Controlling the trajectory of a flexible ultrathin endoscope for fully automated bladder surveillance

Abstract

During cystoscopy, the urologist manually steers a cyst scope inside a patient's bladder to visually inspect the inner surface. Cystoscopies are performed as part of surveillance for bladder cancer, making it the most expensive cancer to treat over a patient's lifetime. An automated bladder scanning system has been devised to reduce workload and cost by relieving the urologist from performing surveillance. Presented here is a proof-of-concept apparatus that controls the motion of a miniature flexible endoscope. Image-based feedback is used to adjust the endoscope's movement so that captured images overlap with one another, ensuring that the entire inner surface of the bladder is imaged. Within a bladder phantom, the apparatus adaptively created and followed a spherical scan pattern comprised of 13 individual latitudes and 508 captured images, while accepting between 60% and 90% image overlap between adjacent images. The elapsed time and number of captured images were sensitive to the apparatus's placement within the phantom and the acceptable image overlap percentage range. A mosaic of captured images was generated to validate comprehensive surveillance. Overall, a robotically controlled endoscope used in conjunction with image-based feedback may permit fully automated and comprehensive bladder surveillance to be conducted without direct clinician oversight.