risk factors, but none have included more than a few thousand patients. The purpose of this study was to define the long-term prognosis of PTMC and to identify the risk factors that predict PTMC-related death in a large patient population to determine which patients need aggressive treatment. METHODS Database A retrospective cohort study was performed by using data from the Surveillance, Epidemiology and End Results (SEER) Cancer Database (available at: www.seer.cancer.gov) maintained by National Cancer Institute. The characteristics and representativeness of SEER database have been discussed in many previous studies. 14,15 The most updated data set was used in the current study, which was released in April 2010, based on the November 2009 submission. Annals of Surgery Volume 254, Number 4, October 2011 www.annalsofsurgery.com 1

Yu et al Annals of Surgery Volume 254, Number 4, October 2011 In brief, the current database collects cancer incidence and survival data from 17 population-based cancer registries representing 26% of the US population. Current SEER registries consist of the states of Connecticut, Hawaii, Iowa, Kentucky, Louisiana, New Jersey, New Mexico, and Utah; the metropolitan areas of Atlanta, Detroit, San Francisco-Oakland, Seattle-Puget Sound, and San Jose-Monterey; and the Alaska Native Tumor Registry, rural Georgia, Greater California, and Los Angeles County. The SEER registries routinely collect data on patient demographics, primary tumor site, tumor morphology, stage at diagnosis, and first course of treatment. The University of Wisconsin institutional review board deemed this project to be exempt from review. Case Definition All patients with primary papillary thyroid cancer diagnosed between 1988 and 2007 were identified by using primary tumor site code of C739 (thyroid) in combination with the International Classification of Disease for Oncology, 3rd Edition, 16 histology codes 8050 (papillary carcinoma, not other specified), 8260 (papillary adenocarcinoma, not other specified), 8340 (papillary carcinoma, follicular variant), 8341 (papillary microcarcinoma), 8342 (papillary carcinoma, oxyphilic cell), 8343 (papillary carcinoma, encapsulated), and 8344 (papillary carcinoma, columnar cell). Only patients with tumor size equal to or less than 1 cm among the papillary thyroid cancer cases were selected in the current study cohort. Patients with more than 1 primary site of malignancy were excluded. Data Analysis After the identification of the PTMC cases in SEER database, we first estimated the 10-year and 15-year overall survival, as well as disease-specific survival, by using Kaplan Meier method. Most of the continuous variables were presented as mean ± standard deviation. Subgroup overall survival curves were estimated by using categorical covariates as dependent variables, including demographic and tumor features. Tests of survival equality were performed during the survival analysis by the log-rank test. For multivariate analysis, Cox proportional hazard model was applied to assess the association between potential predictors and survival. 17 Hazard ratios (HR) with 95% confidence intervals (CI) were used to quantify the strength of the association between predictors and overall survival. Finally, Fisher exact test was used to determine the correlation between PTMC-related death and accumulation of multiple risk factors. All statistical analyses were performed by using SAS 9.2 (SAS Institute, Cary, NC). All tests were 2 sided and a P < 0.05 was defined to be statistically significant. RESULTS Patient Characteristics In the SEER database, there were more than 95,000 thyroid cancer records, among which more than 80% (76,553 cases) were papillary thyroid carcinomas. Our selection criteria identified 18,445 patients in total with PTMC. The mean age of this PTMC cohort was 42 ± 14 years. The majority of the patients were female patients (n = 15,259; 82.7%), and majority were white (n = 15,589; 84.5%). The mean size of the primary tumor was 5.3 ± 3.1 mm. Only 6.3% of PTMCs had extrathyroidal invasion. The incidence of pathological-confirmed lymph node metastases was 12.4%. Although distant metastases for PTMC were fairly rare, 91 patients (0.5%) developed distant metastases in the current cohort. All of the selected patients had surgery as the initial treatment. Most patients (n = 13,563; 73.5%) had total/subtotal thyroidectomy, while 25.7% had lobectomy. Around 30% of the patients had regional lymph node dissection. Univariate and Multivariate Survival Analyses for Potential Risk Factors in Papillary Thyroid Microcarcinoma For the entire cohort of 18,445 patients, the 10-year and 15- year