Squamous cell carcinoma of the bladder: Demographics and outcomes associated with surgery and radiotherapy

Squamous cell carcinoma of the bladder (SCCB) is a rare disease composing 2%−5% of all bladder cancers with no consensus regarding treatment. The present study aims to analyze the outcomes of established treatments, namely chemotherapy, radiation, and surgery, to guide clinical decision‐making for patients with non‐schistosomal SCCB.

subset of urinary bladder cancer composing an estimated 2%−5% of all bladder cancers in Western countries, and while rare in comparison to urothelial carcinomas, is the second most common histologic subtype. 2,3The main cause of SCCB in regions without endemic Schistosoma (known as non-schistosomal or non-bilharzial SCCB) has been cited as chronic irritation, often in the setting of the prolonged indwelling catheters seen in patients with spinal cord injuries. 4The focus of this analysis is on non-schistosomal SCCB in the United States, and reference to SCCB will refer to the nonschistosomal variety for the rest of this paper.
There is currently no consensus regarding the treatment of SCCB, in large part due to the rarity of this disease. 5Traditionally, radical cystectomy has been established as an effective treatment, though many patients present with disease already too advanced for resection.While results are mixed and sample sizes are small-the largest study to date was a retrospective analysis of 27 patients, reported in 2007 by Kassouf et al.-chemoradiation has been proposed as a viable alternative to local surgical treatment. 5However, data for systemic treatments for SCCB is very limited and generally unfavorable; Maia et al. have summarized the various clinical trials that have been performed. 55][6] There is continual interest in using systemic therapies as substitutes or to augment surgical management of SCCB.This study aims to analyze the outcomes of surgery, radiation, and chemotherapy for SCCB, specifically using the Surveillance, Epidemiology, and End Results Registry (SEER) database to guide clinical decision-making regarding SCCB.

| Data analysis
Analysis was conducted on patients with a reported diagnosis of SCCB between 2000 and 2018.Data was processed, including the removal of incomplete entries and unknown values.Data was analyzed in Python (Version 3.1, Wilmington, DE: Python Software Foundation) using packages lime, scikit-learn, and torch.The packages lifelines and pandas were also utilized to assist with statistical tests, graph, and overall survival (OS) curve generation.
After data matricization, multivariable linear regression was completed for each variable in an iterative fashion; subsequent heatmap analysis was visualized using the Python package seaborn.

| Population attributes
A total of 5653 patients with a diagnosis of primary SCCB were identified.
Patient age at diagnosis was reported in 5-year increments, between 15 and 19 and 85+ years.The median age group was 70−74 at diagnosis.The annual incidence in the United States appears to be decreasing since 2000, with 388 patients diagnosed with SCCB in 2000 versus 213 in 2018 (see Figure 1).A breakdown concerning race showed 4887 patients were coded as White (86.9% of the total population), 534 were listed as Black (9.4%), 171 were listed as Asian or Pacific Islander (3.0%), and 29 were listed as American Indian/Alaska Natives (0.5%).This breakdown can be found in Figure 2.
Five-year OS for all patients was approximately 28% (see Figure 3), while 5 year OS broken down by race can be seen in Figure 4.The lowest 5 year OS was seen among American Indian/Alaska Native patients (~10%), while White and Asian or Pacific Islander patients both had the highest 5 year OS at about 35%.The differences in 5 year OS between White and Asian or Pacific Islander patients compared with both Black and American Indian/Alaska Native patients were both statistically significant (p < 0.05 and p < 0.01, respectively).F I G U R E 2 Breakdown of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) by race as listed in the Surveillance, Epidemiology, and End Results (SEER) database from 2000 to 2018 (n = 5653).White patients accounted for 86.9% of all cases, 9.5% were coded as Black, 3.0% were coded as Asian or Pacific Islander, and 0.5% were coded as American Indian/Alaska Native.

