An Epidemiological Analysis of Thyroid Cancer in a Spanish Population: Presentation, Incidence and Survival

Accurate statistics on cancer occurrence and outcome are essential both for the purposes of research and for planning and evaluation programmes for cancer control (Parkin, 2006). Although tumours of thyroid account for only 1% of the overall human cancer burden, they represent the most common malignancies of the endocrine system and pose a significant challenge to pathologists, surgeons and endocrinologists. Among epithelial tumors, carcinomas of follicular cell origin far outnumber those of C-cell origin. The vast majority of carcinomas of follicular cell origin are indolent malignancies with 10 year survivals in excess of 90 %.


2
features, is strongly supported by advances in molecular studies showing the involvement of distinct genes in these four groups, with little overlap (DeLellis & Williams, 2004).

Staging and prognostic factors
Numerous staging systems have been created in an attempt to accurately prognosticate outcomes for individual patients; two careful studies have compared the efficacy of the various staging systems and found that none is superior (Brierley et al., 1997;Sherman et al., 1998).Consequently, the European Thyroid Association (ETA) (Pacini et al., 2006) and the American Thyroid Association (ATA) (Cooper et al., 2009) have recommended the use of the Tumour, Node, Metastasis (TNM) classification of the American Joint Commission on Cancer (AJCC) and the International Union Against Cancer because it is universally available and widely accepted for other disease sites.An interesting feature of the TNM staging system compared to other classifications is the age factor.While the staging of head and neck cancers relies exclusively in the anatomical extent of disease, it is not possible to follow this pattern for the particular group of malignant tumors that arise in the thyroid gland.The effect of age is such significance in behavior and prognosis, that both the histologic diagnosis and the age of the patient are included in the staging system for these tumors.The AJCC classification is based on the TNM system, which relies on assessing three components: (1) extent of the primary tumour (T), (2) absence or presence of regional lymph node metastases (N), and (3) absence or presence of distant metastases (M).The fifth edition (Fleming et al., 1997), (Table 1) was revised as the sixth edition (Greene et al., 2002), (Table 2).A major alteration was the reclassification of tumour staging (T).For differentiated (papillary and follicular) and medullary tumours confined to the parenchyma of the thyroid gland without extrathyroidal extension, there is no evidence to suggest that using a size cutoff of 1 cm provides better prognostic stratification compared with the 2-cm cut-off used for Table 1.AJCC TNM classification for thyroid cancer (fifth edition).T1 -Tumor 1 cm or less in greatest dimension limited to the thyroid.T2 -Tumour more than 1 cm, but not more than 4 cm, in greatest dimension limited to the thyroid.T3 -Tumour more than 4 cm in greatest dimension limited to the thyroid.T4 -Tumour of any size extending beyond the thyroid capsule.T4a -Excluded.T4b -Excluded.Regional lymph nodes are the cervical and upper mediastinal lymph nodes.N1a -Metastasis in ipsilateral cervical lymph node(s).N1b -Metastasis in bilateral, midline, or contralateral cervical or mediastinal lymph node (s).M0no distance metastases; M1distance metastases.Table 2. AJCC TNM classification for thyroid cancer (sixth edition).T1 -Tumor 2 cm or less in greatest dimension limited to the thyroid.T2 -Tumour more than 2 cm, but not more than 4 cm, in greatest dimension limited to the thyroid.T3 -Tumour more than 4 cm in greatest dimension limited to the thyroid or any tumour with minimal extrathyroid extension (extension to sternothyroid muscle or perithyroid soft tissues).T4 -Excluded.T4a -Tumour of any size extending beyond the thyroid capsule to invade subcutaneous soft tissues, larynx, trachea, oesophagus, or recurrent laryngeal nerve.T4b -Tumour invades prevertebral fascia or encases carotid artery or mediastinal vessels.T4a -Intrathyroidal anaplastic carcinoma-surgically resectable.T4b -Extrathyroidal anaplastic carcinomasurgically unresectable.Regional lymph nodes are the central compartment, lateral cervical, and upper mediastinal lymph nodes.N1a -Metastasis to Level IV (pretracheal, paratracheal, and prelaryngeal/Delphian lymph nodes).N1b -Metastasis to unilateral, bilateral, or contralateral cervical or superior mediastinal lymph nodes.M0no distance metastases; M1distance metastases.

4
TNM classification is also used for hospital cancer registries and epidemiologic studies.One of the greatest inadequacies of TNM system is that it is a static representation of the patient's disease at the time of presentation; it does not allow for modification of risk during lifelong follow-up.Most patients with papillary cancer in the TNM system are classified as stage I disease (Hundahl et al., 1998), with an associated mortality rate of 1.7% (Loh et al., 1997).It is important to note, however, that there is a 15% recurrence rate 10 years after initial treatment (Loh et al., 1997).Recurrent or persistent disease, therefore, may necessitate additional therapy and can certainly affect the patient's quality of life.Further limitations of tumour staging include the lack of consideration of tumour histology, extracapsular extension of the tumour or molecular characteristics of the primary tumour.As is well known, these factors can predict poorer outcomes for individual patients.As TNM staging was developed to predict risk of death and not recurrence, the ATA (Cooper et al., 2009) has created a more functional definition of risk stratification for individual patients that is similar to one outlined by the ETA ( Pacini et al., 2006).Patients are classified as low-risk if they have the following characteristics: no local or distant metastases, resection of all macroscopic tumour, no tumour invasion into locoregional tissues, tumour that is not an aggressive histological variant, no vascular invasion, and no uptake outside the thyroid bed on the post-treatment whole body scan (if 131 I is given).Intermediate-risk patients are those with any of the following criteria: microscopic tumour invasion into the perithyroidal tissues at initial surgery, cervical lymph node metastases or 131 I uptake outside the thyroid bed on the initial post-treatment scan, or tumour with aggressive histology or vascular invasion.Finally, high-risk patients have macroscopic tumour invasion, incomplete tumour resection, distant metastases or elevated thyroglobulin out of proportion to what is seen on the post-treatment scan (Cooper et al., 2009).This stratification was designed to help identify patients who are at higher risk for recurrent disease and may benefit from more aggressive postoperative management (Cooper et al., 2009).Such a definition of risk is more intuitive for the management of patients with thyroid cancer and is more in accordance with the clinical behaviour of these tumours.