overall survivals were 94.6% and 90.7%, respectively, while disease-specific survivals were 99.5% and 99.3%. To further investigate the potential risk factors on the overall survival of PTMC, we selected the variables based on several different staging systems, including ptnm 11 and AMES, 18 for well-differentiated thyroid cancer. Descriptive survival statistics and Kaplan Meier curves suggested that all the traditional risk factors for well-differentiated thyroid cancer had an impact on the overall survival of PTMC (Table 1, Fig. 1). As expected, the survival of patients younger than 45 years was much better than the older patients (P < 0.001; Fig. 1A). Male patients had less-favorable prognosis than the female patients (P < 0.001; Fig. 1B). The 15-year survival rate for the male patients was about 86%, while the survival reached 92% for the female patients. Race was another demographic factor, which impacted the overall survival, though it is not commonly included in any of the thyroid cancer staging systems. American African and minorities showed a slightly worse prognosis compared with the white patients (P < 0.001; Fig. 1C). Regarding the tumor features, the presence of extrathyroidal invasion (P < 0.0001; Fig. 1D), lymph node metastases (P = 0.0042; Fig. 1E), and distant metastases (P < 0.0001; Fig. 1F) were all associated with a decrease in overall survival. Cox proportional hazard models were then applied to quantify the prognostic significance of demographic and tumor factors, with adjustment to competing risk factors (Table 2). Patient age significantly impacted the survival, with patients older than 45 years exhibiting a much worse survival (HR = 6.18; P < 0.001). Both sex and race were significant prognostic factors. Male patients (HR = 1.74; P < 0.001) and race of American African (HR = 2.56; P < 0.001) or minority (HR = 1.38; P = 0.037) demonstrated less-favorable prognoses. The presence of distant metastases had a strong predictive TABLE 1. Descriptive Survival Statistics Group 10-year Survival (%) 15-year Survival (%) P* Overall 94.6 90.7 Race White 95.0 91.3 <0.0001 African American 89.7 87.6 Others 93.2 86.8 Unknown NA NA Sex Female 95.5 91.6 <0.0001 Male 90.3 86.2 Age, yr <45 98.5 97.7 <0.0001 45 90.9 83.2 Extrathyroidal invasion No 95.0 91.1 <0.0001 Yes 90.8 86.3 Unknown 75.0 75.0 Lymph node metastases No 94.9 90.9 0.0042 Yes 93.0 89.5 Distant metastases No 94.7 90.8 <0.0001 Yes 68.0 68.0 *P reached statistically significant, which is less than 0.05. NA indicates not applicable, where no events happened to this group of patients. 2 www.annalsofsurgery.com C 2011 Lippincott Williams & Wilkins

Annals of Surgery Volume 254, Number 4, October 2011 Risk Factors in PTMC FIGURE 1. Kaplan Meier overall survival estimates. A, Stratified by age (P < 0.0001). B, Stratified by sex (P < 0.0001). C, Stratified by race (P < 0.0001). D, Stratified by extrathyroidal invasion (P < 0.0001). E, Stratified by lymph node metastases (P = 0.0042). F, Stratified by distant metastases (P < 0.0001). value for unfavorable survival (HR = 3.76; 95% CI, 1.22 11.62; P = 0.021). Extrathyroidal invasion (HR = 1.61; P = 0.002) and lymph node metastases (HR = 1.36; P = 0.011) were also associated with an adverse prognosis in the PTMC patients. Papillary Thyroid Microcarcinoma-Related Death and Accumulation of Potential Risk Factors Although the disease-specific survival was very high in PTMC patients, 49 patients died of thyroid cancer. To evaluate the possible predictive value of the previously described risk factors in PTMCrelated deaths, we examined these patients. Of the patients who died of PTMC (Table 3), 24.6% (12 of 49) were of American African or minority race. Male patients accounted for around one-third (16 of 49) in this specific group, who usually compose less than 20% of whole PTMC population. More than 85% of PTMC-related deaths happened to patients older than 45 years. Twenty patients had extrathyroidal invasion, 29 had pathological-confirmed lymph node metastases, and 10 had distant metastases. On the basis of the previous findings, PTMC-related deaths did not appear to be associated solely with any of the prognostic factors. Therefore, further analyses were performed to evaluate whether including more risk factors could be predictive of disease-specific survival in PTMC. As shown in Table 4, 45 of the 49 patients had at least 2 risk factors. Except for 20 patients who had 2 risk factors, there were 15 and 9 patients, respectively, who presented with 3 or even 4 risk factors. The percentage of patients with 2 or more risk factors was significantly higher in PTMC-related death group as compared with the rest of the cohort (91.8% vs 27.3%; P < 0.001). Furthermore, more than half of the patients in PTMC-related death group presented with 3 or even more risk factors, while only 5.7% of the rest of cohort did so (P < 0.001). C 2011 Lippincott Williams & Wilkins www.annalsofsurgery.com 3

