[Skip to Content]
[Skip to Content Landing]
Figure.
Six-Month Conditional Landmark Inverse Probability of Treatment Weighting–Adjusted Kaplan-Meier Analysis of Overall Survival for Patients Who Received Neoadjuvant Chemotherapy and Radical Cystectomy Followed by Adjuvant Chemotherapy vs Observation for pT3/T4 and/or pN+ Urothelial Carcinoma of the Bladder
Six-Month Conditional Landmark Inverse Probability of Treatment Weighting–Adjusted Kaplan-Meier Analysis of Overall Survival for Patients Who Received Neoadjuvant Chemotherapy and Radical Cystectomy Followed by Adjuvant Chemotherapy vs Observation for pT3/T4 and/or pN+ Urothelial Carcinoma of the Bladder

Data are weighted proportions and not absolute numbers.

Table 1.  
Baseline Characteristics of the Study Patients
Baseline Characteristics of the Study Patients
Table 2.  
Multivariable Logistic Regression Model Determining the Receipt of Neoadjuvant Chemotherapy and Radical Cystectomy Followed by Adjuvant Chemotherapy vs Observation
Multivariable Logistic Regression Model Determining the Receipt of Neoadjuvant Chemotherapy and Radical Cystectomy Followed by Adjuvant Chemotherapy vs Observation
1.
Alfred Witjes  J, Lebret  T, Compérat  EM,  et al.  Updated 2016 EAU guidelines on muscle-invasive and metastatic bladder cancer.  Eur Urol. 2017;71(3):462-475.PubMedGoogle ScholarCrossref
2.
Yafi  FA, Aprikian  AG, Chin  JL,  et al.  Contemporary outcomes of 2287 patients with bladder cancer who were treated with radical cystectomy: a Canadian multicentre experience.  BJU Int. 2011;108(4):539-545.PubMedGoogle ScholarCrossref
3.
Advanced Bladder Cancer (ABC) Meta-analysis Collaboration.  Adjuvant chemotherapy for invasive bladder cancer (individual patient data).  Cochrane Database Syst Rev. 2006;(2):CD006018.PubMedGoogle Scholar
4.
Leow  JJ, Martin-Doyle  W, Rajagopal  PS,  et al.  Adjuvant chemotherapy for invasive bladder cancer: a 2013 updated systematic review and meta-analysis of randomized trials.  Eur Urol. 2014;66(1):42-54.PubMedGoogle ScholarCrossref
5.
Galsky  MD, Stensland  KD, Moshier  E,  et al.  Effectiveness of adjuvant chemotherapy for locally advanced bladder cancer.  J Clin Oncol. 2016;34(8):825-832.PubMedGoogle ScholarCrossref
6.
Zargar-Shoshtari  K, Kongnyuy  M, Sharma  P,  et al.  Clinical role of additional adjuvant chemotherapy in patients with locally advanced urothelial carcinoma following neoadjuvant chemotherapy and cystectomy.  World J Urol. 2016;34(11):1567-1573.PubMedGoogle ScholarCrossref
7.
Sonpavde  G, Pond  GR, Choueiri  TK,  et al.  Single-agent taxane versus taxane-containing combination chemotherapy as salvage therapy for advanced urothelial carcinoma.  Eur Urol. 2016;69(4):634-641.PubMedGoogle ScholarCrossref
8.
Stuart  EA, Azur  M, Frangakis  C, Leaf  P.  Multiple imputation with large data sets: a case study of the Children’s Mental Health Initiative.  Am J Epidemiol. 2009;169(9):1133-1139.PubMedGoogle ScholarCrossref
9.
Rubin  DB.  Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons; 1987.Crossref
10.
Austin  PC.  An introduction to propensity score methods for reducing the effects of confounding in observational studies.  Multivariate Behav Res. 2011;46(3):399-424.PubMedGoogle ScholarCrossref
11.
Lemeshow  S, Hosmer  DW  Jr.  A review of goodness of fit statistics for use in the development of logistic regression models.  Am J Epidemiol. 1982;115(1):92-106.PubMedGoogle ScholarCrossref
12.
Cole  SR, Hernán  MA.  Adjusted survival curves with inverse probability weights.  Comput Methods Programs Biomed. 2004;75(1):45-49.PubMedGoogle ScholarCrossref
13.
Austin  PC.  The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments.  Stat Med. 2014;33(7):1242-1258.PubMedGoogle ScholarCrossref
14.
Clark  PE, Agarwal  N, Biagioli  MC,  et al; National Comprehensive Cancer Network (NCCN).  Bladder cancer.  J Natl Compr Canc Netw. 2013;11(4):446-475.PubMedGoogle ScholarCrossref
Brief Report
February 2018

