Predicting the risk of high-grade bladder cancer using noninvasive data.

ABSTRACT

AIM: To examine the hypothesis that the risk of high-grade bladder cancer can be predicted using noninvasively obtained data. PATIENTS AND METHODS: We retrospectively analyzed the database of 431 patients that had transurethral resection of first-time bladder tumors between June 1998 and December 2009. Pre-operative parameters evaluated were patients' age; gender; sonographic tumor diameter, number and location of tumor inside the bladder; presence of hydronephrosis, and results of urinary cytology. Parameters that showed significance in multivariate analysis were incorporated into the nomogram. RESULTS: Multivariate analysis of the data showed that patient's age, the presence of hydronephrosis, sonographic tumor diameter (risk of a high-grade tumor 14, 29, 43.3, 55.7 and 69.4% at diameters 0.5-1.5, 1.6-2, 2.1-2.5, 2.6-3 and >3 cm, respectively), location of tumor in the bladder (risk of high-grade tumor 28.8, 47, 67.5 and 90.5% in the lateral walls, posterior/base, anterior and dome, respectively), and urinary cytology were all highly significant and independent predictors of high-grade tumors. A nomogram constructed using these variables scored an area of 0.853 in the ROC curve. CONCLUSIONS: The risk of high-grade bladder tumor can be accurately predicted using non-invasively obtained information. This prediction can help to triage patients with newly detected bladder cancer for biopsy.