Epidemiology
Epidemiology has shown the influence of factors such as age and sex on thyroid cancer incidence.Thyroid cancer is rare in children below 16 years, with an annual incidence between 0.02 and 0.3 cases per 100,000 children and occurs exceptionally before age 10.In adults, the mean age of diagnosis is the mid 40´s to early 50´s for the papillary type, 50´s for the follicular and medullary types and 60´s for the less common undifferentiated types.It is well established that thyroid cancer is 2 to 4 times more common in women than in men, although this will differ among countries.Nevertheless, this sex difference is far less pronounced before puberty and after menopause.Several epidemiological studies have examined several reproductive traits, but the cause of this increased prevalence of thyroid cancer in women is unclear.The annual incidence of thyroid cancer varies considerably in different registries, with the highest incidence rates in the world reported in Hawaii and Iceland (Ferlay et al., 2007;Kolonel et al., 1990).In Europe, the highest incidence occurs in Iceland, followed by Finland, while relatively low incidence characterizes the United Kingdom and Denmark (Ferlay et al., 2007,).These differences have been attributed to ethnic or environmental factors, but different standards of health care may also play a role in the efficiency of cancer detection.Although thyroid cancer incidence is low in general 5 when compared with other diseases and tumours, over the last few decades, increasing rates have been reported in several countries, including Europe (Akslen et al., 1993;Colonna et al., 2002 ;dos Santos Silva et al., 1993;Gomez-Segovia et al., 2004;Petterson et al., 1991;Reynolds et al., 2005;Szybinski et al., 2003), the United States (Davies & Welch, 2006;Merhy et al., 2001;Zheng et al., 1996), Canada (Liu et al., 2001), and Australia (Burgess, 2002).
Curiously, this increase has occurred almost exclusively in papillary thyroid cancer, with an epidemic of micropapillary thyroid carcinoma (MPTC) representing up to 43% of operated cancers in the present series (Leenhart et al., 2004a).The reasons for the rise in thyroid cancer incidence are not completely understood and considerable controversy exists now about whether this increase is real or only apparent due to an increase in diagnostic activity (Leenhart et al., 2004a;Leenhart et al., 2004b;Colonna et al., 2007).Recently, some researchers (Colonna et al., 2007;Davies & Welch, 2006;Kent et al., 2007) have suggested that this increase is predominantly due to the increased detection of small, subclinical tumours through the use of medical imaging.Moreover, thyroid surgery is constantly increasing, with more systematic use of total thyroidectomies even for benign pathologies, which makes it easier to detect MPTC.According to the World Health Organization (WHO), the term MPTC is used for a papillary carcinoma of the thyroid no larger than 1 cm in diameter (Hedinger et al., 1988).With the new classification published in 2004, the previous definition of MPTC now includes the additional criteria of being found incidentally (LiVolsi, 2004).MPTC seems to be present in a significant proportion of the general population with large variations in the prevalence rate between different geographic areas (6-35%) (Sampson et al., 1974), which may also be due to differences in the depth of the pathological examination (Martinez-Tello et al., 1993).Although the mortality risk for an individual patient with thyroid cancer is the greatest concern for patients and clinicians alike, most patients have excellent 10-20-year disease specific survival (Hundahl et al., 1998).EUROCARE (European Cancer Registry-based Study on Survival and Care of Cancer Patients) is a collaborative project between European cancer registries (Capocaccia et al., 2003).A major aim of EUROCARE is to estimate and compare cancer survival in European populations.EUROCARE-2 (Teppo & Hakulinen, 1998) was the first publication on thyroid cancer survival in Europe.This study included all malignant thyroid tumors (excluding lymphomas) in patients 15 or older.Relative survival was analyzed using population-based EUROCARE -2 data from 1985-1989.The overall 5-year relative survival rate, standardized by age (Table 3), was 67% for men and 78% for women across Europe.Substantial variation in this 5-year rate was observed between countries ranging from 56% in Slovenia to 100% in Austria (men), (Teppo & Hakulinen, 1998).Higher than average survival rates were observed in Finland, Iceland, The Netherlands and Sweden.Relative survival was higher in the younger population group.In the age group 15 -44 years, for men the rate was at least 86% and for women at least 94 %.In contrast, much lower rates were seen in the the group of older population (75 + years).EUROCARE-3 study (Sant et al., 2003) analyzed the survival of adult cancer diagnosed from 1990 to 1994 in 22 European countries and followed them until the end of 1998.Neoplasms in situ were collected but not included in the analysis of survival.The overall relative survival of patients diagnosed with thyroid cancer in this period was 83% at 5 years.Austria, Finland, France, Iceland, Italy, Norway, Malta, Spain, Switzerland and Sweden had rates above the European average.Most of these countries also had high survival for this cancer in EUROCARE-2.Denmark, Germany, The Netherlands, England, Scotland, Wales and the countries of Eastern Europe had survival below the European average (Table 3).Again, the most favorable outcomes were observed in patients aged 15-44 years; for the oldest patient's survival was five times lower.Part of variation in thyroid cancer survival was attributed to variations in the distribution of histological types.
Other likely factors contributing to this are differences in the stage distribution and varying efficacy of treatment (Sant et al., 2003;Teppo & Hakulinen, 1998 In the U.S., the National Cancer Data Base (NCDB) represents a national electronic registry system of incident cancers.Between 1985 and 1995, NCDB captured demographic, patternsof-care, stage, treatment, and outcome information for a sample of 53,856 thyroid carcinoma cases.The 10-year overall relative survival rates for U. S. patients with papillary, follicular, Hürthle cell, medullary, and undifferentiated/anaplastic carcinoma was 93%, 85%, 76%, 75%, and 14%, respectively (Hundahl et al., 1998).Relative survival, the survival analogue of excess mortality, is commonly used in population-based studies of cancer survival although its utility is not restricted to this area.Relative survival is the ratio of the observed survival in a group of patients to the survival probability estimated over the same period in a group of people in the general population of similar age and sex.It is usual to estimate the expected survival proportion from nationwide population life tables stratified by age, sex, calendar time, and, where applicable, race (Berkson & Gage, 1950).In order to be comparable between different populations, relative survival figures must be either agespecific or age-adjusted.A major advantage of relative survival is that information on cause of death is not required, thereby circumventing problems with the inaccuracy or no www.intechopen.com An Epidemiological Analysis of Thyroid Cancer in a Spanish Population: Presentation, Incidence and Survival 7 availability of death certificates (Percy et al., 1981).However, our interest is typically in net survival rather than all-cause survival, that is, we are interested in mortality due to cancer.Cause-specific survival is commonly estimated in cancer clinical trials and only those deaths which can be attributed to the cancer in question are considered to be events, while all other deaths are considered censorings.Using cause-specific survival to estimate net survival requires that reliably coded information on cause of death is available.The distinguishing feature of survival analysis is that at the end of the follow-up period the event (such as death due to cancer) will probably not have occurred for all patients.For these patients the survival time is said to be censored, indicating that the observation period was cut off before the event occurred.For example, a person who had the cancer and died 10 years later of car accident would be censored at death, having contributed 10 person-year of survival to the analysis.A person who had the cancer and died 10 years later of the cancer would contribute an event, a death due to the cancer, having also contributed 10 person-years of survival time.A 90 % cancer specific survival at 10 years would mean that 90 % of patients had not died from their cancer, while 10 % had died from their cancer (Kaplan, 1958).Calculation of cause-specific survival is especially important when studying diseases with a favorable prognosis, as is the case at hand, where the patients live long enough to be exposed to other causes of death.The indolent course of thyroid cancer requires very large cohorts of patients followed over several decades to confirm significant differences in prognostic factors and treatment efficacy.Neither randomized clinical trials nor metaanalysis are available and evidence is based on a number of retrospective studies with multivariate for mortality risk factors or data from national cancer registries (Gilliland et al., 1997;Hundahl et al., 1998).Unfortunately, very remarkable differences in patient's selection, staging systems, and clinical management affect the available studies.In particular, radioiodine treatment is not routinely carried out in a standard manner and outcome results of different studies are thus not comparable (Sciuto et al., 2009).Since scarce data exist on the epidemiology of thyroid cancer in Spain, the main aim of this study was to analyze changes in thyroid cancer presentation, incidence, prevalence and survival in South Galicia (north-western Spain) over a 24-year period  and compare these results with those described in the leading international series.The people of this region are homogeneous in terms of ethnicity.This period spans the population's transition from mild iodine deficiency to iodine sufficiency after beginning iodine prophylaxis in 1985 (Garcia-Mayor et al., 1999;Rego-Iraeta et al., 2007).As a high incidence of thyroid cancer owing to improved screening procedures is generally associated with an elevated proportion of small carcinomas, we have specifically considered the impact of MPTC on thyroid cancer incidence and trends in tumour s i z e o v e r t i m e a s a n i n d i c a t o r o f e n h a n c e d m e d i c a l procedures for thyroid cancer.We have also studied the proportion of our population undergoing thyroid surgery over the study period and the percentage of thyroid cancers found per thyroidectomy performed.

