You are viewing a javascript disabled version of the site. Please enable Javascript for this site to function properly.

# Impact of Facility Radiation Patient Volume on Overall Survival in Patients with Muscle Invasive Bladder Cancer Undergoing Trimodality Bladder Preservation Therapy

### Background:

There is an association between higher hospital and urologist case volume and improved survival outcomes for patients with bladder cancer. Less data on facility volume and outcomes exists following trimodality bladder preservation with TURBT followed by definitive chemoradiation (CRT).

### Materials and Methods:

The National Cancer Database was queried for patients from 2004–2013 with urothelial bladder cancer (cT2-4aN0M0) receiving definitive CRT after TURBT. We compared OS between high- and low- bladder preservation case volume (BPCV) centers using Cox proportional hazard models. BPCV was dichotomized into high versus low volume at the 70th percentile level (i.e. high volume defined as top 30th percentile of cases and low volume defined as bottom 70th percentile of cases). Propensity matching was performed to match high- and low-volume centers.

### Results:

A total of 666 treatment facilities treating 1,635 patients with bladder preservation were identified with a median follow-up of 26 months (range, 2–136 months). A 70th percentile cutoff identified 497 patients that received treatment at 64 high-volume facilities. Median OS of patients treated at high BPCV centers was 37.0 months (95% CI, 27.3–46.7) versus 32.3 months (95% CI, 27.9–33.6) for patients treated at low BPCV centers (P = 0.004). High BPCV facilities were independently associated with a decreased hazard of death (HR, 0.86, 95% CI, 0.75–0.98; P = 0.031). In the propensity score matched cohort, median OS of patients treated at high BPCV was 36.1 months (95% CI, 26.5–45.8) versus 28.1 months (95% CI, 23.9–32.3) for patients treated at a low BPCV facility (P = 0.016).

### Conclusions:

In this observational cohort, treatment at a high BPCV facility was associated with improved OS. Causal factors for this finding may include superior bladder specific clinician expertise, technology, and multidisciplinary-care.

## INTRODUCTION

Bladder cancer is the second most common genitourinary malignancy in the United States [1, 2]. Recommended definitive treatment options for patients with localized muscle-invasive bladder cancer (MIBC) include neoadjuvant chemotherapy followed by radical cystectomy or bladder preservation therapy with maximal transurethral resection of bladder tumor (TURBT) followed by chemoradiation (CRT) [3]. Multiple studies have associated surgeon and hospital volume as independent predictors of survival outcomes in surgically treated cancer patients [4–9]. Studies comparing hospital or treatment center volume effects on survival in radiation treated cancer patients suggest similar outcomes [4, 10, 11].

Treatment of MIBC with trimodality bladder preservation therapy requires intensive multidisciplinary coordination between medical oncologists, urologists, radiologists, pathologists, and radiation oncologists. Multiple chemotherapy regimens and protocols are available, and the planning and delivery of pelvic radiation therapy is quite complex. Additionally, CRT for bladder cancer is frequently associated with adverse events that are challenging to manage and an experienced care team is needed to ensure optimal patient outcomes. Given the complexity and multidisciplinary effort required for bladder preservation therapy with definitive CRT, we conducted a National Cancer Database (NCDB) analysis to compare the overall survival of MIBC patients treated with trimodality bladder preservation therapy in high versus low volume facilities. It was our hypothesis that treatment centers with more experience with trimodality bladder preservation therapy would have improved overall survival outcomes compared to lower volume centers.

## MATERIALS/METHODS

### Data source and study population

The NCDB Participant User File was queried for de-identified data from 2004 and 2013 for patients 18–90 years of age diagnosed with urothelial bladder carcinoma [International Classification of Disease for Oncology, 3rd edition (ICD-0-3) codes 8120 and 8130]. The NCDB is a joint program of the American College of Surgeons and the American Cancer Society. Data from approximately 70% of patients diagnosed at Commission on Cancer (CoC) accredited cancer centers is included and includes patient, tumor, and treatment characteristics. Data elements are collected and submitted to the NCDB from commission-accredited oncology registries using standardized coding and data item definitions such as radiotherapy dose/technique, chemotherapy use/timing, and comorbidity. The Participant User File contains de-identified patient and center information and was exempt from Institutional Review Board review.

