Background
Thyroid cancer is the most common endocrine carcinoma, and papillary thyroid carcinoma (PTC) is the 7
th most common cancer among women in the world [
1]. There are numerous risk factors for PTC, including radiation exposure, insufficient iodine intake, hormonal factors, and family history [
2]. It is noteworthy that molecular analyses have indicated that lymphocytic infiltration is frequently observed in PTC, suggesting that immunological factors might be involved in tumour progression [
3,
4]. Hashimoto’s thyroiditis (HT) is the most common inflammatory thyroid disease and the most common cause of hypothyroidism in the United States, affecting 22 of every 100,000 individuals. In China, the rates are even higher, with approximately 0.4-1.5% of the population affected, which accounts for 20-25% of the thyroid disease cases in this country [
5‐
7].
Since Dailey et al. first proposed a relationship between HT and PTC in 1955 [
8], there have been many reports on this topic, with conflicting conclusions. Certain authors have agreed that there may be a positive correlation between HT and PTC [
8‐
14], whereas others have not [
15‐
19].
We performed a retrospective analysis of the relationship between HT and PTC using a large population sample from an area with a high HT prevalence.
Discussion
Both HT and PTC have a high prevalence worldwide, and these conditions may have a close relationship. Thus, there may be many patients who have both diseases simultaneously. Whether PTC represents a reactive response to HT or HT is a prerequisite tumourigenic event still remains to be determined.
In our population, 653 patients suffered from HT during the period from 2008 to 2010. Of these patients, 58.3% had co-occurring PTC, which was a higher rate than in previous reports [
8‐
19]. If the patients were subdivided by gender, the result was statistically significant for the females with HT. Compared with the females without HT, the women with HT were found to be approximately 30% more likely to have co-occurring PTC. Our findings in this patient population support the previous studies linking HT with PTC [
10,
13,
20].
At the molecular level, the evidence shows that HT may be a precursor of PTC. It is well known that RET was the first activated receptor-tyrosine kinase to be identified in papillary thyroid cancer. The proto-oncogene, located on chromosome 10q11.2, encodes a transmembrane receptor-tyrosine kinase with four cadherin-related motifs in the extracellular domain [
21]. The reported overall prevalence of RET/PTC rearrangements in papillary carcinomas varies from 3% to as high as 85%, depending on the detection methods and geographical location of the patients; a reasonable range is 13–43% [
22‐
27]. Moreover, RET/PTC rearrangements have also been reported in the thyroids of patients with the autoimmune inflammatory condition known as Hashimoto’s thyroiditis [
28‐
31]. However, these findings are controversial and might reflect either PCR contamination or the presence of microscopic nodules of papillary carcinoma, as studies that have excluded these possibilities do not confirm the data [
32,
33]. Whether RET/PTC rearrangements induce HT to PTC remains unknown. The connection of RET/PTC rearrangements in HT and PTC requires further investigation.
However, several retrospective and prospective studies have found no correlation between these two diseases. Interestingly, our univariate analysis found that HT is a risk factor for PTC, whereas HT played a protective role in the multivariate analysis, which is consistent with previous reports. It is difficult to explain this phenomenon. It has been hypothesised that serum TSH levels may play an important role. Since Boalart et al. found that elevated serum TSH can be a risk factor for PTC [
34], many studies have replicated their results [
15,
35‐
41]. Thus, when considering whether HT is a risk factor for PTC, it is important to evaluate the serum TSH level. In our study, most of the patients’ serum TSH levels were evaluated before their surgeries. As Table
2 and Table
3 show, the HT patients had higher TSH levels than the patients without HT. Moreover, the patients with co-occurring HT and PTC had the highest serum TSH levels among all of the sub-groups.