Identification of thyroid cancer cases
Data on thyroid cancer incidence in the period from 1978 to 2001 (inclusive) were obtained from the Pathology Registry of the University Hospital of Vigo which belongs to the Spanish public health system and collects data on about 97% of the cancerous lesions verified by microscopic examination.This ensures virtually complete ascertainment for all the new cases of thyroid cancer diagnosed in our population during the study period.Over the observation period, a total of 329 cases of thyroid cancer were registered.Seven cases (six lymphomas and one Angiosarcoma) were excluded from the study based on rarity.The remaining 322 cases of primary thyroid cancer were assigned to one of the four major diagnostic categories: papillary thyroid carcinoma; follicular thyroid carcinoma, including Hürthle carcinomas; medullary thyroid carcinoma; and anaplastic thyroid carcinoma, diagnosed according to the WHO classification (Hedinger et al., 1988).Original histology slides for all cases of follicular carcinomas (53 cases) were reviewed by two hystopathologists blinded to the original diagnosis.Nine of them were reclassified as papillary carcinomas and 44 cases were classified as true follicular carcinomas.All tumour stages were classified according to fifth edition of AJCC (Fleming et al., 1997) since most studies reported having used this classification.In the present study all papillary carcinomas of the thyroid <1 cm in diameter were classified as MPTC (Hedinger et al., 1988).
All thyroid cancer cases were also characterized by sex, date of birth, and date of diagnosis.
We also recorded data on number of thyroidectomies recorded in the registry which were almost exclusively performed by two senior surgeons during the study period.Near-total thyroidectomy has been used as standard treatment protocol for thyroid cancer and comprises neck dissection if confirmed lymph node involvement; one course of ablative radioiodine treatment with 100 mC, further radioiodine therapies with 100 mC if needed, with an interval of 6 months-1 year and thyrotropin-suppressive thyroid hormone therapy with levothyroxine lifetime.

Follow up the vital status of patients
Active follow-up of patients was carried-out through searches in medical records and phone contacts.A detailed review of the medical record to ascertain the cause of death was made.Mortality data were taken into account only when primary cause of death was directly related to thyroid cancer and all other deaths were considered censorings.Cause-specific 1-, 5-, 10-, 20-and 25 year survival rates were used as measures of survival.

Study population
The studied population had an average of 500,000 inhabitants.Corresponding population data by size, age, sex, and year were available from official statistics.Data during the period 1978-2001 show that Vigo's population increased by 6.3%.The male to female ratio remained stable at about 0.92.The people of the region are homogeneous in terms of ethnicity.For studies of genetic characteristics, the Spanish Galician region is considered a relatively isolated European population at the westernmost continental edge (Salas et al., 1998).

Statistical analysis
Trends in age, sex, histological type, and tumour size (differentiated thyroid carcinoma) at diagnosis were analyzed.Data on number of thyroidectomies performed were also recorded.The general descriptive analyses were performed using Microsoft Excel and SPSS 12.0 software (SPSS, Inc., Chicago, IL).Data were analyzed using the chi-square test for nonparametric data.A p value below 0.05 was considered to be statistically significant.
Results were expressed as mean± standard deviation of the mean (mean±SD) for quantitative variables.Data were analyzed using the Student t test for normally distributed variables and the chi-square test for nonparametric data.Crude incidence rates, expressed per 100,000 inhabitants each year were calculated.In order to compare incidence rates between populations that differ with respect to age (since age has such a powerful influence on the risk of cancer), age standardized incidence rates were also calculated; standardization was performed using the World Standard Population (direct method) (Bray, 2002).For the whole group of thyroid cancer, the overall incidence by sex for each year from 1978 to 2001 was calculated.Due to the small sample size, which produces unstable rates for individual years, rates were calculated for several years combined (1978 to 1985, 1986 to 1993, and 1994 to 2001).Incidence trends for each of the distinct histological categories, including MPTC incidence, were also examined.The prevalence of thyroid cancer was defined as the number of persons in our defined population whom have been diagnosed of thyroid cancer, and who were still alive in three cross-sectional surveys performed in December 1985, December 1993, and December 2001.The prevalence rates have been reported per 100,000 inhabitants.A 95% confidence interval (CI) for the rates was determined to compare incidence and prevalence rates.Survival from the data of initial surgery to each endpoint, i.e. cancer specific survival, was estimated by the Kaplan-Meier product-limit method at 1, 5, 10, and 20 and, in some cases, at 25 years of diagnosis.The log-rank test was used to assess difference between subgroups.Age at diagnosis was grouped into the same five categories used by previous EUROCARE studies: 15-44, 45-54, 55-64, 65-74 and 75-99 years.We used multivariate Cox analysis to calculate those independent variables related to the survival of differentiated thyroid cancer.

