Bladder cancer (BC) is the second most common urogenital cancer. The initial site of BC metastasis is usually the pelvic lymph nodes, but this cancer often spreads to other organs, most commonly the lungs and bones, through lymph and blood channels. BC has a poor prognosis and is rarely cured. Without treatment, bone metastasis (BM) may lead to the skeletal-related event (SRE), such as pathological fracture, spinal cord compression, and malignant tumor hypercalcemia.[2,3] SRE is still the leading cause of mortality and morbidity and often leads to a decline in patients quality of life. At present, surgery is the first choice for BC; however, most BC patients with BM receive palliative treatment instead of undergoing a radical surgery. The latter may be related to the fact that surgery does not necessarily prolong the survival time in patients. Of course, surgical treatment is not suitable for patients with multiple metastases or those with poor general conditions. All in all, there is no particular effective treatment for BC patients with BM.
Thus far, there are few studies that investigated the risk factors of developing BM in patients with primary BC. Therefore, it is necessary to analyze the epidemiological characteristics of BC with BM comprehensively and identify the risk factors of BM. The TNM staging system of the American Joint Committee on Cancer (AJCC) is publicly recognized and widely used to predict the metastatic risks and prognosis in patients with various cancers. While these staging systems have provided useful estimates for recurrence risks and survival outcomes, the heterogeneity in tumor biology and patient characteristics within each prognostic group lead to significant variations. This highlights the ultimate limitations of the categorical risk grouping models because specific risk factors are defined in a manner that includes patients with varying degrees of risk. In addition, TNM staging does not include other factors such as age, gender, comorbidity, previous treatment, imaging, and molecular characteristics. The TNM staging system is convenient to use, and it is used as a common language to communicate with patients and describe their illness. However, clinicians often combine the TNM staging system with personal experience to predict the prognosis of cancer patients. In view of the inaccuracy of personal judgment, modern statistical methods and computer prediction models should be incorporated into clinical decision-making more frequently.
As a statistical tool, a nomogram can solve the above problems in a more accurate way. A large number of studies showed that a nomogram can predict the prognosis of some malignant tumors. Compared with the traditional AJCC TNM staging system, the nomogram is simpler and more accurate tool and is a good substitute for the TNM staging system. As a result, a number of cancer-related nomograms have been developed. For example, BH et al established and verified a nomogram to predict the risk of recurrence after radical cystectomy for BC. A good nomogram can be used to predict personal results, which is beneficial to both patients and clinicians. However, to the best of our knowledge, there is no research that constructed a nomogram to estimate the risk of BC with BM. This study had 2 main objectives: to identify the significant variables, mainly the risk factors that may affect BC with BM, and to construct a prediction model based on these variables. Therefore, the purpose of this study was to construct and validate the prediction model of BC with BM by analyzing the patient data extracted from the database of Surveillance, Epidemiology, and End Results (SEER).
2.1 Ethics statement
We obtained approval from the Ethics Committee of the Affiliated Hospital of Chengde Medical University before carrying out this study. The content of this study did not involve human subjects or personal privacy; hence, informed consent from patients was not required in this study.
2.2 Patients and data collection
The SEER database is a program of the National Cancer Institute. In brief, it contains data related to cancer incidence and survival outcomes from population-based cancer registries, covering 28% of the US population. In our study, we included patients who were newly diagnosed with BC from 2010 to 2016 in the SEER database. Patients meeting the following criteria were included in our analysis:
- 1. site record: trigon of bladder, dome of the bladder, bladder wall (lateral wall of the bladder, anterior wall of the bladder, posterior wall of the bladder), bladder neck, ureteral associated tissue (ureteric orifice, urachus), overlapping lesion of the bladder, and bladder, NOS according to the Third Edition of International Classification of Diseases for Oncology (ICD-O-3);
- 2. pathological type: transitional cell carcinoma (TCC), squamous cell carcinoma (SCC), adenocarcinoma;
- 3. complete information preservation about T and N classification.
Exclusion criteria were as follows:
- 1. information on survival rate, follow-up time or cause of death was missing or insufficient,
- 2. diagnosis was based on only autopsy results or certificates of death,
- 3. multiple primary tumors and patients with unknown metastatic status.