Yu et al Annals of Surgery Volume 254, Number 4, October 2011 TABLE 2. Multivariate Analysis of Potential Risk Factors on PTMC Overall Survival Variables No. Patients HR 95% CI P Race Caucasian 15,589 1.00 African-American 899 2.56 1.88 3.47 <0.001* Others 1823 1.38 1.02 1.87 0.037* Unknown 134 0.00 0.955 Sex Female 15,259 1.00 Male 3186 1.74 1.44 2.11 <0.001* Age, yr <45 7736 1.00 45 10,709 6.18 4.80 7.97 <0.001* Extrathyroidal invasion No 17,262 1.00 Yes 1085 1.61 1.19 2.19 0.002* Unknown 98 1.89 0.61 5.82 0.268 Lymph node metastases No 16,151 Yes 2294 1.36 1.07 1.73 0.011* Distant metastases No 18,354 Yes 91 3.76 1.22 11.62 0.021* *P reached statistically significant, which is less than 0.05. TABLE 3. Distribution of PTMC-related Deaths According to Demographic and Tumor Characteristics Variables No. Patients (%) Race White 37 (75.5) Africa American 2 (4.1) Others 10 (20.4) Unknown 0 (0.0) Sex Female 33 (67.3) Male 16 (32.7) Age, yr <45 7 (14.3) 45 42 (85.7) Extrathyroidal extension No 21 (42.9) Yes 20 (40.8) Unknown 8 (16.3) Lymph node metastases No 20 (40.8) Yes 29 (59.2) Distant metastases No 39 (79.6) Yes 10 (20.4) Papillary Thyroid Microcarcinomas With Multiple Risk Factors Benefited From Total Thyroidectomy The impact of accumulating multiple risk factors on PTMC s outcome was further analyzed. There were 5033 patients with 2 or more risk factors in our study cohort. The overall survival rate for these patients was significantly worse than those with less than 2 risk factors (15-year survival rate, 83.0% vs 93.4%; P < 0.0001; Fig. 2A). One thousand and forty two patients with 3 or more risk factors showed more unfavorable prognosis compared with those without (15-year survival rate, 75.8% vs 91.5%; P < 0.0001; Fig. 2B). The significant differences for disease-specific survival were also found between PTMCs with multiple risk factors and those without (both P < 0.0001). To evaluate whether different thyroid surgeries could have variant impact on patients outcome, we did multivariate analysis, including the 6 risk factors described before, together with thyroid surgery type, for PTMCs with only multiple risk factors. It was found that PTMCs with 2 or more risk factors had better prognosis if total thyroidectomy was done (HR = 0.70; 95% CI, 0.51 0.96; P = 0.026) compared with either lobectomy or near-total thyroidectomy. In addition, total thyroidectomy benefited the patients with 3 or more risk factors (HR = 0.49; 95% CI, 0.25 0.97; P = 0.042). DISCUSSION The incidence of PTMC is rising rapidly. While most PTMCs behave in a relatively nonaggressive fashion, some patients with this cancer will have poor outcomes. 9,19 22 Recently, a few studies have analyzed data from selected institutions and tried to stratify PTMC patients by using their own defined criteria. 4,13,23 These studies have advanced our understanding of PTMC. However, there is a lack of consistency among the criteria to identify high-risk PTMC patients. Besides, the relative small sample size and short follow-up time TABLE 4. Distribution of PTMC-related Deaths According to Number of Potential Risk Factors No. Risk Factors No. patients (%) 0 1 (2.04) 1 3 (6.12) 2 20 (40.82) 3 or more 25 (50.02) FIGURE 2. Kaplan Meier overall survival estimates. A, Stratified by 2 or more risk factors (P < 0.0001). B, Stratified by 3 or more risk factors (P < 0.0001). 4 www.annalsofsurgery.com C 2011 Lippincott Williams & Wilkins

Annals of Surgery Volume 254, Number 4, October 2011 Risk Factors in PTMC may also create biases due to the general indolent nature of PTMC. Therefore, it is necessary to get a complete view of PTMC from a large population-based study. The present study, using a national database containing more than 18,000 patients, identified predictors for survival. In our study, the 10-year and 15-year overall survival for PTMC were 94.6% and 90.7%, respectively, which were quite similar with the survival rate of papillary thyroid carcinoma. 