 Efallai, hefyd, fel oedaf i'n gweithio'r ysgrifennu. Felly, rwy'n meddwl hwn i'r Christy, i'n meddwl pan am y maen nhw'n meddwl yn ynghylch wedi defnyddio, rwy'n meddwl hwn i'n meddwl hwnnw yn meddwl hynny. Rwy'n meddwl mewn meddwl hwn ymlaen na'r gweithio'r stwyffau i'r gweithio'r gweithio, rwy'n meddwl hynny'n meddwl hefyd. Efallai, roeddwn i'n meddwl hwnnw'n meddwl hwnnw'n meddwl Being at the Christie for about 18 months now and specialist in positron om phenomena which is what I'm going to introduce you to this morning. Now you may be familiar with the idea of antimatter, that there is an opposite version of the stuff that we're all made up of and there is an idea that appears in science fiction novels quite a lot, that you could meet your antiself and that you shut hands with them you disappear in a burst of energy. Well, that actually potentially could happen. Einstein, in the early 1900s, published a number of seminal papers and one of them included the idea of energy mass equivalence. So this equation equals Nc squared essentially tells us that energy and matter are two forms of the same thing. So if you were to basically dissolve stuff it would turn into energy. For those trekkies amongst us I think I'm fairly safe ground presuming there will be some in the room. You will know that the warp core in the Starship Enterprise is powered by a switcher mechanism although how they actually keep the anti-matter and matter separate is still a bit of a mystery to me if you want to enlighten me over lunch that'd be great. This process of pulling matter out of energy is actually why we're all here. At the beginning of the universe we had a big bang, a big burst of energy and somehow out of all that matter stuff that makes a pause, that makes up everything around us emerge dominant over anti-matter. It didn't just have equal amounts of matter and anti-matter and it all dissolved back into energy. Matter won out. It's one of the big questions that's been asked at CERN at the moment is why that was, why we don't have an anti-matter universe, why we have a matter universe. So actually they're looking at what happened right down here, right at the sort of part just before the big bang kind of really exploded and the universe began to take shape. This all sounds a bit abstract and a bit weird and we're back in the distant past so why am I talking about using anti-matter emitters in hospitals? Well it's because we can actually see this stuff in real life. Its existence was predicted by possibly the grumpiest physicist ever to have lived, Paul Dirac. That's him smiling, it's not very convincing is it really. He came up with this spectacular equation which I won't go into any depth about because basically I haven't touched it since first year of physics degree and it basically tried to resolve electromagnetism and quantum theory and it predicts the existence of the electron, so all the properties of the electron it predicts. But it also predicts an opposite number to the electron so something with equal mass, exactly opposite charge and several other similar properties to the electron. So of course when a theoretician like Dirac goes there must be a thing experimentalists go we must find the thing and Carl Anderson got cracking and a mere four years later discovered the positron. So this is his experimental evidence. I've actually seen this photographic plate and I spent rather a lot of time in front of it. It was slightly embarrassing. It's kind of one of my first dates with my current partner and went to an exhibition at the Science Museum and it was almost like he stood me in front of a cathedral window. I was like gawping. In fact he's still with me, it's a good sign I think after that experience. So this is a cloud chamber, a picture of a cloud chamber and cloud chambers work when particles pass through them and they leave like a condensation track through the cloud chamber and if you put a magnetic field across a cloud chamber a particle with a charge will bend in a certain direction and how much it bends depends on how heavy it is. And this bend here is exactly what you'd expect from an electron but the other way around. So if you had basically this bend going exactly the opposite way that would be an electron. So this is a positron because it's behaving exactly the same way as an electron but a mirror image. It's kind of like if you were Newton sitting under a tree and instead of having an apple landing on your head it kind of flew off somewhere in the opposite direction. That's the kind of effect we're kind of seeing here in this cloud chamber experiment. Anderson got the Nobel Prize for this in 1936. I would love to say that this proves that if you want a Nobel Prize you should be an experimental physicist like me but the Higgs boson kind of disproves that theory. So where do we find this antimatter? This is a table which adorns many science laboratories up and down the country otherwise known as a Segre chart you may have come across it under that name where we call it table of the isotopes because that's what it is. Each row is the same chemical element and it has the same number of protons so it has the same number of electrons, the chemistry is exactly the same but a variable number of neutrons making up all these different variants of carbon for example on that row. And the black ones in the middle are the stable elements. Now what you need to know about isotopes is generally they just want a quiet life. They just want to sit there and be. They don't want to be doing all this radioactive decay nonsense. They want to get back to this line of stability. So either above the line they've got too many protons so they'll try and convert a proton to a neutron and get back down to the line. That way, below the line, they've got slightly too many neutrons so they try and convert a neutron to a proton to try and