A total of 35,506 BC patients were included in this study, 796 (2.24%) of whom had BM and 34,710 (97.76%) had no BM.
2.3 Statistical analysis
All statistical analyses in our research were carried out using R software (Version 3.6.1). Receiver operating characteristic curve analysis was used to convert age data into categorical data, and the cut-off was determined based on the maximum of Youdens index. To evaluate the risk factors of BC with BM, the difference in continuous variables between patients with BM and those without BM was compared using Student t test. All BM patients were randomly divided into training and validation cohorts at a ratio of 7:3. In addition, the risk factors of classification variables were identified using the Chi-Squared test or Fisher extraction test. Multivariate logistic analysis included variables with P < .05 in univariate analysis. Then, the independent risk factors of BC with BM in patients were determined, and based on these independent risk factors, a nomogram was established using and RMS packet in R software. Harrell's concordance index (C-index) represented the discrimination of the nomogram. Furthermore, the nomogram was evaluated using calibration curve analysis and decision curve analysis (DCA).
3.1 Baseline characteristics of the study population
According to our criteria, 35,506 BC patients whose records were extracted from the SEER database were included. The patients were divided into the training cohort (24,856) and the validation cohort (10,650). As shown in Table 1, in the training cohort, 70.45% of patients were aged ≤80, mostly white (89.55%). Differentiation in grades III-IV (88.34%) was the most common among tumor classifications. T1-2 (83.65%) and N0-1 (93.90%) phases were common. Distant metastasis was observed in the following cases: 510 (2.05%) had tumor metastasis to the lung, 324 (1.30%) to the liver, and 46 (0.19%) to the brain.
3.2 Risk factors for developing BM
As shown in Table 2, grade, T stage, N stage, brain, liver and lung metastases, age, race, histologic type, and primary site were related to BC with BM. The variables with P value < .05 in univariate analysis were included in multivariate logistic regression analysis to determine the risk factors of BC with BM. The results showed that grade, T stage, N stage, brain metastasis, lung metastasis, liver metastasis, race, age, histologic type, and primary site were independent predictors of BC with BM (Table 2).
3.3 Diagnostic nomogram development and validation
The risk assessment model of the nomogram was based on logistic regression analysis (Fig. 1). The C-index of the nomogram reached 0.812 and 0.806 in the training and validation cohorts, respectively, showing better discrimination ability. The calibration curve showed a high degree of consistency between the observed and predicted results (Fig. 2). In addition, DCA showed that the nomogram had an excellent performance in clinical practice (Fig. 3).
It is estimated that 12,500 people die due to metastatic BC every year in the United States. Lymph nodes are the most common metastatic site of BC, but studies have shown that bones can be even considered as the most common sites for distant metastasis in BC, and approximately 30% to 40% of metastatic BC patients have BM.[8,9] SRE caused by these metastases directly affect the prognosis in BC patients. Therefore, it is crucial to identify the risk factors of BC with BM in patients and to carry out early preventive intervention in patients with a high risk of BM. However, there were few detailed studies on risk factors of BC with BM, and there was no study that established a BM prediction model with a nomogram. BC is a heterogeneous disease. For each patient, there may be many possible treatment methods and prognostic outcomes. Therefore, the TNM staging system alone cannot predict the BM risks individually, visually, and quantitatively. As a reliable graphical calculation model, the nomogram is used to integrate all risk factors of tumor occurrence and predict individual risks of specific events.[10,11] This is a tool that can evaluate the possibility of metastasis progress, tumor specificity, mortality rate, and long-term quality of life accurately. Therefore, nomogram is an important tool to assist clinicians during patient consultation and decision making on treatment options.