| Tumor metastasis and tumor size analysis
Analysis of patients with metastatic disease at diagnosis revealed 145 patients with metastatic disease at one of the following sites: bone, brain, liver, or lung.These patients represented 2.6% of our total patient population.The breakdown of the specific site is as follows: one patient with brain metastasis, 51 patients with liver metastasis, 69 patients with bone metastasis, and 102 patients with lung metastasis.The one patient with brain metastasis also had lung metastasis, while 27 of the bone metastasis patients also had either lung or liver metastasis, and 23 of the 102 patients with lung metastasis also had liver metastasis.It was also noted that 31 patients in the entire sample (5653) had distant lymph node (LN) metastasis; 18 of these patients had metastatic disease at one or more of the aforementioned organ sites, with 12 of these patients also having lung metastasis.
The SEER database refers to tumor size as a part of the collaborative stage (CS) system. 7Median primary tumor size was 5 cm for the 1792 patients with a tumor size recorded.Primary tumor size and OS were not well-correlated (R 2 = 0.002).Furthermore, the presence or absence of distant LN metastasis based on the American Joint Committee on Cancer Cancer Staging Manual, Seventh Edition, 8 in which LN above the aortic bifurcation are considered distant LN, showed a strongly significant impact on OS (see Figure 6, p < 0.01).
While all patients considered for LN metastases died within 3 years, patients without distant LN metastases lived about 1.5 years longer.
Patients not considered for distant LN metastases had a 5 year OS of 36% (denoted incomplete entries in Figure 6).Heatmap analysis of all SEER database variables demonstrated a significant correlation between tumor size and distant metastases (r = 0.898; refer to Supporting Information S1: Figure 2).Patients that received radiotherapy had a 5 year OS of approximately 18% versus 32% for those that did not receive radiotherapy (p < 0.01; Figure 8A).A similar trend was seen concerning the administration of chemotherapy, with patients receiving chemotherapy having a 20% survival versus 32% for those that did not receive any kind of chemotherapy (p < 0.01; see Figure 8B).The addition of radiation treatment to chemotherapy among nonsurgical patients did not yield a statistically significant 5 year OS benefit (Figure 8C).Patients that underwent surgical treatment for SCCB saw an increased 5 year OS of 40% versus 21% for those that did not undergo surgery (p < 0.01; Figure 9A).There was a statistically significant decrease in 5 year OS in patients that underwent both surgery and EBRT (14% vs. 35%, for those that only underwent surgery, respectively; p < 0.05; see Figure 9B).

| Outcomes by treatment modality
Finally, patients that underwent a combination of radiotherapy, chemotherapy, and surgery saw a decreased 5 year OS of 20% versus 25% for patients that underwent chemotherapy and surgery but did not receive radiotherapy (p < 0.05; see Figure 9C).Analysis of 5 year OS stratified by the sequence of systemic therapy in comparison to surgery showed no difference in survival among patients that received chemotherapy before versus after surgery (see Supporting Information S1: Figure 3A).Similar analysis regarding radiotherapy in relation to surgery showed patients that received radiation before surgery (and those that did not receive either radiotherapy or cancer-directed surgery) had an increased 5 year OS of 30%, compared to 16% for those that received radiotherapy after surgery (p < 0.05; see Supporting Information S1: Figure 3B).

| Subanalysis of surgical treatments
Further analysis of surgical interventions was undertaken with respect to stage at time of diagnosis, coded according to the CS system. 7Comparison between 5 year OS was undertaken between the following surgical methods: cystectomy, local surgery, and other (sublocal surgery, including laser treatments, intra-vesicular surgery, F I G U R E 5 Breakdown of all cases (n = 5653) of SCCB by histologic grade.15.7% were listed as well differentiated; Grade I, 35.5% were listed as moderately differentiated; Grade II, 32.8% were listed as poorly differentiated; Grade III, and 16.0% were listed as undifferentiated; Grade IV.SCCB, squamous cell carcinoma of the bladder.| 653 and electrocautery).Without accounting for stage no difference was seen between these three categories (see Figure 10).Stratifying patient stage at diagnosis into localized, regional, and distant disease 9 and repeating the same analysis revealed patients with local disease (primary cancer limited to the urinary bladder) only saw significant benefit with local surgery, though treatment with cystectomy approached significance (p = 0.073; see Supporting Information S1: Figure 4A).Patients with regional disease saw statistically significant benefit from both local surgery and cystectomy compared to those that did not receive any surgery (11%, p < 0.01; 39%, p < 0.01, respectively; see Supporting Information S1: Figure 4B).In this analysis, only one patient was coded as being treated with a procedure in the "other" category.Patients with distant disease saw no improvement in 5 year OS with any surgical procedure (p > 0.05 for all categories; see Supporting Information S1: Figure 4C).

