Johns Hopkins Biomedical Engineering

Transcript

For the last fifty years or more, we have revolutionized the world around us in engineering the external world. In the next fifty years, it's about engineering the internal world. Imaging is this incredible measurement of the state of a patient. In my lab, at Hopkins BME, we develop imaging systems for the operating room. Minimally invasive approaches that are safer and more effective. We want to produce systems that use x-ray sources that are optimal for a particular task and detectors, and we want to arrange those in new hardware configurations that are well-suited to very low-dose imaging. And so over the last 15 years, we've produced several of these completely new imaging systems that are ideal to a particular task in surgery. When you're dealing with x-ray imaging, you have to make an image that sufficient at the lowest possible dose, and that's really the main challenge. Image-guided procedures where you need to produce a quality image at less than a tenth of the dose of a typical diagnostic exam, and you need to do it fast within a very rapid workflow of surgery. So we completely reconfigured the imaging chain on these systems to allow low-dose 3-D imaging in the OR. We've done similar things in radiation therapy, where we're able to image the target at the time of treatment, account for any motion that the target may have been going through, and adjust the treatment so that we're delivering radiation to the tumor and sparing normal, healthy tissues. There's a wealth of image information that we don't fully extract, even in an expert radiologist screening. One of the areas we're looking to next is how to apply big data methods within imaging. You're teaching these algorithms, these neural networks, by way of large amounts of training data, and the machine learns to identify that pathology and find a particular feature that we think is important. At Hopkins, our laboratories are in the heart of the hospital. That's a really unique advantage to a biomedical engineer to be able to go from science to application. The real lightbulb moments tend to come when we learn from a surgeon or radiologist what the real problems are. Being able to extract knowledge from image data is huge part of informing how we diagnose and will treat patients in the future.