Clinicopathological study of organ metastasis in endometrial cancer

Endometrial cancer (EC) is the most common carcinoma of the female reproductive tract in developed countries, and its incidence rate shows a significant upward trend [1]. In the USA, approximately 63,000 cases were newly diagnosed in 2018, making it the fourth most common cause of cancer and the seventh most common cause of death in women [2]. EC is a kind of epithelial malignant tumor which occurs in the endometrium [3]. According to its clinical features and biological behavior characteristics, the pathological types can be divided into endometrioid carcinoma and nonendometrioid carcinoma [4].

Surgery, chemotherapy and radiotherapy are the main treatments for patients with EC [5,6]. The 2009 International Federation of Gynecology and Obstetrics (FIGO) and the tumor-node-metastasis classifications are the most commonly used classifications based on surgical staging [7]. Because most patients with EC are in the initial diagnosis stage, the tumor is still limited to the uterine body, surgical treatment has a good prognosis [7]. Therefore, surgery is the preferred treatment of EC patients [8]. For early stage (stage I and II) EC patients, open surgery, traditional laparoscopy, robot and other surgical methods can be used to completely remove the lesions [9–11]. For advanced stage (stage III and IV) EC patients, surgical removal of cancer lesions is aimed at no residual lesion, postoperative combined radiotherapy and chemotherapy can improve the survival rate of patients [12]. EC patients (15–25%) are advanced at the time of diagnosis, and many clinical and pathological factors affect the prognosis and survival time of EC patients, among which site of metastasis is one of the most important factors affecting survival time [2].

Lymph node metastasis is the main metastatic pathway of EC. The most common metastasis site of EC are pelvic lymph nodes, such as common iliac, internal iliac and external iliac lymph nodes [13], followed by retroperitoneal lymph nodes. Distant metastasis such as lung and brain metastasis is rare [14]. In our study, we used data from the Surveillance, Epidemiology and End Results (SEER) cancer registration database to analyze organ metastasis in EC patients diagnosed from 2010 to 2015.

Patients & methods

Database

The SEER database is currently the largest publicly available cancer database covering approximately 28% of the US population [15]. The SEER database includes demographic information and cancer characteristics, such as diagnosis age, year of diagnosis, race, marital status, insurance status, income status, primary tumor location, tumor grade and stage, histological type, tumor-node-metastasis stage, treatment modality and survival time. In addition, the SEER database began publishing metastasis information related to liver, lung, bone and brain in 2010. The National Cancer Institute’s SEER*Stat software (version 8.3.5; SEER 18 Regs Custom Data (with additional treatment field), November 2017 Sub (1973–2015 varying) database) was used in our study, we identified all 74,573 patients diagnosed with EC between 1 January 2010 and 31 December 2015.

Exclusion criteria in our study were as follows: patients under 18 years of age (n = 6); unknown survival time (n = 100); unknown bone metastasis (n = 1,568); unknown brain metastasis (n = 28); unknown liver metastasis (n = 34); unknown lung metastasis (n = 66); unknown metastasis site (n = 128); unknown insurance status (n = 1,286); unknown income (n = 3); other ICD-O-3 histology codes (8000/3, 8010/3, 8013/3, 8020/3, 8041/3, 8045/3, 8050/3, 8071/3, 8072/3, 8082/3, 8246/3, 8313/3, 8370/3, 8510/3, 8574/3, 8575/3, n = 2,327) [16]. Finally, we left 69,027 eligible patients diagnosed with EC.

Study variables

Variable definition information about diagnostic age, year of diagnosis, race, marital status, insurance status, income status, tumor grade, histological type, lymph node involvement and survival time can be found in the SEER database. Overall survival (OS) and cancer-specific survival (CSS) were the primary study end points. Three-year OS, 5-year OS, 3-year CSS and 5-year CSS were also other interest outcomes in our study.

For the insurance status, patients were divided into three group: ‘Insured group’, ‘medicaid group’ and ‘uninsured group’. The ‘insured’ and ‘insured/no specifics’ were defined as ‘insured group’, ‘any medicaid’ was defined as ‘medicaid group’, and ‘uninsured’ was clustered as ‘uninsured group’.

For the marital status, patients were divided into ‘married group’, ‘unmarried group’ and ‘unknown marital status group’. Unmarried patients include ‘single’, ‘separated’, ‘divorced’ and ‘widowed’.

