Advertisement

Association Between Treatment at a High-Volume Facility and Improved Survival for Radiation-Treated Men With High-Risk Prostate Cancer

Open AccessPublished:December 14, 2015DOI:https://doi.org/10.1016/j.ijrobp.2015.12.008

      Purpose

      Although the association between higher hospital volume and improved outcomes has been well-documented in surgery, there is little data about whether this effect exists for radiation-treated patients. We investigated whether treatment at a radiation facility that treats a high volume of prostate cancer patients is associated with improved survival for men with high-risk prostate cancer.

      Methods and Materials

      We used the National Cancer Database (NCDB) to identity patients diagnosed with prostate cancer from 2004 to 2006. The radiation case volume (RCV) of each hospital was based on its number of radiation-treated prostate cancer patients. We used propensity-score based analysis to compare the overall survival (OS) of high-risk prostate cancer patients in high versus low RCV hospitals. Primary endpoint is overall survival. Covariates adjusted for were tumor characteristics, sociodemographic factors, radiation type, and use of androgen deprivation therapy (ADT).

      Results

      A total of 19,565 radiation-treated high-risk patients were identified. Median follow-up was 81.0 months (range: 1-108 months). When RCV was coded as a continuous variable, each increment of 100 radiation-managed patients was associated with improved OS (adjusted hazard ratio [AHR]: 0.97; 95% confidence interval [CI]: 0.95-0.98; P<.0001) after adjusting for known confounders. For illustrative purposes, when RCV was dichotomized at the 80th percentile (43 patients/year), high RCV was associated with improved OS (7-year overall survival 76% vs 74%, log-rank test P=.0005; AHR: 0.91, 95% CI: 0.86-0.96, P=.0005). This association remained significant when RCV was dichotomized at 75th (37 patients/year), 90th (60 patients/year), and 95th (84 patients/year) percentiles but not the 50th (19 patients/year).

      Conclusions

      Our results suggest that treatment at centers with higher prostate cancer radiation case volume is associated with improved OS for radiation-treated men with high-risk prostate cancer.
      Summary
      After a median follow-up of 81 months, the overall survival of 19,565 radiation-treated high-risk prostate cancer patients was positively associated with the prostate cancer radiation case volume of the treating facility. Each increment of 100 patients was associated with improved survival after adjustment for baseline characteristics (adjusted hazard ratio: 0.97; 95% confidence interval: 0.95-0.98; P<.0001), suggesting that, similar to the volume effect in surgically treated patients, higher radiation case volume of the treating facility is associated with improved outcomes.

      Introduction

      Numerous studies have reported that hospital volume as well as surgeon volume are independent predictors of long-term outcomes in surgically treated cancer patients (
      • Liu C.J.
      • Chou Y.J.
      • Teng C.J.
      • et al.
      Association of surgeon volume and hospital volume with the outcome of patients receiving definitive surgery for colorectal cancer: A nationwide population-based study.
      ,
      • Luchtenborg M.
      • Riaz S.P.
      • Coupland V.H.
      • et al.
      High procedure volume is strongly associated with improved survival after lung cancer surgery.
      ,
      • Schrag D.
      • Cramer L.D.
      • Bach P.B.
      • et al.
      Influence of hospital procedure volume on outcomes following surgery for colon cancer.
      ,
      • Bach P.B.
      • Cramer L.D.
      • Schrag D.
      • et al.
      The influence of hospital volume on survival after resection for lung cancer.
      ,
      • Trinh Q.D.
      • Bjartell A.
      • Freedland S.J.
      • et al.
      A systematic review of the volume-outcome relationship for radical prostatectomy.
      ). This association was attributed to the hospital and surgeon's experience which yielded improved rates of achieving negative margins (
      • Cornish J.A.
      • Tekkis P.P.
      • Tan E.
      • et al.
      The national bowel cancer audit project: The impact of organisational structure on outcome in operative bowel cancer within the United Kingdom.
      ,
      • Gietelink L.
      • Henneman D.
      • van Leersum N.J.
      • et al.
      The influence of hospital volume on circumferential resection margin involvement: Results of the Dutch surgical colorectal audit.
      ,
      • Luryi A.L.
      • Chen M.M.
      • Mehra S.
      • et al.
      Positive surgical margins in early stage oral cavity cancer: An analysis of 20,602 cases.
      ) and higher rates of dissecting involved lymph nodes (
      • Cornish J.A.
      • Tekkis P.P.
      • Tan E.
      • et al.
      The national bowel cancer audit project: The impact of organisational structure on outcome in operative bowel cancer within the United Kingdom.
      ,
      • Siemens D.R.
      • Mackillop W.J.
      • Peng Y.
      • et al.
      Lymph node counts are valid indicators of the quality of surgical care in bladder cancer: A population-based study.
      ,
      • Siemens D.R.
      • Mackillop W.J.
      • Peng Y.
      • et al.
      Processes of care and the impact of surgical volumes on cancer-specific survival: A population-based study in bladder cancer.
      ). However, there is limited data about whether this volume effect exists in radiation-treated cancer patients (
      • Wuthrick E.J.
      • Zhang Q.
      • Machtay M.
      • et al.
      Institutional clinical trial accrual volume and survival of patients with head and neck cancer.
      ), and whether treatment at a radiation facility that treats a high case volume of cancer patients is associated with improved survival (
      • Wang E.H.
      • Rutter C.E.
      • Corso C.D.
      • et al.
      Patients selected for definitive concurrent chemoradiation at high-volume facilities achieve improved survival in stage III non-small-cell lung cancer.
      ). To explore this possibility, we conducted a nationwide cancer database analysis to specifically compare the overall survival of radiation-treated high-risk prostate cancer in high versus low radiation case volume (RCV) facilities.