move upwards to that line of stability. And all of these ones, these are examples of some of the positron emitting isotopes that we can use. So these ones decay by converting a proton to a neutron and it's got to do something with that positive charge. So it goes, I'll create a positive electron. That'll work and chokes the positive electron out of the nucleus and that's how we get a positron decay. So it's conserving charge. It has to have the same amount of charge before and after that interaction and we get a positron generated in that particular form of decay for slightly proton rich isotopes. You'll notice that these are quite handy. Carbon, nitrogen, oxygen, fluorine. These are the kind of atoms that occur in biological molecules the kind of things that we're made up of. And you can stick these atoms in place of something that is naturally occurring. This is glucose. So instead of a hydrogen here, we can stick a fluorine on and fluorine is quite a small atom similar to electron configuration on the outside to a hydrogen. That will behave exactly the same as a non-radioactive form of hydrogen. So this is fluorideoxyglucose. It's one of the most common traces we use in positron emission tomography. So if you're coming for a scan using this particular type of tracer we'd inject you with positron emitting sugar and watch where it goes. So you want to see how it accumulates in the very metabolically active areas of your body and use that as a measure of how active either disease or healthy tissue was in those particular parts. So we basically got somebody who's emitting positrons after this injection so how do we actually go about detecting those? And in order to demonstrate this I have a small illustration at the front. For this I need someone with serious anger management issues. Has anyone had a bad summer? Is that home with the kids a bit too much? I'm going to pick on somebody if you don't volunteer. Bryn, do you want to go just grab someone? No, he's not going to sign up. Bring some glamorous assistants. You need at least two glamorous assistants. So either friends or people you don't want to be friends with after this. It could go either way. So let's pull this forward a tiny bit. We need two people to assist. You can do that later. I need you to do this bit first. So this was designed by a colleague of mine Richard shortly before his retirement. There is a diagram available for anyone who wants to make one in the comfort of their own home. So it has ping pong ball guns on each end and a couple of little catches so that when you smack the door knob in the middle it basically releases the ping pong balls. Now, what do I mean by why the heck have I got ping pong balls and door knobs and all this kind of stuff? Well, as you know, electrons are blue because when you pull a plug out the wall of electrons fall out and it's... No, that's okay. So electric... This is actually a kind of audience where this joke actually works. I love it. This is electrons are blue. So obviously positrons are red. So we have Richard's gamma ray gun complemented by Heather's positron hammer. It is secured by a gaffer tape because I gave it to a very angry physics teacher once and that didn't work out very well. So if you could apply suitable force here without actually breaking it again, that would be lovely. So positron meets the electron, what happens? I am going to need a couple of catches if you could stand here. That would be lovely. I'm going to stand well back so no one plays for Hebdenbridge cricket team then. I can't gather by that. Again, again. Okay, so that's where your taxi catch is. So if you just poke through the actual barrel of the gun, there's a little rod there. And then put it on the catch for me. If you pull it, it can just kind of disengage a little bit. Okay. Okay, so we'll do it a slightly different angle this time. So bear that in mind. Catch us please. In the hat! Oh mate! See what? One more, one more, one more. It will become clear by the way if you just think I'm just playing with toys for lolls up here. I will explain what I'm doing. Oh, we know. It's not a bad way to spend a Friday morning, is it? Let's be honest. Right, okay, off you go. So I can deduce from that that you're not particularly efficient radiation detectors. But apart from that, thank you very much for your enthusiasm. Thank you very much. So what was all that about? Well, we have in current generation PET scanners, we have this big ring of detectors. So, hula hoop. Anyone hold this? Let's go on, Bryn. This is my youngest Bryn. You just want to hold that for me, just so we can see it. Could you hold it up for me? That'd be ace. So what we have is a big ring of these detectors, and the patient goes in the middle, as you'll see as all physicists like to oversimplify things. Patients are always pink ellipses of uniform density. So we have a radioisotope's decaying. It's emitting positrons. The positron runs into the electron somewhere inside the patient's body. Matter meets antimatter. The two cancel out in give us energy, and we have the energy emerging from the patient as two 511 kV gamma rays in opposite directions. Now, the 511 kV is important because that's the amount of energy that you have bound up in one of these positrons or one of these electrons. So that's kiloelectron volts. It's a very tiny, tiny amount of energy, but it's enough to get out of the body for us to detect it. And these pairs of gamma rays, if we can be looking at our detector, this is where it would have been helpful to disentangle these before I started, but never mind. See what we can do. There we go. So we've got a detector that's looking for these gamma ray pairs all the time. So we have two, these gamma rays arrive in a very small window of time. Thank you, Brian. So you have one detected at one side. We have another detected at the other side. We have... I'll