24 While the common conception is that PTMC is not lethal, we found that the disease-specific survival was 99.5%, which means that 0.5% of the patients died of PTMC. This further emphasizes the importance to stratify the patients into different risk levels. Our analyses found that demographic factors, including age, race, and sex, can significantly affect survival. Among these 3 factors, age had an overwhelmingly high HR of 6.18, which suggests that older patients need to be considered for more aggressive treatment. Unique to differentiated thyroid cancer is the fact that age is a key prognostic indicator for these patients. A recent report from the SEER program showed that the death rate from thyroid cancer is 2.7 of 100,000 in patients older than 65 years, while it is only 0.1 of 100,000 in patients younger than 65 years, a greater than 25-fold difference. This phenomenon is reflected in several staging systems, including AJCC/pTNM and AMES staging systems, in which stage is determined predominantly by a patient s age at diagnosis. 11,18 Sex and race are other 2 risk factors that need to be taken into treatment consideration. In our study, the 10-year survival difference was around 5% between 2 sexes or 2 races. These 2 factors have predictive value for survival in PTMC patients, though the reason remains unclear. In addition to patient factors, distant metastases and extrathyroidal invasion, both indicators for tumor aggressiveness, were identified as predictors for survival in PTMC. The incidence rate of distant metastases of PTMC ranges from 0% to 3% according to different studies. 7,9,21,25 27 The incidence we observed in current cohort was 0.5%. These can be the most representative data since these are reflected by a large population. Despite the low incidence of distant metastases in PTMC, the 10-year survival for patients who had distant metastases was 68%, which was about 30% less than for patients who did not have metastatic disease. Our finding on the prognostic value on extrathyroidal invasion is consistent with AJCC staging system, in which PTMC is classified as T3 category if the tumor has any extrathyroidal extension. However, before our study, few studies have illustrated the clinical significance of these 2 factors, probably due to the overall low incidence in PTMC cohort. A variety of studies have been focused on the importance and prognostic value of lymph node metastases in PTMC. We reported that around 12% of all PTMC patients had nodes involvement since lymph node dissection was not routinely performed in the current cohort. The rate can be as high as 64% if all of the PTMC patients undergo lymph node dissection at the time of thyroidectomy. 8 The correlation between lymph node metastases and tumor recurrence has been confirmed in different aspects by many studies. 23,28,29 The controversy regarding the prognostic value of lymph node involvement is mainly on whether it can predict survival. 30,31 Inadequate sample size and relative short follow-up time could contribute to the discrepancies among various observations. In this study, we showed that PTMC patients with lymph node metastases had poorer prognosis than those without. Although the difference between these 2 groups was not huge, it was significant. Therefore, we agree with most of the contemporary literatures that routine therapeutic lymph node dissection at the time of thyroidectomy is warranted in PTMC patients with palpable lymph nodes, but prophylactic lymph node dissection is not necessary in patients with N0 disease. Our study highlights another important finding about PTMC mortality. We found that the presence of 2 or more risk factors was associated with PTMC mortality. More important, total thyroidectomy was found in our study to benefit the survival for PTMCs with multiple risk factors. The rationale to develop a predictive model on multiple risk factors has been universally accepted and implanted into many staging systems. Regarding the treatment for the newly defined high-risk patients, we will recommend total thyroidectomy for patients with 2 risk factors, with a close monitoring during the follow-up. This suggestion is also based on the safety of the total thyroidectomy as well as a high recurrence rate being reported in patients who had only thyroid lobectomy. 32,33 Whether other treatments are needed depends on which 2 risk factors are present. For patients with 3 or more risk factors, total thyroidectomy alone is not enough from our point of view. In this study, there were total of 1082 patients with 3 or more risk factors, among which 25 died of their thyroid disease. The mortality is not as high (just more than 2%) as in other cancers, but for a treatable disease like PTMC, the adequate selection of initial treatment may help to improve the survival. Total thyroidectomy is necessary for these patients with 3 or more risk factor. Neck dissection may not be applied to the patients if they do not present with lymphadenopathy. The indications for radioactive iodine treatment vary from one institute to another. It is