Development and validation of a nomogram to predict Bladder cancer in patients with different metastasis: a SEER database analysis

Purpose: Population-based data on the clinical correlates and prognostic value of the pattern of metastases among patients with bladder cancer are needed. Patients and methods: Surveillance , epidemiology and end results(SEER)database has been explored through SEER program. For each of four distant metastatic sites (bone, brain, liver, lung ),relevant correlation with baseline characteristics were reported. Survival analysis has been conducted through Kaplan-Meier analysis , and multivariate analysis has been conducted through a cox proportional hazard model. The precision of the nomogram was evaluated and compared using concordance index(C-index),and the area under receiver operating characteristic curve(AUC). Results: A total 2715 patients with metastatic bladder cancer were identidied from 2005-2019.Patients with medium risk have the best overall survival,followed by patients with low risk followed by patients with high risk metastases. Multivariate analysis showed that patients who were older than 80 at the time of diagnosis, single, no Chemotherapy, no complete cystectomy and bone metastases were associated with poor survival. A nomogram based on 7 independent risk factors has a good predictive power for the 12-month, 24-month and 36-month prognosis of patients with bladder cancer. The C index of the nomogram has high consistency in evaluating the survival rate of bladder cancer patients(C index=0.722,95% CI=0.712-0.732).The values of AUC for 12-month, 24-month, and 36-month were 0.775, 0.73, and 0.692.Conclusion: Build Nomogram prediction method validation of bladder tumor, according to the results of the Nomogram has good capability of identication and correction, suggests that these column chart.


Introduction
As one of the most common urinary system tumors, bladder cancer has the second highest incidence of urinary system tumors in my country, and the seventh highest in male tumor incidence [1][2] . In Western countries, the incidence and mortality of male bladder cancer are secondly only to In prostate cancer. [3][4][5] . The SEER program shows that in recent decades, the total incidence of bladder cancer has increased by 40%. Metastasis and recurrence are the primary reasons for its poor prognosis. Lymph nodes are the most common metastatic site for bladder cancer, and they have undergone radical cystectomy and pelvic resection. About 25%-30% of bladder cancer patients treated with lymph node dissection are found to have lymph node metastasis [6][7][8][9][10] . Therefore, lymph node metastasis is an important factor affecting the prognosis of bladder cancer patients. The traditional TNM staging has certain limitations for better predicting the prognosis of patients. So we are trying to establish a simpler and more clinically relevant method to assess the prognosis of bladder cancer.
The nomogram is a graphical representation of complex mathematical formulas, and is a common tool for evaluating the prognosis of oncology and medicine. In recent years, nomograms have played an increasingly important role in predicting the prognosis of liver cancer, lung cancer and other cancers [11][12][13] . The nomogram uses biological and clinical variables, such as patient age, gender, tumor grade, treatment methods, etc.To graphically depict statistical prognostic models that produce clinical practice probability, such as cancer recurrence or death. Compared with the conventional TNM staging, it is easier and faster to help doctors predict the prognosis of patients and promote clinical decision-making. The SEER database is a database of cancer patients based on 20 regions of the United States established by the National Institutes of America, covering about 28% of the population of the United States and capturing about 97% of cancer cases in the United States [14][15][16] . Therefore, this study uses the SEER database to extract clinical data of bladder cancer patients and establish a nomogram to guide clinical diagnosis and prognostic evaluation.

Materials And Methods
We collected the clinical data of 2715 patients with bladder cancer diagnosed from 2005 to 2019 in the database of the National Cancer Institute in the United States, collated and analyzed tumor-related clinical pathological parameters, demographic information, treatment methods, etc., and used COX proportional risk The regression model was used for statistical analysis to establish a nomogram to predict the prognosis of bladder cancer at 12 months, and use the consistency index (C-index) to verify the performance of the nomogram through the internal cohort to test the accuracy of the nomogram.
Inclusion criteria: 1) Histopathological examination was diagnosed as bladder cancer; 2) Complete dates, follow-up records, known survival months and cause of death; 3) Su cient/consistent information on variables, including age, gender, race, marriage, Tumor grade, TNM staging, lymph node metastasis, surgical treatment, radiotherapy and chemotherapy, etc. Exclusion criteria: 1) record controversial cases; 2) patients with secondary bladder cancer; 3) missing follow-up records or cause of death;

Statistical analyses
All data were processed by SPSS24.0 (IBM) and R software (version 3.4.5; http://www.r-project.org). Use SPSS24.0 to carry out relevant statistical description of demographic characteristics, and use the survival package of R software to carry out univariate and multivariate COX regression analysis of variables to estimate the mortality risk ratio. All statistical data are two-way tests, P < 0.05 considered the difference to be statistically signi cant.