collect that later. So if we have another one collected over here, and another one over there, and another one, the observant of you will notice that the ribbons coordinate bits, they would be cooler than the ping pong balls. We have another one over here. And another gamma ray detected there. So you can see if you did this over and over and over again, we just raised it up a tiny bit more branches. You see that those lines would tend to overlap most at the points which were sources of the positrons. So we collect all of these overlapping lines. We overlap them in what we call imagery construction, which is basically a load of mats. And then we can build up this image of the distribution where the positron came from. All right, thanks, son. How has this actually done in practice? I mean, I think whoever decided let's inject people with radioactivity with antimatter emitters deserves a large glass of champagne to start with. But how we actually do it is that it starts in 1952 when you have something that looks like they came out of a B movie, maybe a horror B movie. You have these scintillation detectors that turn gamma radiation into light on each side of the patient and you have a perspex grid. And you just step them across this perspex grid and you have a chart recorder. Who remembers chart recorders? Yeah, I caught the tail end of chart recorders. And you end up putting a darker mark where you've got a higher signal. And you end up creating this, you draw around the patient's head and you take measurements of all these different points and you say, we've got slightly more signal here. That means when this example is slightly more blood flow at this point. And that moved into a multi detector system. So basically they just added a ray of these things so it was a bit quicker in the 60s. In 1968 we ended up with this. This is the electronics. Those of you who have a laptop screen and keyboard attached to each other or not. Will know that this is vastly less powerful than what you've probably got in your phone. This is the electronics that was reading out all of these individual detectors. And we could actually take these, we call the gamma ray pairs as we rotated these detectors around. So you stick the patient's head in there and the body coming out there and rotate this thing around them. Probably give them an eye mask because they weren't too aware of what was going on. Looks a bit macabre that machine. And this is a brain. You kind of can see that's a brain. Kind of a cauliflower type thing. Taking up this radio tracer shining nice and bright. And if you look on the left hand side there's less blood flow on the left than there is on the right. In this particular example. This is what we're using now with commercial systems. It's actually made up of these things. I'll pass this round so you can have a look because I know you'll appreciate it. So it's actually got scintillation crystal elements in here in lots and lots of tiny ones. And it's backed by photomultiplier tubes. So we have the light being generated in the crystal being funneled down towards the PM tubes that convert the light into electricity into a measurable current. And the grid of PM tubes can map exactly where the gamma ray came in. So we can say within this tiny little square a few millimetres by a few millimetres. And this is the detector area that I worked on for my PhD. We can say within a few millimetres where this big detector ring. So let me pass that round. It's heavier than you think and that's all in the crystal. So remember that number. 12,096. This is the Scannerach Commission for Manchester Running Firmory. I loved it very much. It's just been decommissioned and it's a very sad day. But they've upgraded so I'm very happy for them. Anyone want to bet how many detector elements we've got in this? We had 12 in the one that I started with in 2000. This was installed probably 2007. You're a bit generous there. Not quite that exciting. Slightly lower. This number is obviously not impressive enough for you guys. I'm happy that I've basically got three times more detector elements within said five, seven years with this detector. So it's about 20 centimetres long. And this scanner has actually got a CT scanner on the front. So that you have this map of where your trace has gone showing how active all these processes are in the body. What process you see depends on what tracer that you give. And that you can map that directly onto the structure of the body because you can take a CT scan with a patient in exactly the same position. So most hospitals now will have a combined PET CT system. So the kind of big ring on the front is the CT scanner. So 3D X-ray image. And the back is the PET 3D radiatrace map at the back. I've given you the kind of short version and I'm aware I'm going to have to hurry up because that already took the best part of half an hour. But just to skip through and just highlight a few of the challenges that we have with our systems. Obviously the positron travels some distance before it meets the electron. And that means that this introduces a degree of positional uncertainty in the system that's dependent on the energy of the positron. So actually depending on which radioisotrace you choose you can get a different spatial resolution, which is weird. But fortunately fluorine has got quite a short positron range so this is a minimal problem for most of our studies. Also positrons and electrons are completely stationary when they meet. So you have some residual momentum that means that the gamma rays are quite back to back. And again that introduces some positional uncertainty. And all that's dependent on how you build your detector to start with and reconstruct the images. What I've told you about so far is true coincidences back to back gamma rays. We can also have ones where one of the gamma rays is scattered and