The Cancer of the Bladder Risk Assessment (COBRA) score: Estimating mortality after radical cystectomy

• Published in: Cancer Vol. 123; no. 23; pp. 4574 - 4582
• Main Authors: Welty, Christopher J., Sanford, Thomas H., Wright, Jonathan L., Carroll, Peter R., Cooperberg, Matthew R., Meng, Maxwell V., Porten, Sima P.
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01-12-2017
• Subjects: Aged; Aged, 80 and over; and End Results (SEER); Bladder; Bladder cancer; Cancer; Clinical trials; Cohort Studies; Confidence intervals; Cystectomy - mortality; Dissection; Epidemiology; Female; Follow-Up Studies; Hazard assessment; Health risks; Humans; Invasiveness; Lymph; Lymph Node Excision - mortality; Lymph nodes; Lymphatic system; Male; Medical prognosis; Medical research; Middle Aged; National Cancer Data Base (NCDB); Neoplasm Grading; Neoplasm Staging; Oncology; Patients; Prognosis; Risk Assessment; Risk levels; Surveillance; Survival; Survival Rate; Urinary bladder; Urinary Bladder Neoplasms - mortality; Urinary Bladder Neoplasms - pathology; Urinary Bladder Neoplasms - surgery; Urothelial carcinoma; risk assessment; National Cancer Data Base (NCDB); Surveillance; bladder cancer; and End Results (SEER); Epidemiology; prognosis
• ISSN: 0008-543X; 1097-0142
• DOI: 10.1002/cncr.30918

BACKGROUND
Risk stratification of patients with urothelial carcinoma of the bladder (UCB) after cystectomy has important clinical and research implications. The authors assessed the relative effect of tumor stage and lymph node status on cancer‐specific survival (CSS) after cystectomy and developed a simplified risk‐assessment tool.
METHODS
In total, 14,828 patients who underwent cystectomy with lymph node dissection for UCB were identified from the Surveillance, Epidemiology, and End Results database (1988‐2011). The relative importance of tumor stage and lymph node status with regard to CSS was assessed using stratified Kaplan‐Meier and Cox proportional‐hazards analyses. The patients were split randomly into development and validation cohorts. Additional validation using overall survival was performed on 19,362 patients from the National Cancer Data Base. The Cancer of Bladder Risk Assessment (COBRA) tool was created using a Cox model incorporating age, tumor stage, and lymph node density. Performance was validated using observed versus expected survival plots and the Harrell concordance index.
RESULTS
Patients with muscle invasive (T2), lymph node‐positive disease had a survival curve similar to that in patients with extravesical (T3 and T4), lymph node‐negative disease (2‐year CSS, 67% and 70%, respectively). Each point increase in the COBRA score (range, 0‐7) was associated with a 1.61‐fold increase (95% confidence interval, 1.56‐fold to 1.65‐fold increase) in the risk of bladder cancer death in the development cohort. The model accurately stratified patients across risk levels in the development cohort and the 2 validation cohorts (C‐index, 0.712, 0.705, and 0.68, respectively).
CONCLUSIONS
The COBRA score offers a straightforward, validated risk‐stratification tool that incorporates the relative contribution of tumor stage and lymph node involvement to patient prognosis after cystectomy for UCB. Cancer 2017;123:4574‐4582. © 2017 American Cancer Society.
The authors have developed the Cancer of the Bladder Risk Assessment (COBRA) score to provide a straightforward, multivariable risk‐stratification tool that incorporates the relative contribution of tumor stage and lymph node involvement to predict prognosis after cystectomy. This provides a simplified prognostic tool of clinical utility to counsel patients, improve risk stratification, and identify early candidates for clinical trials of adjuvant therapy for aggressive bladder cancer.