Abstract
PURPOSE: To analyze prognostic factors associated with ureteral stent failure and to develop a prediction model for malignant ureteral obstruction (MUO) in patients with non-urological cancers. MATERIALS AND METHODS: We retrospectively reviewed patients with non-urological cancers who underwent ureteral stenting or percutaneous nephrostomy (PCN) for MUO between 2006 and 2014. Variables predicting stent failure were identified using Cox regression analysis. RESULTS: Of the 743 patients, 468 (63.0%) underwent ureteral stenting only, and 275 (37.0%) underwent PCN owing to technical (n=215) or functional (n=60) stent failure. The median overall survival was 4 [interquartile range (IQR) 1-11] months, and the median interval duration to stent failure was 2 (IQR 0-7) months. In univariate analysis, lower gastrointestinal cancer, previous radiotherapy to the pelvis, bladder invasion, lower ureteral obstruction, and low previous estimated glomerular filtration rate (eGFR) (<30 mL/min/1.73 m²) were significantly associated with a decreased survival rate. In multivariate analysis, bladder invasion and previous eGFR were significant predictors. With these two predictors, we divided patients into three groups based on their presence: low-risk (neither factor; n=516), intermediate-risk (one factor; n=206), and high-risk (both factors; n=21). The median stent failure-free survival rates of patients in the low-, intermediate-, and high-risk groups were 26 (8-unreached), 1 (0-18), and 0 (0-0) months, respectively (p<0.001). CONCLUSION: In cases of ureteral obstruction caused by non-urological cancers, patients with bladder invasion and a low eGFR showed poor stent failure-free survival. Therefore, PCN should be considered the primary procedure for these patients.