Results
A total of 322 cases of primary differentiated thyroid cancer were diagnosed in our area between 1978 and 2001.The mean age at diagnosis was 46.6 years (range, 8-91 years).Eight patients were younger than 18 years at diagnosis.The female to male ratio was 3.6/1.

Histological distribution
Out of 322 cases of primary thyroid cancer, papillary was the predominant tumour type with 245 cases (76%), followed by follicular with 44 cases (13.7%), medullary with 23 cases (7.1%), and anaplastic with 10 cases (3.1%), (Table 1).The papillary to follicular ratio in the entire period was as high as 5.8; when MPTC cases were excluded, this ratio was 2.

Age and sex distribution
The youngest age at presentation corresponded to medullary and papillary cancers of the thyroid.Anaplastic cancer and Hürthle cells occurred at older ages.Of the total of thyroid cancers, 78.3% of the cases were females and 21.7% outstanding men.This female predominance is maintained in all histologic types (Table 4).

Pathologic Tumor-Node-Metastases (pTNM) distribution
Altogether, 73% of the primary tumours presented with T1 to T3 tumor size; 15 % were locally invasive to extrathyroidal soft tissues (T4), 22% had metastatic involvement of cervical lymph nodes and 4.7% had distant metastases.Of 11.8% of cases the tumor size was unknown.Among all tumors, medullary and papillary carcinomas were the most commonly presented with cervical lymphadenopathy while follicular carcinoma the most often presented distant metastasis (  1, illustrates the distribution of pTNM stage and histologic subgroup of thyroid cancer patients.

Trends in thyroid cancer incidence
As shown in Fig. 2 and Table 6, incidence rates were considerably lower for males than for females.Overall crude incidence of thyroid cancer in women increased significantly from 1.61 per 100,000 each year (1978 to 1985) to 4.43 (1986 to 1993) and 10.29 (1994 to 2001).These figures in men were 0.35, 1.31, and 3.24, respectively.Age-standardized incidence rates (ASR) over this period show the same tendency, with a significant increase in females: 1.56 per 100,000 each year (1978 to 1985) to 3.83 (1986 to 1993) and 8.23 (1994 to 2001); and males: 0.33, 1.19, and 2.65, respectively (  3 displays the overall (males and females) crude incidence rates of thyroid cancer in relation to the histological types; the increase in the incidence of thyroid cancer over the three periods of time was primarily due to an increase in papillary cancer incidence.After the second period, the incidence of follicular cancer decreased and there was no significant change in the incidence of MTC and anaplastic cancer.Table 7 shows that the increase in the incidence of PTC was the result of an increased incidence of both MPTC and papillary measuring more than 1 cm (Papillary non-MPTC).This occurred both in males and females.

Thyroid cancer survival
We followed a total of 321 cases of thyroid cancer.The median follow-up was 7.7 years, ranging between 4 and 27.8 years.We recorded a total of 43 deaths, of which 30 (70%) were directly related to thyroid cancer, yielding a cancer-specific mortality rate of 9. 3 % for the whole cohort.Over 4 %( 4.3) of cancer -specific deaths was represented by patients with differentiated thyroid carcinomas.Among the remaining 13 deaths not attributable to thyroid cancer, 9 (69%) were due to second malignancies (three breast cancer case, 1 prostate cancer case, 1 case of sigmoid colon cancer, 1 case of liver cancer, 1 case of glioblastoma multiform, 1 case of pancreatic cancer , 1 case of multiple myeloma) and 4 (31%) were attributed to other causes.Overall survival of patients diagnosed with thyroid cancer in the period 1978-2001 was 88 % at 25 years, being 90 % for women and 80% for men; although survival was higher in women, there were no significant differences between both genders (p = 0, 097), (Table 9).When excluding MPTC, we observed a decrease in thyroid cancer survival.Thus, the overall survival of thyroid cancer was 84% at 25 years, being 87% in women and 76% in men, again without significant differences between genders (p = 0.15), (Table 10).

Cause -specific survival according to histological type
As known, histologic type is a strong determinant of thyroid cancer survival.In our series, papillary thyroid cancer patients had 25-year specific-survival greater than 93 %, even when excluding MPTC.The survival of MPTC was 100% at 25 years in the present study.Follicular and medullary carcinoma patients had lower survivals (83% at 25 years and %at 20 years, respectively).However, the prognosis was is ominous for anaplastic thyroid carcinoma (Table 12 and Figure 5).12. Cause-specific survival of thyroid cancer according to histological type .

Cause -specific survival according to p TNM stage distribution
Stage at diagnosis is a strong prognostic factor for thyroid cancer survival.Thus, cause specific-survival vas 100% at 25 years of follow-up in stage I.At more advanced stages survival decreases progressively (Table 13).  .