still controversial whether the administration of radioactive iodine treatment can improve the survival of high-risk PTMCs. However, when considering about the more aggressive phenotype of cancer these patients may have, as well as the possible mortality (even it is not high), radioactive ablation should be a choice that clinician need to provide, not to mention the benefit to prevent tumor recurrences. Despite the previous findings, this study does possess some limitations. The SEER database, as a large administrative database, cannot capture every subtle factor, some of which may be critical for clinicians. One of the factors not included was the presence of multifocality of the tumor, which may impact outcomes. 4,34 Other important details like thyroid hormone suppressive therapy and radioactive iodide ablation were failed to be captured in SEER database. In addition, we had little information on disease-free survival due to the lack of recurrence records in the database. With more and more evidences showing molecular markers like BRAF mutations can be useful to predict PTMC aggressiveness, 35,36 future studies are urged to investigate the prognostic value of these markers and utility in decision-making algorithm. Taken together, though PTMC is generally associated with an excellent prognosis, 0.5% patients may die of PTMC. The presence of 2 or more risk factors is strongly associated with cancer-related mortality and can help to identify patients who should be considered for total thyroidectomy and even more aggressive management. REFERENCES 1. National Cancer Institute. SEER Cancer Statistics Review 1975 2007. http://seer.cancer.gov/csr/1975_2007/browse_csr.php?section = 26&page = sect_26_table.05.html. Accessed March 6, 2011 2010. 2. Davies L, Welch HG. Increasing incidence of thyroid cancer in the United States, 1973 2002. JAMA. 2006;295:2164 2167. 3. Cramer JD, Fu P, Harth KC, et al. Analysis of the rising incidence of thyroid cancer using the Surveillance, Epidemiology and End Results national cancer data registry. Surgery. 2010;148:1147 1152; discussion 1152 1153. 4. Haymart MR, Cayo M, Chen H. Papillary thyroid microcarcinomas: big decisions for a small tumor. Ann Surg Oncol. 2009;16:3132 3139. 5. Cooper DS, Doherty GM, Haugen BR, et al. Management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid. 2006;16:109 142. C 2011 Lippincott Williams & Wilkins www.annalsofsurgery.com 5

Yu et al Annals of Surgery Volume 254, Number 4, October 2011 6. Pacini F, Schlumberger M, Dralle H, et al. European consensus for the management of patients with differentiated thyroid carcinoma of the follicular epithelium. Eur J Endocrinol. 2006;154:787 803. 7. Ito Y, Uruno T, Nakano K, et al. An observation trial without surgical treatment in patients with papillary microcarcinoma of the thyroid. Thyroid. 2003;13:381 387. 8. Wada N, Duh QY, Sugino K, et al. Lymph node metastasis from 259 papillary thyroid microcarcinomas: frequency, pattern of occurrence and recurrence, and optimal strategy for neck dissection. Ann Surg. 2003;237:399 407. 9. Pellegriti G, Scollo C, Lumera G, et al. Clinical behavior and outcome of papillary thyroid cancers smaller than 1.5 cm in diameter: study of 299 cases. J Clin Endocrinol Metab. 2004;89:3713 3720. 10. So YK, Son YI, Hong SD, et al. Subclinical lymph node metastasis in papillary thyroid microcarcinoma: a study of 551 resections. Surgery. 2010;148:526 531. 11. Sherman SI. Thyroid carcinoma. Lancet. 2003;361:501 511. 12. Greene FL, Page DL, Fleming ID, et al. AJCC Cancer Staging Mannual. 6th edition. New York, NY: Springer-Verlag, 2002. 13. Sugitani I, Toda K, Yamada K, et al. Three distinctly different kinds of papillary thyroid microcarcinoma should be recognized: our treatment strategies and outcomes. World J Surg. 2010;34:1222 1231. 14. Nathan H, Pawlik TM. Limitations of claims and registry data in surgical oncology research. Ann Surg Oncol. 2008;15:415 423. 15. Gaur P, Leary C, Yao JC. Thymic neuroendocrine tumors: a SEER database analysis of 160 patients. Ann Surg. 2010;251:1117 1121. 16. Fritz AG. International Classification of Disease for Oncology: ICD-O. 3rd ed. Geneva, Switzerland: World Health Organization, 2000. 17. Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996;15:361 387. 18. Cady B, Rossi R. An expanded view of risk-group definition in differentiated thyroid carcinoma. Surgery. 1988;104:947 953. 19. Chow SM, Law SC, Chan JK, et al. Papillary microcarcinoma of the thyroid prognostic significance of lymph node metastasis and multifocality. Cancer. 2003;98:31 40. 20. Hay ID, Grant CS, van Heerden JA, et al. Papillary thyroid microcarcinoma: a study of 535 cases observed in a 50-year period. Surgery. 