this dotted line is what we record and that's wrong. If we're working at very high radioactive concentrations we can have two positrons meeting two electrons at the same time which also needs to be corrected for. We also need to correct for the fact that some of the gamma rays don't make it out with a patient at all because they get either absorbed or scattered out of the detector. There's all kinds of clever data corrections that need to be done and these are all actually now built into the image reconstruction loop. So actually if you're looking for a computing challenge medical image reconstruction is a field that's growing very very very quickly and can save a lot of radiation dust to patients and a lot of time in the clinic. One thing that's been particularly exciting over the last few years is that our detectors have got super super fast. In order to do this conventional detection that I've told you about where you can say that one gamma ray, the two gamma rays came from the same positron meeting the electron you need a timing resolution of about four to six nanoseconds so that's four to six thousandths of a millionth of a second. Our scanners are now down to less than half a nanosecond so what we can say is that positron met the electron there with a certain degree of uncertainty really improves the accuracy with which we can pinpoint where the signal is coming from. It directly improves image contrast and quality and allows us to scan patients with low radiation dust and also a lot quicker. So this is our current standard if you're going shopping this is your menu of current pet scanners. Ignore Philips the rubbish. This is being recorded. The whole pet community in the UK would agree with me so I'm safe ground there. Some reason Philips have really taken the ball off the eye of the ball with pet scanners in the last five years is real disappointment because at the moment it's basically a head between G and Siemens and they're pretty close to be honest. Fast insulation crystals you can scan about twenty centimetres of the patient it won't go. Really fast response and pretty good spatial resolution. But there's only so good you can get with pets. For those of you with any image reconstruction nerds in the room I'll move on. The next leap up from that is actually to use solid state detectors rather than to use insulation crystal and photomultiplier tube detectors which are a lot quicker. Philips kicked this off and again I don't know what happened to them there. He tried to match them and didn't quite manage it. Siemens played a blinder about 18 months ago and brought us down to 214 picoseconds. What that means in reality is we can tell with a Siemens system which is what now they've got a Manchester own infirmary where a positron metal electron within three centimetres. That's how good that system is. So we've got this system. We've got these lovely images of these bright spots of a lung tumour, the white bit there. If you calibrate your pet scanner properly and it's actually quite simple to do you measure how much radioactivity you've got and put it into an object with so many millilitres hence kilobeckles per mill. You scan it and you get an intensity of pixel values that you can then say that corresponds to so many kilobeckles per mill you calibrated your scanner. Once you've done that you can actually measure how much trace you've got in different bits of the body. That's incredibly powerful because you can quantify how active all these processes are, how much of this trace was actually going through these different bits of the body. So just some example images to finish off. This is the first scan we did of a patient with epilepsy at Manchester own infirmary and it turned out to be beautiful. It shows what you want on your first scan. So this is a brain. It looks a bit better than one I showed you in the previous picture, doesn't it? It's a colour scale that goes from purple with not very good uptake of the traces through to bluey green and then yellow and red and then into white for lots and lots of uptake. So lots of this glucose is going to the outside of the brain which is the most metabolically active bit of the brain. In epilepsy when you have an epileptic fit you get loads and loads of glucose in that really active area. But in between fits, which is when we scan because that's when the patient stays still, we can see an actually lower glucose uptake in the same area. So we're looking for a lower than normal. Can anyone spot it comparing right to left are there any bits that you think are actually a bit lower on one side than the other? How about there? So this is your temporal lobe and actually epilepsy of the temporal lobe is fairly common if you're going to have epilepsy it tends to affect that bit of the brain. So we were able to say yes it's this bit of the brain that's affecting this particular patient, the medication they were on wasn't working. They actually had part of their temporal lobe surgically removed to control their symptoms. So yeah he's doing a lot better because we were able to say take this bit of the brain out and it'll help. Some heart scans, because we've got a detector that's all around the patient all the time, we can take pictures of this radiotrace distribution as it's changing. So we can inject someone and see where it goes. So in this example you've got the blood that comes into the heart in the middle and it distributes that into the heart muscle. We're looking at the delivery of blood to the heart muscle which is often a problem that underpins a lot of heart conditions. And we can get diagrams that look a lot like this. So we can look at the heart under stress conditions, heart under rest conditions and see how well the blood flow is ramped up in stress conditions to compensate. And because we know what's going into this system we can create a fairly simple