It is well known that serum TSH has a trophic effect on follicular-cell-derived thyroid cancer growth. Unfortunately, the present study could not provide evidence concerning the malignant transformation of thyroid follicular cells in the TSH-cAMP pathway. Recently, genome-wide association studies (GWAS) and carefully designed candidate gene approaches have determined that the FOXE1 genetic variant, which is downstream of the TSH-cAMP pathway, is the suspected risk factor for follicular-cell-derived thyroid cancer [
42‐
44]. In the future, genetic studies may provide stronger evidence supporting the theory that high TSH levels cause thyroid cancer. When considering HT, it should be recognised that long-term HT can cause hypothyroidism, which elevates the serum TSH level. In China, where iodine intake is increasing, there is a high prevalence of HT and hypothyroidism in subjects with euthyroidism [
7]. In our study, the rates of co-occurring PTC in the HT patients were higher than those in previous reports [
8‐
19]. Considering these results, we hypothesise that long-term HT causes elevated serum TSH, which is the real thyroid cancer risk factor. This hypothesis may explain why HT was a risk factor in the univariate analysis, whereas in the multivariate analysis, which controlled for TSH levels, HT became a protective factor. Furthermore, this hypothesis may explain why several of the prospective studies that have examined the association between HT and thyroid cancer have found negative results (i.e., because the HT patients in these studies did not have high serum TSH levels for a long period). Since TSH is a growth factor for PTC, the serum TSH could be an individual marker for each HT patient. In clinic, doctors could control the serum TSH level under a lower level in the normal range of TSH for HT patients [
45].
With regard to the clinical course, thyroid tumours associated with HT do not appear to be any more aggressive than those not associated with HT [
20,
46]. In our series of studies, we found that PTC patients with HT had less aggressive tumours than the patients with PTC only. Interestingly, the PTC patients with HT had less CLNM than those without HT (Table
3). This result conflicts with those of a previous report [
14]. In Table
5, TSH is not a factor associated with an increased risk of central lymph node metastases in patients with PTC, which is not in agreement with Kim’s study [
47]. The Kim’s study included only 554 patients, which was less than our patients. This might cause our different result. Furthermore, we take TSH as a constants variable, while Kim’s study considered TSH as a dichotomous variable.
Recently, Hanahan et al. proposed a new hallmark of cancer genesis: “Tumor-Promoting Inflammation.” [
48] This new hallmark could play an important role in the pathogenic process of co-occurring PTC and HT. In certain cases, it is evident that inflammation is present at the earliest stages of neoplastic progression and capable of fostering the development of incipient neoplasia into cancer and causing the invasion and metastasis of existing cancer [
49,
50]. It is certain that in the near future, the relationship between inflammation and cancer invasion or metastasis will be further elucidated.
Consistent with previous reports [
14,
15], our study also found that a nodule >1 cm plays a protective role in our study population (Table
4). This finding may have been due to the selection of patients for surgery. Patients with the small benign nodules are typically not sent to surgery, whereas patients with larger benign nodules, especially >4 cm, are more likely to be candidates for surgical treatment. However, in PTC patients, a primary tumour > 1 cm is a risk factor for neck metastasis (Table
5). Also, there was an interesting phenomenon that the cancer size of “PTC with benign disease” is larger than “PTC only” or “ PTC with HT”, while “PTC with benign diseas” shows less or similar invasion. This might be due to the different genetic origin of different types of PTC. The genesis of "PTC with benign nodule" might come from the benign nodule, which is less invasion [
51]. The genesis of “PTC only” might come from the somatic mutation, which is more invasion.
In the morphological examination, an extensive relationship was shown between HT and PTC (Additional file
1: Figure S1-S4). Thus, based on the histomorphological features, laboratory findings and clinical data, the co-occurrence of PTC and HT is an interesting topic concerning thyroid cancer research and cancer hallmarks.
We also have some limitations in this study. Since our study is a retrospective cross-section study, there could be a selection bias in the incidence of PTC in HT. The number of HT patients who develop nodules was not given. Next, we could not estimate the mean time for the HT patient who would suffer from PTC. It is necessary that the basic research on raised TSH causing PTC in HT patients should be done in the future, which could provide more powerful evidence for the correlation of HT and PTC.
