Bone metastasis in esophageal adenocarcinoma and squamous cell carcinoma: a SEER-based study

Esophageal cancer is a common worldwide disease with a higher mortality rate. Studies on esophageal cancer patients with bone metastasis are rare. Our study focused on the clinicopathological features of patients with bone metastasis using the Surveillance, Epidemiology and End Results (SEER) database to further explore the risk factors and survival for bone metastasis. Esophageal cancer patients with bone metastasis were extracted from the SEER database. Univariable analysis and multivariable logistic regression were used to study the risk factors for bone metastasis. Univariable analysis and multivariable Cox regression were performed to reveal the survival and prognostic factors for bone metastasis. The competitive risk model was made to compare the association with bone metastasis among different causes of death. Propensity score matching was used to reduce the bias. Male, middle esophagus, with brain metastasis, without lung metastasis and without liver metastasis were major independent risk factors of bone metastasis. Older age, poorly differentiated and undifferentiated, with brain metastasis and with liver metastasis were major independent prognostic factors of bone metastasis. Patients with bone metastasis had a worse prognosis before and after propensity score matching than patients with other metastasis. Esophageal cancer patients with male sex, middle esophagus and brain metastasis were more likely to have bone metastasis. Compared to patients with other metastatic sites such as liver, brain and lung, patients with bone metastasis had a worse prognosis. Our findings provide recommendations about clinical guidelines for esophageal cancer patients with bone metastasis.


Introduction
Esophageal cancer (EC) is the seventh most common cancer and the sixth leading cause of death in the world [1]. A study based on Cancer Incidence in Five Continents (CI5) shown that the highest incidence rates of EC were in Malawi, South Africa, and Iran and the highest mortality rates were in South Africa and Kazakhstan [2]. A higher incidence of EC was found in males and common histological types included adenocarcinoma (EAC) and squamous cell carcinoma (ESCC). The incidence of ESCC declined in western countries due to reducing the prevalence of smoking and drinking [3][4][5]. The incidence of EAC increased. Gastroesophageal reflux disease (GERD), H. pylori infection and obesity were main known risk factors for EAC [6]. Although the overall incidence of EAC and ESCC displayed Ya Qin, Jiannan Mao and Xiao Liang have contributed equally to this work. 1 3 a downward trend in many studies, EC imposed a tremendous public health burden globally.
EC is a tumor that is prone to distant metastases. Over fifty percent of patients have unresectable or metastatic disease at initial diagnosis [7]. At this stage, there were very few treatment choices which can only bring dismal survival advantages for patients [8]. The overall five-year survival rate for patients with distant metastases was only 5% [9]. Another study even indicated the median survival time of EC patients with metastatic disease was only 6 months and 2-year survival rate was only 11.8% [10]. Bone is a common site of metastases in malignant tumors [11][12][13]. Bone metastasis (BM) is the third most common site in patients with EC [14,15]. Compared to other metastatic sites such as liver, brain and lung, BM had worse overall survival (OS) in EC [16]. Few articles systematically analyzed BM in EC, so it requires a deeper analysis.
Our study aimed to reveal the clinical characteristics and the prognosis of EC with BM based on SEER database. SEER database registers four main metastatic organs of tumors: liver, lung, bone and brain. It is a population-based cancer registry covering about 30% of the US population, bringing much reliable information [17].

Patients selected
Information was collected from 18 population-based cancer registries with additional treatment fields in SEER database (version 8.3.8). A total of 28,243 patients with EC from 2010 to 2016 were identified. And 5175 patients with metastatic cancer were selected according to the inclusion and exclusion criteria. Inclusion criteria: patients diagnosed with EC between 2010 and 2015; histological types were EAC and ESCC; IV stage EC; first malignant primary cancer. Exclusion criteria: patients underwent surgery, survival time was unknown, not stage metastatic EC or unknown stage EC; EC patients diagnosed in 2016. The process was presented as a flow process chart in Fig. 1. For at least 1-year follow-up, we selected before 2016.

Data collection
Histological types were confirmed by the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3). EAC included 8140-8389 and ESCC included 8050-8089. EC patients were divided into five groups: upper esophagus including upper third and cervical, middle esophagus including middle third, lower esophagus including lower third and abdominal, overlapping and NOS including NOS and thoracic. Other clinical features were also collected including age, sex, race, organs of metastasis and survival time.

Statistical methods
The risk factors were investigated by univariable and multivariable analysis. Chi-square test was used for univariable analysis. Multivariable logistic regression analysis further confirmed the risk factors for EC patients with BM. Survival estimates were performed by the Kaplan-Meier method and compared by the log-rank test. We brought factors which were statistically significant into multivariable analysis. And to avoid missing important factors, we expanded P-value to 0.1. Cox regression model was used for multivariable analysis. The competitive risk model was made by R package cmprsk. Propensity score matching (PSM) was used by R package MatchIt. Statistical analyses were performed using SPSS 25.0 and R.3.6.3. Pictures in our study were drawn by GraphPad Prism 7.0, and the distribution of metastatic sites was drawn by Oliveros, J.C. (2007-2015) Venny (https:// bioin fogp. cnb. csic. es/ tools/ venny/ index. html). Statistical significance was set at two-side P < 0.05.

Epidemiological trends
The incidence of EC declined from 2010 to 2016 (Fig. 2a). For both male and female, the incidence showed a similar trend (Fig. 2b). Among the histological types, the incidence of ESCC declined, but the incidence of EAC still rose (Fig. 2c). For all EC patients with BM from 2010 to 2015, prevalence was 7.11% (1503/21128). It peaked in patients between 31 and 40 and then displayed a downward trend. The prevalence of BM in metastatic esophageal cancer (mEC) was 23.5% (1503/6395). It rose with age and held steady in patients between 31 and 70 (Fig. 2d).