Therefore, for the first time, we have established the predictive nomogram for BC with BM based on the retrospective analysis of data of BC patients from the SEER database. We determined the risk factors that may lead to BC with BM in patients, including age, race, T and N stage, grade, lung, liver and brain metastases, primary site, and histologic type. Among them, histologic type and primary site have not been examined as BM risk factors in previous similar studies. For example, Zhang et al studied and analyzed the risk factors of BC with BM, including age, race, marital status, insurance status, T and N stage, tumor grade, liver, lung, and brain metastases. SCC accounts for only 2% to 5% of BC, but bladder SCC has the characteristics of high malignant degree and high recurrence rate. SCC has a faster disease progression than TCC in BC patients with stage III or IV. Therefore, there may be differences in BC with BM between the 2 tissue types. Previous studies by Weiner et al have shown that triangular and bladder neck tumors were associated with higher lymph node involvement rates, indicating that they had higher invasion and metastasis potential. Animal experiments showed that urothelial stem cells were mainly located in the bladder triangle and bladder neck, and cancer stem cells may be highly distributed in these areas of the bladder. Therefore, BM risks in different locations may also be different, and BM risk is added as a study variable. Fortunately, our study results also confirmed that histological type and primary site were BM risk factors; SCC is indeed more likely to cause BM than other BC types, and the possibility of BM was highest when the tumor was located in the bladder neck. Other BM risk factors included age, race, T and N stage, grade, lung, liver, and brain metastases, which were the same as those reported in the literature. Most importantly, we have successfully established the prediction model of BC with BM, which was done for the first time in the field for the BC with multiple BM risk factors. Second, Zhang et al have only studied the risk factors of BC with BM, but no prediction model has been established. We believe that our study is more accurate and more convenient to apply in clinical work.
In addition to the newly discovered risk factors, we have also studied several other risk factors and analyzed the possible reasons for them to become risk factors. The increase in T and N stage and grade was an independent risk factor of BC with BM. As shown in previous studies, the increase in T and N stage in patients with malignant tumors may mean the increase in tumor volume and the involvement degree and range of adjacent tissues and lymph nodes, while the increase in grade may mean the increase in the malignant degree of a tumor, which are all manifestations of further progression of the malignant tumor. Among BC patients with cancer progression, 40% of those with advanced disease will have BM.[17,18] Therefore, effective treatment as early as possible and intervention with tumor progression to prevent or delay the increase of T and N stage and grade are essential means to prevent BM in BC patients. Black race has been proven to be one of the risk factors for BM diagnosis, which was consistent with other reported cancer types.[19–21] Previous studies also reported that black patients showed higher stage disease and worse disease-specific survival rate than white race.[22,23] There are also studies showing that SCC is more common in blacks than in whites, while our study and previous studies showed that SCC is more malignant, leading to a higher risk of BC with BM, which, we believe, may be the reason why black BC patients were more prone to develop BM. Finally, age as a continuous variable had an odds ratio value of 0.983 < 1 in our study, indicating that the lower the age of the BC patients, the higher was the risk of BM. This also indicates the importance of paying close attention to BC patients in low age group in clinical work. Finally, we also confirmed that BC patients were more likely to have BM if accompanied by liver, lung, brain, and other metastases. Most people think that once a tumor metastasizes to a distant area in an organ, it may accelerate metastasis in other parts, which is in agreement with our observations in this study. Therefore, controlling tumor metastasis in other body parts is also important for preventing BM. To sum up, the prediction model that we have established is of great help to predict BC with BM risks in patients in the clinical setting and to formulate treatment plans according to the patients physical conditions.
The nomogram developed in this study can be used as an auxiliary graphical tool for patients with BC to help clinicians evaluate the risks of BC combined with BM and predict prognosis. Verification and application in an independent population showed that the prediction model had excellent performance and clinical application value.
We would like to thank all the staff in the Department of Minimally Invasive Spine Surgery, Affiliated Hospital of Chengde Medical University, for their contribution on our research.
Conceptualization: Zhiyi Fan, Zhangheng Huang, Chengliang Zhao.
Data curation: Chuan Hu, Yuexin Tong.
Formal analysis: Zhangheng Huang, Chuan Hu.
Methodology: Chuan Hu.
Supervision: Zhiyi Fan, Chengliang Zhao.
Writing – original draft: Zhiyi Fan, Zhangheng Huang.
Writing – review & editing: Zhiyi Fan, Zhangheng Huang, Chuan Hu.
. Kaufman DS, Shipley WU, Feldman AS. Bladder cancer. Lancet 2009;374:239–49.
. Rosen LS, Gordon D, Tchekmedyian S, et al. Zoledronic acid versus placebo in the treatment of skeletal metastases in patients with lung cancer and other solid tumors: a phase III, double-blind, randomized trial--the Zoledronic Acid Lung Cancer and Other Solid Tumors Study Group. J Clin Oncol 2003;21:3150–7.
