Risk factors and prognostic factors for inammatory breast cancer with bone metastasis: A population-based study

Background: Inammatory breast cancer (IBC) is a rare type of breast cancer with poor prognosis. IBC patients with bone metastasis (BM) often suffer from many complications. This study was to identify risk factors with strong capability of predicting high BM risk for IBC patients and nd prognostic factors of IBC patients Methods: The Surveillance, Epidemiology and End Results (SEER) database was used to collect the clinicopathological and survival information of IBC patients. 966 IBC patients diagnosed between 2010 and 2015 were to study the risk factors for developing BM by using Multivariable logistic regression. 194 and 176 patients were to analyze independent prognostic factors for overall survival (OS) and cancer specic survival (CSS) of IBC patients with BM by performing Cox proportional hazard model. Results: Of the 966 IBC patients, 194 (20.1%) patients were with BM. IBC patients of unmarried, double breast tumor, N1 stage, N3 stage, and liver metastases had higher risk of BM, while those of uninsured status and triple negative breast cancer (TNBC) were less likely to have BM. The survival analysis showed that TNBC subtype and liver metastases were independently signicantly associated with poorer OS and CSS of BM patients, while chemotherapy could serve as an independent prognostic factor for better OS and CSS of BM patients. Conclusions: The risk factors for developing BM could provide potential guidelines for screening BM in IBC patients. The independent prognostic factors for survival outcome of IBC patients with BM could help doctors precisely treat those patients.

symptoms of BM. Hence, IBC patients always diagnose with BM when they have already had severe skeletal-related events (SREs). Therefore, it is important to nd reliable clinical risk factors to screen BM for IBC patients. In clinical conditions, factors utilized to indicate BM mainly include hypercalcemia and alkaline phosphatase (9). However, there has been no investigation studying factors with strong capability of predicting high BM risk for IBC patients. Besides, the study for prognostic factors of IBC patients with BM also needs more investigation.
In this study, the Surveillance, Epidemiology, and End Results (SEER) database was used to study the risk factors for developing BM in IBC patients. To further understand outcomes of IBC patients with BM, we also assessed the survival outcome and independent prognostic factors for IBC patients with BM between 2010 and 2015.

Patients
The SEER database was utilized to establish the study cohort. The SEER program owns population-based data, which covers about 30% of the population in the US (10,11). It mainly included the information of cancer patients' incidence, clinicopathological factors, survival outcomes, as well as treatment methods.
Patients initially diagnosed with IBC between 2010 and 2015 were included. IBC was diagnosed on the basis of International Classi cation of Disease for Oncology, 3rd edition (8530/3). Those who were under 18 years old, without positive histology, with unknown survival time, and with unavailable BM status were excluded from the study cohort. A total of 966 IBC patients were included to analyze the risk factors for developing BM. After excluding those without BM, 194 BM patients were left to study prognostic factors for overall survival (OS) of IBC patients with BM. Then, those dead from other reasons were excluded. 176 patients were used to investigate prognostic factors for cancer speci c survival (CSS) of IBC patients with BM. The procedure of the study cohort selection was exhibited in Fig. 1.

Statistical analysis
All of the data in this study were downloaded from SEER*Stat Software version 8.3.5. SPSS Version 23.0 software (IBM Corporation, Armonk, NY, USA) were utilized to analyze the data. The risk factors for BM of IBC was found by performing Multivariable logistic regression. Survival analysis for OS and CSS was conducted by using Kaplan-Meier analysis (log-rank test). Cox proportional hazard model was used to assess independent prognostic factors OS and CSS for IBC patients with BM. A p-value < 0.05 < 0.05 indicated statistical signi cance.

Results
Clinical characteristics of patients with bone metastases

Analysis of prognostic factors for OS of patients with bone metastases
In the survival analysis, we found that marital status ( Fig. 2A), liver metastases (Fig. 2B), breast subtype

Discussion
On the basis of SEER database, our study found that 20.1% of IBC patients had BM. Hence, it is important to nd risk factors of developing BM in IBC patients and the prognostic factors of IBC patients with BM.
In the previous study, Zheng Wang et al. found BM could be independent prognostic factor for IBC patients (12). However, there has been no investigation focusing on the cohort of IBC patients with BM. Those BM patients who suffered from the pain often had poor physical status. Therefore, it is vital to nd risk factors that can promote BM in IBC patients and prognostic factors that can independently predict the prognosis of IBC patients with BM.
In our study, we found that IBC patients of unmarried, double breast tumor, N3 stage, and liver metastases were more likely to develop BM. Therefore, physicians need to pay great attention to IBC patients with those risk factors for the high risk of BM. Besides, a routine bone scanning is highly advised for IBC patients with these risk factors. Interestingly, our study showed that those with Medicaid had higher risk of BM than those uninsured patients. We inferred that this might be the reason that those with Medicaid have more chance of having a thorough check of their body than those uninsured patients. Thus, IBC patients with Medicaid were easier to detect BM than uninsured IBC patients. Besides, our study found TNBC patients had lower risk of BM compared with HR+/HER2 + patients in the IBC cohort. TNBC is a highly aggressive subtype of breast tumor that is with high rate of metastasis and poor survival outcome (13). However, in IBC patients, the property of bone invasion of TNBC subtype seems to weakened. This might be due to the different molecular and biological mechanisms between IBC and non-IBC patients, which needs further investigations in the future.
Besides, we also investigated prognostic factors for IBC patients with BM. Our study showed that TNBC subtype and liver metastases could serve as independent prognostic factors for poorer OS and CSS of BM patients, and chemotherapy was independently signi cantly associated with better OS and CSS of IBC patients with BM. On the basis of these prognostic factors, physicians were able to effectively make a survival estimation for IBC patients with BM in clinical conditions. In addition, these prognostic factors can also serve as clinical guidelines for doctors. Our study indicated that despite the factor that patients of TNBC had lower risk of BM compared with HR+/HER2 + patients, those with TNBC had worse survival outcome than those with HR+/HER2 + in BM patients. More importantly, we found chemotherapy was the most suitable treatment methods for IBC patients with BM. A previous study showed that patients of stage IV IBC had an improved survival outcome after receiving surgery of the primary tumor (14).
However, there has been no study evaluating the treatment methods of IBC patients with BM. Therefore, this study is the rst one that reports chemotherapy rather than radiation therapy or surgery of primary site could signi cantly improve the OS and CSS of IBC patients with BM. In addition, this study showed that BM patients suffering from liver metastasis exhibited a worse survival outcome, whereas brain metastases and lung metastases had no in uences on BM patients' prognosis. Therefore, oncologist should pay great attention to liver metastases of IBC patients with BM, considering the poor prognosis of BM patients with liver metastases.
Despite of astonishing discoveries of the study, some limitations existed in this investigation. First, part of the unknown data was excluded. Besides, some useful clinicopathological factors were not included in the SEER database. In addition, considering this is a retrospective study, inherent bias is unavoidable.

Conclusions
In conclusion, this study identi ed risk factors for developing BM including unmarried status, double breast tumor, N3 stage, and liver metastases, while IBC patients with Medicaid and TNBC were less likely to develop BM. We also identi ed prognostic factors including TNBC subtype, liver metastases, and chemotherapy could serve as independent prognostic factors for OS and CSS of IBC patients with BM.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available either online or from the corresponding author on reasonable request.

Figure 1
Flowchart for selecting IBC patients with BM.