Clinicopathological and Survival Outcomes of Well-Differentiated Thyroid Carcinoma Undergoing Dedifferentiation: A Retrospective Study from FUSCC

Background Recently, several studies have reported that dedifferentiation occurs in fatal well-differentiated thyroid cancer (WDTC) cases. This study aimed at investigating the clinicopathological characteristics of WDTC undergoing dedifferentiation. Methods A total of 63 WDTC patients harboring dedifferentiated phenotype were enrolled in the study. The Kaplan-Meier method and Cox regression analysis were used to perform survival analyses. Harrell index of concordance (C-index) and Akaike information criterion (AIC) were calculated to compare the predictive value for prognosis among several prognostic classification systems. Results The median cause-specific survival (CSS) of patients was 138 months, with the CSS rate of 64.0% and 53.3% at 5 and 10 years, respectively. Presence of the anaplastic thyroid cancer (ATC) phenotype significantly increased the risk of poor CSS (P = 0.033), and age was the only independent risk factor for disease progression (P = 0.015). The C-index and AIC of the age, grade, extent, size (AGES) prognostic classification system for the CSS were 0.723 and 59.937, respectively. Conclusions The presence of dedifferentiated phenotypes can be responsible for the poor outcomes in WDTC patients. The AGES system demonstrates to be an optimal prognostic system for WDTC undergoing dedifferentiation.


Introduction
In the United States, thyroid cancer has become the fifth most common cancer expected to occur in women in 2017 [1]. Among women before the age of 30 years, thyroid cancer is the most commonly diagnosed cancer in China [2]. Almost the entire change has been attributed to the rapid increase in incidence of well-differentiated thyroid cancer (WDTC), which comprises the vast majority (>90%) of all thyroid cancers [1,3,4], including papillary thyroid cancer (PTC) and follicular thyroid cancer (FTC). Though WDTCs usually have good prognoses, some can still develop into aggressive diseases. Recently, several studies have reported presence of poorly differentiated phenotypes in fatal WDTC cases [5][6][7].
Anaplastic thyroid cancer (ATC) is one of the most aggressive malignancies with a median survival of 5 months and near 100% mortality rate [8] while poorly differentiated thyroid cancer (PDTC) is associated with a 5-year survival rate of 40-70% [5,9,10]. ATC and PDTC represent a major clinical challenge because they are refractory to radioiodine therapy and insensitive to traditional chemotherapy and radiotherapy. It has been generally accepted that ATC and PDTC usually develop from the dedifferentiation of differentiated thyroid cancer (DTC), which is supported by that histological concurrence of PDTC/ATC and DTC, and molecular inheritance from DTC to PDTC/ATC has been confirmed in various studies [11][12][13][14][15][16].
The clinicopathological features of WDTC patients undergoing dedifferentiation are poorly characterized. In the past decade, we have found a cohort of WDTC patients harboring PDTC/ATC at Fudan University Shanghai Cancer Center (FUSCC). Therefore, the present study aimed at investigating the clinicopathological and survival outcomes of WDTC undergoing dedifferentiation.

Methods
Each patient provided a written informed consent for his/her information to be used for research and stored in the hospital database. All the procedures performed in our study were in accordance with the ethical standards of the FUSCC research committee and the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
2.1. Patients. We retrospectively searched for the patients with histopathological diagnosis of WDTC harboring dedifferentiated components, who received surgeries or pathology consultation at FUSCC from January 2004 to December 2016. In the present study, dedifferentiated components were defined as presence of poorly differentiated follicular cell-derived carcinoma or anaplastic carcinoma. The patients included in our study met the following criteria: (1) histopathological confirmation of presence of WDTC, (2) harboring dedifferentiated components at primary site or metastatic sites, and (3) availability of adequate medical records and follow-up outcomes.

Histopathological
Review. All hematoxylin and eosin slides of the selected patients were subjected to evaluation by expert pathologists. Tumor size was measured as the maximum dimension of the resected tumor specimen. Extrathyroidal extension (ETE) was divided into two categories: absence of ETE and presence of microscopic or gross ETE. The histological types of tumor were classified according to World Health Organization criteria (2004). PDTC was defined on the basis of the Turin proposal [17]: (1) presence of a solid/trabecular/insular pattern of growth, (2) absence of the conventional nuclear features of papillary carcinoma, and (3) presence of at least one of the following features: convoluted nuclei, mitotic activity ≥ 3 × 10 HPF, and tumor necrosis. In the present study, WDTC undergoing dedifferentiation was defined as primary WDTC with dedifferentiated components at primary or metastatic sites.  stage (the 7th edition) as previously described [18,19]. The 8th edition of TNM thyroid cancer staging system published by the American Joint Committee on Cancer (AJCC) [20] was also added in our study to be compared with the above staging systems in a prognostic value. The CSS was defined as the time between the date of carcinoma diagnosis by operation and the date of death from cancer. The disease-free survival (DFS) was considered as the duration from the date of carcinoma diagnosis by operation to the date of disease recurrence or progression or death. Patients with no event were censored on the date of the last follow-up.