. Saad F, Gleason DM, Murray R, et al. A randomized, placebo-controlled trial of zoledronic acid in patients with hormone-refractory metastatic prostate carcinoma. J Natl Cancer Inst 2002;94:1458–68.
. Selvaggi G, Scagliotti GV. Management of bone metastases in cancer: a review. Crit Rev Oncol Hematol 2005;56:365–78.
. Shabsigh A, Bochner BH. Use of nomograms as predictive tools in bladder cancer. World J Urol 2006;24:489–98.
. Bochner BH, Kattan MW, et al. International Bladder Cancer Nomogram Consortium. Postoperative nomogram predicting risk of recurrence after radical cystectomy for bladder cancer. J Clin Oncol 2006;24:3967–72.
. Advanced Bladder Cancer (ABC) Meta-analysis Collaboration. Neoadjuvant chemotherapy in invasive bladder cancer: update of a systematic review and meta-analysis of individual patient data advanced bladder cancer (ABC) meta-analysis collaboration. Eur Urol 2005;48:202–5. discussion 205-6.
. Shinagare AB, Ramaiya NH, Jagannathan JP, et al. Metastatic pattern of bladder cancer: correlation with the characteristics of the primary tumor. AJR Am J Roentgenol 2011;196:117–22.
. Witjes JA, Compérat E, Cowan NC, et al. EAU guidelines on muscle-invasive and metastatic bladder cancer: summary of the 2013 guidelines. Eur Urol 2014;65:778–92.
. Kluth LA, Black PC, Bochner BH, et al. Prognostic and prediction tools in bladder cancer: a comprehensive review of the literature. Eur Urol 2015;68:238–53.
. Balachandran VP, Gonen M, Smith JJ, et al. Nomograms in oncology: more than meets the eye. Lancet Oncol 2015;16:e173–80.
. Zhang C, Liu L, Tao F, et al. Bone metastases pattern in newly diagnosed metastatic bladder cancer: a population-based study. J Cancer 2018;9:4706–11.
. Abol-Enein H, Kava BR, Carmack AJ. Nonurothelial cancer of the bladder. Urology 2007;69: (1 Suppl): 93–104.
. Izard JP, Siemens DR, Mackillop WJ, et al. Outcomes of squamous histology in bladder cancer: a population-based study. Urol Oncol 2015;33:
. Weiner AB, Desai AS, Meeks JJ. Tumor location may predict adverse pathology and survival following definitive treatment for bladder cancer: a national cohort study. Eur Urol Oncol 2019;2:304–10.
. Nguyen MM, Lieu DK, deGraffenried LA, et al. Urothelial progenitor cells: regional differences in the rat bladder. Cell Prolif 2007;40:157–65.
. Saad F, Lipton A. Zoledronic acid is effective in preventing and delaying skeletal events in patients with bone metastases secondary to genitourinary cancers. BJU Int 2006;97:1351–2.
. Coleman RE. Metastatic bone disease: clinical features, pathophysiology and treatment strategies. Cancer Treat Rev 2001;27:165–76.
. Guo X, Zhang C, Guo Q, et al. The homogeneous and heterogeneous risk factors for the morbidity and prognosis of bone metastasis in patients with prostate cancer. Cancer Manag Res 2018;10:1639–46.
. Sun M, Abdollah F, Liberman D, et al. Racial disparities and socioeconomic status in men diagnosed with testicular germ cell tumors: a survival analysis. Cancer 2011;117:4277–85.
. Roetzheim RG, Pal N, Tennant C, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst 1999;91:1409–15.
. Martin AM, Cagney DN, Catalano PJ, et al. Brain metastases in newly diagnosed breast cancer: a population-based study. JAMA Oncol 2017;3:1069–77.
. Yee DS, Ishill NM, Lowrance WT, et al. Ethnic differences in bladder cancer survival. Urology 2011;78:544–9.
. Prout GR Jr, Wesley MN, Greenberg RS, et al. Bladder cancer: race differences in extent of disease at diagnosis. Cancer 2000;89:1349–58.