Papillary Thyroid Carcinoma in Denmark,1996–2008: Outcome and Evaluation of Established Prognostic Scoring Systems in a Prospective National Cohort

Abstract

Background:

Regional as well as national series show an increasing incidence of thyroid cancer largely small size papillary thyroid carcinoma (PTC). Prognostic scoring systems have been developed, but these do not take into account the rapidly changing case mix, and adjustments may be required. The purposes of this study were to evaluate treatment outcomes and to analyze the value of older prognostic scoring systems tested on a relatively new, unselected national cohort of PTC patients.

Methods:

This was a national prospective cohort study conducted in Denmark, which has a population of 5.5 million.

Results:

A total of 1350 patients were diagnosed with PTC during 1996–2008, and the median follow-up time was 7.9 years. The 10-year recurrence-free survival rate was 90.2%, and the 10-year crude and cause-specific survival (CSS) rates were 83.7% and 93.8% respectively. By multivariate Cox regression, it was possible to confirm age, metastases (distant and nodal), extrathyroidal extension, and tumor size as predictors of mortality, whereas only nodal metastases, extrathyroidal extension, and tumor size were predictors of recurrence. In analyses of older prognostic scoring systems, a significant correlation between the risk group ranks was found for survival as well as recurrence. The c-index for CSS was highest for MACIS (0.92) and lowest for AMES (0.80). In the TNM, MACIS, and EORTC systems, most patients were classified as stage 1, and for these patients, the 10-year CSS rate was approximately 99.5%, confirming the generally excellent survival.

Conclusion:

This national study provides further evidence that a favorable prognosis is to be expected for patients diagnosed with PTC. Also, it was possible to confirm age, metastases, extrathyroidal extension, and tumor size as predictors of mortality, whereas only nodal metastases, extrathyroidal extension, and tumor size were predictors of recurrence. All the scoring systems evaluated were able to produce a highly significant risk group stratification, showing that in spite of the changes in the case mix of PTC, these systems are still applicable, and in fact contain valuable prognostic information useable for treatment planning.

Introduction

Thyroid carcinoma is the most common endocrine cancer, and about 80% of thyroid carcinomas are of the papillary type (1). Regional as well as national series show an increasing incidence, primarily of papillary thyroid carcinoma (PTC) <2 cm, in particular papillary microcarcinomas (≤10 mm) (1,2), making PTC a growing challenge for the surgeon as well as the endocrinologist. Most patients have a favorable prognosis. However, aggressiveness can differ significantly, and patients still die from local invasive disease and metastases (3,4). The differences in prognosis have been known for decades, and several prognostic scoring systems have been developed (5 –8). However, since most of these are based on older retrospective single center series, they do not take into account the rapidly changing case mix with increasing numbers of especially small PTC, and the possibility for an overall alteration of prognosis (4,9). New assessments of potential prognostic indicators are therefore needed in order to assess whether the existing systems should be improved. Furthermore, the literature and the published treatment guidelines indicate lack of treatment consensus regarding the smaller carcinomas, with debate over the extent of thyroidectomy, use of neck dissection, and radioiodine ablation (10).

In Denmark, all inhabitants have their own personal identification number (obligatory when using the healthcare system), which makes it feasible to trace patients over time at a national level. Using this system, a clinical national database was established in 1996. The database includes information about patient characteristics, diagnostics, treatment, pathology, recurrence, survival, and causes of death for all registered Danish thyroid cancer patients. The purposes of this study were to evaluate treatment outcomes and to analyze the value of older prognostic scoring systems in a relatively new unselected national cohort of PTC patients.