patients characteristics
A total of 2715 patients with bladder cancer and distant metastases were identi ed in the period from 2005 to 2019 and were included into the analysis. Table 1 summarizes the distribution of different distant metastatic sites for included patients. 940 patients were diagnosed with isolated bone metastases, 72 patients were diagnosed with isolated brain metastases, 473 patients were diagnosed with isolated liver metastases, 863 patients were diagnosed with isolated lung metastases. Statistically signi cant correlations between different baseline characteristics and different sites of metastatic disease are shown in Table 1. The following associations were noted between baseline characteristics and speci c sites of metastases.: Bone metastasis were more commonly associated with age at diagnosis < 70 years(p < 0.001), Male gender(P < 0.001), node positive(P < 0.001),T2 stage(P = 0.048).
⋅Lung metastasis: Lung metastasis were associated with papillary TRC and Node positive(p < 0.0001).
The above associations have,however ,to be interpreted cautiously given the presence of "unknown"category in many of the variables which may have confounded the Chi-square association testing. Lung metastasis were associated with papillary TRC and Node positive(p < 0.0001).

Survival outcomes
The overall survival (OS) of bladder cancer patients analyzed by utilizing Kaplan-Meier survival curves. OS analysis was performed by stratifying different risks of bladder cancer. Statistically signi cant differences were identi ed with regard to risks (high risk vs. Medium risk, P < 0.001 ; Fig .1).

Prognostic factors
A cox proportional hazards regression model was constructed to evaluate predictors of OS (Table 2). Univariate analysis of OS revealed the risk of mortality was signi cantly higher for patients that were In the multivariate analysis (Table 3) (Fig. 2).The performance of the nomogram was internally validated by discrimination and calibration methods. The nomogram (Fig. 4)

Discussion
As one of the most common urinary system tumors, bladder cancer has high morbidity and mortality. The primary cause of poor prognosis is metastasis and recurrence. The common ones are lymph node metastasis, bone metastasis, brain metastasis, liver metastasis and lung metastasis [17][18] . Mainly originated from bladder epithelial tissue and interstitial tissue. Smoking, occupational exposure (insecticides, especially, dyes and other industrial chemical products) are the main carcinogenic risk factors [19] . The histological types can be divided into bladder urothelial carcinoma, bladder non-urothelial carcinoma (squamous cell carcinoma, glandular Cancer, small cell carcinoma, etc.), some of which are mixed, and 90% of patients are bladder urothelial carcinoma (transitional cell carcinoma) [20] . The main cause of patient death is tumor metastasis and recurrence. Previous studies have shown that about 25-30% of bladder cancer patients undergoing radical cystectomy and pelvic lymphadenectomy are found to have lymph nodes and other distant sites Patients with metastases and distant metastases have a higher rate of tumor recurrence after surgery. The prognosis of patients with bladder cancer varies depending on the site of metastasis. Therefore, accurate judgment of whether there is metastasis and the location and extent of metastasis has an important reference value for evaluating the prognosis of patients with bladder cancer, and is one of the foundations for formulating treatment plans and judging the prognosis.
We conducted a deeper study on the large sample data of the SEER database. By extracting the clinical data of bladder cancer patients in the SEER database that t this study, we found that gender, age, T stage, and N stage are related to bone metastases. Prognosis is relevant. In brain metastases, N stage is related to its prognosis. In liver metastasis, tumor grade, T stage, and N stage are related to its prognosis.
In lung metastasis, histological type and N stage are related to its prognosis. In summary, we have developed an accurate and intuitive tool to predict the survival rate of patients with bladder cancer in 12 months, 24 months, and 36 months. This chart can provide corresponding treatment assistance for patients with bladder cancer and provide basis for clinicians to make treatment decisions. However, whether the model is applicable to patients in other countries and regions is still subject to a large number of clinical studies to verify the treatment.

Declarations
Ethical approval This article does not contain any studies with human participants or animals performed by any of the authors.Informed consent: as this study is based on a publicly available database without identifying patient information, informed consent was not needed. Data Availability Statement:The datasets generated during and/or analysed during the current study are available in the SEER database's website link(https://seer.cancer.gov/data) repository,or can be obtained by contacting the corresponding author.
Competing interests The authors declare that they have no competing interests Author contribution Ning Xiao Topic selection, data collection and fund support.
Yongfu Long Topic selection, data collection and fund support.
Weijian Lin Processing data, making graphs and tables and writing papers.
Qi Tang Processing data.
Sheng Zhu Processing data.
Bin Jin Paper modify.