1992;112:1139 1146; discussion 1146 1147. 21. Baudin E, Travagli JP, Ropers J, et al. Microcarcinoma of the thyroid gland: the Gustave-Roussy Institute experience. Cancer. 1998;83:553 559. 22. Page C, Biet A, Boute P, et al. Aggressive papillary thyroid microcarcinoma. Eur Arch Otorhinolaryngol. 2009;266:1959 1963. 23. Mercante G, Frasoldati A, Pedroni C, et al. Prognostic factors affecting neck lymph node recurrence and distant metastasis in papillary microcarcinoma of the thyroid: results of a study in 445 patients. Thyroid. 2009;19:707 716. 24. Hundahl SA, Fleming ID, Fremgen AM, et al. A National Cancer Data Base report on 53,856 cases of thyroid carcinoma treated in the U.S., 1985 1995 [see comments]. Cancer. 1998;83:2638 2648. 25. Sugitani I, Yanagisawa A, Shimizu A, et al. Clinicopathologic and immunohistochemical studies of papillary thyroid microcarcinoma presenting with cervical lymphadenopathy. World J Surg. 1998;22:731 737. 26. Appetecchia M, Scarcello G, Pucci E, et al. Outcome after treatment of papillary thyroid microcarcinoma. J Exp Clin Cancer Res. 2002;21:159 164. 27. Lo CY, Chan WF, Lang BH, et al. Papillary microcarcinoma: is there any difference between clinically overt and occult tumors? World J Surg. 2006;30:759 766. 28. Besic N, Zgajnar J, Hocevar M, et al. Extent of thyroidectomy and lymphadenectomy in 254 patients with papillary thyroid microcarcinoma: a singleinstitution experience. Ann Surg Oncol. 2009;16:920 928. 29. Lombardi CP, Bellantone R, De Crea C, et al. Papillary thyroid microcarcinoma: extrathyroidal extension, lymph node metastases, and risk factors for recurrence in a high prevalence of goiter area. World J Surg. 2010;34:1214 1221. 30. Sellers M, Beenken S, Blankenship A, et al. Prognostic significance of cervical lymph node metastases in differentiated thyroid cancer. Am J Surg. 1992;164:578 581. 31. Mazzaferri EL, Young RL. Papillary thyroid carcinoma: a 10 year follow-up report of the impact of therapy in 576 patients. Am J Med. 1981;70:511 518. 32. Pearce EN, Braverman LE. Papillary thyroid microcarcinoma outcomes and implications for treatment. J Clin Endocrinol Metab. 2004;89:3710 3712. 33. Cheema Y, Olson S, Elson D, et al. What is the biology and optimal treatment for papillary microcarcinoma of the thyroid? J Surg Res. 2006;134: 160 162. 34. Lin JD, Chao TC, Hsueh C, et al. High recurrent rate of multicentric papillary thyroid carcinoma. Ann Surg Oncol. 2009;16:2609 2616. 35. Xing M. BRAF mutation in papillary thyroid microcarcinoma: the promise of better risk management. Ann Surg Oncol. 2009;16:801 803. 36. Melck AL, Yip L, Carty SE. The utility of BRAF testing in the management of papillary thyroid cancer. Oncologist. 2010;15:1285 1293. J. Olson (Durham, NC): The authors report nearly 19,000 patients with small papillary thyroid cancer followed in the SEER database for a period of 10 to 15 years. During this period, a total of 49 cancer-related deaths occurred, representing well less than 1% of all patients in the series. From their analysis, they have identified 6 factors that were associated with a higher risk for recurrence. These factors have been described in the context of thyroid cancer before and remain significant even in this low-risk population. Nevertheless, based on very low mortality in this disease, it is difficult to know how to apply the knowledge of the identified high-risk factors to the management of any given patient with an incredibly low-risk chance of recurrence and death. I have 3 questions. First, did you analyze the impact of surgery on the outcome of this disease, specifically, the choice of thyroid lobectomy versus total thyroidectomy, and was it significant? If not, how can you justify your conclusion that patients with more risk factors should be aggressively treated? Based on your data, would you advocate completion thyroidectomy and radioiodine for a 60- year-old man with an incidental 5-mm papillary cancer discovered after resection of a goiter? Second, with a very low mortality and disease recurrence, even in the highest-risk cohorts, there is a concern that treatment can be worse than the disease. Can and should a policy of aggressive surgical resection and adjuvant treatment in these low-risk patients be justified across the spectrum of surgical practice? Finally, as we have emerged to the molecular era, there are new available assays testing molecular profiles of these tumors, including BRAF analysis. Should these biomarkers be incorporated into any decision tree for the management of what otherwise would be considered a low-risk disease? To address your first question, we