Discussion
The main objective of epidemiological studies is to measure the frequency of disease.
Prevalence measures are particularly useful in the healthcare planning.Furthermore, incidence reflects the "flow" from health to illness in populations and therefore constitutes the basis of causative research.On the other hand, survival measures are an indicator of the global efficiency of healthcare services.These questions are also fundamentally important in thyroid cancer, since an understanding of the basic causes and related risk factors may lead to novel interventions and preventive measures.The reason why we carried out this work was the paucity of data on the major epidemiological features of thyroid cancer in our country (Spain) and particularly in our region (Galicia).In this study we conducted an epidemiologic survey in our community, to evaluate time trends in presentation, incidence, prevalence and survival of thyroid cancer between January 1978 and December 2001; a period that spans our community transition from mild iodine deficiency to iodine sufficiency after beginning iodine prophylaxis (iodized salt) in 1985.Several factors could have an impact on the epidemiology of thyroid cancer in our area.On the one hand, the eradication of iodine deficiency in our population in the last decades (Garcia-Mayor et al., 1999;Rego-Iraeta et al., 2007;Rodriguez I et al., 2002), on the other hand, the progressive increase in the use of diagnostic techniques and the preference for carrying out near-total thyroidectomy since the nineties, (compared to the "lumpectomy" or hemithyroidectomy) which is known to increase the likelihood of detecting microscopic carcinomas incidentally, mainly papillary lineage.In Galicia, like in the rest of Spain, 97% of the population receives health care through the public health system, with other kinds of medical care being negligible (Etxabe & Vazquez, 1994).This ensures virtually complete case ascertainment for diagnosed thyroid cancer in the population.Given that practically the whole population of Galicia is registered with the social security system, our study is representative of the Galician population (North-western Spain).
The general characteristics of our patients with thyroid cancer such as the mean age at diagnosis (46.6 years old) and the predominance of females to males (3.6/1) were similar to those reported in other studies around the world (Blanco Carrera et al., 2005;Gilliland et al., 1997;Sant et al. 2003;Scheiden et al., 2006;Sciuto et al., 2009).Agree with this, the youngest age at presentation corresponded to medullary and papillary thyroid carcinomas, followed by follicular carcinoma, Hürthle cell cancer and finally by the anaplastic carcinoma.The histological distribution of thyroid cancer in the present series was similar to that reported in iodine-sufficient areas, with a higher proportion of papillary thyroid cancer (76% over overall thyroid carcinomas).It represents the same pattern that was reported in the USA (Hundahl et al., 1998;Scheiden et al., 2006;Schlumberger et al., 2008) and some European countries (Blanco Carrera et al., 2005;Farahati et al., 2004;Sant et al. 2003;Scheiden et al., 2006).Likewise, in our series the ratio papillary/follicular was high, which is the predominant pattern reported in the Western world (Sant et al., 2003;Teppo & Hakulinen, 1998).This happened even in the first period of our study when there was a mild iodine deficiency in our population (median iodine of 60μg/l) (Garcia-Mayor et al., 1999).This ratio increased over time, even when MPTC were excluded from the calculation.We speculate that the amelioration of iodine nutrition, which happened in our population over the last decades (Garcia-Mayor et al., 1999;Rego-Iraeta et al., 2007), may explain this finding to some degree.However, an increase in the incidence of papillary thyroid cancer has also been reported in Luxemburg, considered an iodine deficient area (Sant et al., 2003), and in Tasmania (Australia), in spite of the recurrence of mild iodine deficiency (Burgess et al., 2000).This phenomenon may be related with a dose-threshold effect for iodine nutrition and modulation of tumour genesis (Burgess et al., 2000), or alternatively, environmental factors other than iodine nutrition may be contributing.
In our study, most of patients (75 %) had low pathological p-TNM stages (stages I and II) at presentation.These results are similar to those described in a cohort of more than 2700 patients with thyroid cancer at the Mayo Clinic of Rochester who underwent thyroidectomy from 1940 to 1997 (Schlumberger et al., 2008).Among papillary carcinoma cases, this series reported extrathyroidal invasion in 15 % (range 5 to 34%), and clinically evident lymphadenopathy at presentation in about one third of cases.Only 1 to 7% of papillary carcinomas had metastases at diagnosis.However, it is noteworthy that about 35 to 50 % of removed neck nodes have histologic evidence of involvement (Schlumberger et al., 2008), so that our low rate of lymph node involvement may reflect the fact that node dissection is not performed routinely in our environment.Regional lymph node metastases from follicular carcinomas are uncommon (4 to 6% of patients).Indeed, wherever they are observed, other alternative diagnoses, should be considered.Around 5 to 20% of these tumors have distant metastases at presentation (Schlumberger et al., 2008).In our series, 12.5% of follicular carcinomas and 8% of Hürthle cell carcinomas had lymph nodes at diagnosis.In accordance with previous observations, we found distant metastases in 19% of follicular carcinomas, although we found no distant metastases in the group of the Hürthle cell carcinomas.In relation to staging at presentation, we compared our results with the case material of the Mayo Clinic, between 1940 and 1997.
The distribution of pathological stages I, II, III and IV for papillary carcinoma at the Mayo Clinic was respectively 60, 22, 17 and 1%, very similar to the corresponding in our study, which was 63, 18, 12 and 1.2% for the same stages.With regard to follicular carcinomas, the Mayo Clinic study found that the distribution of stages I, II, III and IV was 22, 53, 4 and 17% for follicular carcinoma and 17, 69, 9 and 5% for the Hürthle cell carcinoma.Curiously, stage I was more frequent among our follicular carcinomas (the distribution of stages I, II, III and IV for follicular carcinoma was 37, 28, 6.2 and 15.6% and for Hürthle carcinoma 25, 58, 8 and 0%; respectively).With reference to medullary carcinoma, the Mayo Clinic cohort data revealed a predominance of stage I (stage I, II, III and IV; 57, 19, 22 and 2% respectively), while in our study only 4.3 % of patients presented with stage I.This is probably due to earlier introduction of RET proto-oncogene testing at the Mayo Clinic.
Thyroid cancer has increased dramatically in most countries in the last 30 years (Kilfoy et al., 2009), excluding countries such as Iceland, Sweden and Norway (Engholm et al., 2009).In the present study, the incidence of thyroid cancer is increasing over time.When comparing incidence rates among different populations there are two points to bear in mind; first, since a rapid increase in thyroid cancer incidence is seen, it is important to consider the period of time the rate refers to, and secondly due to the differences in the age distribution among different populations, it is necessary to display an age-standardized rate of incidence (ASR).Most of the series, report ASR referred to world population.In the present investigation, both crude incidences and ASR show an increasing trend over time.In comparison with other European countries, our ASR in the final period of time, 1994-2001, (8.2 per 100,000-year in women and 2.65 in men) is similar to the reports from our neighbouring countries such as Portugal, France and Italy and is higher than that reported by the IARC for Spain (period 1997-1999) (Ferlay et al., 2004).A previous study in our community (Garcia-Mayor et al., 1997), reported a decrease in the number of patients requiring surgical treatment with an increase in the frequency of malignancy in the surgical specimens after the introduction of FNAB in the management of nodular thyroid disease in 1991.However, when we studied the total number of thyroid surgeries performed in our population, we found that the rate of population undergoing thyroid surgery significatively increased over time with an increase in the ratio of total thyroidectomy for other kind of thyroidectomy.Undeniable, this trend makes it easier to detect MPTC.In fact, 43 % of our operated cancers were MPTCs in the latter period studied versus a 16.7 % in the first period.
A similar trend has been reported in France where there was an increase in thyroid cancer incidence, mainly due to the papillary type, with an epidemic of microcarcinomas (43% of operated cancers, for the period 1998-2001) (Leenhardt et al., 2004a).This trend has been reported in many other studies (Chow et al., 2003;Colonna et al., 2007;Scheiden et al., 2006;Verkooijen et al., 2003).The improvement in diagnostic tools (image procedures and fineneedle aspiration biopsy) (Colonna et al., 2007;Scheiden et al., 2006) and greater extensiveness and number of thyroidectomies performed, which makes it easier to detect MPTCs (Leenhardt et al., 2004a), has been suspected to be of etiological importance in the observed increase of papillary thyroid cancer.One study (Kovacs et al., 2005) estimated the prevalence of thyroid microcarcinomas found at autopsies is 100-1000 times higher than in clinical cancer; they were not related to iodine intake and were exclusively of the papillary type (MPTC).It suggests that a large proportion of the population probably lives with undetected thyroid cancer and fits with the hypothesis of an apparent increase in thyroid cancer incidence.Any interpretation of reports of the incidence of papillary thyroid carcinoma must take into account the remarkably high prevalence of MPTC in thyroids removed for reasons other than thyroid cancer and in autopsy series (Hedinger & Sobin 1988).In this sense, it is noteworthy, that many cancer registries do not specify the contribution of MPTC to the incidence of thyroid cancer, so differences in the inclusion criteria can cause mistakes in the comparison of the incidences (Teppo & Hakulinen 1998).
For these reasons we have separately analyzed the incidence of MPTC and the incidence of papillary cancer not including MPTC (Papillary non MPTC).In the present investigation, we found 245 cases of papillary cancer, of which 95 cases (38.7 %) were MPTC carcinomas (pT1).Remarkably, most of these tumours (91%) were detected incidentally after thyroid surgery performed for reasons other than thyroid cancer.Although the incidence of MPTC is increasing in our population, also an increase in the incidence of tumours greater than 1 cm (Papillary non MPTC) was evident in both sexes.Similar findings have been reported in studies performed by Burgess in Australia (Burgess, 2002;Burgess & Tucker 2006).We also observed an increase in the ratio of total thyroidectomy for other kind of thyroidectomy over time.A recent study performed by Mitchell et al. in USA (Mitchell et al., 2007), examined trends in surgical therapy for thyroid cancer.They hypothesized that if a true increase occurs in the incidence of thyroid cancer, then thyroidectomy, as the primary treatment for thyroid cancer, should also increase during the same period.This study reported a regional difference in the incidence of thyroid cancer with an increase in Northeastern and Southern and an actual decrease in the Midwest United States.Newly papillary thyroid cancer accounted for most of this increase.Furthermore, thyroidectomy, in these areas seemed to mirror their respective regional changes in incidence.Supporting the hypotheses of recent changes in medical practice as the cause of the increase in thyroid cancer incidence, some authors (Davies & Welch, 2006;Kent et al., 2007) found a shift in the tumour size distribution of thyroid cancer toward smaller papillary cancers in recent years, suggesting an apparent (not real) increase in thyroid cancer incidence due to increased detection of subclinical tumours.However, besides MPTC, we did not observe a significant change in tumour size over time in differentiated thyroid carcinomas at presentation with a percentage of pT 2 , pT 3 and pT 4 lesions which remain stable over time.Moreover, there are three questions to be in mind; firstly, the increase in thyroid cancer incidence in our area has happened equally in the incidence of MPTC and in the incidence of PTC greater than 1 cm.; secondly, there has not been a shift over time in thyroid cancer tumour size, besides MPTC, and thirdly, there is no similar increase in the incidence of other histological types of thyroid cancer (Rego-Iraeta et al., 2009).Interestingly, Kent and colleagues (Kent et al., 2007) found that the incidence of medium-sized tumours (2-4 cm) remained stable over time, but were surprised to discover a slight increase in large tumours (larger than 4 cm).Several others papers from U.S. support the notion that the increase in incidence is not entirely due to increased screening and detection.Thus, Enewold and colleagues found, among white women, the rate of increase for cancers larger than 5 cm.almost equaled that for the smallest papillary cancers.Chen et al. similarly reported an increase in differentiated thyroid cancer of all sizes with the most rapid increase occurring in females.Cramer et al. showed an increase in the incidence of papillary thyroid cancers with a significant increase in all size categories.A report by Morris & Myssiorek drew similar conclusions based on data demonstrating significant rises in the incidence of large (>4 cm), and well-differentiated cancers with clinically significant pathological adverse features (Chen et al., 2009;Cramer et al., 2010;Enewold et al., 2009;Hodgson et al., 2004;Morris & Myssiorek, 2004).Improved detection has undoubtedly occurred and may explain much of the increase in small welldifferentiated cancers.However, the most important evidence that increased diagnostic activity is not the sole cause for this increase is that large and more advanced cancers are increasing as well as small tumours.This trend suggest than some environmental factor, besides increased diagnostic activity, may be contributing to the increase in thyroid cancer incidence.Thyroid cancer can be induced in experimental animals directly by mutagenic carcinogens and indirectly through hormone imbalance.
