Yale plastic surgeon
Derek Steinbacher, DMD, MD combines machine-learning algorithms with digital modeling for complicated facial plastic surgeries, creating pre-surgical 3-D renderings that are precise and personalized. This video discusses this type of 3-D planning and shows how Dr. Steinbacher and his colleagues now take it a step further. there's so many functional considerations that go along with the face. This is what somebody is wearing front and center. This is what everybody sees. This is how they're interacting at school, at work, when social engagement. So it's vitally important to somebody. It's an important to their identity most of the time that it's this form function following one another similar to architecture where if something is not well proportioned or doesn't look right or is imbalanced, then most often there is a functional component that goes along with that in the head and face, a neck area that we deal with that's frequently being able to eat and speak and chew or breathe. I'm very interested in computer assisted design and how we can manipulate patient data in the three D. Space to better understand their pre existing or preoperative diagnoses or situations, manipulating that in a way that will help us reposition tissues and structures to help us get the best result that we can we start with the patient their own data. They're customized data that represents their facial bones. Then we take that those images were able to digitally render them uh and then we can manipulate them basically performing the surgery in digital space. Then from that virtual space we can generate three D prints that we can either utilize in the operating room to help us manipulate and reposition structures or that are sometimes part of the actual reconstruction themselves. We have a machine learning model now that's based on about 4000 normal individuals that this has come about from some of our collaborations with people nationally and internationally that are all interested in three D morphy metrics or three D shapes of faces. I contributed over 100 of my actual before and after or thematic jaw surgery patients to look at how we can better model based on normals and based on the after the postoperative scans, how we can achieve that ultimate endpoint based on large amounts of data sets. This will help minimize the number of variables that we have to think about inter operatively or that we're eyeballing based mostly on subjective criteria. We don't wanna leave it to the hands of a computer to make every single decision as to how we're gonna move bone and what the end result will be. But we can use this model to get us three quarters of the way in terms of what the bone relationship will be and how we need to move the bone to achieve what our facial result is. I think incorporating this motto into our planning process is going to help us get reproducible, high fidelity accurate, vary, aesthetic results.