| Subanalysis of treatment with chemotherapy or radiotherapy
Patient 5 year OS curves with respect to receipt of chemotherapy stratified by disease stage can be found in Supporting Information S1: On the other hand, patients with local and regional disease at diagnosis saw decreased 5 year OS with receipt of beam radiotherapy (refer to Supporting Information S1: Figures 6A and 6B, respectively).
Patients diagnosed with local disease and treated with beam radiotherapy saw a 5 year OS of 10%, compared to a 5 year OS of 48% for patients that did not receive beam radiotherapy (p < 0.01).This difference was less profound among patients with regional disease, with patients treated with beam radiotherapy having a 5 year OS of F I G U R E 7 Breakdown of patients that underwent permutations of surgery, radiotherapy, and/or chemotherapy for squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 based on the Surveillance, Epidemiology, and End Results (SEER) database (n = 5653).Analysis of treatment for SCCB revealed 664 patients received beam radiation (11.7%), 59 refused radiotherapy (1.0%), 30 were recommended to receive radiotherapy but administration was unknown (0.5%), five received radiation with no specified type (~0.0%), and two received a combination of beam radiation and implants or isotopes (~0.0%).Regarding chemotherapy for SCCB, 1081 patients (19.1%) were listed as having received chemotherapy, with the remainder (4572; 80.9%) listed as either having not received chemotherapy or receipt was unknown.Concerning surgical treatment of SCCB, 4852 patients underwent surgery (85.8%), 608 patients were not recommended to undergo surgery (10.8%), 159 patients were recommended to undergo surgery but ultimately refused (2.8%), and 27 patients were listed as "Unknown" (0.5%).
F I G U R E 8 Five-year overall survival (FYOS) of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 stratified by administration of radiotherapy (A), chemotherapy (B), or chemotherapy with or without radiotherapy (C).Patients receiving radiotherapy saw a decreased FYOS of approximately 18% versus 32% for those that did not receive any type of radiotherapy (p < 0.01).Patients receiving chemotherapy saw a decreased FYOS of approximately 9% versus 18% for those that did not receive any type of chemotherapy, which approached statistical significance.No statistically significant difference in FYOS was seen between patients that received both chemotherapy and radiotherapy versus those that only received chemotherapy (FYOS was 12% for both).Lighter outline around each survival curve denotes a 95% confidence interval.
12% compared to 20% for those that did not receive radiotherapy (p < 0.01).Finally, no survival difference was observed among patients diagnosed with distant disease who either received or did not receive radiotherapy (refer to Supporting Information S1: Figure 6C).

Stratification of patients by presence of LN metastases versus
absence, and removing blank or incomplete entries, revealed 32 patients with LN metastasis at diagnosis (see Table 1) compared to 734 without positive LN metastasis (see Table 2).No difference in F I G U R E 9 Five-year overall survival (FYOS) of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 stratified by surgical treatment (yes/no; A), broken down by patients treated by surgery with or without radiotherapy (B), and stratified by patients treated with both surgery and chemotherapy with/without radiotherapy (C).Patients that underwent surgery saw an increased, statistically significant (p < 0.01) FYOS of approximately 40% versus 21% for those that did not receive any type of surgery.Patients that underwent surgical treatment only, with no radiotherapy or chemotherapy, saw an increased FYOS of 35% compared to 14% for patients treated with both radiation and surgical treatment (p < 0.05).All patients that were treated with both surgery and a combination of beam radiation with implants or isotopes died within 1 year of treatment.Patients that underwent surgical treatment and received chemotherapy only, with no radiotherapy, saw an increased FYOS of 25% compared to 20% for patients treated with radiation, chemotherapy, and surgical treatment.This was statistically significant (p < 0.01).All patients that were treated with chemotherapy, surgery, and a combination of beam radiation with implants or isotopes died within 2.5 years of treatment.Lighter outline around each survival curve denotes a 95% confidence interval.
F I G U R E 10 Five-year overall survival (FYOS) of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 for patients treated with surgery and stratified by surgical treatment method.Stage is not accounted for in this graph.Patients treated with surgical methods grouped as "other" (including sublocal surgery, including laser treatments, intra-vesicular surgery, and electrocautery) had a 5 year OS of about 44%, compared to 30% for both local surgery and cystectomy.No statistically significant difference in FYOS was seen between surgical methods.Lighter outline around each survival curve denotes a 95% confidence interval.
T A B L E 1 Frequency of systemic therapy and radiotherapy for patients with positive lymph node (LN) metastasis at diagnosis of primary squamous cell carcinoma of the bladder.

