Background: Bladder cancer is one of the most common malignant diseases of the urinary tract. Cystoscopy is currently commonly performed for diagnostic evaluation and follow-up. This examination allows reliable assessment of the urinary bladder but involves an invasive procedure and may be associated with discomfort for patients. Non-invasive examinations are already available as complementary methods, for example the examination of cells from urine. However, the diagnostic performance of these methods is currently not sufficient to reliably replace cystoscopy. Research question: This study investigates whether different information and non-invasive examinations can be meaningfully combined to predict the presence of bladder cancer as reliably as possible. For this purpose, information on known risk factors, the detection of blood in urine, an additional urine test, and computer-assisted analysis of cells from urine are evaluated together. Methods: Patients who are scheduled to undergo cystoscopy as part of their routine clinical care, independently of the study, may participate. Before cystoscopy, participants complete a short questionnaire and provide a urine sample. Several examinations are performed using this urine sample. In addition, cells from the urine are examined, and the resulting images are scientifically analyzed using artificial intelligence. The results of these additional examinations are then compared with the result of cystoscopy. If cystoscopy reveals an abnormal finding and a tissue sample is subsequently examined, the result of this tissue examination is also taken into account. The additional examinations are performed exclusively for scientific purposes and have no influence on the further medical care of the participants.
Patients who undergo a cystoscopy as part of routine clinical care
- No means of obtaining a urine sample - Minors and adults incapable of giving consent - No histopathological findings available in cases of suspected tumor
What: Sensitivity of the exploratory multimodal model for the detection of urothelial carcinoma of the urinary bladder in an outpatient cystoscopy cohort. When: Assessed after completion of data collection and determination of the clinical reference standard. How: Comparison of the model-based prediction with the clinical reference standard. In patients with an unremarkable cystoscopy, the cystoscopic finding serves as the negative reference standard. In patients with an abnormal or tumor-suspicious cystoscopy, the histopathological finding is used as the reference standard.