Artificial intelligence to improve cytology performance in urothelial carcinoma diagnosis: results from validation phase of the French, multicenter, prospective VISIOCYT1 trial

Category	Detail
Condition	Bladder cancer, including non-muscle invasive bladder cancer (NMIBC)
Key Mechanisms	AI-based deep learning automated image processing analyzing urothelial cell morphology in voided urine samples
Target Population	Patients older than 18 years undergoing bladder endoscopy for suspected bladder cancer or lower urinary tract exploration
Care Setting	Multicenter clinical settings with cytology, cystoscopy, and histopathology capabilities

VisioCyt® uses AI to analyze morphological changes in urothelial cell nuclei from urine samples to detect bladder tumor cells.
The VISIOCYT1 trial validated VisioCyt® as a noninvasive diagnostic tool with improved sensitivity over standard cytology, especially for low-grade tumors.
VisioCyt® digitizes slides and performs automated analysis blinded to histopathology, reducing interobserver variability inherent to cytology.

Combine urinary cytology with white-light cystoscopy as current gold standard for bladder cancer diagnosis and surveillance.
Consider VisioCyt® testing as a noninvasive adjunct or alternative to improve detection sensitivity, particularly for low-grade tumors.
Exclude patients with urinary tract infections, urinary lithiasis, prior pelvic radiotherapy, or renal transplants from VisioCyt® testing.

Use cystoscopy with biopsy or transurethral resection for histopathological confirmation and staging following positive diagnostic tests.
Incorporate VisioCyt® results alongside cytology and cystoscopy findings to guide clinical decision-making.

Perform follow-up cytology and cystoscopy at 6 and 12 months post-diagnosis as per standard of care.
Use VisioCyt® testing on voided urine samples collected prior to cystoscopy for ongoing surveillance.

Cystoscopy is invasive and may cause urinary tract infections, dysuria, hematuria, or bladder wall perforation.
Cytology has high interobserver variability and moderate sensitivity, especially for low-grade tumors.
VisioCyt® requires adequate urothelial cell count (≥15 cells) for reliable analysis.

Adults undergoing evaluation for suspected or recurrent bladder cancer without contraindications such as infections or prior pelvic radiotherapy.

VisioCyt® provides a noninvasive, AI-driven diagnostic option that enhances detection accuracy and may reduce reliance on invasive cystoscopy.

Ensure urine samples are collected prior to cystoscopy and properly fixed and prepared according to VisioCyt® protocol.
Train pathologists in standardized cytology criteria (Paris System 2016) to reduce variability in cytology interpretation.
Exclude samples with fewer than 15 urothelial cells to maintain diagnostic accuracy.
Use VisioCyt® as a complementary tool to standard cytology and cystoscopy rather than a standalone diagnostic method.

Clinical Scorecard: Enhancing Cytology Accuracy in Urothelial Carcinoma Diagnosis through Artificial Intelligence: Findings from the Validation Phase of the Multicenter French VISIOCYT1 Study