We limited our analysis to patients with muscle invasive disease and clinically negative lymph nodes (cT2-4a N0) who received radiation therapy to the bladder or pelvis (Radiation Treatment Volume codes 29 and 34) in conjunction with concurrent chemotherapy (including both single and multi-agent therapy given within 21 days of radiation). Patients with metastatic disease and incomplete treatment data were excluded. All patients underwent TURBT for pathologic diagnosis prior to definitive treatment. Total radiation dose received was defined as the sum of the regional dose delivered plus any boost dose. Patients with total radiation dose <5040 cGy or > 7500 cGy, and less than 20 fractions were excluded from analysis as these radiation doses likely represented either palliative treatment or potential coding errors. Flow chart of patient inclusion is shown in Fig. 1.

##### Fig.1

Consort diagram of inclusion and exclusion criteria. TURBT, transurethral resection of bladder tumor; XRT, radiation therapy.

Other covariates included in the analysis included clinical variables such as clinical tumor stage (cT), Charlson Deyo comorbidity, treatment facility type, radiation dose, radiation fractionation, number of chemotherapy agents; and socioeconomic variables such as age, sex, race, year of diagnosis, treatment facility location, insurance status, population setting, household income, and education level.

The primary endpoint of the study was the association between overall survival (OS) and BPCV. To adjust for potential confounding covariables between high and low BPCV centers, OS was also calculated in a propensity score matched cohort. Analysis for patients treated with standard curative fractionated radiation therapy (59.4–64.8 Gy in 33–36 fractions) was also performed.

### Statistical analysis

Overall survival was calculated from diagnosis until death, censoring at last follow-up for patients who were alive. The Kaplan-Meier method was used to estimate overall survival probabilities and the log-rank test was used to distinguish survival differences. In addition to these analyses, the clinical and socioeconomic covariables described above were used in univariable analysis. The χ2 test and Fisher’s exact test were used to evaluate contingency tables, as appropriate. Variables with p-values < 0.05 on univariate testing were entered into the multivariable analyses using the Cox proportional-hazards model. Propensity score analysis was performed to correct for baseline differences between treatment groups. A (1:1) matching algorithm with 0.2 caliper [12] including the variables used in univariate analysis was used. Significance was considered at a value of p < 0.05. SPSS Statistics v.23 (IBM Corporation; Armonk, New York) was used for all statistical analyses.

## RESULTS

### Demographics, Patient, Tumor, and Treatment Characteristics

A total of 666 treatment facilities treating 1,635 patients with bladder preservation were identified with a median follow-up of 26 months (range 2–136 months). A 70th percentile cutoff identified 497 patients that received treatment at 64 high-volume facilities that treated 6 or more cases in the time period examined. 1,138 patients therefore were treated at 602 low-volume facilities (i.e.<70th percentile or <6 cases per 10-years). 228 facilities only treated one case in the time period examined. A complete summary of demographic and patient characteristics by treatment volume is provided in Table 1. An additional dichotomization at the 90th percentile cutoff (> = 10 cases per facility in examined time period) identified 172 patients treated at 15 high-volume facilities and 1,463 patients treated at 651 low-volume facilities (Supplemental Table 1).