Adjuvant Chemotherapy vs Observation for Patients With Adverse Pathologic Features at Radical Cystectomy Previously Treated With Neoadjuvant Chemotherapy

Author Affiliations
  • 1Center for Surgery and Public Health, Division of Urological Surgery, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
  • 2Center for Outcomes Research, Analytics, and Evaluation, Vattikuti Urology Institute, Henry Ford Health System, Detroit, Michigan
  • 3Department of Urology, Pitié Salpétrière Hospital, Assistance Publique des Hôpitaux de Paris, Pierre and Marie Curie University, Paris, France
  • 4Department of Radiation Oncology, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
  • 5Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts
JAMA Oncol. 2018;4(2):225-229. doi:10.1001/jamaoncol.2017.2374
Key Points

Question  What is the role of adjuvant chemotherapy for patients with adverse pathologic features after neoadjuvant chemotherapy and radical cystectomy?

Findings  In this cohort study of 788 patients with pT3/T4 and/or pN+ urothelial carcinoma of the bladder, the receipt of adjuvant chemotherapy after neoadjuvant chemotherapy and radical cystectomy was associated with an overall survival benefit.

Meaning  Adjuvant chemotherapy after neoadjuvant chemotherapy and radical cystectomy may prolong overall survival among patients with pT3/T4 and/or pN+ urothelial carcinoma of the bladder.

Abstract

Importance  Despite existing evidence of a benefit associated with cisplatin-based adjuvant chemotherapy (AC) after radical cystectomy (RC) for chemotherapy-naive patients with pT3/T4 and/or pN+ urothelial carcinoma of the bladder (UCB), to our knowledge, no studies have addressed the effectiveness of AC in those who received neoadjuvant chemotherapy (NAC) before surgery.

Objective  To assess the comparative effectiveness of AC vs observation for patients with pT3/T4 and/or pN+ UCB previously treated with NAC and RC.

Design, Setting, and Participants  This observational cohort study used the National Cancer Data Base (January 1, 2006, through December 31, 2012) to identify individuals who received NAC and RC followed by AC or observation for pT3/T4 and/or pN+ UCB.

Main Outcomes and Measures  After multiple imputation was used to handle missing data, inverse probability of treatment weighting (IPTW)–adjusted Kaplan-Meier and Cox proportional hazards regression analyses were performed with a 6-month conditional landmark to compare overall survival (OS) among patients who received NAC and RC followed by AC vs observation. In addition, exploratory analyses were conducted to examine the heterogeneity of the treatment effect according to age (continuous), sex (female vs male), Charlson comorbidity index (≥1 vs 0), pT/N stage (pT3/T4N0 vs pTanyN+), and surgical margin status (positive vs negative) by testing interaction terms within the IPTW-adjusted Cox proportional hazards regression model.