      Methods and Materials

       Data source and study population

      Our study population was derived from the National Cancer Database (NCDB), a joint program of the Commission on Cancer (CoC) and the American Cancer Society (ACS). NCDB is a nationwide hospital-based database capturing 70% of newly diagnosed cases and those cases are diagnosed and treated at CoC accredited cancer programs (
      • Mohanty S.
      • Bilimoria K.Y.
      Comparing national cancer registries: The National Cancer Data Base (NCDB) and the Surveillance, Epidemiology, and End Results (SEER) program.
      ). Patients diagnosed with non-metastatic prostate adenocarcinoma (site code: C61.9, International Classification of Disease for Oncology code, 3rd edition: 8140) and treated only with external beam radiation therapy (EBRT) and/or brachytherapy in 2004 to 2006 were identified (N=114,530). In our analysis, we excluded patients with unknown tumor stage (n=2889), Gleason score (n=4238) or prostate-specific antigen (PSA) (n=5562) level to ensure precisely defined National Comprehensive Cancer Network (NCCN) high-risk patients. Patients whose radiation therapy was delivered at multiple places or whose facility information was unknown were further excluded (n=9304). In the final study population, there were 92,537 patients and 1099 facilities. Patients were followed until 2012. The institutional review board approved this study. Figure 1 summarizes the study population selection process.
      Figure thumbnail gr1
      Fig. 1Flow chart for patient selection. Abbreviations: NCDB = National Cancer Database; RT = radiation therapy.

       Primary endpoint and covariates

      The primary endpoint is the effect of RCV on overall survival. To determine the RCV of each facility, the cumulative radiation-treated prostate cancer patients from all risk groups at each facility during 2004-2006 was calculated. We first analyzed RCV as a continuous variable. Then for illustrative purposes, we dichotomized RCV into high versus low RCV at the 80th percentile of numbers of patients treated per facility. We chose the 80th percentile as the cutoff because the top 20% of facilities treated 54% of all the prostate cancer cases and 50% of the high-risk disease, meaning that about half of patients were treated at high-volume versus low-volume facilities under this definition. We also performed sensitivity analyses using various cutoffs for percentiles of RCV (50th, 75th, 90th, or 95th). After the RCV of each facility was defined, we investigated the association between RCV and overall survival specifically in high-risk prostate cancer (N=19,565). We chose to study exclusively high-risk prostate cancer because these are the patients who are most likely to die from prostate cancer and in whom we would be most likely to observe a difference in survival over the follow-up time of 81 months.
      Other covariates included in the analysis were sociodemographic variables such as age (coded as a continuous variable), race and insurance status, household income, residence type, percentage of education levels less than high school for each patient's area of residence; clinical variables included Charlson-Deyo comorbidity score (0, 1, ≥2), tumor stage (AJCC 6th Cancer Staging Manual, American Joint Committee on Cancer, 2002), prostate-specific antigen (PSA, coded as a continuous variable), Gleason score (≤6, 7, and 8-10), radiation type (external beam radiation therapy [EBRT] only, brachytherapy only, EBRT plus brachytherapy), use of androgen deprivation therapy (ADT), and hospital setting (academic/research program was defined as academic; comprehensive community cancer center, community cancer center and others were defined as a non-academic setting). Income and education data were based on 2000 US Census data (

      National Cancer Data Base Data Dictionary PUF 2012, http://ncdbpuf.facs.org/content/median-household-income-2000.

      ,

      National Cancer Data Base Data Dictionary PUF 2012, http://ncdbpuf.facs.org/content/without-high-school-degree-2000.

      ), and residence type was determined using the 2003 US Department of Agriculture Economic Research Service (

      National Cancer Data Base Data Dictionary PUF 2012, http://ncdbpuf.facs.org/content/urban-rural-continuum-2003.

      ), and all data were provided by NCDB.