| Population characteristics
The population characteristics in our cohort identified in the most recent SEER database from between 2000 and 2018 were consistent with other population data reported in the literature. 1,3The median age range in our cohort was 70−74 years with a median OS of 13 months.There is a paucity of data outside of database analyses for comparison of survival statistics, however, there are several small, single-institution studies that demonstrated comparable survival statistics. 3,10The largest single-institution study published to date, Kassouf et al., had a sample size of 27 with a 5 year OS of 40% 5 ; this is in comparison to the 28% 5 year OS (see Figure 3) we found in our study of 5653 patients.
Our 5 year OS data broken down by race demonstrates both Asian/Pacific Islander and White patients have a longer 5 year OS compared with both Black and American Indian/Alaska Native patients.This data must be interpreted with the caveat that most patients in this analysis (86.9%) were coded as White.

| Analysis of grade and tumor size
Disease grade in general has important implications for disease survival, particularly with metastatic disease, and this is true of SCCB in particular: poor grade has been found to directly impact OS in patients with metastatic bladder cancer. 11The only significant variation in histology across all racial demographics was histologic breakdown of patients identified as American Indians/Alaska Natives compared to all other patients.The higher proportion of patients in this subset coded as having Grade II SCCB, when compared to 35% for all other patients, may have implications for treatment and may theoretically suggest a protective factor, in the sense that higher Grade generally confers a lower OS. 11mor size and OS were not well correlated (R 2 = 0.002), likely reflecting the multifactorial nature of OS for SCCB.However, significant correlation between tumor size and distant metastases was seen via heatmap analysis (r = 0.898).3][14] Distant LN metastasis was also seen to decrease 5 year OS, in keeping with the literature. 5,15,16It is likely patients being evaluated for the presence of distant metastases are at increased risk for actually having such metastases, as compared to those patients in which this was not considered or assessed, explaining the trend of patients without distant LN metastases still dying in under 3 years from diagnosis.

| Treatment modality outcomes and surgical treatment subanalysis
Traditionally, surgical treatment of SCCB has been the most common treatment approach, though the development of newer chemotherapy drugs and the potential benefits of chemoradiation have stimulated discussion concerning the treatment of SCCB with a multimodality approach. 4,6,7The specifics of the chemotherapy and radiotherapy patients received are not recorded in the SEER database.Our analysis showed that surgery, both in general, and specifically surgical methods not including anterior exenteration, results in longer 5 year OS compared to any combination of radiotherapy and chemotherapy.Furthermore, the addition of radiotherapy and/or chemotherapy does not appear to result in a significant increase in 5 year OS.Based on the results evidenced in Figure 9C, the addition of radiotherapy to both chemotherapy and surgery may even decrease 5 year OS.
While no difference was seen in 5 year OS among cystectomy, local surgery, and other surgical methods (sublocal surgery, including laser treatments, intra-vesicular surgery, electrocautery, etc.), further stratification revealed that patients with lower stage, locoregional disease seemed to benefit more from surgery in general.This is likely a result of being at an earlier stage to receive less-invasive surgery.
Furthermore, in our analysis of 5 year OS with respect to both CS and Intraoperative rad with other rad before/after surgery 2 Sequence unknown, but both were given 0 surgical methods, patients with regional disease at diagnosis saw increased survival with local surgery, cystectomy, and the "other" category of surgical methods.This result was less pronounced with patients that had local disease, in which only local surgery showed a significant increase in survival.