Risk factors for BM
Univariable analysis (Table 1) showed that there were six factors with statistical significance (P value < 0.05): age, sex, primary tumor location, brain metastasis, lung metastasis and liver metastasis. BM was more common among 51-60 years old. Male patients were more likely to have BM. For tumor location, middle esophagus had higher percentage of BM. Patients with brain metastasis had a higher risk of having BM. However, EC patients with lung or liver metastasis were less likely to develop BM.

Survival and prognostic factors for BM
Univariable analysis showed that there were eight factors including age, race, tumor location, histology, grade, brain metastasis, lung metastasis and liver metastasis that were statistically significant with OS. Then we put the above eight factors into the Cox regression model. The results confirmed that only four factors were significantly associated with OS: age, tumor grade, brain metastasis and liver metastasis (Table 3).
The cumulative incidence function curve (Fig. 7) showed that compared to EC patients without BM, the cause-specific mortality was higher in BM. And the mortality rates of other causes were not statistically significant between the two groups.

Discussion
Metastasis is one of the most important characteristics of solid tumors. Metastasis often means that malignant tumors are advanced and incurable. BM is relatively common in solid tumors, and it brings a series of skeletal-related events (SREs): pain, pathological fractures, nerve compression syndromes and hypercalcemia [18,19].
In our study, the metastatic rate to bone from EC patients at the time of diagnosis was 5.77%. However, the actual incidence of EC patients with BM may be underestimated. Because bone imaging was not a routine imaging test, methods of examination for BM including bone scan, MRI, PET-CT and so on, which were relatively expensive. We do not recommend all patients with EC undergo these tests. Therefore, it is necessary to determine whether the EC patients have risk factors for BM. Our study showed that patients with brain metastasis were more likely to have BM. However, patients with age 71-80 years, female, lower esophagus, with lung metastasis, with liver metastasis, were relatively less likely to have BM. During development and bone remodeling that occurs in the adult, osteoclasts and osteoblasts regulate bone modeling [20]. Malignancies may break the balance of bone physiology, and which will result in an environment that promotes metastasis [13]. BM include osteolytic metastasis and osteoblastic metastasis [19]. Osteolytic metastasis is the most common form of BM, and is associated with increased osteoclast activity and reduced osteoblast activity. In contrast, osteoblastic metastasis is rare, and was reported in prostate, breast, colon, cervical cancer and so on [18]. Imual et al. reported that in their study, multiple osteolytic BM commonly occurred in the axial skeleton in EC patients with BM, and nearly 91.4% patients suffered SREs [21]. In previous studies of EC, Ai et al. showed that in EC patients, there was no significant difference in OS among the four sites of distant metastases (liver, lung, bone and brain) [15]. Tanaka et al. also reported that there was no difference in the median survival among the three metastatic sites (liver, lung and bone) [22]. However, our study showed that in EC patients, BM was associated with poorer OS (4 vs 6 months, respectively), which was consistent with a previous study [16].Why is the OS of BM shorter than that of other metastatic sites? The mechanism is not yet clear. We think that BM is different from visceral metastases. Most patients with BM suffered SREs, which would reduce physical function and quality of life. They may bring a series of complications, and the PS score of patients would be elevated. These factors could affect patients to receive further treatments, thus shortening survival time. Similarly, Ulas et al. showed that in non-small cell lung cancer patients with BM, SREs were detected in 72.8% patients, and the median OS time for patients with SREs were shorter than patients without SREs (7 VS 12 months, respectively) [23]. Our study further found that for EC patients with BM, older age, poorly differentiated and undifferentiated, with brain metastasis and with liver metastasis had poorer OS.
With the progress of medical science, there are many treatments for EC, including surgery, chemotherapy, radiotherapy (RT), targeted therapy, immunotherapy and so on, which will improve survival [24,25]. Because of the limitations of SEER database and that surgery is not encouraged in EC patients with BM, and to reduce the bias, we did not study treatments including surgery, chemotherapy and radiotherapy. In 2020, ESMO released clinical practice guidelines about bone health in cancer. Multidisciplinary management including palliative RT, radionuclide therapy and bone-targeted agents (BTAs) were recommended [26]. D'Oronzo et al. found that in patients with BM other than breast or prostate cancer, zoledronate would reduce the incidence of SREs, which were associated with worse survival [27]. Nowadays, bone health is becoming increasingly important in the management of cancer. We need to further research the mechanism and clinical characteristics of BM to provide guidance for clinical treatment.  There were fewer clinical articles about BM in EC, our study identified risk factors and prognostic factors for BM in EC patients. However, we should acknowledge that our study has certain limitations. First of all, only five sites of metastasis were provided in SEER database, including lung, liver, bone, brain and distant lymph node, whether patients had other metastatic sites were unknown. It would result in bias in our results. Moreover, information of treatment was too simple, we did not further study the effect of treatment on the prognosis for EC patients with BM. Nevertheless, our study will be helpful for diagnosis of EC patients with BM. Author contributions YQ and JM offered the idea of this study, analyzed the data, and drafted the manuscript. XL, NW and MY conducted the statistical analysis and revised the manuscript. JZ involved in data collection. DW and QW were the supervisors of this study and wrote the manuscript. All authors read and approved the final manuscript.
Funding This study was sponsored by grants from the General Project of Jiangyin Municipal Health Commission (No. S201803).

Data availability
The authors declare that the data supporting the findings of this study are available in this manuscript.

Conflict of interest None declared.
Ethics approval and consent to participate Not applicable.
Informed consent Not applicable.