The only well established risk factor for thyroid cancer in humans is ionizing radiation.Sex hormones, iodine deficiency and other factors (nutritional, volcanoes) have been proposed as risk factors for thyroid cancer, but the findings are inconsistent (Nagataki & Nystrom, 2002).The increase in thyroid cancer risk could be attributed to ionizing radiation exposure.
Studies of individuals living in the Chernobyl areas have shown an increased risk among those exposed as children (Cardis et al., 2006).In our study, only 8 cases of thyroid cancer in people of 18 years of age or less could be identified, so ionizing radiation exposure cannot explain the recent increase in incidence of thyroid cancer in our community.However, a longer latency period for low doses of radiation could not be ruled out (Yamashita, 2006).Some authors have suggested that iatrogenic exposure to radiation during imaging by computed tomography, especially in children when radiation sensitivity of the thyroid gland is greater, could contribute to the increase in thyroid cancer (Baker & Bhatti, 2006); however this link remains unproven at the moment.Nutritional factors such as a low fruit and vegetable and selenium consumption have been linked to thyroid cancer and cancer in general (Clark et al., 1996;Franceschi et al, 1990;Rayman, 2000).The role of Brassica vegetable in cancer protection (Keck & Finley, 2004;Verhoeven et al., 1996) has also been reported.Remarkably, our soils are acidic with low selenium content and the consumption of these Brassica vegetables (cabbage), traditional in our community of the North-West of Spain, has decreased over the last decades due to the globalization of our diet and loss of the popularity of these foods because of their goitrogenous potential.
Contrary to what happened with reports on the incidence of thyroid cancer, reports on prevalence are scarce.In the USA, the prevalence estimate for thyroid cancer was 310,000 in 2001 (Sherman & Fagin 2005), averaging about 105 cases 100,000 population.In our study, we observed an increase in the prevalence of thyroid cancer from 8 cases per 100,000 population in 1985 to 83 cases in 2001.The prevalence varied for both sexes, with the figure being greater in women (128 cases per 100,000 in 2001).The increase in thyroid cancer incidence seen in our area together with the good prognosis of this neoplasia can explain the increase in the prevalence of thyroid cancer.These data should be taken into account when planning health resources for the management of these patients.
In the present study, we also performed an analysis of cause-specific survival in our patients diagnosed of thyroid cancer between 1978-2001.In the case of deceased patients, we investigated the exact cause of death by reviewing the medical records to carry out the calculation of cause-specific survival.Throughout the follow-up period, we recorded in our series a total of 43 deaths, of which 70% were directly related to thyroid cancer.Among the remaining deaths not attributable to thyroid cancer, 69% were due to second malignancies.A high percentage of secondary neoplasms as the cause of death in thyroid cancer patients is also reflected in other series, for example, a Norwegian and a Dutch study found a 38% and 58% of deaths, respectively, attributable to other malignancies (Akslen et al., 1991;Eustatia-Rutten et al., 2006).In Europe we have data on thyroid cancer survival from the EUROCARE database.As the first publication on cancer survival in Europe, EUROCARE-1 (1978EUROCARE-1 ( -1985) ) (Berrino et al., 1995) did not involved thyroid cancer, EUROCARE-2 (1985EUROCARE-2 ( -1989) (Teppo et al., 1998) was the first publication on thyroid cancer survival.We also have more recent data from the EUROCARE-3 (1990EUROCARE-3 ( -1994) ) (Sant et al., 2003).Both studies were population-based and used relative survival, i.e. an estimate of excess mortality.Five-year relative survival collected for Spain in the EUROCARE-3 (85.7% women and 82% in men) places it slightly above the European average (81.4% females and 71.8% in men).In our study, we observed an overall thyroid cancer cause-specific survival after excluding MPTC (91% in women, 89% in men) that is better than those previously reported over European countries as was reflected in the EUROCARE 2 and 3.As expected, we observed that survival decreases gradually with age in our study.We found a 5-year survival of 65% in the group older than 75 years (59% when excluding the MPTC).With regard to the histologic distribution, cause-specific survival at 25 years in our series is 95% for papillary thyroid carcinoma (93% if we exclude MPTC) and 83% for follicular carcinoma (including Hürthle cell).In the case of medullary carcinoma, the 20-year survival rate was 63%.As expected, anaplastic carcinoma was an ominous prognosis with a 5-year survival rate of 10%.We can compare these results with those of the cohort of patients with thyroid cancer at the Mayo Clinic (Schlumberger et al., 2008), where the cause-specific survival at 25 years for papillary thyroid carcinoma was 95%, wich was significantly higher than the rates found for medullary, Hürthle and follicular thyroid carcinoma, wich were 79, 71 and 66%, respectively.Curiously, patients with medullary thyroid carcinoma at the Mayo Clinic have similar or better outcomes than those patients with non papillary follicular thyroid carcinoma; as more patients have been diagnosed by genetic testing, most of them have curable disease and better survival.With regard to survival by histologic type, we found very similar results to those published in the U.S. series (Hundahl et al., 1998 ) where 10year overall relative survival rates for patients with papillary, follicular, Hürthle cell, medullary, and undifferentiated/anaplastic carcinoma were 93, 85, 76, 75 and 14%, respectively (Hundahl et al., 1998).In our series, 10-year cause-specific survival for papillary thyroid carcinoma was 96% (95% if we exclude the MPTC, 87% for follicular carcinoma (including Hürthle cell), 70 % for medullary carcinoma and 0% for anaplastic carcinoma.A similar distribution of histologic types and pTNM stages may be one explanation for these similar results.With regard to staging, as expected, we found a survival of 100% at 25 years for tumors which were presented in stage I. Accordingly; we did not record any deaths due to thyroid cancer in our 95 cases of MPTC.Survival gradually worsened with more advanced staging at presentation, being only 15% at 10 years for stage IV tumors.In our study, the stage of tumors at presentation was generally favorable, which may have contributed positively to overall survival.Despite being a relatively benign disease, a continual decline in cancer-specific survival is noted in all tumor stages at successive followup intervals.This underscores the need of life-long surveillance for thyroid cancer patients.
Another objective of this study was to describe the prognostic factors associated with thyroid cancer survival.Because these variables are often strongly interrelated, we have identified risk factors associated with mortality from differentiated thyroid cancer using the Cox regression analysis.This was determined only in the group of differentiated thyroid carcinomas, as the rest of the tumors represent a little large in total thyroid cancers and they have a different clinical behavior.Age, follicular histotype, local tumor extension as well as distant metastases were found to have a significant negative influence on survival in the univariate analysis.However, in the multivariate analysis, only age and distant metastases were found to retain their independent prognostic values.Multiple studies have identified several prognostic factors, but overall the findings have been inconsistent, possibly due to bias introduced by the use of different institutional series with different distributions of histologic types and differences in follow-up and histologic classification of disease.Therefore, we do not know the relative importance of each of these features as prognostic factors and whether the findings in one population can be generalized to other populations (Gilliland et al., 1997).Although the majority of studies reported the effect of age on the prognosis of patients with differentiated thyroid carcinoma, data from some studies (Elisei et al. 2010;Gilliland et al., 1997;Hundahl et al., 1998;Sciuto et al. 2009) also suggest an effect on other thyroid histologies as well.The association between age and survival is not explained by differences in stage at diagnosis, differentiation, socio-demographic variables, or treatment.It has been speculated that other age-dependent factors such as nutritional or immune status, or differences in the spectrum of genetic alterations in tumors in the elderly, may play a role in survival (Gilliland et al., 1997).In this sense, several studies have found that survival does not differ for patients of similar ages and stages diagnosed with papillary or follicular carcinoma (Thoresen et al. 1989;Torres et al., 1985).Children and people younger than 20 years tend to present with higher stage disease and greater likelihood of locoregional and distant metastases, despite it, children generally have excellent survival rates, the exception to this rule is the disease in children aged < 10 years (Sipos & Mazzaferri, 2010).In the same line as most of the studies (Beenken et al., 2000;Eichhorn et al., 2003;Gilliland et al., 1997;Lerch et al., 1997;Mazzaferri, 1999;Shah et al., 1992), we found that the presence of distant metastases at presentation is an independent risk factor in the prognosis of differentiated thyroid carcinoma.The main cause of death from differentiated thyroid cancer is distant metastases (Mazzaferri, 1993).Mortality is high with distant disease, with 50 % survival at 3.5 years according to one recent study (Sampson et al., 2007).However, survival is improved in younger patients (Sampson et al., 2007), patients with microscopic rather than macroscopic disease (Durante et al. 2006), and patients with iodine-avid tumours (Durante et al. 2006;Sampson et al., 2007).Furthermore, the ability to achieve a negative post-treatment scan after multiple doses of radioiodine was associated with 92% overall 10-year survival, compared with 19% survival for patients who did not achieve a negative post-treatment scan (Durante et al. 2006).Numerous factors affect outcome of patients with thyroid cancer; in spite of these various factors, only a few are considered in the currently recommended TNM staging system.The clinician must therefore have a complete understanding of the various prognostic factors and how they contribute to the outcome, so that the patient can be counselled accordingly about treatment and longterm surveillance decisions (Sipos & Mazzaferri, 2010).

