 Hi guys, my name is Dan Crawford and I'm the Founder and Chief Strategy Officer here at Axial 3D. And today what I'm going to talk to you about is a little bit of background in the company. What we do in medical 3D printing and how it impacts the hospitals that we work with. Not only that, I'll talk a little bit about machine learning and the AWS cloud is helping us make this technology more readily available with all of the hospitals around the world. Ultimately as a company what we're trying to do is make 3D printing and medicine completely routine by easily transforming everything traditionally 2D in medicine like CT and MRI scans into physical objects which provide huge benefits over traditional medicine. Making surgeries faster for surgeons, cheaper for hospitals but more importantly actually safer for patients. And what this practically means is the process of taking these medical images and delineating the anatomical structures that you actually need from them and reconstructing these in the one-to-one scale physical models of actually what's in your body. When you boil Axial 3D down into one concept it's about making patient data real. We live in a 3D world and our bodies are made up of thousands of complex three-dimensional structures so if you think about it your surgeon or your acting clinician is there a huge disadvantage when they actually go to treat you because traditionally in medicine a lot of these technologies are two-dimensional images viewed on 2D screens and they have to create a conceptual understanding of actually what's going on with their patients which more often than not can be difficult to create or unsafe to interpret. And we've been 3D printing in the medical space now for around five years and each and every case that we've worked with has a huge impact on a patient's outcomes whether it be a change in diagnosis, an approved outcome or enabling a surgeon to actually learn something completely new about a patient that they hadn't seen on traditional two-dimensional images. And just as an example I wanted to introduce you to Jonah so Jonah's a case that we would see pretty much on a daily basis now in Axial 3D one that's hard to diagnose, hard to plan and so Jonah's a 14-year-old boy who suffered from scoliosis all his life and he's gone through a number of surgeries to recently went through was to have the ears that were anchoring rods that were keeping his spine straight to be replaced and unfortunately over time these metal rods that were anchoring on his pelvis had started to wear away and break down over time as Jonah had started to grow meaning that the rods that were actually keeping his spine straight were becoming compromised and a surgeon, Mr. O'Brien, had approached us I mean he was trying to make a picture in his head and attempt to make a plan with not only his own team but a team of neurosurgeons that were in charge of keeping Jonah's spinal cord intact and from these images he had zero surgery on the day around where the pins lay in Jonah's anatomy and actually the extent of the damage that the rods had had to Jonah's pelvis was effectively going in blind we were able to take these images and make them in the single model of Jonah's spine and using this Mr. O'Brien was actually able to see the exact extent of the damage to his pelvis and also able to find three new fractures on Jonah's rib cage that had occurred as a result of the metal work that they hadn't seen in 2D images and this allowed him to prepare all of the equipment that he needed to treat all of these pathologies well in advance of stepping foot into the operating theatre which represented a huge cost saving and time saving and if you think about it, a minute in surgery time can cost up to around $100 three additional rib fractures he was able to bring all the equipment that he needed to treat Jonah to do all the operations on one sitting as opposed to having to do additional operation when he found the rib fractures from the first surgery meaning that Jonah and his parents only had to go through one operation and all the ensuring ICU and recovery time that's associated with this type of operation so it's a huge benefit to Jonah and his family so some of you may not know this but 3D printing has been around since 1984 and in medicine it's been around since the early 90s with the recent desktop revolution, printers have dropped in price by around 99% which opens up this to a much wider market so you might be asking yourself with all the benefits why is 3D printing not routine so from taking a step back we can see three main bottlenecks for this technology not being adopted by every hospital around the world so getting access to the raw images from a hospital and getting an accurate prescription of exactly what is required from the 3D printed medical model is a clumsy and disjointed process which can take weeks to actually do so we fixed this by creating a web application that's hosted on the AWS cloud to allow surgeons anywhere in the world to upload their diagram images exactly how they want the model to be made what anatomy they want to see or be highlighted what material they want to be in and when they actually need it for the operation and this simple application allowed us to make medical models and as little as 48 hours for surgeons cutting the process that could traditionally take four to six weeks so we were able to create an easy and scalable way of collecting data from customers but still the biggest issue in processing the data was scaling and before I tell you a little bit about how we solved that issue I wanted to take another step back and give you another example of why we actually wanted to address this biggest bottleneck in the market which was to do with scalability of our processes and while we began to do up machine learning and SageMaker in the processes is actually because of this patient so Kiri, I wanted to introduce you to she's a 22 year old mother based in Balthast Kiri's gone through renal problems most of her life meaning