Clinicopathological characteristics and survival outcomes in Paget disease: a SEER population-based study.

Abstract The objective of this study was to investigate the clinicopathological characteristics and survival outcomes of Paget disease (PD), Paget disease concomitant infiltrating duct carcinoma (PD‐IDC), and Paget disease concomitant intraductal carcinoma (PD‐DCIS). We identified 501,631 female patients from 2000 to 2013 in the Surveillance, Epidemiology, and End Results (SEER) database. These identified patients included patients with PD (n = 469), patients with PD‐IDC (n = 1832), and patients with PD‐DCIS (n = 1130) and infiltrating ductal carcinoma (IDC) (n = 498,076). Then, we compared the clinical characteristics of these patients with those who were diagnosed with IDC during the same period. The outcomes of these subtypes of breast carcinoma were different. Based on the overall survival, the patients with PD‐IDC had the worst prognosis (5‐year survival rate = 84.1%). The PD‐DCIS had the best prognosis (5‐year survival rate = 97.5%). Besides, among patients with Paget disease, the one who was married had a better prognosis than who were not. And, according to our research, the marital status was associated with the hormone receptor status in patients with PD‐IDC. Among three subtypes of Paget disease, patients with PD‐IDC had the worst prognosis. Besides, patients who were unmarried had worse outcomes. And the marital status of patients with PD‐IDC is associated with hormone status. The observation underscores the importance of individualized treatment.

Introduction ductal tumor and then move along the lactiferous ducts to the nipple. And the other theory suggested that the cells were in situ in the major lactiferous sinuses.
Characterized by malignant crusting or ulceration of the nipple, Paget disease can present in one of three ways. The first one is in conjunction with an underlying invasive cancer. The second one is in conjunction with underlying ductal carcinoma in situ (DCIS). The last one is alone without any underlying invasive breast carcinoma or DCIS [6]. The Paget disease can be treated by central lumpectomy with breast conservation. However, the prognosis of the PD is not well. IDC is the most common breast carcinoma subtype during the world. Recent study has suggested that patients with Paget disease conjunction with invasive cancer had worse prognosis [7]. Nevertheless, study about all these three kinds of PD is not being researched. And study on relationship between PD and the IDC is rare. Previous study described that Paget disease alone without an underlying cancer is rare, and it presents utmost 8% of patients with Paget disease [8].
Married persons enjoy overall better health and increase life expectancy compared the unmarried (divorced, separated, and never married) [9,10]. Previous studies have indicated a survival advantage for married persons living with cancer [11][12][13]. And a research found that married men and women with cancer to have a 15% reduced risk of death [14]. We compared with unmarried men and women in different subtypes of Paget disease. Besides the different outcomes in unmarried patients, we found the correlation between the marital status and the hormone status and the human epidermal growth factor receptor II, which can guide the individualized treatment in clinic.

Ethics statement
We obtained permission to access the SEER research data. The data downloaded from the SEER do not require informed patient consent. Besides, our research was approved by the Ethical Committee and Institutional Review of Fudan University Shanghai Cancer Center (FDUSCC). The methods were performed in accordance with the approved guidelines.

Data source
We examined the data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program, which contains the population-based central cancer registries of 18 geographically defined regions. For this study, we use the November 2014-18 submission.

Patient selection
We use the histopathology codes from the International Classification of Disease for Oncology third edition (ICD-O-3) to select female patients. In the ICD-O-3, the codes are defined as follows: code 8500 (ductal carcinoma), code 8540 (mammary Paget disease), code 8541 (Paget disease with infiltrating ductal carcinoma), and code 8543 (Paget disease with intraductal carcinoma). According to the ICD-O-3, we defined and choose the patients who had the PD (ICD-O-3 code 8540/3), PD-IDC (ICD-O-3 code 85413), PD-DCIS (ICD-O-3 code 8543/3), and IDC (ICD-O-3 code 8500/3). In this study, women who were diagnosed as all three kinds of PD and ICD between 2000 and 2013 were included (n = 501,631). And these identified patients included patients with PD (n = 469), patients with PD-IDC (n = 1832), and patients with PD-DCIS (n = 1130) and infiltrating ductal carcinoma (IDC) (n = 498,076).