##### Table 1

Demographics and clinical characteristics

 Low volume High volume P-value Total patient # (%) 1138 (69.6) 474 (30.4) ––––––– Age Mean (yr) (range) 76 (24–90) 74 (32–90) ––––––– Sex 0.267 Male 831 (73.0) 371 (74.6) Female 307 (27.0) 433 (26.5) Race 0.051 White 1,028 (90.3) 449 (90.3) Black 77 (6.8) 29 (5.8) Other 26 (2.3) 7 (1.4) Unknown 7 (0.6) 12 (2.4) Clinical T stage 0.244 T2 920 (80.8) 419 (84.3) T3 128 (11.2) 45 (9.1) T4a 90 (7.9) 33 (6.6) Charlson Deyo Comorbidity 0.645 0 751 (66.0) 330 (66.4) 1 270 (23.7) 123 (24.7) >1 117 (10.3) 44 (8.9) Year of Diagnosis 0.999 2004–2008 506 (44.5) 221 (44.5) 2009–2013 632 (55.5) 276 (55.5) Facility location <0.0001 Central 495 (43.5) 143 (28.8) Northeast 259 (22.8) 131 (26.4) South/Southeast 207 (18.2) 132 (26.6) West 177 (15.6) 91 (18.3) Setting 0.614 Metropolitan 871 (76.5) 393 (79.1) Urban 190 (16.7) 78 (15.7) Rural 14 (1.2) 5 (1.0) Unknown 63 (5.5) 21 (4.2) Facility Type 0.027 Academic/Research Program 272 (23.9) 142 (28.6) Non-academic 866 (76.1) 355 (71.4) Insurance Status 0.618 Medicaid 32 (2.8) 13 (2.7) Medicare 868 (76.3) 363 (73.0) Not insured 21 (1.8) 8 (1.6) Other government 18 (1.6) 8 (1.6) Private 199 (17.5) 105 (21.1) Household Income 0.850 < $30,000 130 (11.4) 53 (10.7)$30,000–34–999 217 (19.1) 91 (18.3) $35,000–46,000 308 (27.1) 144 (29.0) >$46,000 437 (38.4) 185 (37.2) Unknown 46 (4.0) 24 (4.8) Education Level 0.280 ≥29% 152 (13.4) 52 (10.5) 20–28.9% 272 (23.9) 106 (21.3) 14–19.9% 287 (25.2) 135 (27.2) <14% 381 (33.5) 180 (36.2) Unknown 46 (4.0) 24 (4.8) Chemotherapy 0.571 Single agent 642 (56.4) 270 (54.3) Double agent 409 (35.9) 192 (38.6) Agent # unknown 87 (7.6) 35 (7.0) Radiation Dose 0.181 Median dose (cGy) (range) 6480 (5040–7480) 6480 (5040–7380) Median # fractions 36 (20–44) 36 (20–41)

High volume = 6 or more cases per 10-year period (>70th percentile). Low volume = 5 or fewer cases per 10-year period (<70th percentile).

Patients treated at facilities with higher BPCV were more likely to receive treatment at academic/research centers (28.6% versus 23.9%, P = 0.027) and less likely to be located in the central part of the United State (28.6% versus 43.3%, P < 0.0001). No differences in age, clinical T-stage, Charlson-Deyo comorbidity, radiation dose, number of chemotherapy agents utilized, or other socioeconomic variables were different between the groups.

### Survival outcomes

The median OS of patients treated at high BPCV centers was 37.0 months (95% CI, 27.3–46.7) versus 32.3 months (95% CI, 27.9–33.6) for patients treated at low BPCV centers (Fig. 2a; log rank P = 0.004). After controlling for demographic and clinical factors in Cox proportional hazards modeling, treatment at high BPCV facilities was independently associated with a decreased hazard of death (Hazard ratio (HR), 0.86, 95% CI, 0.75–0.98; log rank P = 0.031) (Table 2). Academic affiliation was not associated with OS (HR: 0.92, 95% CI, 0.79–1.06) (Table 2). When BPCV was analyzed in this same model as a continuous variable, an increasing number of bladder preservation cases per institute was also independently associated with improved OS (HR: 0.98, 95% CI, 0.97–0.99; log rank P = 0.044) (Table 2). In sensitivity analysis, multivariable Cox proportional hazard models in which BPCV was dichotomized at other cutoffs, improved OS was observed at the 75th, 80th, 85th, and 90th percentiles (Table 3). At the 60th and 50th percentile cut points, there was no OS difference. The median OS of patients treated at very high BPCV centers (90th percentile;≥10 cases) was 37.7 months (95% CI, 24.4–34.7) versus 31.9 months (95% CI, 29.3–34.7) for patients treated at low BPCV centers (Fig. 2b; P = 0.038). In a subgroup analysis, there were 1,213 patients receiving standard curative fractionated radiation therapy (59.4–64.8 Gy in 33–36 fractions) with 897 patients receiving treatment at low BPCV centers and 316 patients at high BPCV centers. The median OS of this subgroup of patients treated at high BPCV centers was 39.1 months (95% CI, 29.3–48.8) versus 30.7 months (95% CI, 27.6–33.7) for patients treated at low BPCV centers (log rank P = 0.009). After controlling for demographic and clinical factors in for this population, treatment at high BPCV facilities remained independently associated with a decreased hazard of death (Hazard ratio (HR), 0.83, 95% CI, 0.73–0.96; log rank P = 0.037)

##### Fig.2

Kaplan Meier overall survival curves for: (a) patients treated at high-BPCV centers (>70th percentile;≥6 cases) versus low-BPCV; (b) patients treated at very high-BPCV centers (>90th percentile;≥10 cases) versus low-BPCV. BPCV, bladder preservation case volume.