Results  Of the 788 patients with pT3/T4 and/or pN+ UCB (mean [SD] age, 65.3 [9.4] years; 603 [76.5%] male and 185 [23.5%] female), 184 (23.4%) received NAC and RC followed by AC and 604 (76.6%) received NAC and RC followed by observation. The 6-month conditional landmark, IPTW-adjusted Kaplan-Meier curves showed that median OS was significantly longer for NAC and RC followed by AC (29.9 months; interquartile range, 15.1-85.4 months) vs NAC and RC followed by observation (24.2 months; interquartile range, 12.9-58.9 months) (P = .046). The 5-year IPTW-adjusted rates of OS were 36.8% for NAC and RC followed by AC vs 24.7% for NAC and RC followed by observation. In the IPTW-adjusted Cox proportional hazards regression analysis, NAC and RC followed by AC was associated with a significant OS benefit (hazard ratio, 0.78; 95% CI, 0.61-0.99; P = .046). Interaction term analyses indicated that the OS benefit of NAC and RC followed by AC decreased significantly with age (hazard ratio, 0.97; 95% CI, 0.95-0.99; P = .02), whereas no significant interaction was observed with sex (P = .82), Charlson comorbidity index (P = .51), pT/N stage (P = .95), and surgical margin status (P = .29).

Conclusions and Relevance  This study found that AC after NAC and RC may be associated with an OS benefit for patients with pT3/T4 and/or pN+ UCB. The present findings should be considered as preliminary evidence to conduct a randomized clinical trial to address this association.

Introduction

On the basis of level 1 evidence, neoadjuvant chemotherapy (NAC) followed by radical cystectomy (RC) constitutes the standard of care for individuals with localized muscle-invasive urothelial carcinoma of the bladder (UCB).1 Despite the well-established downstaging and survival benefits associated with this approach, approximately 20% of patients harbor adverse pathologic features at surgery.2

Of interest, current evidence suggests that the receipt of cisplatin-based adjuvant chemotherapy (AC) after RC for pT3/T4 and/or pN+ UCB may decrease the risk of recurrence and ultimately prolong survival.3-5 However, these studies were limited to individuals who did not receive NAC before undergoing RC; the role of AC after NAC and RC for non–organ-confined disease at definitive surgery, to our knowledge, has never been evaluated in a randomized setting. To date, a single observational study6 with a small sample size has addressed this topic. Although the investigators6 found no difference in recurrence-free and cancer-specific survival between patients who received AC vs observation after NAC and RC, others7 have shown a potential benefit for salvage chemotherapy in those previously treated with platinum-based regimens for metastatic UCB. As such, we hypothesized that AC after NAC and RC for pT3/T4 and/or pN+ disease may provide an overall survival benefit relative to observation.

Methods

From a population of 313 040 individuals diagnosed with bladder cancer from January 1, 2006, through December 31, 2012 (International Classification of Diseases of Oncology, Third Edition codes C67.0-C67.9), in the National Cancer Data Base (NCDB), we identified 788 patients who received NAC and RC followed by AC or observation for pT3/T4 and/or pN+ UCB (eFigure 1 in the Supplement). First, multiple imputation using chained equations was performed to handle missing data that were assumed to be missing at random for all covariates8; we generated 15 imputed data sets by using a sequential regression method. In all subsequent analyses, Rubin’s rules were applied to summarize the effect estimates and variances from the 15 different analyses across multiple imputed data sets.9 Second, to account for selection bias, observed differences in baseline characteristics between the 2 groups were controlled for with inverse probability of treatment weighting (IPTW)–adjusted analyses.10 The goodness-of-fit statistic of the propensity score model, including linear or nonlinear covariates categorized with clinically relevant cutoffs, was assessed using the method described by Lemeshow and Hosmer.11 Covariate balance was evaluated by using the standardized differences approach and Kernel density plots. Third, to account for immortal-time bias, summary 6-month conditional landmark, IPTW-adjusted Kaplan-Meier curves were calculated to compare overall survival between patients who received NAC and RC followed by AC vs observation.12 We further fitted an IPTW-adjusted Cox proportional hazards regression model to compute the corresponding hazard ratios (HRs).13 A post hoc power analysis was performed to evaluate our ability to detect an association between treatment and overall survival. Fourth, we conducted exploratory analyses to examine heterogeneity of treatment effect according to age (continuous), sex (female vs male), Charlson comorbidity index (≥1 vs 0), pT/N stage (pT3/T4N0 vs pTanyN+), and surgical margin status (positive vs negative) by testing interaction terms within the IPTW-adjusted Cox proportional hazards regression model. A waiver was obtained before the study was conducted from the Brigham and Women's Hospital, Harvard Medical School, Institutional Review Board in accordance with institutional regulation when dealing with deidentified, previously collected data.