Statistical analysis
Overall survival (OS) was measured from the date on which the first-time definite diagnosis was made until the date of death, the date last known to be alive, or September 2013. Disease-specific survival (DSS) was measured from the date of diagnosis to the date of death which is associated with breast carcinoma. The National Cancer Institute's SEER*Stat software package (version 6.1.4; built on April 13, 2005) was used to calculate incidence rates. Baseline patient demographic characteristics and tumor information were compared using the Pearson's chi-square test for categorical variables. Survival curves were plotted according to the Kaplan-Meier method and compared using the log-rank test in a univariate analysis. Cox regression analysis was performed to compute hazard ratios and 95% confidence intervals (95% CIs) and to evaluate the effects of confounding factors. All the tests were two sided, and P values less than 0.05 were considered statistically significant. All the statistical analyses were performed using SPSS statistical software, version 22.0 (IBM Corp, Armonk, NY). demographics and clinicopathological characteristics of PD, PD-IDC, and PD-DCIS were compared with IDC. And the results are summarized in Table 1. Using the Pearson's chi-square test, for PD and IDC, the significant variables were age (P < 0.001), marital status (P < 0.001), laterality (P < 0.001), tumor size (P < 0.001), lymph node status (P < 0.001), Grade (P < 0.001), AJCC stage (P < 0.001), ER (estrogen receptor) status (P < 0.001), PR (progesterone receptor) status (P < 0.001), HER2 (human epidermal growth factor receptor 2) status (P < 0.001), and whether had radiation treatment (P < 0.001). For PD-IDC and IDC, the significant characteristics were race (P = 0.011), marital status (P < 0.001), tumor size (P < 0.001), lymph node status (P < 0.001), Grade  (P < 0.001), AJCC stage (P < 0.001), ER status (P < 0.001), PR status (P < 0.001), HER2 status (P < 0.001), and whether had radiation treatment (P < 0.001). For PD-DCIS and IDC, the considerable characteristics were age (P < 0.001), marital status (P < 0.001), tumor size (P < 0.001), Grade (P < 0.001), AJCC stage (P < 0.001), ER status (P < 0.001), PR status (P < 0.001), HER2 status (P < 0.001), and whether had radiation treatment (P < 0.001). Table 2 presents the distribution of characteristics of women with breast cancer stratified by marital status.

Clinicopathological characteristics of PD
For patients with PD, the clinicopathologic characteristics were age at diagnosis (P = 0.002), race (P = 0.027), laterality (P = 0.004), tumor size (P < 0.001), lymph node status (P = 0.001) and radiation situation (P < 0.001). The hormone status did not have statistical significance. However, according to the analyses, patients who were diagnosed with PD-IDC had different statistical factors. The hormone status had statistical significance-ER status (P = 0.01), PR status (P = 0.006), and HER2 status (P = 0.025). Meanwhile, for patients with PD-DCIS, the associations were different again. Among the three   hormones, only HER2 had statistical significance (P = 0.01). Other characteristics were age (P < 0.001), race (P = 0.012), and AJCC stage (P < 0.001). Be differ from the other two subtypes, the marital status of patients with PD-DCIS had no significant correction with the radiation status.

Comparison of survival between three subtypes of Paget disease and IDC
Utilizing the Kaplan-Meier method, we analyzed all these four subtypes (PD, PD-IDC, PD-DCIS, and IDC) of mammary carcinoma. On the basis of the OS, the  I  I  I  II  II  II  III  III  III  IV  IV Figure 1. Patients with PD-DCIS had the best prognosis with a 5-year OS 83.6%. The one worse than the PD-DCIS was IDC. The 5-year OS of patients with IDC was 81.1%. Then, the next one was PD. The 5-year OS of patients with PD was 72.9%. The one with worst outcomes was PD-IDC, whose 5-year OS was 71.4%. Then, we analyzed the cases utilizing the DSS, and the comparison of different kinds of mammary cancer is shown in Figure 2. The patients with PD-DCIS had the best prognosis with a 5-year survival rate of 98.2%. The worse one was patients with PD. Its 5-year survival rate was 92.4%. The survival rate of patients with IDC was 91%. And patients who were diagnosed with PD-IDC had the worst outcomes.
Its 5-year survival rate was 84.1%. Apparently, the results of the analyses based on the OS and DSS had a little difference. Based on the OS, the results showed that the prognosis of PD was worse than IDC. However, based on the DSS, the outcome of the IDC was worse than PD. Meanwhile, the prognostic indicators can be found during the univariate analysis.