       Statistical analysis

      Descriptive statistics were used to present the baseline characteristics. Categorical variables were assessed with χ2 test; continuous variables were compared with Student t test or Mann-Whiney U test as appropriate. We first modeled RCV as a continuous variable and used Cox-regression analysis to determine the association between RCV and overall survival. For illustrative purposes, RCV was dichotomized at the 80th percentile into high RCV versus low RCV. We presented the effect of RCV on overall survival using Kaplan-Meier curves before and after propensity score-based adjustment using inverse propensity score weighting (IPSW). The inverse-propensity score weight was first calculated as 1/(propensity score) for patients treated at high-volume facility and 1/(1-propensity score) for patients treated at low-volume facility and then further stabilized the weight with a constant as described by Robins et al (
      • Robins J.M.
      • Hernan M.A.
      • Brumback B.
      Marginal structural models and causal inference in epidemiology.
      ). Propensity score was generated on a multivariate logistic regression model to estimate the probability of being treated in a high- or low-RCV facility. The covariates included in the propensity model were all the patient baseline characteristics as provided in Table 1.
      Table 1Patient baseline characteristics
      CharacteristicHigh volumeLow volumeP value
      UnadjustedIPTW adjusted
      Total patient number (%)9817 (50.2)9748 (49.8)
      Median patient number (IQR)223 (161-348)76 (47-103)
      Mean age (95% CI)69.2 (69.0-69.3)69.7 (69.5-69.9)<.0001.88
      Race (%)<.0001.97
       Non-Hispanic whites7303 (74.4)7469 (76.6)
       African American1595 (16.3)1533 (15.7)
       Hispanic whites356 (3.6)378 (3.9)
       Others387 (3.9)262 (2.7)
       Unknown176 (1.8)106 (1.1)
      Insurance status<.0001.99
       None112 (1.1)158 (1.6)
       Private3453 (35.2)2632 (27)
       Medicaid184 (1.9)265 (2.7)
       Medicare5751 (58.6)6252 (64.1)
       Others160 (1.6)238 (2.4)
       Unknown157 (1.6)203 (2.1)
      Charlson comorbidity index
       08892 (90.6)8750 (89.8).105.98
       1803 (8.2)852 (8.7)
       2+122 (1.2)146 (1.5)
      Median PSA (IQR)14.5 (27.7)20.0 (34.0)<.0001.90
      T stages<.0001.99
       T14511 (46.0)4311 (44.2)
       T23654 (37.2)3933 (40.4)
       T31568 (16.0)1424 (14.6)
       T484 (0.9)80 (0.8)
      Gleason score<.0001.66
       ≤61651 (16.8)1906 (19.6)
       72072 (21.1)2206 (22.6)
       8-106094 (62.1)5636 (57.8)
      RT type<.0001.82
       EBRT6784 (69.1)7520 (77.1)
       Brachytherapy1355 (13.8)1432 (14.7)
       EBRT plus brachytherapy1678 (17.1)796 (8.2)
      ADT use.58.89
       Yes7307 (74.4)7222 (74.1)
       No2510 (25.6)2526 (25.9)
      Hospital setting<.0001.75
       Academic/research program3749 (38.2)2074 (21.3)
       Comprehensive community cancer program5892 (60.0)5684 (58.3)
       Community cancer program176 (1.8)1983 (20.3)
       Other types0 (0)7 (0.1)
      Household income<.0001.99
       <$30,0001124 (11.5)1468 (15.1)
       $30,000-34,9991466 (14.9)1976 (20.3)
       $35,000-45,9992498 (25.5)2702 (27.7)
       ≥$46,0004352 (44.3)3233 (33.2)
       Unknown377 (3.8)369 (3.8)
      Education level<.0001.97
       ≥29%1478 (15.1)1718 (17.6)
       20%-28.9%2008 (20.5)2350 (24.1)
       14-19.9%2235 (22.8)2345 (24.1)
       <14%3719 (37.9)2965 (30.4)
       Unknown377 (3.8)370 (3.8)
      Residence<.0001.81
       Metropolitan8240 (83.9)7081 (72.6)
       Urban1084 (11.0)2053 (21.1)
       Rural147 (1.5)289 (3.0)
       Unknown346 (3.5)325 (3.3)
      Abbreviations: ADT = androgen deprivation therapy; EBRT = external beam radiation therapy; IPSW = inverse propensity score weighting; IQR = interquartile range; PSA = prostate-specific antigen; RT = radiation therapy.
      Percentage may not add up to 100 due to rounding. Education level is the percentage of residents with education level < high school within the same ZIP code of the patient's residence. Education level and household income were derived from 2000 US Census data. Residence was derived from 2003 US Department of Agriculture Economic Research Service. Household income, education level, and residence type were determined at the county level.
      We compared the results from the IPSW-adjusted model to multivariate Cox proportional hazard models. All statistical analyses were performed using SAS version 9.4 software (SAS Institute Inc, Cary, NC) and figures were plotted with STATA version 13.1 software (StataCorp, College Station, TX). We used a two-sided P value of <.05 in all analyses as criteria for statistical significance.

