Computer Vision News - April 2019

23 Computer Vision News Parvin Mousavi “I think we have a few more years, but hopefully not too far, and hopefully before I retire!” Women in Science to the anatomy. You need additional information to tell you where to biopsy. That additional information is starting to come from MRI (Magnetic Resonance Imaging). However, MRI is done prior to biopsy and in order to integrate that information into the biopsy process, image registration is required. Your readers will know all about registration of preoperative and intraoperative images, which is challenging to start with. Otherwise, we wouldn’t have registration tracks and papers at MICCAI for so many years. Also, during biopsy, patients are awake. Every poke you make with a needle, the patient will jump. Then the registration will be lost. It is very important that what you have guidance technology that can make up for this. Most of my research in ultrasound in prostate cancer is focused on a new technology called Temporal enhanced Ultrasound (TeUs). With this technology, we take consecutive images for a short time of the tissue without intentional motion of the tissue or the imaging probe. From this time series, we are able to detect signatures of tissue that are associated with pathology. How could this help the biopsy? You could still use MRI information that you have taken pre-operatively. Now, with the information that we create in real-time during biopsy from TeUS, we can provide a color map. The color map shows how likely the patient is to be cancerous. If you were to biopsy, and even if the patient moved, the color map can be regenerated in real-time. You can actually compensate for the dissociation of your pre-operative images because you have real-time description of the tissue. We are really excited about this work. We are now doing multi-center feasibility trials. I am really looking forward to see if this can make a difference in patients’ lives, and that is what my lab strives for. How long will it take to turn this into a clinical reality? It takes a long time to see anything, research-wise, get adopted. For now, we are starting our prospective studies. We have done, so far, retrospective studies. It’s important to show that prospectively this can work. Another important item to show is that it works in different centers equally well. Our first trials were at the National Institute of Health, their clinical center in Bethesda, Maryland. Now we are extending this to the west coast and east coast of Canada. It is very important to show that it works equally well with different machines, in different centers, in real clinical conditions, not in a research center. I think we have a few more years, but hopefully not too far, and hopefully before I retire! Read more interviews like this Photo: Bernard Clark