The survival analyses in subtypes of Paget disease
According to the Kaplan-Meier method and compared utilizing the log-rank test, we analyzed the Paget disease and its indicator which were associated with the prognosis. The results of the analyses are shown in Table 3. For PD, indicators which had significance were age at diagnosis (P < 0.001), marital status (P < 0.001), tumor size (P < 0.001), lymph node status (P < 0.001), and AJCC stage (P < 0.001). For PD-IDC, the significant indicators were age at diagnosis, marital status, tumor size, lymph node status, Grade, AJCC stage, and ER status. Meanwhile, the significant indicators of PD-DCIS were age at diagnosis (P < 0.001), marital status (P < 0.001), tumor size (P < 0.001), lymph node status (P < 0.001), AJCC stage (P < 0.001), HER2 status (P < 0.001), and radiation or not (P = 0.007).
Using Cox regression analysis was performed to compute hazard ratios and 95% confidence intervals. Choosing the variates which were significant in the univariate analyses, the multivariate analysis was performed. And the results are shown in Table 4

The association between Paget disease and patient's marital status
According to the Kaplan-Meier method and compared using the log-rank test, we analyzed the Paget disease and the marital status. And Figure 3 presents the correlation. For patients with PD (Fig. 3A), the married patients had the best prognosis with a 5-year OS of 85.6%. The unmarried patients (included single patients who never married, widowed, divorced, and separated patients) had worse outcomes with a 5-year OS of 65.2%. Patients whose marital status was unknown had the worst diagnosis with a 5-year OS of 48.7%. And the difference between them had statistical significance (P < 0.001). For patients who were diagnosed with PD-IDC (Fig. 3B), the married patients had the best prognosis with a 5-year OS of 78.7%. The next was patients who were unmarried with a 5-year OS of 64.1%. For this subtype, the patients whose marital status was unknown had the almost similar 5-year OS of 64.9%. And the difference was statistically significant as well (P < 0.001). For patients with PD-DCIS (Fig. 3C), the 5-year OS was 90.8% (married), 76.3% (unmarried), and 76.2% (unknown).

Discussion
Previous study had reported that patients who were diagnosed of Paget disease with underlying invasive cancer had poor tumor characteristics [15]. A previous research showed that the Paget disease with underlying invasive cancer had tumors with Grade 3 histology [8]. In 1881, Thin observed that the nipple lesion contained malignant cells which were correlated to the underlying cancer [16]. And this observation suggested the process of intraductal extension of cancer through the major lactiferous sinuses. We call it "pagetoid spread" nowadays. Histologically, Paget cells are large cells with pale, clear cytoplasm. It has enlarged nucleoli located within the epidermis and along the basal layer. The most widely accepted hypothesis to explain the origin of Paget cells is the epidermotropic theory. And this theory considered that Paget cells are derived from an underlying mammary adenocarcinoma [17]. Evidence supporting the epidermotropic theory is based on studies showing that Paget disease is associated with an underlying breast carcinoma in most patients [5,[18][19][20]. Binding of heregulin to its receptor on Paget cells can induce chemotaxis of these breast cancer cells, and the cells eventually migrate into the overlying nipple epidermis [21]. It is noteworthy that Paget cells and the underlying associated ductal carcinoma share the same immunohistochemical profile [22] and the same patterns of gene expression.
In allusion to different subtype of Paget disease, we found that the significantly associated indicators were different. Unmarried patients of PD, including those who were widowed, divorced, and never married, were at significantly great risk of existing lymph node metastasis. Meanwhile, for patients of PD-IDC, we found that the hormone status was related to the human epidermal growth factor receptor II. However, for the patients with PD-DCIS, only human epidermal growth factor receptor II had statistical significance. The association between marital status and these indicators was significant for every malignancy evaluated. Previous studies have linked marriage to improvements in cardiovascular, endocrine, and immune function, and marriage may be a determinant of the magnitude and presence of this effect [23,24]. Cortisol levels seem to be lower in patients with cancer who have adequate support networks, and diurnal cortisol patterns have been linked with natural killer cell count and survival in patients with cancer [25,26], potentially providing a physiologic basis for the psychologically based data described previously [27]. Further investigations on this subject are warranted.
However, the study also had some limitations. The SEER database did not give us enough information about the lymphovascular invasion which can be regarded as the prediction of lymph node metastasis. Besides, the followup of many patients was limited. And the information of systemic therapy of the patients was lack according to the SEER system. Based on the SEER database, the HER2 status was tested from 2010; however, the cases were from 2000 to 2013. Apparently, analyses of the HER2 were limited. And it made us unable to explore the clinical significance of HER2 status. Therefore, our study was limited by lack of some information. Besides, there is potential for misclassification of marital status. We did not take into account changes of marital status which may have occurred during the follow-up period. And this phenomenon may have influenced our results. Thus, our findings may underestimate the protective effect that marriage has on breast cancer outcome. We defined that the single category contained divorcees, widows, and never married women. However, previous studies had found that there may be some difference among groups of unmarried women. Although the difference existed, the unmarried women fare worse than the married counterparts.
In conclusion, our study showed patients with PD-IDC have the worst prognosis. Among all these three kinds of Paget disease, unmarried patients had worse outcomes. And the marital status of patients with PD-IDC is associated with hormone status and HER2 status. The observation underscores the importance of individualized treatment.