Materials and Methods

This was a prospective national cohort study. In Denmark, which has a population of 5.5 million, all patients diagnosed with thyroid cancer since January 1, 1996, have been registered in a validated national database, DATHYRCA (11). Registration takes place at the five centers for head and neck oncology, located at Copenhagen University Hospital, Herlev Hospital, Odense University Hospital, Aarhus University Hospital, and Aalborg University Hospital. Contributors to the database include head and neck surgeons, pathologists, and oncologists. The DATHYRCA database is part of the Danish Head and Neck Cancer Group (DAHANCA) and prospectively registers clinical, surgical, histopathological, and follow-up data from all Danish thyroid cancer patients on a national level. Due to a unique 10-digit personal identification number (CPR-number, which also contains birth date and sex information), it is possible to trace the individual patient through all their contacts with hospital and clinical outpatient services, thus ensuring a high level of database completeness. To identify patients who might have been missed at initial contact, in early 2010, the DATHYRCA database was cross-checked against three national governmental databases: the Danish Cancer Registry, the Danish Pathology Databank, and the Danish National Patient Register. All cancers in Denmark are reported to the Danish Cancer Registry, and since 1987, reporting of data has been mandatory. The Danish Pathology Databank is a national database that includes all histological and cytological reports in Denmark. The Danish National Patient Register includes codes for clinical diagnoses and surgical procedures for all hospitalizations and outpatient procedures in Denmark.

CPR-number lists from the DATHYRCA database were compared to lists from the Danish National Patient Register and the Danish Cancer Registry according to the ICD-10 thyroid cancer code (C73.9), and those patients not found in the DATHYRCA database were identified. Furthermore, a cross-reference based on topological and morphological codes was made to the Danish Pathology Databank. For patients missing from the DATHYRCA database, medical records were retrieved and reviewed, resulting in the retrospective inclusion of 185 thyroid cancer patients.

This study is based on data available from the DATHYRCA database on May 21, 2012. The inclusion criteria were PTC verified by cytological or histological examination; diagnosed between January 1, 1996, and December 31, 2008; no prior history of thyroid cancer; and alive at the time of diagnosis.

The overall database consisted of 2037 thyroid carcinoma patients, including papillary (67%), follicular (18%), medullary (7%), or undifferentiated (anaplastic) type (8%). A total of 1350 patients met the inclusion criteria.

Predictor variables were chosen from the clinical and histopathological data in the DATHYRCA database and consisted of age, sex, date of diagnosis, TNM status (5th ed.), tumor diameter, and information about the number of tumors at histological examination.

Outcome variables were time to death and time to recurrence. If a patient was registered as being alive, the information was checked in the Central Office of Civil Registration, and when deceased, the information was updated. The follow-up period ended on March 1, 2012, and patients were censored in the event of emigration from Denmark.

Four of the best-known and most used prognostic scoring systems were chosen for evaluation: TNM staging (5th ed.), MACIS, EORTC, and AMES (see Supplementary Table 1; Supplementary Data are available online at www.liebertpub.com/thy). All were applicable to our cohort. The 5th edition of the TNM classification was chosen for evaluation because this was the only version consistently used for all patients in the DATHYRCA database. TNM staging uses the anatomical extent of the primary tumor, involvement of regional lymph nodes, and distant metastasis in combination with patient age, and assigns four stages (8). MACIS is a staging system by which patients are given a score and divided into four risk groups on the basis of metastases (distant), age, complete resection, invasion, and size (7). EORCT assigns the patient with a composite score calculated using age, invasion of the thyroid capsule, and presence of distant metastases. Based on the score, patients are allocated into five risk groups (5). AMES stands for age, metastases (distant), extent, and size, and AMES designates patients as low- or high-risk cases (6).

To estimate the most precise classification, the T, N, and M stages were based on a synthesis of clinical and pathological information made available immediately following the conclusion of treatment. Thus, information obtained by clinical examination, imaging, surgery, and pathology could be included. If a discrepancy was observed, the pathological state was used. Most of the distant metastases were diagnosed solely by imaging.

In the database, recurrences were defined as persistent disease or occurrence of disease after the end of primary treatment. This could be confirmed by the following modalities: histology, cytology, and/or imaging.