      Results

       Baseline characteristics stratified by high versus low radiation case volume facilities

      There were 19,565 radiation-treated high-risk prostate cancer patients in the study population. 9817 (50.2%) were treated at high RCV facilities and 9748 (49.8%) were treated at low RCV facilities based on the 80th percentile dichotomization. The median number of radiation-treated high-risk prostate cancer patients was 223 (interquartile range: 161-348) in high RCV facility and 76 (interquartile range: 47-103) in low RCV facility during 2004 to 2006. The median follow-up time was 81 months (range: 1-108 months). Patient sociodemographic and clinical characteristics stratified on RCV are summarized in Table 1. Compared with low RCV facilities, high RCV facilities were more likely to be an academic setting (38.2% vs 21.3%, respectively) and more likely to be in a metropolitan area (83.9% vs 72.6, respectively). They also treated a higher proportion of Gleason score 8 to 10 patients (62.1 vs 57.8, respectively) and T3 patients (16.0 vs 14.6, respectively) and were more likely to give combination therapy of EBRT with brachytherapy (17.1% vs 8.2, respectively), although the median PSA was lower (14.5 vs 20.0, respectively; all P<.0001). ADT use (P=.58) and Charlson comorbidity score (P=.105) did not differ by high- versus low-RCV facility. After IPSW adjustment, all baseline characteristics were distributed evenly without statistical significance (all P>.05) between the high- and low-RCV facilities.

       Association between radiation case volume and survival

      When RCV was analyzed as a continuous variable, each increment of 100 patients treated was associated with improved overall survival (adjusted hazard ratio [AHR]: 0.97, 95% confidence interval [CI]: 0.95-0.98, P<.0001) after adjusting for known confounders (Table 2).
      Table 2Multivariate Cox regression analysis
      VariableAHR (95% CI)P value
      Increment of 100 patients0.97 (0.95-0.98)<.0001
      Increment of 1-y age1.05 (1.04-1.05)<.0001
      Increment of 1-unit PSA1.004 (1.003-1.005)<.0001
      Race
       Black1.00 (0.91-1.09).93
       Hispanic-white0.70 (0.59-0.84)<.0001
       Other0.78 (0.65-0.92).004
       Non-Hispanic whiteRefNA
      Insurance status
       Private insurance0.97 (0.71-1.31).83
       Medicaid1.41 (0.99-2.00).05
       Medicare1.04 (0.77-1.40).82
       Other1.26 (0.87-1.81).22
       UninsuredRefNA
      Charlson comorbidity score
       2+2.28 (1.90-2.73)<.0001
       11.50 (1.37-1.64)<.0001
       0RefNA
      Tumor stage
       T43.06 (2.46-3.79)<.0001
       T31.41 (1.30-1.54)<.0001
       T21.10 (1.03-1.17).0033
       T1RefNA
      Gleason score
       8-101.77 (1.60-1.95)<.0001
       71.23 (1.11-1.37).0001
       <=6RefNA
      Hospital setting
       Academic0.97 (0.90-1.03).32
       NonacademicRefNA
      Use of ADT
       ADT0.97 (0.90-1.04).40
       No ADTRefNA
      Radiation type
       EBRT plus brachytherapy0.67 (0.61-0.74)<.0001
       Brachytherapy only0.83 (0.76-0.92).0002
       EBRT onlyRefNA
      Household income
       ≥$46,0000.81 (0.73-0.91).0003
       $35,000-45,9990.86 (0.78-0.95).003
       $30,000-34,9991.00 (0.91-1.09).92
       <$30,000RefNA
      Education level
       <14%0.82 (0.73-0.92).001
       14%-19.9%0.87 (0.79-0.96).005
       20-28.9%0.94 (0.86-1.03).18
       ≥29%RefNA
      Residence
       Rural1.03 (0.86-1.23).75
       Urban1.01 (0.94-1.10).74
       MetropolitanRefNA
      Abbreviations: ADT = androgen deprivation therapy; AHR = adjusted hazard ratio; EBRT = external beam radiation therapy; NA = not applicable; PSA = prostate-specific antigen.
      The Kaplan-Meier curves illustrate overall survival stratified by high and low RCV before and after IPSW adjustment (Fig. 2, Fig. 3). In unadjusted analysis, the 7-year overall survival was statistically higher in high RCV facilities (77% vs 73%, respectively; log-rank test P<.0001) (Fig. 2). This survival benefit was smaller but persisted after IPSW adjustment (76% vs 74%, respectively, log-rank test P=.0005) (Fig. 3).
      Figure thumbnail gr3
      Fig. 3Inverse-propensity score weighting adjusted overall survival.
      Table 3 presents overall survival hazard ratios comparing high with low RCV facilities from the unadjusted Cox regression, multivariate Cox regression, and IPSW-adjusted models. On multivariate Cox regression, high versus low RCV was associated with improved OS (AHR: 0.91, 95% CI: 0.86-0.97, P=.002), and this was confirmed with nearly identical results for IPSW analysis (AHR: 0.91, 95% CI: 0.86-0.96, P=.0005). In sensitivity analysis, in models in which radiation case volume was dichotomized at other cutoffs, improved overall survival was observed at the 75th, 90th, and 95th percentiles, corresponding to >37, >60, and >84 prostate patients, respectively, treated at the facility per year. However, at the 50th percentile (>19 patients per year), there were no survival differences (AHR: 0.99, 95% CI: 0.92-1.07, P=.78) (Table 4).
      Table 3Overall survival hazard ratios comparing high versus low case volume facilities
      Continuous variable ModelAHR (95% CI)P value
      Increment of 100 patients0.97 (0.95-0.98)<.0001
      Binary modelsCrude HR (95% CI)P value
      Unadjusted0.83 (0.79-0.88)<.0001
      AHR (95% CI)P value
      Multivariate Cox regression model0.91 (0.86-0.97).0021
      IPSW-adjusted Cox regression model0.91 (0.86-0.96).0005
      Abbreviations: AHR = adjusted hazard ratio; HR = hazard ratio.
      Models were adjusted for age, race, insurance status, tumor stage, Gleason score, prostate-specific antigen level, Charlson comorbidity score, use of androgen deprivation therapy, radiation modalities, household income, education level, hospital type, and residence type.
      Table 4Sensitivity analysis
      Binary model percentile for cutoff (number of RT-treated patients)% of high-risk patients treated in high-volume facilityAHR (95% CI)P value
      95th (>84 vs ≤ 84 per year)200.90 (0.84-0.98).013
      90th (>60 vs ≤ 60 per year)330.90 (0.85-0.96).002
      80th (>43 vs ≤ 43 per year)500.91 (0.86-0.97).001
      75th (>37 vs ≤ 37 per year)580.91 (0.86-0.97).002
      50th (>19 vs ≤ 19 per year)830.99 (0.92-1.07).78
      Abbreviation: AHR = adjusted hazard ratio.