Basic Characteristics of Dedifferentiated Patients.
We retrospectively collected and analyzed the records of consecutive WDTC patients with dedifferentiated phenotype confirmed by histopathology who received surgeries or histopathology consultation at FUSCC. The flow graph of inclusion and analysis of study subjects in our study was shown in Figure 1. As shown in Table 1

Comparison of Different Prognostic Classification Systems in Dedifferentiated
Cases. The analyses of evaluation for several prognostic systems including EORTC, AGES, AMES, MACIS, the 7th edition TNM, and the 8th edition TNM were performed to determine whether any of them had a better predictive effect of CSS in dedifferentiated patients. Table 4 showed the CSS rate of each group of the prognostic systems at 5 years and 10 years and C-index and AIC values for each system. The AGES demonstrated to be a better survival pre  Figure 4 showed the Kaplan-Meier curves of CSS according to the above prognostic systems.

Association of Dedifferentiated Features with
Clinicopathological Factors. The further analysis of possible factors affecting dedifferentiation components indicated that age ≥ 55 significantly correlated with the presence of the ATC phenotype (Table 5). Additionally, we found that five cases of papillary thyroid microcarcinoma (PTMC) (7.9%) presented dedifferentiation in the thyroid (2/5, 40.0%) and lymph node (3/5, 60.0%), of which 2 cases (40.0%) were observed to be fatal (Table 6).

Discussion
The present study revealed the dedifferentiation characteristics, the risk factors associated with CSS and DFS, and the optimal prognostic classification system for the 63 WDTC patients harboring dedifferentiated phenotype from FUSCC. The 63 patients were retrospectively selected from a large cohort of thyroid cancer patients at the department of head and neck surgery from FUSCC. To our knowledge, this is the first study that has reported the clinicopathological features and survival outcomes of WDTC harboring dedifferentiated phenotype, which allows us to explain aggressive behaviors and fatal outcomes of some WDTCs in part from the perspective of histological phenotypes. Consistent with the findings in the previous studies [5,6,13], our study confirms that dedifferentiation occurs in both primary tumor and metastatic regions in the histological forms of PDTC and ATC. Considering the significant correlations of age with ATC phenotype and metastatic tumor dedifferentiation, age demonstrates to be a risk factor for aggressive dedifferentiation in this study. And as a result, age is confirmed to be a risk factor for poor DFS in multivariate analysis and decreased CSS in univariate analysis in WDTC patients harboring dedifferentiated phenotype, which is supported by the fact that age is an independent prognostic factor for DTC [20]. The presence of ATC phenotype is an independent risk factor for poor CSS, suggesting that ATC phenotype plays a critical role in determining the patients' prognoses although there is no statistical significance (P = 0 137) in the analysis of association between the ATC phenotype and DFS, with the significant effect size (HR = 1.941). In addition, both dedifferentiation of metastatic sites and DM were risk factors for poor CSS and DFS in univariate analysis but could not be statistically significant in multivariate analysis. It is interesting to find that adjuvant therapy after surgery significantly correlates with decreased CSS, which may be explained by insensitivity and increased complications of therapy. Other clinicopathological factors including gender, tumor size, ETE, and histological subtypes of WDTC and LNM are not significantly associated with CSS and DFS in our study.
Due to the widespread use of high-resolution ultrasound and ultrasound-guided fine-needle aspiration, increasing numbers of PTMC patients are diagnosed [21]. Although PTMC usually presents an indolent course, some PTMC patients can still develop to fatal outcomes [7,22]. In the present study, we found five PTMC cases with dedifferentiation, and 2 patients of them died of  [18,20,23]. Considering the difference between WDTC harboring dedifferentiated phenotype and DTC, we further analyzed the predictive value of several common classification systems to provide an optimal prognostic system for WDTC patients with dedifferentiated phenotype. As expected, the 8th edition TNM stage shows better predictive value than the 7th edition TNM stage. The AGES system demonstrates to be an optimal system in comparison with the 8th edition TNM stage, EORTC, AMES, and MACIS, which can attribute to the inclusion of histological grade score in the AGES system.
Due to the limited number of WDTC patients harboring dedifferentiated phenotype included in our study, the analyses of clinicopathological outcomes and survival can have statistical bias. Therefore, a large multicenter cohort with long follow-up will be necessary to demonstrate the above outcomes in the future study. Furthermore, the further investigation of molecular biology will be significant to get insights into the mechanism of dedifferentiation of WDTC.

Conclusions
The presence of dedifferentiated phenotypes can be responsible for the poor outcomes in WDTC patients. The AGES system demonstrates to be an optimal prognostic system for WDTC undergoing dedifferentiation.