Descriptive statistics were derived according to data type, that is, categorical variables were displayed as frequencies and respective percentages, whereas continuous variables were analyzed by means, medians, and ranges. Multivariate regression using Cox method was applied to examine predictor variables, under the assumption of proportional hazards and confirmed by test of Schoenfeld residuals. The Kaplan–Meier method was used to evaluate survival. In order to analyze the prognostic scoring systems, log-rank testing was applied, and Harrell's C calculated. The C index estimates the probability of concordance between predicted and observed responses. A value of 0.5 indicates no predictive discrimination, and a value of 1.0 indicates perfect separation of patients with different outcomes (12). The level of significance was 5% (two-sided). The database and analysis system Medlog was used for data registration, and STATA/IC 11 (StataCorp LP, College Station, TX) was used for statistical analyses.

The project was approved by the Regional Ethics Committee and by the National Danish Data Protection Agency.

Results

The study population consisted of 1350 patients. Table 1 shows available data by variables (patient, tumor, surgery) including “unknown,” which is an option if information has been sought but found not to exist. From Table 1, it is evident that the data have a very high degree of completeness, ranging from 97% to 100% in all variables. Median follow-up for the study was 7.9 years (range 0.0–16.2). Mean age was 47.5 years, and the male:female ratio was 1:2.8. In 406 patients, a microcarcinoma was histologically confirmed (≤10 mm). These are described in detail in a separate publication (13).

Table 1.

Baseline Number and Percent of Variables for Patients with Histologically or Cytologically Diagnosed Papillary Thyroid Carcinoma in Denmark Between 1996 and 2008

Variable	n	%
Center
1	359	26.7
2	300	22.2
3	220	16.3
4	159	11.8
5	155	11.5
6	157	11.6
Unknown	0	0
Age (years)
≥45	652	48.3
<45	698	51.7
Unknown	0	0
Sex
Male	353	26.1
Female	997	73.9
Unknown	0	0
T-classification 5th ed.
T0	10	0.7
T1	423	31.3
T2	558	41.3
T3	119	8.8
T4	232	17.2
Tx	5	0.4
Unknown	3	0.2
N-classification 5th ed.
N0	858	63.6
N1a	319	23.6
N1b	156	11.6
Nx	9	0.7
Unknown	8	0.6
M-classification 5th ed.
M0	1274	94.4
M1	36	2.7
Mx	19	1.4
Unknown	21	1.6
Tumor diameter
0–10 mm	411	30.4
11–20 mm	26	1.9
20–40 mm	626	46.4
>40 mm	241	17.9
Unknown	46	3.4
Multifocality
Yes	892	66.1
No	423	31.3
Unknown	35	2.6
Thyroid surgery
No surgery	7	0.5
Diagnostic biopsy	11	0.8
Tumorectomy	4	0.3
Istmusectomy	7	0.5
Unilateral resection	5	0.4
Bilateral resection	3	0.2
Hemithyroidectomy	227	16.8
Subtotal thyroidectomy	50	3.7
Total thyroidectomy	1035	76.7
Unknown	1	0.1
Nodal surgery
No surgery	661	49.0
Lymph node extirpation	216	16.0
Modified neck dissection	422	31.3
Classic neck dissection	47	3.5
Unknown	4	0.3

Recurrence

The database contained 123 cases of recurrences, and thus 9.1% of patients experienced a recurrence during follow-up. Figure 1 shows the pattern of failure, and it can be seen that a nodal site (N) is most commonly involved (65% of cases). In three cases, the site of recurrence was not reported.

FIG. 1.

Venn diagram of pattern of recurrence of papillary thyroid carcinoma diagnosed in Denmark between 1996 and 2008.

The five-year recurrence-free survival (RFS) was 92.2%, and the 10-year RFS was 90.2%.

Mortality

A total of 202 cases of mortality occurred in the follow-up period. Of these, 70 died from or had active thyroid cancer, which means the study's overall cause-specific mortality was 5.2%.

The five-year crude and cause-specific survival (CSS) rates were 90.1% and 95.9% respectively, and the 10-year crude and CSS rates were 83.7% and 93.8% respectively.

Variables affecting outcome

Univariate analysis showed significance of all predictor variables, except for multifocality. This variable was subsequently omitted from the multivariate testing.