The ICD-0-3 site/histology validation list was used to distinguish endometrioid histology (endometrioid/adenocarcinoma/mucinous, 8380–8383/3, 8140/3, 8210/3, 8211/3, 8560/3, 8260/3, 8262/3, 8263/3, 8570/3, 8261/3, 8480–8482/3 [17]; serous, 8441/3, 8460/3, 8461/3; carcinosarcoma, 8950/3, 8951/3, 8980/3, 8981/3; clear cell, 8310/3, and mixed epithelial, 8323/3, 8255/3).

Grade was defined by the following code; well differentiated (Grade I); moderately differentiated (Grade II); poorly differentiated (Grade III); undifferentiated (Grade IV) and unknown grade.

Statistical analysis

χ² analysis was performed to evaluate clinical characteristics with organ metastasis. Kaplan–Meier curve was used to estimate the OS and CSS in different groups, and the difference between the curves was analyzed by log-rank test. Univariate and multivariate Cox regression models were performed to estimate the hazard ratios (HR) and 95% CI to analyze independent prognostic factors. Statistical Package for the Social Sciences software (version 20.0; SPSS, IL, USA) was used for all statistical analyses. A p ≤ 0.05 was considered statistically significant.

Results

Demographic & clinical characteristics

A total of 69,027 EC patients who were eligible for our inclusion and exclusion were included in our study cohort from 1 January 2010 to 31 December 2015. Among them, there were 1048 (1.5%) patients with lung metastasis, 541 (0.8%) patients with liver metastasis, 388 (0.6%) patients with bone metastasis and 105 (0.2%) patients with brain metastasis. Demographic and clinical characteristics of EC patients with different organ metastasis were shown in Table 1.

Lung metastasis

Among the four types of organ metastasis, lung metastasis had the highest number and proportion of patients (1048, 1.5%). Over time, the number of lung metastasis increased from 284 (1.3% in 2010–2011) to 394 (1.6% in 2014–2015; p = 0.037). For race, non-Hispanic black patients had the highest rate of lung metastasis (188, 2.9%). Compared with unmarried patients, married patients had a lower rate of lung metastasis (1.3%). Insured and medicaid patients had lower percentage of lung metastasis, while uninsured had highest lung metastasis rate (2.6%). In terms of income, high income corresponds to low metastasis rates. For the tumor grade, undifferentiated tumors had the highest rate of lung metastasis (3.7%). Among the histological types, the lung metastasis rate of carcinosarcoma was significantly higher (4.5%; p < 0.001) than other histological types.

Liver metastasis

Over time, the number of liver metastasis changed from 163 (0.8% in 2010–2011) to 202 (0.8% in 2014–2015), but there was no statistical significance. Patients with liver metastasis were significantly younger than those without liver metastasis. Non-Hispanic whites and Hispanics had significantly lower rates of liver metastasis than non-Hispanic blacks (p < 0.001). The rate of liver metastasis in unmarried or uninsured patients was higher than that in married or insured patients (All p < 0.001). In addition, the liver metastasis rate of poorly differentiated and undifferentiated tumors was higher, while that of well differentiated tumors was lower.

Bone metastasis

The characteristics of patients with bone metastasis were similar to those of liver metastasis. Bone metastasis are predisposed to elderly, non-Hispanic black, unmarried, uninsured, clear cell carcinoma type and undifferentiated patients (All p < 0.001).

Brain metastasis

Among the four types of organ metastasis, the number and proportion of patients with brain metastasis were the lowest (132, 0.2%). There was no significant difference in year of diagnosis, race, marital status and income between patients with and without brain metastasis (All p > 0.05). Similarly, brain metastasis was more likely to occur in elderly, uninsured, carcinosarcoma histological type and undifferentiated patients.

Combination of metastasis patterns

In our study, we found that some patients had multiple site metastasis. Table 2 summarizes all possible combinations of these four metastasis sites. Among the single organ metastasis, the proportion of lung metastasis was the highest (681, 0.98%), followed by liver metastasis (297, 0.43%), bone metastasis (170, 0.25%) and finally brain metastasis (41, 0.06%). The most common combination of two-site metastasis was lung and liver (150, 0.22%). Of the three sites of metastasis, lung, liver and bone were the most common (49, 0.07%). Only 8 (0.01%) patients had all four sites metastasis.