All statistical analyses were performed using STATA software, version 14.0 (StataCorp) (eAppendix in the Supplement). Two-sided statistical significance was defined as P < .05.

Results
Patient Characteristics

Of the 788 patients with pT3/T4 and/or pN+ UCB (mean [SD] age, 65.3 [9.4] years; 603 [76.5%] male and 185 [23.5%] female), 184 (23.4%) received NAC and RC followed by AC and 604 (76.6%) received NAC and RC followed by observation (eFigure 1 in the Supplement). Unweighted and weighted baseline characteristics of eligible patients, stratified according to the receipt of NAC and RC followed by AC vs observation, are reported in Table 1. Standardized differences of unweighted comparisons showed that both treatment groups differed significantly with respect to most clinical, socioeconomic, demographic, and tumor characteristics of interest.

Determinants of Receiving NAC and RC Followed by AC vs Observation

Results of multivariable logistic regression analysis determining the receipt of NAC and RC followed by AC vs observation with adequate goodness of fit are reported in Table 2. The odds of receiving NAC and RC followed by AC vs observation remained stable over time (annual percentage change, 0.00%; 95% CI, −0.02 to 0.02%; P = .93) (eFigure 2 in the Supplement).

Treatment Effect of NAC and RC Followed by AC vs Observation

After IPTW adjustment, all standardized differences were less than 10%, which indicated that patients who received NAC and RC followed by AC vs observation were subsequently comparable (eFigure 3 in the Supplement). Propensity score distribution between the treatment groups achieved adequate balance after IPTW adjustment (eFigure 4A and B in the Supplement).

The median follow-up in the weighted population was 45.7 months (interquartile range [IQR], 31.2-67.8 months). The 6-month conditional landmark, IPTW-adjusted Kaplan-Meier curves (Figure) showed that median overall survival was significantly longer for NAC and RC followed by AC (29.9 months; IQR, 15.1-85.4 months) vs observation (24.2 months; IQR, 12.9-58.9 months) (P = .046). The 5-year IPTW-adjusted rates of overall survival were 36.8% for NAC and RC followed by AC vs 24.7% for NAC and RC followed by observation. In IPTW-adjusted Cox proportional hazards regression analysis, NAC and RC followed by AC was associated with a significant overall survival benefit (HR, 0.78; 95% CI, 0.61-0.99; P = .046).

The post hoc power calculation showed that we had adequate power to detect a clinically significant HR. Specifically, with our sample size of 788 patients, we had 80% power with a 5% significance level to detect an HR of 0.78 for NAC and RC followed by AC vs observation using a log-rank test (20% alive in the observation group at the end of the study).

Interaction term analyses indicated that the overall survival benefit of NAC and RC followed by AC decreased significantly with age (HR, 0.97; 95% CI, 0.95-0.99; P = .02), whereas no significant interaction was observed with sex (P = .82), Charlson comorbidity index (P = .51), pT/N stage (P = .95), and surgical margin status (P = .29).

Discussion

Although evidence supporting the role of AC for advanced UCB treated with RC is contentious, the use of a cisplatin-based regimen is generally advocated in sufficiently healthy individuals who did not receive NAC before surgery.1,14 However, less is known about the association of AC with survival among patients with pT3/T4 and/or pN+ UCB who received NAC and RC. As such, we sought to investigate the role of this treatment strategy in a large NCDB sample that included nearly 800 individuals. Of interest, with a median follow-up of approximately 4 years, our IPTW-adjusted analysis showed a significant overall survival benefit for AC when adverse pathologic features are found after NAC and RC. Specifically, individuals who received AC were more than 20% less likely to die of any cause after NAC and RC compared with their observation counterparts; this translated into a 5-month overall survival benefit.