| Chemotherapy and radiotherapy subanalysis
While patients treated with chemotherapy were noted to have a decreased 5 year OS compared to those that did not receive chemotherapy, regardless of stage at diagnosis, the results for patients that received radiotherapy are more nuanced.The relative difference in 5 year OS for patients that received radiotherapy was greatest for those with the least advanced disease (local SCCB), with a 38% lower 5 year OS for patients that received radiotherapy compared to those that did not.This difference narrowed to 8% for patients with regional disease and was not present for patients with distant disease.It is unclear whether the increased mortality from treatment of SCCB without surgery is the result of toxicities related to these other treatment modalities or related to the potentially more advanced disease these patients present with.Taken in the context of our analysis of tumor size, metastases, and 5 year OS, while radiation may decrease tumor size, it does not appear to decrease the likelihood of tumor metastasis and ultimately does not improve OS.
Furthermore, when radiotherapy is applied, patients that receive radiotherapy before surgery tend to fare better at the 5-year mark than those that receive radiotherapy after, which could be the result of greater control of disease but may also be a reflection of patient health.Finally, concerning the impact of chemotherapy and radiotherapy timing for patients with metastatic disease at diagnosis, the specific timing of these therapies does not appear to increase 5 year OS based on our analysis.

| Future work
The SEER database holds a multitude of data, but it is classically limited in many regards.Key demographics including patient age (currently reported as a range) and markers of socioeconomic status such as income and education level are missing.Additionally, histories of exposures or illnesses that may predispose patients to specific cancers (e.g., HPV or previous schistosomal infections in the case of this article) are not recorded.Social history factors, especially tobacco and alcohol use, have been linked to SCCB and squamous cell carcinoma in general; these factors have the potential to impact prognosis and treatment, though this information is not present in the SEER database. 14,17In the case of SCCB in our analysis we assume that all cases are nonschistosomal given that schistosomal infection is non-endemic in the United States, and less than 2% of SCCB in the United States has schistosomal infection implicated as a risk factor. 18However, a history of schistosomal infection is not reported in SEER and this is therefore not known with certainty.Regarding treatment outcomes compared to tumor size and spread, the SEER database is limited in the available information concerning patient overall health and access to healthcare services; thus, it is possible confounding factors may be at play.Limited information on the specific timing and sequence of therapies is available in the SEER database, and specific information such as the days between therapies, presence of positive margins, and so forth is lacking as a result.Finally, specific information concerning the treatments patients underwent is limited.Regarding surgery, no information concerning the specific procedure patients underwent or any complications associated with the procedure is available.The same is true regarding specific chemotherapeutic agents and radiation regimens.Future work may involve addressing these limitations and expanding on this current work.

| CONCLUSION
Bladder SCC is a rare disease with limited information in the literature and limited consensus on treatment.Our analysis of the SEER database regarding patients with primary SCCB is the largest analysis of this disease to date and supports the limited research that surgical management of SCCB is the most effective treatment, particularly among patients with a lower stage at diagnosis.Furthermore, this analysis casts doubt on the efficacy of both radiotherapy and chemotherapy for treatment of SCCB.However, further research is needed into the use of these treatment modalities and socioeconomic factors not accounted for in the SEER database.
This is a retrospective cohort study utilizing the SEER database sponsored by the National Cancer Institute.This registry provides deidentified, detailed disease course data including approximately 28% of the US population.Incidence and population data associated by age, sex, race, year of diagnosis, and geographic area are available through this database, and new research data are added annually every Spring.The registry was queried for all cases of "squamous cell neoplasms" (variable name Histology Recode-broad groupings) occurring in the bladder (variable name Site recode ICD-O-3/WHO 2008).Histologic codes 8050−8089 were included.Exclusion criteria included cancers of the urethra or upper or lower ureters, epithelial histology not otherwise specified, and non-primary tumors.Specifically, cases with histologic code 8120 representing urothelial carcinoma with squamous cell differentiation were excluded.Variables selected for review included patient demographics such as race and age, and tumor characteristics including grade, tumor type, and histopathologic characteristics.Treatment modalities, such as surgical resection, radiation therapy, and systemic therapy were considered in conjunction with Kaplan Meier survival outcomes for various combinations and orders of the aforementioned treatment types.Outcomes based on tumor grade and histological characteristics were also analyzed.
All cases of SCCB fell within the broad histologic grouping of codes 8050−8089, based on the World Health Organization (WHO) International Classification of Diseases for Oncology, 3rd edition (ICD-O-3).
Figure 5 shows a breakdown of all cases of SCCB by histologic grade, excluding cases with incomplete or unknown entries.The most common grade at diagnosis was recorded as moderately differentiated (Grade II) and poorly differentiated (Grade III) at 35.5% and 32.8%, respectively.21.1% of patients had an incomplete or unknown histologic grade from the entire database.Further analysis of patient primary tumor histologic grade at diagnosis, stratified by race, can be seen in Supporting Information S1: Figure1.Among American Indian/Alaska Native patients, 62% presented with tumors noted as Grade II, compared with 35% for all patients (p < 0.05).This excluded those with incomplete entries for either race or Grade.All other subanalyses based on race and grade breakdown were statistically insignificant.