Multivariate Cox regression showed that only extrathyroidal extension, tumor size, and nodal metastases were predictors of recurrence, with nodal metastasis as the strongest predictor after adjusting for other covariates.

For cause-specific mortality, being over 45 years of age was the strongest predictor, but nodal metastases, distant metastases, extrathyroidal extension, and tumor size were also significant predictors. Table 2 shows the results of the regression analyses.

Table 2.

Result of Cox Regression Analysis from 1265 Cases of Papillary Thyroid Carcinoma

Variable	HR_CSM	CI_CSM	p_CSM	HR_Recurrence	CI_Recurrence	p_Recurrence
Age >45 years	18.9	[5.8–61.0]	<0.0001	1.5	[1.0–2.2]	0.04
Sex	0.8	[0.5–1.3]	0.29	0.0	[0.5–1.2]	0.30
Extrathyroidal extension	2.0	[1.1–3.7]	0.02	2.1	[1.3–3.3]	0.001
Nodal metastasis	5.0	[2.6–9.8]	<0.0001	3.7	[2.4–5.8]	<0.0001
Distant metastasis	2.4	[1.2–4.8]	0.01	1.3	[0.6–2.7]	0.55
Tumor size (cm)	1.4	[1.2–1.5]	<0.0001	1.2	[1.1–1.3]	<0.0001

HR, hazard ratio; CI, confidence intervals; CSM, cause-specific mortality.

Prognostic scoring systems

The 1350 patients diagnosed with PTC were classified according to risk stages in the chosen prognostic systems. Not all patients could be classified in all four systems, and in particular, the completeness of resection in the MACIS system led to exclusion of cases because of “unknown” status.

Table 3 shows the allocation into risk groups in the four systems. Log-rank testing showed the significance of the higher risk group in all systems, even for the occurrence of recurrence, for which they have not been developed. The C index was calculated for CSS for all four systems; 1172 patients who could be classified in all four systems were included. MACIS had the highest score—0.92 [CI 0.88–0.96]—followed by TNM—0.86 [CI 0.79–0.92]—and EORTC—0.86 [CI 0.78–0.94]. AMES had the lowest score—0.80 [CI 0.72–0.89]. In the TNM, MACIS, and EORTC systems, most patients were classified as stage 1, and the 10-year CSS in the systems were 99.6%, 99.4%, and 99.4% respectively, confirming the excellent survival for these patients. However, of these stage 1 patients, 6.2%, 4.8%, and 6.8% experienced recurrence during follow-up.

Table 3.

Crude Survival, Cause-Specific Survival, and Recurrence-Free Survival, with Confidence Intervals, According to Four Different Staging Systems

