Location: Monday, 19 March 2018, 15:45 – 17:15, Blue Area, Room 4 (Level 0)
Thank you I’m going to be talking about bladder cancer imaging. So there are limitations to what we can see, what the camera catches. Those limitations are because there are biophysical limitations to the eye and also because there are neurophysiological limitations to image perception. However there’s much more that is captured by the camera and there are some very simple image enhancements that could be made to standard video capture, and you can see in this picture here on the left. We’ve got a picture of the bladder lining and on the right an application of some of these simple enhancements taking into account the particular shortcomings we have with the human eye perception. As part of a study with an industrial collaborator funded by Innovate UK we were able to video film 123 cystoscopies in a haematuria clinic, bladder cancer follow-up cystoscopy clinic. Out of those there were 30+ abnormalities and 13 were referred for biopsy and the histology was noted. We then had a look at the imaging and the effective imaging. We compared students and we compared consultants, before and after enhancement.
The data is summarised right there on the graph what we see is that consultants are actually quite good at doing cystoscopy which is a relief. And the enhancement added very little to that. On the left you see the novices actually do have a shortfall in their ability to detect lesions and when we use the enhancement it does improve with a relative rate at 11% of their performance. So in conclusion what we have here is a video capture mechanism. It allows inexperienced more junior staff to make a referral with that video image capture. If they have the enhancement it helps them make the right call if they’re still uncertain they can ask for a more senior judgement on that.
It’s a stepping stone moving forward to perhaps more significant ambitions. What this imaging allows us to do is to create a 2D picture. So this 3D mapping can be resolved into a 2D picture and we’re generating something that we’re calling a cystograph. It’s very much akin to the idea of a radiograph. We know bladder cancer is incredibly expensive to manage because of its surveillance, most of that cost comes from consultant delivered cystoscopy. If we now have technicians delivering a bladder map using a panoramic type stitching together of an image into a 2D map, a cystograph, a consultant can batch process this in a single session looking at maybe 50+ cases. Moreover the image enhancement forms the basis for algorithms for artificial intelligence and artificial intelligence aided diagnoses and potentially even pre-screening. Very similar to what we see in the histological screening of cervical carcinoma pap smears. Thank you.