##### Table 2

Univariate and multivariate analysis

 Analysis Univariate Analysis Multivariate Analysis Variable HR (95% CI) P-value HR (95% CI) P-value Facility volume As a continuous variable* 0.98 (0.96–0.99) 0.005 0.98 (0.97–0.99) 0.044 Low-volume Reference group Reference group High-volume 0.82 (0.72–0.94) 0.004 0.86 (0.75–0.98) 0.031 Age Continuous 1.03 (1.02–1.03) <0.0001 1.03 (1.02–1.04) <0.0001 Sex Male Reference group ––– Female 1.05 (0.92–1.20) 0.473 ––– Race White Reference group ––– Nonwhite 1.09 (0.89–1.34) 0.409 ––– Clinical T stage T2 Reference group Reference group T3 1.38 (1.14–1.65) 0.001 1.33 (1.11–1.61) 0.003 T4a 1.27 (1.02–1.59) 0.033 1.31 (1.05–1.64) 0.019 Charlson Deyo Comorbidity: 0 Reference group Reference group 1 1.21 (1.05–1.40) 0.008 1.22 (1.06–1.41) 0.007 >1 1.77 (1.46–2.15) <0.0001 1.76 (1.45–2.14) <0.0001 Year of Diagnosis 2004–2008 Reference group ––– 2009–2013 1.05 (0.92–1.19) 0.485 ––– Facility location Central Reference group ––– Northeast 1.11 (0.90–1.31) 0.581 ––– South/Southeast 1.07 (0.86–1.11) 0.675 ––– West 1.05 (0.94–1.09) 0.751 ––– Setting Metropolitan Reference group ––– Urban 1.04 (0.88–1.22) 0.631 ––– Rural 1.19 (0.69–2.05) 0.535 ––– Unknown 1.05 (0.97–1.21) 0.135 ––– Facility Type Academic/Research Program 0.90 (0.77–1.00) 0.050 0.93 (0.80–1.07) 0.301 Non-academic Reference group Reference group Insurance Status Medicaid Reference group ––– Medicare 1.04 (0.92–1.11) 0.682 ––– Not insured 1.68 (0.80–3.33) 0.178 ––– Other government 1.50 (0.82–2.79) 0.222 ––– Private 0.92 (0.75–1.29) 0.685 ––– Unknown 1.88 (0.75–4.25) 0.189 ––– Household Income < $30,000 Reference group –––$30,000–34–999 1.07 (0.88–1.29) 0.503 ––– $35,000–46,000 1.05 (0.78–1.14) 0.541 ––– >$46,000 1.05 (0.76–1.11) 0.385 ––– Unknown 1.26 (0.84–1.91) .266 ––– Education Level ≥29% Reference group ––– 20–28.9% 1.03 (0.84–1.26) 0.804 ––– 14–19.9% 0.84 (0.70–1.02) 0.086 ––– <14% 0.97 (0.79–1.19) 0.738 ––– Unknown 1.30 (0.85–1.99) 0.223 ––– Chemotherapy Single agent Reference group ––– Double agent 0.91 (0.72–1.15 0.443 ––– Agent # unknown 1.03 (0.82–1.29) 0.810 –––

High volume = 6 or more cases per 10-year period (>70th percentile). Low volume = 5 or fewer cases per 10-year period (<70th percentile). *Cox proportional hazard modeling using facility volume as a continuous variable and excluding the variable low-volume and high-volume in analysis. All other values in the table exclude facility volume as a continuous variable.