In addition, we compared the risk of bone and brain metastases between EC patients with or without lung or liver metastases. We found that patients with lung metastases had a significantly higher risk of bone metastasis (17.6 vs 0.3%) or brain metastasis (15.7 vs 0.4%; Figure 1). A similar phenomenon was also found in patients with liver metastasis.

Survival

In our study, Kaplan–Meier curve was used to analyze the influence of related factors on OS and CSS of patients with EC (Table 3). Kaplan–Meier analysis showed that age at diagnosis, race, marital status, insurance status, income, tumor grade and organ metastasis were also significantly correlated with OS and CSS. The 5-year OS was 11 versus 79% (p < 0.001) and 5-year CSS was 18 versus 88% (p < 0.001) for patients with or without lung metastasis (Figure 2A and Supplementary Figure 1A), 8 versus 78% (p < 0.001) and 18 versus 87% (p < 0.001) for patients with or without liver metastasis (Figure 2B and Supplementary Figure 1B), 9 versus 78% (p < 0.001) and 16 versus 87% (p < 0.001) for patients with or without bone metastasis (Figure 2C and Supplementary Figure 1C), 10 versus 78% (p < 0.001) and 18 versus 87% (p < 0.001) for patients with or without brain metastasis (Figure 2D and Supplementary Figure 1D), respectively.

Table 2 shows the median OS, median CSS, 5-year OS and CSS for patients with different metastasis combinations. In patients with single organ metastasis, the median OS for patients with lung, liver, bone and brain metastasis was 11, 10, 8 and 5 months, respectively. Similarly, patients with lung, liver, bone and brain metastasis had median CSS at 14, 15, 11 and 8 months, respectively. Patients with brain metastasis alone had the worst OS and CSS (Figure 3A & B).

In the two-site metastasis, the combination of lung and bone metastasis had the longest median OS (9 months), 5-year OS (18%) and 5-year CSS (14%; Figure 3C & D). The combination of liver and brain metastasis had the shortest median OS (1 months), median CSS (1 months), 3-year OS (0%) and 3-year CSS (0%). In multi-sites metastasis, the survival time of the patients was very short, with both 3-year OS and CSS 0% (Figure 3E & F). In addition, OS and CSS decreased with the increase of metastatic sites (Figure 4A & B).

Identification of prognostic factors of OS in EC patients

Univariate and multivariate cox regression were used to analyze the factors associated with OS and CSS in EC patients. As shown in Table 4 & 5, age at diagnosis, race, marital status, insurance status, income, histological type, grad and whether metastasis are factors that affect the OS and CSS in patients with EC. Multivariate cox regression showed no lung metastasis (vs lung metastasis; HR = 0.21, 95% CI: 0.19–0.23; p < 0.001), no liver metastasis (vs liver metastasis; HR = 0.31; 95% CI: 0.28–0.35; p < 0.001), no bone metastasis (vs bone metastasis; HR = 0.41; 95% CI: 0.36–0.47; p < 0.001); no brain metastasis (vs brain metastasis; HR = 0.36; 95% CI: 0.29–0.45; p < 0.001) were associated with better OS (Figure 5). Similarly, in terms of CSS, multivariate Cox regression analysis also indicated whether organ metastasis was a prognostic factor for EC patients (no lung metastasis vs lung metastasis, HR = 0.18; 95% CI: 0.17–0.21; p < 0.001; no liver metastasis vs liver metastasis, HR = 0.32; 95% CI: 0.28–0.37; p < 0.001; no bone metastasis vs bone metastasis, HR = 0.38; 95% CI: 0.32–0.44; p < 0.001; no brain metastasis vs brain metastasis, HR = 0.36; 95% CI 0.28–0.47; p < 0.001; Supplementary Figure 2).

Discussion

Our study reported the different organ metastasis of EC patients through a large population-based case study. Our study showed that in patients with EC, distant organ metastasis was more common in lung metastasis (1.5%), followed by liver metastasis (0.8%) and bone metastasis (0.6%), and brain metastasis was rare (only 0.2%). In the single organ metastasis, lung metastasis has the longest survival time, and brain metastasis has the shortest survival time. In two-site metastasis, the combination of lung and bone metastasis had the longest median OS (9 months), and the combination of liver and brain metastasis had the shortest median OS (1 month) and median CSS (1 month). In multi-site metastasis, the survival time of the patients was very short, with both 3-year OS and CSS 0%. In addition, OS and CSS decreased with the increase of metastasis site.