To our knowledge, the present study represents the first sizeable comparative effectiveness assessment of AC after NAC and RC. Indeed, the only observational report6 in the literature included 80 patients with pT3/T4 and/or pN+ disease after NAC and RC, 29 of whom further received AC that was not independently associated with recurrence-free or cancer-specific survival. However, that analysis had limited power to detect a significant benefit, especially after adjusting for potential confounders in multivariable models.

Limitations

Our results need to be interpreted within the limitations of the observational study design. Of note, the present analyses are subject to selection bias, which we attempted to correct by using an IPTW-adjusted approach. Nonetheless, several unmeasured confounders, including performance status or renal function, may have affected the receipt of NAC and RC followed by AC vs observation in this study. In addition, the detailed chemotherapy regimen administered at the time of NAC or AC as well as the completeness and number of chemotherapy cycles in both contexts are not recorded in the NCDB; as a result, we could not comment on the specific treatment sequence that would provide the greatest overall survival benefit. That said, from a biological perspective, it may be appropriate to sequentially deliver different combination regimens with complementary cytotoxic effects to target different cell populations in the primary tumor.6,7

Conclusions

To summarize, we observed that AC after NAC and RC was associated with an overall survival benefit for pT3/T4 and/or pN+ UCB. The present findings should be considered as preliminary evidence to conduct a randomized clinical trial to address this association.

Back to top
Article Information

Corresponding Author: Quoc-Dien Trinh, MD, Center for Surgery and Public Health, Division of Urological Surgery, Brigham and Women’s Hospital, Harvard Medical School, 45 Francis St, ASB II-3, Boston, MA 02115 (qtrinh@bwh.harvard.edu).

Accepted for Publication: June 9, 2017.

Published Online: August 24, 2017. doi:10.1001/jamaoncol.2017.2374

Author Contributions: Drs Seisen and Trinh had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Seisen, Leow, Lipsitz, Sun, Bellmunt, Choueiri, Trinh.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Seisen, Jamzadeh, Cole, Lipsitz, Choueiri.

Critical revision of the manuscript for important intellectual content: Seisen, Jamzadeh, Leow, Rouprêt, Lipsitz, Kibel, Nguyen, Sun, Menon, Bellmunt, Choueiri, Trinh.

Statistical analysis: Seisen, Leow, Lipsitz.

Obtained funding: Kibel, Choueiri.

Administrative, technical, or material support: Kibel, Nguyen, Menon, Choueiri.

Study supervision: Rouprêt, Kibel, Sun, Bellmunt, Choueiri, Trinh.

Conflict of Interest Disclosures: Dr Trinh reported receiving support from an unrestricted educational grant from the Vattikuti Urology Institute, a Clay Hamlin Young Investigator Award from the Prostate Cancer Foundation, and a Genentech BioOncology Career Development Award from the Conquer Cancer Foundation of the American Society of Clinical Oncology. Dr Rouprêt reported having a consulting or advisory role at Sanofi and Ipsen. Dr Kibel reported having a consulting or advisory role at Dendreon, Sanofi, Tokai Pharmaceuticals, MTG Biotherapeutics, and Profound Medical. Dr Bellmunt reported having a consulting or advisory role at Pierre Fabre, Astellas Pharma, Pfizer, Merck, Genentech, and Novartis; receiving research funding from Millennium Pharmaceuticals and Sanofi; and receiving travel, accommodations, and expense funds from Pfizer and MSD Oncology. Dr Choueiri reported receiving honoraria from the National Comprehensive Cancer Network and UpToDate; having a consulting or advisory role at Pfizer, Bayer AG, Novartis, GlaxoSmithKline, Merck, Bristol-Myers Squibb, Genentech, Eisai, Prometheus Labs, Foundation Medicine Research, Cerulean Pharma, AstraZeneca, and Peloton; and receiving funding from Pfizer, Novartis, Merck, Exelixis, TRACON Pharmaceuticals, GlaxoSmithKline, Bristol-Myers Squibb, AstraZeneca, Peloton Therapeutics, and Genentech. No other disclosures were reported.