##### Table 3

Sensitivity analysis

 Approximate percentile for cutoff Number of Patients (%) High-volume Hazard Ratio P-value 90th (≥10 vs < 10 cases) 172 (10.5%) 0.82 (0.67–0.98) 0.030 85th (≥9 vs < 9 cases) 241 (14.7%) 0.82 (0.68–0.97) 0.027 80th (≥8 vs < 8 cases) 276 (16.9%) 0.83 (0.70–0.98) 0.033 75th (≥7 vs < 7 cases) 369 (22.6%) 0.86 (0.74–0.99) 0.046 70th (≥6 vs < 6 cases) 497 (30.4%) 0.86 (0.75–0.98) 0.031 60th (≥5 vs < 5 cases) 635 (38.8%) 0.90 (0.79–1.02) 0.086 50th (≥4 vs < 4 cases) 818 (50%) 0.92 (0.81–1.05) 0.176

### Propensity matched cohort

Propensity score matching yielded 487 pairs of patients who were treated at high and low BPCV centers (974 total patients). After matching, there were no statistically significant differences in any treatment or socioecomonic variables analyzed, including association with academic/research center. The median OS of patients treated at high BPCV was 36.1 months (95% CI, 26.5–45.8) versus 28.1 months (95% CI, 23.9–32.3) for patients treated at a low BPCV facility (Fig. 3; log rank P = 0.016). Cox proportional hazards modeling of the propensity score matched cohort showed a similar OS benefit with treatment at high BPCV (HR: 0.82, 95% CI, 0.70–0.96; P = 0.016).

##### Fig.3

Kaplan Meier overall survival curves for the propensity score matched cohort comparing patients treated at high-BPCV centers (>70th percentile;≥6 cases) versus low-BPCV. BPCV, bladder preservation case volume.

## DISCUSSION

There are multiple causal explanations to support our results of improved overall survival with more experience treating MIBC with trimodality bladder preservation. Bladder preservation requires a large multidisciplinary team and coordination between multiple oncologic disciplines [25–27]. A high volume facility is more likely to have experts in each of the required disciplines and more frequent multidisciplinary urologic tumor conference to coordinate care. This may translate into patients receiving care in a more consistent approach, including work-up, staging, and treatment, at a high volume center [28–30]. Additionally, high volume facilities have been reported to have higher rates of protocol compliance that translates into improved cancer outcomes [31, 32]. Bladder preservation therapy is also associated with significant toxicities during and after treatment [33], including radiation-induced cystitis, bowel injury, and hematologic toxicities that require comprehensive support from members of the allied health services, which may be more prevalent at higher volume centers and associated with better outcomes. Also, in terms of chemotherapy, it has been shown that patients treated with systemic therapy at lower volume centers may have a higher mortality rate [34].

There are limitations in our analysis that may contribute to improved survival seen at high BPCV centers. Survival outcomes may be influenced by selection bias and referral patterns, with patients who have greater access to economic and social resources seeking care at higher volume facilities. In our analysis, however, we found no significant differences in insurance status, income, or education level between high and low BPCV institutes. There is also the possibility that patients initially seen at lower volume centers were referred to institutes with more experience with bladder preservation and such patients are not identifiable in our analysis. Additionally, our results are in contrast to a NCDB study published in 2018 which investigated the impact of academic facility type and case volume on survival for bladder cancer patients undergoing curative intent radiation therapy [38]. These authors found that both academic facility type and case volume were not associated with survival outcomes [38]. In their study, however, nearly half (48%) of the patients did not receive concurrent chemotherapy. The differences in outcome between their study and ours could be related to the use of systemic therapy, as well as inherent differences in study design and definition of high-volume.

Although the strengths of our study include the large number of patients treated in the modern era of bladder preservation, there are other notable limitations to this retrospective study. Firstly, there are a fairly limited number of bladder preservation cases in the NCDB and even the highest volume center in the analysis only treated 20 cases over the 10-year period. Additionally, in the current study both cause of death and cancer-specific survival are not available. This is particularly problematic given the multiple comorbidities often facing non-surgical bladder cancer patients and contributes to overall survival. Also, details of the chemotherapeutic agents used and their dosing are not available in the NCDB. Lastly, there is likely a selection bias with respect to treatment center that cannot be fully adjusted for on a multivariable analysis.

## CONCLUSION

In our analysis, we found that treatment at facilities with a higher volume of trimodality bladder preservation patients was associated with improved overall survival. After adjustment for covariables and using propensity matching, this association persisted. As with other disease sites, these findings are likely a result of more tumor site-specific oncologists, greater availability of a multidisciplinary team, more extensive supportive services, higher protocol compliance, and better radiation plan quality.

## CONFLICT OF INTEREST

The authors have no conflict of interest to report.