did look at the treatment of total versus lobe for this disease, and it was statistically different. Total thyroidectomy versus lobectomy for thyroid cancer has been debated for decades, and only recently have there been clear, conclusive data to show that total thyroidectomy is the preferred operation for patients with papillary thyroid cancers greater than 1 centimeter. Perhaps, it will take several more decades to answer the question with regard to the microcarcinomas because their prognosis is even better. The first step might be looking at whether total thyroidectomy reduces the recurrence rate in patients with microcarcinomas. As for a patient with 3 risk factors, I would certainly consider doing a completion thyroidectomy in these patients. With regard to your second question regarding complication rate, we would all agree that safety is the number 1 concern when we balance the aggressiveness of the treatment versus the outcome of the disease and the complication rate. Each surgeon must realize what their own complication rate is. If you read the literature, the added complication for a total thyroidectomy would be hypoparathyroidism, which ranges from 1% to more than 50%. If you are a surgeon who has a 50% complication rate, you should consider sending that patient to a surgeon who has a 1% complication rate. Lastly, you asked about molecular markers, specifically BRAF mutations, which have been shown to be correlated with a more 6 www.annalsofsurgery.com C 2011 Lippincott Williams & Wilkins

Annals of Surgery Volume 254, Number 4, October 2011 Risk Factors in PTMC aggressive thyroid cancer and can be tested in either fine-needle aspiration biopsies or from the histology. I agree that we should utilize molecular markers, but we cannot utilize them in everyone. If we applied it to all 18,000 patients to detect 50 patients who had more aggressive disease, the cost of that would be tremendous. What we can do is focus on the patients with 3 or more adverse factors about 1000 patients and figure out which of these thousands have the BRAF mutation and, therefore, should be treated more aggressively. K. Bland (Birmingham, AL): I am a little confused about the choice of the operation, because all of us use extremely high-quality ultrasonography and can find a subcentimeter node in the neck and also find a subcentimeter nodule in the thyroid. For example, if you preoperatively determine that a male African American older than 45 years has a less than 1- centimeter intrathyroidal lesion with negative nodes, what would you prospectively plan the operation to be? Do you think that is a great enough risk to do bilateral total thyroid lobectomies if the contralateral lobe is negative? In that particular patient you described, ultrasonography is important to dictate what we do. If there were contralateral nodule disease, we would be advocating total thyroidectomy. But in your specific patient who has nothing in the other lobe, given these risk factors, I would certainly consider doing a total thyroidectomy after discussing the data with them. Depending on the comorbidities, and the patient s wishes, you have to balance taking additional thyroid hormone replacement versus the need to surveil the contralateral lobe for the remainder of the patient s life. Those are the factors that I take into consideration when I see every patient and offer those 2 options. If things are equal, I definitely push them to a total thyroidectomy for a biopsy-proven microcarcinoma. R. A. Prinz (Chicago, IL): A common goal for all surgeons is to prevent death from cancer. In this study, you are trying to approach that goal by selecting those patients at risk of dying from their thyroid cancer. But is it practical or cost-effective in this group of patients where you start with 18,445 patients and end up with only 49 who died from their disease? Even if you use your 3 or more signs or factors that are indicative of a poorer prognosis, you come up with 25 patients. So you are missing half of the patients. If you look at the 18,396 patients who had but did not die of thyroid cancer, your abstract states that 5.7% of these patients had 3 or more factors. In other words, there were 1049 patients who had 3 or more factors but did not die of this disease. It is very difficult to apply these factors clinically, and 3 of them metastases, presence of lymph nodes, and local invasion are the things that are more likely to be found on pathology rather than preoperative evaluation of the patient. Please comment on this. Second, the issue of markers such as BRAF and other mutational changes in these tumors has been brought up. You commented that