that she'd previously received a kidney transplant and because of this she was on immunosuppressant drugs and unfortunately for Kiri her previous kidney transplant was beginning to fail and because she was on immunosuppressant drugs it meant that her donor pool was extremely limited to basically only family members but luckily for her when her family were tested her father William was a perfect match meaning that Kiri had one last chance of actually getting another kidney and having a long fruitful life so William went through the standard processes for evaluating the transplantation his vitals were tested and he was scanned to ensure that he had a good anatomical match for Kiri but from the work up William's surgical team found a previously undiagnosed problem William actually had a tumor on his left kidney which meant that Kiri's only kidney that she could get was deemed untransplantable initially given that it was too close given that the tumor was actually too close to all the vital structures within the kidney so Mr Brian the surgeon who was in charge of the transplantation process from William to Kiri approached us get a full picture of what was actually going on with the kidney and if anything could be done so we were able to take the images directly from the hospital through a web platform and create a model of William's kidney on the same day created using 302 dimensional images to create this a physical one-to-one scale model of William's kidney and with this Mr Brian was actually able to conceptualise exactly where the tumor lay which you see in black here against the other vital renal arteries, veins and ureter within the kidney and from having this Mr Brian was actually instantly able to plan an approach to harvesting enough of William's kidney to make it a viable transplant for his daughter Kiri and a procedure called an X-FIVO partial nephrectomy effectively cutting out a portion of the kidney but still leaving enough to be fully functional and not only that, he was able to actually bring the kidney into theatre with him to use as an operative guide to cut out the perfect margins of William's kidney to ensure that he was removing just the right amount of the entire tumor and not to hit any of the vital structures within the organ and after the procedure the pathologist was able to actually confirm with Mr Brian and in fact he'd carried out a perfect nephrectomy so he'd cut the exact margin to remove the tumor within William's kidney on the first try so something that had previously never been done before anywhere in the world making it a world first so this is something, just another one of those examples of hundreds of cases that are being done internally at Axial 3D that has an amazing story behind it made possible by giving surgeons access to this 3D printing technology so back to the biggest problem why is this technology not being used in every hospital around the world? For a hospital or a company to make patient specific data and patient specific 3D printed medical models from scans the biggest bottleneck to all of this is a process called segmentation and this process put simply is the when which you take medical images and annotate them to pull out the anatomy that you want to be printed it requires a knowledge of medical physics it requires a knowledge of anatomy and the engineering principles behind 3D printing to make sure that the images are segmented in the correct way that they can be printed and with each data set being made up of hundreds of two dimensional images this process can take hours and if we look at the levels of complex surgeries and injuries and pathologies that are traded across the world it would take thousands of years of manual processing by skilled personnel so we looked at this and it's just completely unfeasible for 3D printing to be part of everyday practice and critical care if this isn't fixed and this is the main reason medical 3D printing is not being used by every hospital around the planet so we're fixing this by creating machine learning models to automatically process this data based on a verified ground trust database of hundreds of thousands of medical images and these convolutional neural networks that we've built and trained on the AWS cloud and SageMaker are able to process thousands of cases in parallel enabling us to create tools to allow every hospital across the world to actually adopt this technology and what this looks like in practice is we're able to take these data sets like Jumas that would have traditionally taken four or five hours to process by hand and run these through an inference of ground trust data to automatically segment structures from the CT scans in minutes not hours to create 3D models that are able to be verified and sent for printing pretty much instantly and orthopedics represents our biggest use case right now but for this to be routine for every hospital around the world we've developed algorithms for a number of different specialties including cardiology, trauma oncology and maxillofacial surgery and an internal database of ground trust data that is constantly growing and being fine to speed up our processes and we've been creating these machine learning models for over two years now and seeing the huge benefits that they bring to our processes and our scalability but internally we focus on the patient benefits and the reason we do what we do and why we're developing the tools is the impact 3D printing actually has on patients we've helped diagnose and treat over thousands of difficult pathologies and helps save countless hours of surgeries and in some cases like the ones mentioned in this presentation we've actually saved lives by making this technology routine practice in medicine and these newly developed tools along with the scalability that the IWS cloud provides Axial 3D is on a mission now to make 3D printing readily available for every hospital around the world allowing every surgeon to gain access to the technology to improve the lives of their patients Thanks