Staging system	Number of patients	Deaths	5-year CS [CI]	10-year CS [CI]	Death from PTC	5-year CSS [CI]	10-year CSS [CI]	Recurrences	5-year RFS [CI]	10-year RFS [CI]
TNM 5th ed.	1306
I	798	38	96.9 [95.3–97.9]	95.4 [93.4–96.7]	2	99.9 [99.1–100]	99.6 [98.2–99.9]	45	95.5 [93.7–96.7]	93.8 [91.7–95.4]
II	218	41	91.0 [86.2–94.1]	79.6 [72.2–85.3]	6	98.1 [95.1–99.3]	96.6 [92.5–98.5]	10	96.7 [93.2–98.4]	96.1 [92.3–98.0]
III	262	90	76.1 [70.3–81.0]	59.7 [51.9–66.6]	44	85.4 [80.3–89.3]	78.3 [70.8–84.1]	51	79.8 [73.9–84.5]	77.0 [70.6–82.2]
IV	28	24	38.7 [21.0–56.1]	12.9 [2.9–30.5]	16	59.4 [37.3–76.0]	29.7 [11.4–50.6]	10	80.7 [56.0–92.4]	40.4 [8.2–71.9]
AMES	1320
Low risk	1103	100	95.2 [93.7–96.3]	90.8 [88.5–92.5]	17	98.9 [98.0–99.4]	98.1 [96.7–98.8]	62	95.5 [94.1–96.7]	93.9 [92.1–95.3]
High risk	217	100	63.7 [56.7–69.8]	44.5 [35.7–52.8]	53	79.6 [73.3–84.6]	67.5 [57.8–75.4]	58	72.6 [65.2–78.6]	67.5 [59.2–74.4]
EORTC	1350
1	678	21	98.3 [97.0–99.1]	96.7 [94.8–98.0]	3	99.7 [98.7–99.9]	99.4 [97.9–99.8]	44	94.9 [92.9–96.3]	93.2 [90.8–94.9]
2	354	35	94.2 [91.1–96.3]	88.1 [83.1–91.7]	6	99.4 [97.6–99.9]	96.6 [92.1–98.5]	19	95.5 [92.7–97.3]	93.7 [90.1–96.0]
3	204	63	79.0 [72.6–84.2]	67.0 [58.5–74.1]	22	90.4 [85.4–93.8]	87.8 [80.9–92.3]	34	84.1 [77.9–88.8]	81.7 [75.0–86.8]
4	108	78	47.5 [37.4–57.0]	16.2 [8.1–26.7]	36	70.9 [60.1–79.3]	49.9 [34.9–63.2]	24	75.4 [63.7–83.8]	67.6 [51.6–79.3]
5	6	5	33.3 [4.6–67.6]	0	3	44.4 [6.6–78.5]	0	2	60.0 [12.6–88.2]	0
MACIS	1175
1	900	54	96.6 [95.2–97.6]	93.6 [91.5–95.2]	4	99.6 [98.8–99.9]	99.4 [98.3–99.8]	40	96.7 [95.2–97.7]	95.2 [93.4–96.5]
2	125	28	87.0 [79.2–92.0]	78.6 [68.5–85.8]	5	97.5 [92.6–99.2]	95.5 [87.3–98.5]	16	89.4 [82.0–93.9]	84.1 [75.0–90.1]
3	67	29	73.6 [60.9–82.8]	52.2 [35.8–66.2]	12	84.2 [72.5–91.2]	74.5 [54.5–86.7]	17	70.7 [56.7–80.9]	70.7 [56.7–80.9]
4	83	56	48.2 [36.7–58.9]	15.8 [7.0–27.9]	33	65.6 [53.4–75.3]	43.0 [25.5–59.3]	22	72.4 [59.6–81.8]	58.3 [38.5–73.8]

CS, crude survival; CSS, cause-specific survival; RFS, recurrence-free survival.

Discussion

This national and prospective study provides further evidence that a favorable prognosis is to be expected for patients diagnosed with PTC, with a 10-year CSS of 93.8%. Moreover, it was possible to confirm age, metastases (distant and nodal), extrathyroidal extension, and tumor size as predictors of mortality, whereas only nodal metastases, extrathyroidal extension, and tumor size were predictors of recurrence. Sex and multifocality were not found to be predictors. All the evaluated scoring systems were able to produce a highly significant risk group stratification, showing that in spite of changes in the case mix of PTC, these systems are still applicable and contain valuable prognostic information that can be used for treatment planning.

The DATHYRCA database is national, and it has several contributors. This could potentially be a weakness of the study, since the structure allows for some inconsistency in the way data are registered. However, cross-checks to governmental databases were performed in order to secure sufficient catchment, and validation of reported data shows overall good κ values regarding reliability (11), indicating that the DATHYRCA database is valid. Furthermore, the database covers all patients in Denmark, and is unselected and unbiased with regard to social class. Conversely, Denmark is a population of primarily ethnic Caucasians, and the database does not represent a diversity of ethnicity. To our knowledge, this is the first national study of prognosis of PTC from a national prospective, clinical database.