these are not readily available. Do you think that ultrasonographic findings may be more useful now, since almost all of these patients have ultrasonography and things like local invasion and suspicious lymph nodes can be seen on the ultrasonography? The follow-up in the SEER database can be a limitation of your study. As we follow patients more and more, we find that there can be very late deaths from this disease. Please comment on this. If we had an effective way of really identifying these patients at risk of dying, I would like your thoughts on whether it would be appropriate to do central lymph node dissection in them, since that is a very hot topic. You are correct; the vast majority of patients with these risk factors will not die of the disease. You mentioned that there are factors that we can use preoperatively, such as age, sex, and race. It may be useful to utilize those who have 3 of 3 positive for the preoperative evaluation and move them toward a consideration of a more aggressive therapy. Ultrasonography can certainly play a role. Local invasion seen by ultrasonography would be a reason to be more aggressive. In the SEER database, they do not study multifocality, so we cannot really comment on that. But most of us in practice would tend toward a total thyroidectomy with multifocal disease. The follow-up in this study is fairly short because we chose to study those patients who were in the SEER database about a decade ago. As we all know, the 20- or 30-year disease-specific survival is very good with thyroid cancer, so with longer follow-up, we may have more patients dying. Lastly, with regard to lymph node dissection, I do not do a prophylactic central lymph node dissection for papillary thyroid cancer greater than a centimeter, given that there are no demonstrated data that this affects survival or recurrence rate. I would not advocate it for patients with microcarcinoma. Q. Y. Duh (San Francisco, CA): I have 2 questions for you. Another risk factor that you did not take into account here is whether the cancer was discovered incidentally for other reasons. If you just take out 100 thyroids randomly in the United States, you will find 5 to 10 microcancers. The problem is confusing incidental cancer with clinical cancers. A patient who presents with big lymph nodes in the neck associated with a microcarcinoma is very different than somebody with a follicular tumor and who incidentally has a 2-mm cancer. My second question relates to size. There is good literature to show that cancers smaller than 5 millimeters have a different prognosis than those that are between 5 mm and a centimeter. The SEER does not distinguish whether or not the thyroid cancer is incidentally discovered. Perhaps, a way to do that study would be to look at the SEER database before we were doing routine ultrasonography, because most of the patients in that time period probably were not diagnosed incidentally. If you looked at those in our database more recently, many were routinely but incidentally diagnosed by using ultrasonography. Addressing your second question with regard to size, we did specifically look at a group of patients who had tumors less than 5 millimeters or greater than 5 millimeters, and there was no difference. D. J. Winchester (Madison, WI): Just a comment. You have a large study, very few events, and you make the conclusion that perhaps the patients with more risk factors should have more aggressive therapy. I guess my problem with that is, how do you even know that more aggressive therapy has an impact on these patients survival? Second, going back to Dr Olson s comment in terms of the surgical treatment of these patients, you identified positive lymph nodes as a risk factor. What percentage of these patients actually had lymph nodes sampled during their surgery? C 2011 Lippincott Williams & Wilkins www.annalsofsurgery.com 7

Yu et al Annals of Surgery Volume 254, Number 4, October 2011 I agree with you; it is difficult to determine whether or not aggressive therapy would make a difference in those patients who had multiple risk factors. Three quarters of the patients in the SEER database had a total thyroidectomy. And only about 1000 of these patients had 3 or more risk factors. So, the majority of these patients did not need a total thyroidectomy, and perhaps, the only ones that did were the more aggressive ones. Thereby, you could decrease the rate of total thyroidectomy. With regard to lymph node metastases, in this series, 12.4% of patients had positive lymph nodes. We do not know how many patients had lymph node sampling, though. 8 www.annalsofsurgery.com C 2011 Lippincott Williams & Wilkins