 Savas is the perfect example of why finding a really good PhD student is both a blessing and a curse. It's a curse because he comes to my office three times a day and normally he doesn't leave unless I kick him out. That's also a blessing because what he wants to talk about is really really high level, very interesting scientific stuff, which is a very welcome distraction when you're waiting yet another boring report that no one will ever read. He's originally a civil engineer, believe it or not, and then he decided to take our masters, which is where we met. He took our module, my module, and was asking questions all the time, so I can't even say that I wasn't pre-worn, but he must have liked it because he then decided to do a PhD with us, which is doing on the application of our imaging methods to tissue engineering, which is expected to play a very important role in the future of medicine. So during the project, not only has he developed and honed his physics and imaging skills to a very high degree, but he's also developed this amazing ability to interact with the medical community and you don't need me to tell you how important that is for the sort of research that we do in the department. So if you want to take him along to your next meeting with the medics, maybe you can borrow him from us for a very small fee. And with that, I'll hand over to Savas and his new medical, new imaging applications for tissue engineering. Thank you, Sandra, very much. I mean, you oversold me a bit, so anything I present now is not going to be adequate up to the standard that people are expecting in the audience, unfortunately. I'm going to try my best, though. So today I'm going to present to you one of the chapters of my PhD thesis, which titles guiding the tissue engineering of Anisophagus using X-rays. This is a multi-disciplinary project which is supervised by Professor Sandro Alivo and Professor Paolo, the copy in the Institute of Child Health. So the case of isophageal atucia is when the isophagus goes to a planet that is still being connected to the stomach. So visualize it as if there is a gap along the length of the isophagus and if that gap is short enough then the clinicians can actually pull the two sides together and rejoin the isophagus. However, if this gap is long there is no current therapeutic option. The possible solution to this problem is to have a tissue engineer construct which can be transplanted in the isophagus and therefore bridge the gap between the two sides of the isophagus. The general requirements in order to have a tissue engineer construct are cells, the scaffolds of interest, growth factors and then you combine them together, you put them in incubation in a controlled environment and you have your final product. The work in ICH is currently in in vitro stages and we are working with piglets. So we have the native samples extracted from a piglet which is then desaleralized so it's stripped out of its cells in order to have the scaffold sample. Then the appropriate cells found in the isophagus are seated in the scaffold. It's put in incubation for 10 days and we have the final resaleralized in vitro matured organ. And as a final step, before imaging we critically point right in our samples. Of course the current gold standard process in the construct is histology. However, as you know histology is destructive, it's limited to a single orientation of dissection and it can actually go really long with the waxing and if you get the wrong, you lose information. It doesn't allow of the extraction of non-destructive volumetric microarchitectural information of the sample. And at the same time it doesn't allow quantitative intraneter sample statistics. So comparison. A potential candidate is X-ray tomography and I'm pretty sure you're familiar with a conventional attenuation based X-rays which relies on the degree at which X's are attenuated in order to create contrast between different tissue in a sample. This is an isophagus by the way tomography lies. However, as you can see and as you already most probably know X-rays suffer from soft tissue contrast. However, if we exploit the fact that X-rays are also phase shifted as it travels through a material and we have a device that can actually capture a shift, we can have soft tissue contrast and at the same time we can have a density map of the material we are imaging. Having a closer look, the phase based image compared to attenuation based you can see maybe not so well on the screen but we can distinguish the four isophageal layers of interest in the image on the left and it's quite challenging on the attenuation based image. And then this is the system developed by Sandro and it sits on the second floor of medical physics called the edge illumination system. Of course, all of the findings that we have with X-rays need to be confirmed with a cutting gold standard. So this is an example of a native sample and that we're actually image first and then cut it out histology, H and E which is the standard histology that the clinicians are currently carrying out. As you can see, all the features that can be identified in the histology are also visible in the X-ray sample and as this is a density map of the actual sample, regions that are cell rich like the muscle layer of the sample are going to appear bright, which means dense and empty regions like the lumen of the sofas are of course going to appear as dark. An example of a scaffold sample, which is cell free and the intensity of the walls now are significantly reduced because now it's desidualized so it's straight out of the cells. And then this is a case of a desidualized sample which I put in a parenthesis half seeded because at this stage the clinicians seeded cells in on the one side of the sample only. This is immunostaining an H and E on the same sample and you can see that this region is populated with cells and it appears bright on the X-rays and then the regions that appear less dense on the X-rays are the ones that have no cells. An example of fully seeded sample where it's evident when comparing with the previous one that this one is fully populated with cells along with a corresponding histology. So what can we do with our images? As I mentioned before histology suffers from a single dissection orientation. In our case you can actually use the different images stack them together and have volumetric information independent of the orientation. So we can extract information about the sample in all plates. Furthermore we can have a quantitative map of the different layers found in the esophagus and in this case following N127 from the outside towards the lumen of the sample we can see the serosa which is a connective tissue then we have a blood vessel which appears way less dense because it's empty. We have the two muscle cells of the esophagus. We have a very narrow peak which is basically a connective tissue. I'm not quite sure if you can see on the screen but it appears as a bright