      Discussion

      To our knowledge, our study is one of the first and largest demonstrating that similar to the hospital and surgeon volume effect on surgically treated patients (
      • Trinh Q.D.
      • Bjartell A.
      • Freedland S.J.
      • et al.
      A systematic review of the volume-outcome relationship for radical prostatectomy.
      ), the higher radiation case volume of the treating facility is associated with improved overall survival in high-risk prostate cancer patients. We believe this case volume effect is likely true for other disease sites besides prostate cancer, as Wang et al (
      • Wang E.H.
      • Rutter C.E.
      • Corso C.D.
      • et al.
      Patients selected for definitive concurrent chemoradiation at high-volume facilities achieve improved survival in stage III non-small-cell lung cancer.
      ) also recently reported improved survival for patients treated at high-volume facilities in the setting of concurrent chemoradiation for stage III non-small cell lung cancer.
      There are multiple potential reasons underlying this effect. One reason may relate to the quality of the radiation that can be delivered by a high case volume radiation center. Radiation requires close collaboration between the treating physicians, dosimetrists, physicists, and therapists to ensure exactly the right targets are outlined, the optimal plan is generated, and right targets are hit every day. Judgment on the part of radiation oncologist is required throughout the contouring process to decide on which areas should be treated and which should be left out (
      • Hong T.S.
      • Tome W.A.
      • Harari P.M.
      Heterogeneity in head and neck IMRT target design and clinical practice.
      ), and expertise gained from working at a higher case volume facility can lead to better judgment about which areas are likely to harbor disease (
      • Wuthrick E.J.
      • Zhang Q.
      • Machtay M.
      • et al.
      Institutional clinical trial accrual volume and survival of patients with head and neck cancer.
      ). Physicians at higher case volume centers may also feel more comfortable with using higher doses, as shown by their increased use of the external beam plus brachytherapy combination, which delivers much higher doses to the prostate than external beam radiation alone and which has been associated with significantly improved outcomes in a recently reported randomized trial (
      • Morris W.J.
      • Tyldesley S.
      • Pai H.H.
      • et al.
      ASCENDE-RT: A multicenter, randomized trial of dose-escalated external beam radiation therapy (EBRT-B) versus low-dose-rate brachytherapy (LDR-B) for men with unfavorable-risk localized prostate cancer (abstr 3).
      ). Interestingly, we observed our effect even after adjusting for the difference in treatment technique and dose, but the findings suggest that higher volume physicians may be generally more comfortable with taking on a higher risk of complications to achieve better results.
      Local treatment with radiation has been shown to deliver a 10% overall survival benefit at 10 years for locally advanced prostate cancer (
      • Widmark A.
      • Klepp O.
      • Solberg A.
      • et al.
      Endocrine treatment, with or without radiotherapy, in locally advanced prostate cancer (SPCG-7/SFUO-3): An open randomised phase III trial.
      ), so it is plausible that the 2% difference at 7 years that we observed in overall survival is partly related to radiation quality, but another reason for the finding could relate to the experience of the providers in other disciplines at centers that treat high volumes of prostate cancer patients with radiation. The care of cancer patients is a sophisticated process involving timely diagnosis, early initiation of guideline-based treatment and post-treatment follow-up. The care of prostate cancer patients necessitates multidisciplinary collaboration (
      • Aizer A.A.
      • Paly J.J.
      • Efstathiou J.A.
      Multidisciplinary care and management selection in prostate cancer.
      ,
      • Fowler Jr., F.J.
      • McNaughton Collins M.
      • Albertsen P.C.
      • et al.
      Comparison of recommendations by urologists and radiation oncologists for treatment of clinically localized prostate cancer.
      ,
      • Bellardita L.
      • Donegani S.
      • Spatuzzi A.L.
      • et al.
      Multidisciplinary versus one-on-one setting: A qualitative study of clinicians' perceptions of their relationship with patients with prostate cancer.
      ,
      • Gomella L.G.
      • Lin J.
      • Hoffman-Censits J.
      • et al.
      Enhancing prostate cancer care through the multidisciplinary clinic approach: A 15-year experience.
      ) among urologists, radiation oncologists, medical oncologists, radiologists, and pathologists, and a high case volume facility has higher likelihood of having experts in all the related specialties onsite. In such high-volume centers, care is likely to be well coordinated between specialists and patients are likely to be treated in a consistent fashion (
      • Wuthrick E.J.
      • Zhang Q.
      • Machtay M.
      • et al.
      Institutional clinical trial accrual volume and survival of patients with head and neck cancer.
      ,
      • Gray P.J.
      • Fedewa S.A.
      • Shipley W.U.
      • et al.
      Use of potentially curative therapies for muscle-invasive bladder cancer in the United States: Results from the National Cancer Data Base.
