Abstract
Background/Objectives: In non-muscle-invasive bladder cancer (NMIBC), the decision for immediate postoperative single-dose intravesical chemotherapy (SI) is based on clinical and presumed pathological features, as a definitive pathology is unknown at the time of surgery. This study aims to assess how accurately urologists can predict the pathological features of bladder tumors based solely on cystoscopic appearance and evaluate their ability to identify patients eligible for SI. Methods: A total of 104 patients with bladder masses were included. Seven senior urologists and four residents participated. Before transurethral resection, both groups predicted tumor stage, grade, and the presence of carcinoma in situ (CIS). Resident predictions were collected for all 104 patients, while senior predictions were collected for 72 patients. Based on these predictions, patient eligibility for SI was determined according to the EAU NMIBC guidelines. After final pathology reports, risk scores were recalculated and compared with the surgeons' predictions. Cohen's Kappa (κ) coefficient was used to assess agreement between predictions and pathology. Positive and negative predictive values were also calculated for both groups. Results: Strong agreement with final pathology could not be demonstrated for stage, grade, or CIS for either group. Urology residents' predictions were slightly more accurate than those of senior urologists. Overall, 19.4% (14/72) (based on senior urologists' predictions) and 18.2% (19/104) (based on resident predictions) of patients were misclassified and either overtreated or undertreated. Conclusions: Cystoscopic visual prediction alone is insufficient for determining eligibility for immediate postoperative intravesical chemotherapy, regardless of the urologist's experience. More objective criteria are needed to improve the selection of appropriate patients for SI.