Conclusions
In conclusion, we have analyzed for the first time, the descriptive epidemiology of thyroid cancer in Vigo, Galicia (Spain), between 1985 and 2001.Long term follow-up ascertainment was practically complete, providing valid information on thyroid cancer prognosis.The results of this study can be summarized as follows: the first point to note is the histologic distribution of thyroid cancer in our population; wich is similar to that found in areas with high iodine intake, with a clear predominance of differentiated thyroid carcinoma and a high ratio of papillary to follicular carcinomas.As in many other regions and countries, the incidence of thyroid cancer is increasing and this trend is primarily caused by an increase in the incidence of papillary type.Our data showing an increase in papillary cancers larger than 1 cm suggest that some environmental factor may be contributing to this trend.There is a significant increase in the prevalence of thyroid cancer over time, especially among women.These data should be taken into account when planning health resources for management of this disease.Cause-specific survival of thyroid cancer in our study is higher than the European average, similar to that found in the U.S. series of thyroid cancer.Possible explanations for these results are: a high proportion of differentiated carcinoma, particularly papillary thyroid carcinoma, and a favorable stage (I and II) of the tumors at presentation.This study has some limitations: the relatively limited number of cases, particularly for selected histologic type; the study is merely descriptive, so it is not possible to give a definitive explanation for the observed increase in the incidence of thyroid cancer.His strength is also limited because we could not continue the study beyond 2001 in order to see if this trend continues or instead, the incidence of thyroid cancer reaches a plateau.Furthermore, information on a number of variables, such as vascular invasion, tumour recurrence and treatment (dose and frequency of I-131) were not controlled and may have influenced on survival.Further studies in these areas seem prudent.
Fig. 3. Time trend of crude incidence rates of thyroid cancer, by histology.