      ,
      • Monson J.R.
      • Probst C.P.
      • Wexner S.D.
      • et al.
      Failure of evidence-based cancer care in the United States: The association between rectal cancer treatment, cancer center volume, and geography.
      ,
      • Cliby W.A.
      • Powell M.A.
      • Al-Hammadi N.
      • et al.
      Ovarian cancer in the United States: Contemporary patterns of care associated with improved survival.
      ,
      • Ho V.K.
      • Damhuis R.A.
      • Hartgrink H.H.
      Adherence to national guidelines for gastric cancer in the Netherlands: A retrospective population-based audit. International journal of cancer.
      ,
      • Merkow R.P.
      • Bilimoria K.Y.
      • Chow W.B.
      • et al.
      Variation in lymph node examination after esophagectomy for cancer in the United States.
      ), including optimal staging and selection of duration of androgen deprivation therapy, leading to better outcomes.
      The primary implication of our study is that similar to patients seeking the best surgical outcomes, patients seeking the best radiation outcomes should also seek out facilities that treat a high volume of patients with radiation. As is the case with urologic surgery (
      • Trinh Q.D.
      • Bjartell A.
      • Freedland S.J.
      • et al.
      A systematic review of the volume-outcome relationship for radical prostatectomy.
      ,
      • Santos F.
      • Zakaria A.S.
      • Kassouf W.
      • et al.
      High hospital and surgeon volume and its impact on overall survival after radical cystectomy among patients with bladder cancer in Quebec.
      ,
      • Yu H.Y.
      • Hevelone N.D.
      • Lipsitz S.R.
      • et al.
      Hospital volume, utilization, costs and outcomes of robot-assisted laparoscopic radical prostatectomy.
      ,
      • Carter S.C.
      • Lipsitz S.
      • Shih Y.C.
      • et al.
      Population-based determinants of radical prostatectomy operative time.
      ,
      • Sammon J.D.
      • Karakiewicz P.I.
      • Sun M.
      • et al.
      Robot-assisted versus open radical prostatectomy: The differential effect of regionalization, procedure volume and operative approach.
      ) and cardiac care (
      • Luft H.S.
      • Bunker J.P.
      • Enthoven A.C.
      Should operations be regionalized? The empirical relation between surgical volume and mortality.
      ,
      • Gauvreau K.
      Reevaluation of the volume-outcome relationship for pediatric cardiac surgery.
      ,
      • Arnold R.
      • Ley-Zaporozhan J.
      • Ley S.
      • et al.
      Outcome after mechanical aortic valve replacement in children and young adults.
      ), our results suggest that care is best when it is clustered at high-volume facilities.
      Although ours is the largest study and the only one to focus on high-risk prostate cancer, evidence of a volume outcome effect could also be gleaned from a secondary analysis of the Radiation Therapy Oncology Group 0129 trial (
      • Wuthrick E.J.
      • Zhang Q.
      • Machtay M.
      • et al.
      Institutional clinical trial accrual volume and survival of patients with head and neck cancer.
      ) and a Surveillance, Epidemiology, and End Results-Medicare study (
      • Chen A.B.
      • D'Amico A.V.
      • Neville B.A.
      • et al.
      Provider case volume and outcomes following prostate brachytherapy.
      ) suggesting that among men with prostate cancer treated with brachytherapy, patients treated by higher volume physicians had a near-significant decrease in all-cause mortality (HR: 0.95/100 cases; P=.05). Although SEER-Medicare only includes men over 65 years of age and captures only 28% of that population, our NCDB captures 70% of the population, and so our study is more generalizable to the greater US population. By focusing on a high-risk population and having a study nearly 4 times larger (19,565 vs 5595 patients, respectively), we had ample power to find a very significant association between radiation case volume and overall survival (P<.0001).
      Certain limitations of our study should be considered. First, NCDB is a hospital-based cancer registry and only cases diagnosed and treated at Commission on Cancer (CoC) accredited programs were reported. Cases from non-CoC-accredited programs were unavailable in NCDB. However, this makes the point that even within CoC-accredited programs, higher radiation case volume was associated with better overall survival. Second, our study design was an observational study, and although propensity-score based comparative effective analysis was adopted, we are still subject to unmeasured confounders. Third, the NCDB only reports the overall survival instead of cancer-specific survival, so there is a possibility that patients who were healthier at baseline are the ones who tend to be treated at higher-volume cancer centers. However, we attempted to account for this limitation with adjustment for the Charlson comorbidity score in the analysis. Fourth, although we adjusted for socioeconomic status with household income and education level, those information was the county level data of a patient's residence and individual socioeconomic status might not be fully accounted for. Patients treated at higher RCV facilities might be of higher socioeconomic status which explained improved overall survival. Fifth, we didn't have information about the duration of androgen deprivation therapy.