Disclaimer: The data used in the study are derived from a deidentified National Cancer Data Base file. The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methods used or the conclusions drawn from these data by the investigator.

References
1.
Alfred Witjes  J, Lebret  T, Compérat  EM,  et al.  Updated 2016 EAU guidelines on muscle-invasive and metastatic bladder cancer.  Eur Urol. 2017;71(3):462-475.PubMedGoogle ScholarCrossref
2.
Yafi  FA, Aprikian  AG, Chin  JL,  et al.  Contemporary outcomes of 2287 patients with bladder cancer who were treated with radical cystectomy: a Canadian multicentre experience.  BJU Int. 2011;108(4):539-545.PubMedGoogle ScholarCrossref
3.
Advanced Bladder Cancer (ABC) Meta-analysis Collaboration.  Adjuvant chemotherapy for invasive bladder cancer (individual patient data).  Cochrane Database Syst Rev. 2006;(2):CD006018.PubMedGoogle Scholar
4.
Leow  JJ, Martin-Doyle  W, Rajagopal  PS,  et al.  Adjuvant chemotherapy for invasive bladder cancer: a 2013 updated systematic review and meta-analysis of randomized trials.  Eur Urol. 2014;66(1):42-54.PubMedGoogle ScholarCrossref
5.
Galsky  MD, Stensland  KD, Moshier  E,  et al.  Effectiveness of adjuvant chemotherapy for locally advanced bladder cancer.  J Clin Oncol. 2016;34(8):825-832.PubMedGoogle ScholarCrossref
6.
Zargar-Shoshtari  K, Kongnyuy  M, Sharma  P,  et al.  Clinical role of additional adjuvant chemotherapy in patients with locally advanced urothelial carcinoma following neoadjuvant chemotherapy and cystectomy.  World J Urol. 2016;34(11):1567-1573.PubMedGoogle ScholarCrossref
7.
Sonpavde  G, Pond  GR, Choueiri  TK,  et al.  Single-agent taxane versus taxane-containing combination chemotherapy as salvage therapy for advanced urothelial carcinoma.  Eur Urol. 2016;69(4):634-641.PubMedGoogle ScholarCrossref
8.
Stuart  EA, Azur  M, Frangakis  C, Leaf  P.  Multiple imputation with large data sets: a case study of the Children’s Mental Health Initiative.  Am J Epidemiol. 2009;169(9):1133-1139.PubMedGoogle ScholarCrossref
9.
Rubin  DB.  Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons; 1987.Crossref
10.
Austin  PC.  An introduction to propensity score methods for reducing the effects of confounding in observational studies.  Multivariate Behav Res. 2011;46(3):399-424.PubMedGoogle ScholarCrossref
11.
Lemeshow  S, Hosmer  DW  Jr.  A review of goodness of fit statistics for use in the development of logistic regression models.  Am J Epidemiol. 1982;115(1):92-106.PubMedGoogle ScholarCrossref
12.
Cole  SR, Hernán  MA.  Adjusted survival curves with inverse probability weights.  Comput Methods Programs Biomed. 2004;75(1):45-49.PubMedGoogle ScholarCrossref
13.
Austin  PC.  The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments.  Stat Med. 2014;33(7):1242-1258.PubMedGoogle ScholarCrossref
14.
Clark  PE, Agarwal  N, Biagioli  MC,  et al; National Comprehensive Cancer Network (NCCN).  Bladder cancer.  J Natl Compr Canc Netw. 2013;11(4):446-475.PubMedGoogle ScholarCrossref
×