F
I G U R E 1 Breakdown of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) by year of diagnosis as listed in the Surveillance, Epidemiology, and End Results (SEER) database from 2000 to 2018 (n = 5653).General downtrend in cases is seen with increasing year (trendline equation: year of diagnosis = −5.6544× [number of cases] + 11 657, with R 2 of 0.5635).
Analysis of treatment for SCCB revealed 664 patients received external beam radiation (EBRT) (11.7%), 59 were offered radiotherapy but refused (1.0%), 30 were recommended to receive radiotherapy but administration was unknown (0.5%), five received radiation with no specified type, and two received a combination of EBRT and implants or isotopes.Regarding chemotherapy for SCCB, 1081 patients (19.1%) received chemotherapy.Concerning surgical treatment of SCCB, 4852 patients underwent surgery (85.8%), 608 patients were not recommended to undergo surgery (10.8%), 159 patients were recommended to undergo surgery but ultimately refused (2.8%), and 27 patients had incomplete entries (0.5%).

Figure 7
Figure 7 demonstrates the treatments patients underwent.

F
I G U R E 3 Five-year overall survival (FYOS) of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 (n = 5653) based on analysis of the SEER database.The FYOS seen was 28%.Lighter outline around the survival curve denotes a 95% confidence interval.SEER, Surveillance, Epidemiology, and End Results Registry.F I G U R E 4 Five-year overall survival (FYOS) of patients diagnosed with squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 broken down by race (n = 5653).The lowest FYOS was seen among American Indian/Alaska Native patients (~10%), while White and Asian or Pacific Islander patients both had the highest FYOS at about 35%.Lighter outline around each survival curve denotes a 95% confidence interval.

F
I G U R E 6 Kaplan−Meier survival curve for 5-year overall survival for patients diagnosed with squamous cell carcinoma of the bladder (SCCB) from 2000 to 2018 based on the Surveillance, Epidemiology, and End Results (SEER) database (n = 5653).Patients positive for distant metastases lived approximately 1.5 years after diagnosis, compared to about 3 years for those that were negative.Approximately 40% of the overall patients diagnosed with SCCB had incomplete entries for this variable in the database, and were plotted using a blue line, with a 5 year OS of approximately 36%.Lighter outline around each survival curve denotes a 95% confidence interval.LARKINS ET AL.

Figure 4 .
Figure 4.No survival benefit was seen among patients that received chemotherapy versus those that did not for patients with local, regional, or distant disease at diagnosis (refer to Supporting Information S1: Figures 5A, 5B, and 5C, respectively).
other rad before/after surgery 0 Sequence unknown, but both were given 1 LARKINS ET AL. | 655 specific therapy timing administered was seen among patients that received chemotherapy (p = 0.22) or radiotherapy (p = 0.31).No trend was seen regarding average tumor size compared to specific chemotherapeutic or radiotherapeutic treatment.
T A B L E 2 Frequency of systemic therapy and radiotherapy for patients with negative lymph node (LN) metastasis at diagnosis of primary squamous cell carcinoma of the bladder.