Methods
Patients
The study subjects were all patients treated from January, 2008, to December, 2010, at the Department of Head and Neck Surgery at the Cancer Hospital of Fudan University in Shanghai, China. All of the patients provided written informed consent for their information to be stored in the hospital database and used for research, and the study was approved by the Ethical Committee of the Cancer Hospital of Fudan University. The patients included in the study met the following criteria: no previous thyroid surgery, no PTC or benign thyroid nodule confirmed by histopathology, and the availability of an adequate medical history. A total of 6109 patients were included in this study.
All patients with thyroid nodules underwent a thyroid US examination in our hospital. In our hospital, we selected patients with thyroid nodules for surgery based on the suspicious features on the ultrasound images, which were defined as micro-calcification, hypo-echoic, increased nodular vascularity, infiltrative margin and taller than wide on transverse view. A suspicious malignant nodule was diagnosed when at least one of the above images was present. Surgical operation was suggested in patients with at least one suspicious malignant nodule.
If more than one nodule were found under US, each nodule in the whole thyroid gland was evaluated by one clinical physician at the same time.
All patients that have not undergone operation were followed with serial US examinations every 3–6 months, especially in those with suspicious malignant nodule but rejected the surgery suggestion considering the relative good prognosis of PTMC and the potential risk of the operation. Surgery was indicated during period of follow-up when there was more than a 50% change in volume of suspicious malignant nodules or presence of at least one suspicious US feature mentioned above in other nodules.
In abstract, there were three criteria for our selecting patients before surgery.
1.
suspect ultrasound images, for example, microcalcifications, hypoechoic, increased nodular vascularity, infiltrative margins and taller than wide on transverse view.
2.
ultrasound images changed in the 3–6 months follow-up
This was described in our previous paper [
52]. For patients with benign nodules, surgery only was done in the nodule larger than 4 cm.
Histopathology
Resected surgical specimens were fixed in 10% formalin and cut at 5-um intervals. One to 3 representative sections of the tumour and all suspicious lesions were submitted for microscopy. All of the sections were fixed in formalin, embedded in paraffin wax, and stained with haematoxylin and eosin. HT was diagnosed by the presence of diffuse lymphoplasmacytic infiltrations with germinal centres, parenchymal atrophy with oncocytic changes, and variable amounts of stromal fibrosis throughout the thyroid gland. We examined and recorded any coexisting HT in all of the patients included in the study.
Serum thyroid function measurement
In our hospital, serum fT4, fT3, and TSH are measured by a chemiluminescent microparticle immunoassay (Abbott Laboratories, Abbott Park, IL 60064, USA) that has interassay coefficients of variations of less than 10% over the ranges 2.63–5.70 pmol/litre, 9.01–19.05 pmol/litre, and 0.35–4.94 μIU/mlitre.
Clinicopathologic variables assessed
The following variables were used in our study: gender, age at diagnosis, maximal microscopic tumour size (the largest diameter; for multifocal lesions, the maximal microscopic tumour size of the largest lesion was used, as outlined by the AJCC 7th edition), primary tumour invasion and central lymph node status. Prophylactic CLND was routinely performed when the results of the frozen section indicated a malignant nodule. Invasion of the primary tumour consisted of gross extrathyroidal extension and microscopic extrathyroidal extension, which were assessed in the final pathology. The final pathology decided the status of the lateral neck lymph nodes and if there was central lymph node metastasis.
Statistical analysis
The results are expressed as the means ± standard deviation (SD). The statistical analysis was performed using a Student’s T test or Mann–Whitney test or chi-square test, when appropriate. ORs with 95% confidence intervals (CIs) for the relationships between the variables and PTC were calculated using binary logistic regression. A p-value of less than 0.05 was considered statistically significant. The statistical analyses were performed using SPSS software, version 12.0 (SPSS, Chicago, IL, USA).
Competing interests
The authors have declared that no competing interests exist.
Authors’ contributions
LZ and HL had equally made substantial contributions to acquisition of data, analysis and interpretation of data and draft the manuscript for important intellectual content. QJ made contribution to the conception and design of this study; give final approval of the version to be published. YZ, ZW, YW, CH and QS made contributions to acquisition of data. DL and YW made contributions to interpretation of data. All authors read and approved the final manuscript.