      Conclusions

      Despite the potential limitations stated above, our study may be the first to report a highly significant association between radiation case volume and improved overall survival in a radiation-treated patient population. As patients often do for their surgical care, consideration should be given to favoring radiation treatment at a center that treats a high volume of patients with a patient's particular disease.

      References

        • Liu C.J.
        • Chou Y.J.
        • Teng C.J.
        • et al.
        Association of surgeon volume and hospital volume with the outcome of patients receiving definitive surgery for colorectal cancer: A nationwide population-based study.
        Cancer. 2015; 121: 2782-2790
        • Luchtenborg M.
        • Riaz S.P.
        • Coupland V.H.
        • et al.
        High procedure volume is strongly associated with improved survival after lung cancer surgery.
        J Clin Oncol. 2013; 31: 3141-3146
        • Schrag D.
        • Cramer L.D.
        • Bach P.B.
        • et al.
        Influence of hospital procedure volume on outcomes following surgery for colon cancer.
        JAMA. 2000; 284: 3028-3035
        • Bach P.B.
        • Cramer L.D.
        • Schrag D.
        • et al.
        The influence of hospital volume on survival after resection for lung cancer.
        N Engl J Med. 2001; 345: 181-188
        • Trinh Q.D.
        • Bjartell A.
        • Freedland S.J.
        • et al.
        A systematic review of the volume-outcome relationship for radical prostatectomy.
        Eur Urol. 2013; 64: 786-798
        • Cornish J.A.
        • Tekkis P.P.
        • Tan E.
        • et al.
        The national bowel cancer audit project: The impact of organisational structure on outcome in operative bowel cancer within the United Kingdom.
        Surg Oncol. 2011; 20: e72-e77
        • Gietelink L.
        • Henneman D.
        • van Leersum N.J.
        • et al.
        The influence of hospital volume on circumferential resection margin involvement: Results of the Dutch surgical colorectal audit.
        Ann Surg. 2015; (PMID: 25790120)
        • Luryi A.L.
        • Chen M.M.
        • Mehra S.
        • et al.
        Positive surgical margins in early stage oral cavity cancer: An analysis of 20,602 cases.
        Head Neck Surg. 2014; 151: 984-990
        • Siemens D.R.
        • Mackillop W.J.
        • Peng Y.
        • et al.
        Lymph node counts are valid indicators of the quality of surgical care in bladder cancer: A population-based study.
        Urol Oncol. 2015; 33: 425.e15-425.e23
        • Siemens D.R.
        • Mackillop W.J.
        • Peng Y.
        • et al.
        Processes of care and the impact of surgical volumes on cancer-specific survival: A population-based study in bladder cancer.
        Urology. 2014; 84: 1049-1057
        • Wuthrick E.J.
        • Zhang Q.
        • Machtay M.
        • et al.
        Institutional clinical trial accrual volume and survival of patients with head and neck cancer.
        J Clin Oncol. 2015; 33: 156-164
        • Wang E.H.
        • Rutter C.E.
        • Corso C.D.
        • et al.
        Patients selected for definitive concurrent chemoradiation at high-volume facilities achieve improved survival in stage III non-small-cell lung cancer.
        J Thorac Oncol. 2015; 10: 937-943
        • Mohanty S.
        • Bilimoria K.Y.
        Comparing national cancer registries: The National Cancer Data Base (NCDB) and the Surveillance, Epidemiology, and End Results (SEER) program.
        J Surg Oncol. 2014; 109: 629-630
      1. National Cancer Data Base Data Dictionary PUF 2012, http://ncdbpuf.facs.org/content/median-household-income-2000.

