Y Lakhman, D Yakar, DA Goldman, SS Katz, HA Vargas, M Miccò, J Zheng, CS Moskowitz, RA Soslow, H Hricak, NR Abu-Rustum, E Sala
PURPOSE: To develop a preoperative CT-based nomogram for predicting overall survival (OS) in patients with non-endometrioid carcinomas of the uterine corpus. METHODS: Waiving informed consent, the institutional review board approved this HIPAA-compliant, retrospective study of 193 women with histopathologically proven uterine papillary serous carcinomas (UPSC), uterine clear cell carcinomas (UCCC), and uterine carcinosarcomas (UCS) who underwent primary surgical resection between May 1998 and December 2011, and had a preoperative CT ≤ 6 weeks before surgery. All CT scans were reviewed for local or/and regional tumor extent, presence of pelvic or/and para-aortic adenopathy, and presence of distant metastases. Univariate survival analysis was performed using log-rank test and Cox regression. Variables shown significant by the univariate analysis were evaluated with the multivariable Cox regression analysis and the results were used to create a nomogram for predicting OS. The predictive accuracy of the nomogram was assessed with the concordance probability index (c-index) and a 3-year calibration plot. RESULTS: Mean patient age was 67.2 years (range 49.0-85.9); histologies included UPSC (n = 116), UCCC (n = 27), and UCS (n = 50). Median follow-up was 38.1 months (0.9-168.5 months). At multivariate analysis, patient age, ascites, and omental implants on CT were significant adverse predictors of OS and were used to build the nomogram. Concordance index for the nomogram was 0.640 ± 0.028. CONCLUSION: We developed a nomogram with a good concordance probability at predicting OS based on readily available pretreatment clinical and imaging characteristics. This preoperative nomogram has the potential to improve initial treatment planning and patient counseling.