As a clinical database, some biological explanatory variables may exist that are not registered. It has been stated that aggressive subtypes may affect prognosis. Variables concerning this aspect have not been established in the DATHYRCA database, and we have not been able to analyze this characteristic. However, in a recent study, Silver et al. reviewed existing literature on the subject, concluding that the evidence for histological subtype as an independent risk factor was weak, and that the aggressiveness of these carcinomas was probably correlated to other, already known factors, associated with worse prognosis (14). Another limitation of our study is its observational nature, by which change in treatment strategy and quality is not included in the analyses, even though these factors might have influenced prognosis. However, since 2001, all Danish thyroid cancer patients have been treated according to national guidelines (15), which reduces the influence from the treatment aspect, and former studies indicate a generally good quality of thyroid surgery in Denmark (16,17).

The data on mortality in this study were considered as worst case. In a few cases, it was not possible to establish the exact cause of death. In these cases, thyroid cancer was considered the cause of death if the patient had any signs of disease at the last follow-up. This could bias the results, especially the value of age as prognostic indicator, since older patients are more prone to die from other causes.

This study found a favorable rate of recurrence for patients diagnosed with PTC. Compared to data reported by Grogan et al., where the 10-year recurrence rate was 19%, it was 9.8% in our study (18). However, the cohort studied by Grogan et al. consisted of highly selected patients and our unselected series may suggest an overestimation in the cohort of these authors, or it may reflect differences in treatment strategies. Toniato et al. previously reported a recurrence rate of 8.3% after a follow-up of 7.8 years (19), which is similar to that found in this study. The mortality rate found was also favorable and in accordance with previous long-term studies from the Mayo clinic (4) and the University of Chicago cohort (18).

For both recurrence and mortality, more events are to be expected with longer follow-up, which imposes another limitation. This study's median follow-up of 7.9 years, with some patients followed for more than 16 years, should ensure that most events have occurred. Late recurrences are seen, however, and the study by Grogan et al. showed that 11% of recurrences and 17% of deaths occurred after 20 years (18). Further follow-up could thus modify some of the conclusions.

It has previously been reported that age, tumor size, extrathyroidal extension, and nodal and distant metastasis were predictors for mortality (20 –25). The data from our study confirm this. For recurrence, it seems that lymph node metastasis is the most significant predictor, at least for Danish patients. Previous studies have shown conflicting data regarding this subject. In the study by Toniato et al. on 950 consecutive patients from a single institution, nodal metastasis was not found to have significant impact on recurrence (19). Grogan et al., however, found this parameter to be significant. The conflicting evidence may be partly explained by the mode of discovery of nodal metastases. In a study from Kuma Hospital, Japan, it was shown that only clinically evident metastases were significant, as opposed to those found only after histological examination (26). In Denmark, it is not customary to perform prophylactic nodal dissection, apart from exploration/dissection of level VI. Thus, neck dissection is reserved for patients with clinically apparent metastases. Recent data also suggest that the significance is modulated by size and number of metastases and location in either central or lateral neck (27). It was not possible to test this in this national cohort, since data on these parameters were not available.

As for the analysis of scoring systems, no greater differences in outcomes were found among EORTC, MACIS, and TNM, which were all significant and had similar C indexes. AMES also showed significance but with lesser ability to discriminate between patients with a favorable outcome and aggressive cases, with only two risk groups and a slightly lower C index. In general, the C index confirms that the scoring systems are able to separate patients with different outcomes. However, overlapping confidence intervals were found, and no system was significantly superior. When compared to the results of the multivariate regression in this study, TNM would be the preferred system, since it is the only one to incorporate lymph node metastases as a prognostic factor, which is in accordance with the current recommendation of the American Thyroid Association (28). Preferably, evaluation should have been undertaken using the 7th edition of the TMN classification. However, this was not possible because the 7th edition was not used in the early years of registration. A study by Meixner et al. comparing the 5th and 7th editions did not show significant superiority of the 7th edition regarding PTC (29). Implementation of the 7th edition is therefore not expected to change the results significantly.

Earlier, Lang et al. found results comparable to ours when comparing staging systems in a cohort of 589 patients diagnosed at a single institution between 1961 and 2001 (30). Thus, it is confirmed that the “good old” prognostic scoring systems are still going strong and that they are applicable in spite of changing demographics.