peak because connective tissue is dense. We have the loose connective tissue called the submucosa and of course the final layer which is the mucosa and epithelium which is bright and dense. In the case of the scaffold material again we can distinguish all the different layers but for this slide I want to highlight that it is very important that we can confirm that volumetrically all the different layers of the samples are preserved post-desertalization. So after we strip out the cells of the sample. This is a deserterized sample and the first thing that strikes your eye is of course that the entire microarchitecture of the sample is lost. We can observe ingredients from R3 to R4 of the cell population so we can confirm from the images that more cells for some reason tend to populate towards the inner side of the sample and that this is explained due to the fact that during the tissue engineering process there is flow of media through the lumen and therefore cells are moving towards the lumen and most importantly the clinician said that it's crucial to know whether this deserterization approach took place uniformly across the volume of the sample and I think that's something we can derive from our volumetric images. Of course to have more detailed statistical analysis on the samples we need to mask out the different samples so isolate from the surrounding. In order to do that we have option number one to do it manually which is going to take ages because I have over 50 samples and over 4,000 slices to do manually or choose the smart option of using a semi-automated approach developed by Alessia Azzeni which is a fellow CUT student and the only thing you need to do is basically input the first segmentation image that you've done manually the last one and then everything in between happens for you automatically. As I mentioned before the actual image is a density map of the sample and if we use the tissue mask that I showed you before we can isolate the sample and then for each slice we can have the average density value of the sample. Of course we can plot the density variation along the volume of the sample and for this example I'm demonstrating a native sample, a resaloralized sample, a scaffold. Both the native and resaloralized samples have higher density values than the scaffold and then this makes sense because these two are cell-rich while this one is just a resaloral matrix with no cells. However we realize that as you go across the volume of the resaloralized sample the density value drops the value of a scaffold. So this raised the flag and we checked out the images at the start of the volume and the end of the volume and as you can see visually it's pretty evident that the last slice has way less cells than the first slice. Then we thought that's the clinicians and the clinicians confirmed to us that that end of the sample was the part that was clamped to the bioreactor and therefore had no cell injections close by and therefore no cells populating the sample. Of course we can do the same thing for all samples but if we want to have a statistical comparison between 50 or more samples we need to express each sample as its own volumetric density value. I think I know everybody knows where this is going. We put them together in different groups and we have statistical testing on the samples. We have the native, the scaffold and the resaloralized and both the native and the resaloralized show significant difference when compared to the scaffold. And then this is some great news because you expect the resaloralized sample to have higher density overall than the scaffold and similar density values to the native. So every time we actually extract the different sample from the biological testing we input in this population here and we see where it's actually going to be sitting in the distribution. Another type of analysis that we can do is to compare the physical size of the samples and of course to do so we need to have a normalized value of each sample because okay it's fair enough to say that all the samples are derived from the same peak type but there is some variation because it naturally occurs. Therefore we decided to express each sample as a isophageal tissue percentage. So how much of the sample is occupied by the tissue? This is a case of three different types of samples. Apologies. This is a case from the same sample at three different stages. So when it was native you had the 70% tissue percentage, a scaffold it became 45% tissue percentage. So that's an indication that the walls of the samples got smaller because now they are cell free and then upon repopulation the tissue percentage increased because now we have populated sample. Again, statistical analysis. Both native and decentralized sample show significant difference when compared to scaffold. At this point, yes, this is indeed a confirmation that the decentralized sample indicates that it's populated with cell because the walls are increasing size but at the same time this indication of the mechanisms at which the scaffold is repopulated. I'm going to explain to you what I mean with this diagram. So have a scaffold then we have some cells injected in the walls of the scaffold and then the cells if I may they have the free will to move outwards or inwards towards the lumen. If we have seven an increase in the percentage which obviously is in decreasing the lumen percentage it implies that the cells tend to move towards the lumen of the sample and then we have the final tissue engineer construct. This was discussed with the clinicians and it was confirmed that it actually makes sense because cells tend to respond to external stimuli and shear stresses and as I mentioned before if we have the flow of growth factors in the center in the lumen of the sample then the cells tend to move towards those shear stresses. Of course they grow and proliferate more at areas where they can find growth factors and at the same time the sample the cells tend to favor regions that they have low density so the cells have the option of either moving towards a connective tissue called the xerosa on the outside of the sample or move towards the empty space which is the lumen and of course they rather move towards the lumen. To conclude we are able to demonstrate that we can integrate our exit technique with the current gold standard technology. We can extract volumetric information which can be vital for the assessment of the tissue engineer construct and scaffolds. We can have visualization of the samples microarchitecture. We can carry out quantitative intra-inter sample assessment and of course we can get an insight in the mechanisms of repopulation. I would like to thank everyone from the Institute of Child Health focusing on this guy Maria Ghirli which is the postdoc that runs the show if I may and from the advanced x-ray image group I would like to thank literally everyone because we're a team where everybody actually it's there and helps everyone else.