Fig. 5 .
Fig.5.Cause-specific survival of thyroid cancer according to histological type.

Table 4 .
Thyroid cancer characteristics at diagnosis.

4 Distribution of pTNM stages of thyroid cancer at diagnosis (1978-2001) Most
of thyroid cancer patients (75 %) presented low pathological tumor-node,-metastases (stages I and II).Most papillary cancers presented with either stage I (63 %) or stage II (18 %).Stage III accounted for fewer than 12 % of cases.Few (1.2 %) patients presented with distant metastases and had stage IV disease.For follicular and Hürthle cancers these figures were 37, 28, 15, 6% and 25, 58, 8 and 0 % respectively.Most patients with medullary thyroid cancer (43.5 %) had stage II; patients with stage I accounted for only 4 %; and stages III and IV, 26 % and 13 % respectively.Figure

Table 5 .
Table5).Spanish Population: Presentation, Incidence and Survival Time trend of thyroid cancer presentation.

Table 6 .
Time trend of crude and age-standardized incidence rates of thyroid cancer, by sex.(*) Age-standardized incidence rate (ASR).

Trends in thyroid cancer prevalence
Time trend of crude incidence rates of thyroid cancer, by histology.Table8shows that prevalence of thyroid cancer increased substantially between 1985 and 2001 in both sexes.Thyroid cancer was significantly more prevalent in female than in male subjects.

Table 8 .
Time trend of thyroid cancer prevalence, by sex.

Table 9 .
Overall cause-specific survival of thyroid cancer.

1 Cause -specific survival according to age
Table11 and Figure 4, reflect the cause-specific survival by age group (excluding MPTC) and emphasizes the influence of age on the prognosis of patients with thyroid carcinoma.As can be seen there is one more striking decline in survival after 55 years of age.

Table 14 .
Risk factors associated with differentiated thyroid cancer mortality were identified by Cox regression analysis.Univariate and multivariate analysis results for thyroid cancer mortality are illustrated in Table14.In the univariate analysis, the following factors were significantly associated with mortality for differentiated thyroid cancer: age, follicular histology, local tumor extension and distant metastases at presentation.Neither sex nor the presence of lymph node metastases contributed to mortality risk.Multivariate analysis confirmed as independent predictor variables of increased risk of cancer mortality-only age and presence of distant metastases.Univariate and Multivariate survival analysis of prognostic factors of differentiated thyroid cancer