Ethics

This study has been approved by the Danish Regional Ethics Committee and by the Danish Data Protection Agency.

Footnotes

Acknowledgments

The authors thank all contributing departments for reporting data to the DATHYRCA database. Funding for the study was provided by Odense University Hospital, University of Southern Denmark, DAHANCA, Becket-Fonden, Else og Mogens Wedell-Wedellsborgs Fond. The study sponsors have had no role in the present study, apart from funding.

Author Disclosure Statement

No competing financial interests exist.

References

Davies

, Welch

. 2006. Increasing incidence of thyroid cancer in the United States, 1973–2002. JAMA, 295:2164–2167.

Londero

, Krogdahl

, Bastholt

, Overgaard

, Pedersen

, Frisch

, Bentzen

, Pedersen

, Christiansen

, Godballe

. 2013. Papillary thyroid carcinoma in Denmark 1996–2008: an investigation of changes in incidence. Cancer Epidemiol, 37:e1–e6.

Elaraj

, Sturgeon

. 2009. Adequate surgery for papillary thyroid cancer. Surgeon, 7:286–289.

Hay

, Thompson

, Grant

, Bergstralh

, Dvorak

, Gorman

, Maurer

, McIver

, Mullan

, Oberg

, Powell

, van Heerden

, Goellner

. 2002. Papillary thyroid carcinoma managed at the Mayo Clinic during six decades (1940–1999): temporal trends in initial therapy and long-term outcome in 2444 consecutively treated patients. World J Surg, 26:879–885.

Byar

, Green

, Dor

, Williams

, Colon

, van Gilse

, Mayer

, Sylvester

, van

. 1979. A prognostic index for thyroid carcinoma. A study of the E.O.R.T.C. Thyroid Cancer Cooperative Group. Eur J Cancer, 15:1033–1041.

Cady

, Rossi

. 1988. An expanded view of risk-group definition in differentiated thyroid carcinoma. Surgery, 104:947–953.

Hay

, Bergstralh

, Goellner

, Ebersold

, Grant

. 1993. Predicting outcome in papillary thyroid carcinoma: development of a reliable prognostic scoring system in a cohort of 1779 patients surgically treated at one institution during 1940 through 1989. Surgery, 114:1050–1057.

Edge

, Byrd

, Compton

, Fritz

, Greene

, Trotti A AJCC Cancer Staging Handbook: TNM Classification of Malignant Tumors. Seventh ed. Springer, New York.

Pathak

, Leslie

, Klonisch

, Nason

. 2013. The changing face of thyroid cancer in a population-based cohort. Cancer Med, 2:537–544.

10.

Hay

. 2010 Managing patients with a preoperative diagnosis of AJCC/UICC stage I (T1N0M0) papillary thyroid carcinoma: East versus West, whose policy is best?. World J Surg, 34:1291–1293.

11.

Londero

. 2014 Completeness and validity in a national clinical thyroid cancer database—DATHYRCA. Cancer Epidemiol, 2014 Aug 14 [Epub ahead of print]; DOI: 10.1016/j.canep.2014.07.009.

12.

Harrell

Jr , Lee

, Mark

. 1996. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med, 15:361–387.

13.

Londero

, Krogdahl

, Bastholt

, Overgaard

, Trolle

, Pedersen

, Bentzen

, Schytte

, Christiansen

, Godballe

. 2013. Papillary thyroid microcarcinoma in Denmark 1996–2008: a national study of epidemiology and clinical significance. Thyroid, 23:1159–1164.

14.

Silver

, Owen

, Rodrigo

, Rinaldo

, Devaney

, Ferlito

. 2011. Aggressive variants of papillary thyroid carcinoma. Head Neck, 33:1052–1059.

15.

National retningslinier for udredning og behandling af thyreoideacancer i Danmark. Available at: www.dahanca.dk/get_media_file.php?mediaid=255 (accessed June 30, 2014 ).

16.