      2. National Cancer Data Base Data Dictionary PUF 2012, http://ncdbpuf.facs.org/content/without-high-school-degree-2000.

      3. National Cancer Data Base Data Dictionary PUF 2012, http://ncdbpuf.facs.org/content/urban-rural-continuum-2003.

        • Robins J.M.
        • Hernan M.A.
        • Brumback B.
        Marginal structural models and causal inference in epidemiology.
        Epidemiology. 2000; 11: 550-560
        • Hong T.S.
        • Tome W.A.
        • Harari P.M.
        Heterogeneity in head and neck IMRT target design and clinical practice.
        Radiother Oncol. 2012; 103: 92-98
        • Morris W.J.
        • Tyldesley S.
        • Pai H.H.
        • et al.
        ASCENDE-RT: A multicenter, randomized trial of dose-escalated external beam radiation therapy (EBRT-B) versus low-dose-rate brachytherapy (LDR-B) for men with unfavorable-risk localized prostate cancer (abstr 3).
        J Clin Oncol. 2015; 33
        • Widmark A.
        • Klepp O.
        • Solberg A.
        • et al.
        Endocrine treatment, with or without radiotherapy, in locally advanced prostate cancer (SPCG-7/SFUO-3): An open randomised phase III trial.
        Lancet. 2009; 373: 301-308
        • Aizer A.A.
        • Paly J.J.
        • Efstathiou J.A.
        Multidisciplinary care and management selection in prostate cancer.
        Semin Radiat Oncol. 2013; 23: 157-164
        • Fowler Jr., F.J.
        • McNaughton Collins M.
        • Albertsen P.C.
        • et al.
        Comparison of recommendations by urologists and radiation oncologists for treatment of clinically localized prostate cancer.
        JAMA. 2000; 283: 3217-3222
        • Bellardita L.
        • Donegani S.
        • Spatuzzi A.L.
        • et al.
        Multidisciplinary versus one-on-one setting: A qualitative study of clinicians' perceptions of their relationship with patients with prostate cancer.
        J Oncol Pract. 2011; 7: e1-e5
        • Gomella L.G.
        • Lin J.
        • Hoffman-Censits J.
        • et al.
        Enhancing prostate cancer care through the multidisciplinary clinic approach: A 15-year experience.
        J Oncol Pract. 2010; 6: e5-e10
        • Gray P.J.
        • Fedewa S.A.
        • Shipley W.U.
        • et al.
        Use of potentially curative therapies for muscle-invasive bladder cancer in the United States: Results from the National Cancer Data Base.
        Eur Urol. 2013; 63: 823-829
        • Monson J.R.
        • Probst C.P.
        • Wexner S.D.
        • et al.
        Failure of evidence-based cancer care in the United States: The association between rectal cancer treatment, cancer center volume, and geography.
        Ann Surg. 2014; 260 (discussion 631): 625-631
        • Cliby W.A.
        • Powell M.A.
        • Al-Hammadi N.
        • et al.
        Ovarian cancer in the United States: Contemporary patterns of care associated with improved survival.
        Gynecol Oncol. 2015; 136: 11-17
        • Ho V.K.
        • Damhuis R.A.
        • Hartgrink H.H.
        Adherence to national guidelines for gastric cancer in the Netherlands: A retrospective population-based audit. International journal of cancer.
        Int J Cancer. 2013; 132: 1156-1161
        • Merkow R.P.
        • Bilimoria K.Y.
        • Chow W.B.
        • et al.
        Variation in lymph node examination after esophagectomy for cancer in the United States.
        Arch Surg. 2012; 147: 505-511
        • Santos F.
        • Zakaria A.S.
        • Kassouf W.
        • et al.
        High hospital and surgeon volume and its impact on overall survival after radical cystectomy among patients with bladder cancer in Quebec.
        World J Urol. 2015; 33: 1323-1330
        • Yu H.Y.
        • Hevelone N.D.
        • Lipsitz S.R.
        • et al.
        Hospital volume, utilization, costs and outcomes of robot-assisted laparoscopic radical prostatectomy.
        J Urol. 2012; 187: 1632-1637
        • Carter S.C.
        • Lipsitz S.
        • Shih Y.C.
        • et al.
        Population-based determinants of radical prostatectomy operative time.
        BJU Int. 2014; 113: E112-E118
        • Sammon J.D.
        • Karakiewicz P.I.
        • Sun M.
        • et al.
        Robot-assisted versus open radical prostatectomy: The differential effect of regionalization, procedure volume and operative approach.
        J Urol. 2013; 189: 1289-1294
        • Luft H.S.
        • Bunker J.P.
        • Enthoven A.C.
        Should operations be regionalized? The empirical relation between surgical volume and mortality.
        N Engl J Med. 1979; 301: 1364-1369
        • Gauvreau K.
        Reevaluation of the volume-outcome relationship for pediatric cardiac surgery.
        Circulation. 2007; 115: 2599-2601
        • Arnold R.
        • Ley-Zaporozhan J.
        • Ley S.
        • et al.
        Outcome after mechanical aortic valve replacement in children and young adults.
        Ann Thorac Surg. 2008; 85: 604-610
        • Chen A.B.
        • D'Amico A.V.
        • Neville B.A.
        • et al.
        Provider case volume and outcomes following prostate brachytherapy.
        J Urol. 2009; 181 (discussion 118): 113-118

      Comments

      Commenting Guidelines

      To submit a comment for a journal article, please use the space above and note the following:

      • We will review submitted comments as soon as possible, striving for within two business days.
      • This forum is intended for constructive dialogue. Comments that are commercial or promotional in nature, pertain to specific medical cases, are not relevant to the article for which they have been submitted, or are otherwise inappropriate will not be posted.
      • We require that commenters identify themselves with names and affiliations.
      • Comments must be in compliance with our Terms & Conditions.
      • Comments are not peer-reviewed.