EC is one of the common malignant tumors in gynecology. The modes of metastasis in patients with EC include intra-abdominal spread, lymph node involvement and distant organ metastasis [18,19]. Distant organ metastasis (lung, liver, brain or bone) is very rare, but it has a great impact on the prognosis and survival of EC patients. There are few studies on distant metastasis in patients with EC, and most of them focus on single metastasis site or specific stage. Liu et al. [20] through the study of 2948 patients with stage IV EC from 2010 to 2014, found that the most common metastasis site of stage IV EC was the lung, while the survival rate of patients with brain metastasis was the worst. Ignatov et al. [21] described the recurrence pattern of EC according to the risk characteristics through a 7-year follow-up of 2177 EC patients. It was found that patients with different EC subtypes had specific recurrence patterns and recurrence time, and high-risk patients needed shorter follow-up time.

Data from current analysis showed that lung metastasis was the most common distant organ metastasis, while brain metastasis is the rarest distant organ metastasis, which is consistent with previous retrospective studies [22–24]. In addition, the analysis of our study was performed on a larger sample size and provided information on the proportion and prognosis time of different metastasis sites and combinations of different metastasis sites. We also found that the pattern of distant metastasis appears to be influenced by the histological type of patients. The histological type of carcinosarcoma had significantly higher rate of lung, liver and brain metastasis, while clear cell type was more likely to have bone metastasis. EC patients with lung or liver metastasis were at higher risk of bone and brain metastases than those without lung and liver metastasis. This is consistent with previous research on other types of cancer [25,26].

Some researchers believe that the combination of liver metastasis and brain metastasis is unusual, because the lungs are often the first filter, and the occurrence of brain metastasis is usually associated with lung metastasis. Through the investigation of 45 patients with brain metastasis of colorectal cancer, Sun et al. [27] found that 19 patients were not complicated with lung metastasis. Our study also found that not all patients with brain metastasis were associated with lung metastasis, which may be due to the metastasis of tumor cells into the brain through the prevertebral venous plexus.

Our analysis found that the percentage of lung, liver, bone and brain metastasis in uninsured patients was significantly higher than that in insured patients. This may be because insured EC patients have more opportunities to receive surgical treatment for a better prognosis [28]. In addition, in terms of race, non-Hispanic black patients are more likely to have metastasis than Hispanic-white patients, which was consistent with previous study [29]. Moreover, married EC patients are less likely to have metastasis, and married patients have a lower risk of OS and CSS, perhaps because spouses can provide social support and encourage patients to seek medical help [30,31].

In our study, we found not only the difference in prognosis of single organ metastasis, but also the difference between single organ metastasis and multiple organ metastasis, which were similar to the data from other solid tumors [32,33]. Abdel-Rahman [34] found that liver metastasis had the worst survival time in renal cell carcinoma single organ metastasis. Qiu et al. [35] found that in gastric cancer, lung and bone metastasis had the longest median OS (9 months) in two-site metastasis, while liver and brain metastasis had the shortest median OS (1 months) and 5-year CSS (0%).

There are limitations to be recognized in this study. First, this study is a retrospective study with obvious limitations. Second, there is no FIGO grading system for EC patients in the SEER database, only the Grade classification, which still requires for conversion. In addition, the SEER database began to release metastasis information related to liver, lung, bone and brain in 2010 and the follow-up period was not long enough. We can only obtain information about organ metastasis, and the sequence of metastasis information is not mentioned in the SEER database. Moreover, the lack of information on the physical condition and complications of patients, both of which are prognostic factors for metastasis in patients with EC.

Information about the impact of distant organ metastasis on the prognosis of patients with EC will be helpful for the proper stratification of advanced patients and for clinicians to determine appropriate follow-up strategies, and guide clinicians to design personalized pretreatment for EC patients.

Conclusion

In our study, we found that organ metastasis is an independent prognostic factor for survival in EC patients. In single organ metastasis, lung metastasis has the best survival time. Assessing distant organ metastasis and metastasis pattern can help clinicians design personalized treatment for patients with advanced EC.