Godballe

, Madsen

, Pedersen

, Sorensen

, Pedersen

, Frisch

, Helweg-Larsen

, Barfoed

, Illum

, Monsted

, Becker

, Nielsen

. 2009. Post-thyroidectomy hemorrhage: a national study of patients treated at the Danish departments of ENT Head and Neck Surgery. Eur Arch Otorhinolaryngol, 266:1945–1952.

17.

Godballe

, Madsen

, Sorensen

, Schytte

, Trolle

, Helweg-Larsen

, Barfoed

, Kristiansen

, Sorensen

, Samuelsen

, Pedersen

. 2014. Risk factors for recurrent nerve palsy after thyroid surgery: a national study of patients treated at Danish departments of ENT Head and Neck Surgery. Eur Arch Otorhinolaryngol, 271:2267–2276.

18.

Grogan

, Kaplan

, Cao

, Weiss

, Degroot

, Simon

, Embia

, Angelos

, Kaplan

, Schechter

. 2013. A study of recurrence and death from papillary thyroid cancer with 27 years of median follow-up. Surgery, 154:1436–1447.

19.

Toniato

, Boschin

, Casara

, Mazzarotto

, Rubello

, Pelizzo

. 2008. Papillary thyroid carcinoma: factors influencing recurrence and survival. Ann Surg Oncol, 15:1518–1522.

20.

Konturek

, Barczynski

, Nowak

, Richter

. 2012. Prognostic factors in differentiated thyroid cancer—a 20-year surgical outcome study. Langenbecks Arch Surg, 397:809–815.

21.

Pelizzo

, Boschin

, Toniato

, Piotto

, Pagetta

, Gross

, Al-Nahhas

, Rubello

. 2007. Papillary thyroid carcinoma: 35-year outcome and prognostic factors in 1858 patients. Clin Nucl Med, 32:440–444.

22.

Bellantone

, Lombardi

, Boscherini

, Ferrante

, Raffaelli

, Rubino

, Bossola

, Crucitti

. 1998. Prognostic factors in differentiated thyroid carcinoma: a multivariate analysis of 234 consecutive patients. J Surg Oncol, 68:237–241.

23.

Eustatia-Rutten

, Corssmit

, Biermasz

, Pereira

, Romijn

, Smit

. 2006. Survival and death causes in differentiated thyroid carcinoma. J Clin Endocrinol Metab, 91:313–319.

24.

Lundgren

, Hall

, Dickman

, Zedenius

. 2006. Clinically significant prognostic factors for differentiated thyroid carcinoma: a population-based, nested case-control study. Cancer, 106:524–531.

25.

Mazzaferri

, Jhiang

. 1994. Long-term impact of initial surgical and medical therapy on papillary and follicular thyroid cancer. Am J Med, 97:418–428.

26.

Ito

, Miyauchi

. 2012. Prognostic factors of papillary and follicular carcinomas in Japan based on data of Kuma Hospital. J Thyroid Res, 2012:973497.

27.

Randolph

, Duh

, Heller

, LiVolsi

, Mandel

, Steward

, Tufano

, Tuttle

. 2012. The prognostic significance of nodal metastases from papillary thyroid carcinoma can be stratified based on the size and number of metastatic lymph nodes, as well as the presence of extranodal extension: ATA Surgical Affairs Committee's Taskforce on Thyroid Cancer Nodal Surgery. Thyroid, 22:1144–1152.

28.

Cooper

, Doherty

, Haugen

, Kloos

, Lee

, Mandel

, Mazzaferri

, McIver

, Pacini

, Schlumberger

, Sherman

, Steward

, Tuttle

. 2009. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid, 19:1167–1214.

29.

Meixner

, Hellmich

, Dietlein

, Kobe

, Schicha

, Schmidt

. 2013. Disease-free survival in papillary and follicular thyroid carcinoma. Comparison between UICC 5th and 7th classifications of T stage, and the prognostic value of primary tumour size. Nuklearmedizin, 52:71–80.

30.

Lang

, Lo

, Chan

, Lam

, Wan

. 2007. Staging systems for papillary thyroid carcinoma: a review and comparison. Ann Surg, 245:366–378.