model of how that distribution is happening. So there's the blood concentration and that will go in and out of the myocardium, the heart muscle. We've got that variation in time. We've got the variation of this in time from all these different bits. And then we can build up a model of exactly how much of this of blood flow is going into the heart muscle in terms of mills per gram per minute. Why is this important? Well, believe it, this is a heart scan. So we see mainly the left ventricle because it's the biggest chunkiest muscle in the heart. You just about see the right ventricle on the side there, that shadow. It's kind of a cup shape. It pumps up like that. So if you cut through it that way, you get a series of donut shapes. If you cut through it that way or that way, your image is. Not for real, fortunately. Without taking the heart out, it's great. So you get U shapes if you go that way. And that looks fairly even, fairly normal uptake of our radiatrace. You would think that was good. And it is. These numbers show that the blood flow under stress conditions is ramping up exactly as it should. There's enough getting through. This one looks similarly good. But the numbers tell us it's not uniformly good. It's uniformly bad. The blood flow is not ramping up to the same extent under stress conditions. And the reason for this, if we do more detailed CT, is there are blockages in all three of the arteries feeding the cardiac muscle. So this means that delivery of the blood throughout the heart is hampered to the same degree all over the heart. And you would miss that if you didn't have the numbers. That's one of the powers of doing PET scanning and going to the effort of getting these numbers out of the scams. This is what I'm working on at the moment at the Christie. We are using PET to not only look at how extensive cancer is, how aggressive it is, and also how it responds to treatment, but also able to image the delivery of the treatment because some of the treatments we use are radioactive as well, and they give off some positrons. So just to show you some example images, this is quite a big chunky tumour in the liver. You can see that on our baseline scans it's quite extensive. We've got the PET in an overlaid in a ready colour scale under the CT that's grey there. Fortunately our therapy seems to have worked because it's a lot smaller on the follow-up scan. These are just different ways of imaging the same thing with different technology. I won't touch on those, but this is an image of the actual distribution of the therapeutic age that we've given. So this is at Iterium 90, which gives off a little bit of positrons. So whilst we get a crummy PET image we can actually see where it's going. So on the projects I'm working on at the moment is to say we've got all this data on the same patient. If we know that this therapeutic dose has been delivered to there, can we predict what that scan is going to look like? So that's one of the things I'm working on at the moment. Getting numbers out of this is a nightmare, by the way. That's why it's holding me up. So what's next? That's what we're doing now. What's next? You can go large. Bigger is better. Let's have a massive PET scanner. So the Americans, to build something large, got to be the Americans, hasn't it? So they've got this absolutely massive PET scanner that they're building in the US. I've nicknamed it Coffin PET because it literally covers the patient from there to there. So if you think we've got claustrophobia issues in our current scanner, I have no idea how our patients are going to hold up in this. But it's really, really quick and you can image the whole of the body in real time. So if you've got disease that's affecting the whole of the body you can model how that is taking up trace. You can model how various bits of the body are behaving differently to others. As a research to it, it's incredibly powerful and the gossip is that Edinburgh are already trying to get the money for one. £10 million, if anyone's got deep pockets. The other thing which I'm particularly interested in with being at the Christie, we've just done proton beam therapy in the last year is that when you fire protons or carbon ions for therapeutic purposes at a patient, you can actually create positrons inside the patient. So if we could put a PET scan on the back of our proton beam therapy machine, we could image how successfully we've delivered that therapy before they even get off the table because we can image where the positrons are being created. So I'm hoping they'll get me to play with proton beam therapy at some point and put one of those on the back. The other thing is that we're getting huge quantities of data out of our PET scanner. So if we could, we need a clever way of processing that huge volume of data. You probably hear a story every week about how radiologists are going to be replaced by computers. You have to train that software as carefully as we would a junior doctor and that is the stumbling block at the moment and that's why it's not good enough to trust it yet. It works in a situation where it's being trained, it can't be generalised to other centres. But there are ways forward where we could perhaps use AI techniques to pick up subtle details that perhaps we would have missed, a clinician would have missed. So it acts as a second reporter, another pair of eyes on what we're already doing. And it allows us to pick up subtleties on poorer quality images which is what this paper is about. Both of those are open access if you want to go and have a look. So that's me, hopefully a little bit more about why we inject people with radioactive stuff now. As I mentioned, I'm a wearer of many hats. But I think we might have time for one question. One more question. But do find me on Twitter and ask me questions about antimatter any time you like. Thanks very much.