Interview with Hamamatsu on PET, CT, MRI and 64 pixels in 1 inch square
Hamamatsu’s Richard Harvey spared a few moments to talk to us during the inaugural MediSens conference, held in London in 2016.
Hamamatsu is a company that makes sensor chips, tell us more…
That’s correct. We make sensor chips but we also make back-in tube detectors which are the precursor to the current technology and we also made complete systems as well for end users.
And what about the topics at the conference that you’ve experienced so far, you’ve been in to many sessions?
I think excellent, they’ve been very good, very enjoyable. I’ve been to all the sessions, and they’ve been very interesting and particularly the ones on PET and the combination of PET with CT and PET with MRI have been particularly interesting for us.
Is that because the sensors that are being talked about in those sessions are relevant to your area of expertise or business?
Absolutely. The detectors which enable you to do PET plus MRI, are silicon photomultipliers (that we’re showing today), at Hamamatsu we call it MPPC. It’s a silicon technology based on avalanche photodiode but you array these up to the common output. That allows you to do photon counting at room temperature as it’s insensitive to magnetic fields.
With an MRI scan you have a very very high magnetic fields (3, 4 or even 5 Tesla) and with the traditional photomultiplier tube technology, that just doesn’t work; you can’t combine that detector with an MRI machine. But with the silicon PMs, because they’re insensitive to magnetive fields you can combine that with the high magnetic field MRI and you have multimodality imaging, so that’s really the key enabler.
You mentioned “at room temperature”, why is temperature a limiting factor?
Particularly, not just in medical imaging but in general applications, when you’re measuring very very low levels of light, higher temperature means higher noise and that means you can basically measure less signals. It’s also your signal to noise ratio; you cool things down and then you can get better signal-to-noise but with the silicon PM now, the signal-to-noise ratio even at room temperature is good enough to do single photon counting.
What sort of devices are these types of sensors going into?
The plan is that they are going to the full scanners, so this is the preclinical scanners and then into human scanners as well but they’re also actually usable in a wide range of applications beyond the medical arena. So, the first people to use these were in the higher energy physics market so it’s as academics in big Collider experiments. Initially it was avalanche photodiodes but now they’re shifting towards this technology, Down at CERN, for example, but also for things like range-finding, so LiDAR in the automotive sector could also use this technology.
Talk us through what you’ve got here…
This is just showing a range of the different packaging options, all of these devices here are just single element detectors. They would conventionally be used in, say something like fluorescence detection, it could be used in range-finding, the actual devices which we use in medical imaging are such as this, so this is an 8 by 8 array (see right) silicon photomultiplier.
So, rather than having one individual sensor, you’ve actually got 64 sensors.
Absolutely, yes, that basically allows you to build up the camera effectively, so these are effectively pixels in a camera but they’re just very very large and maybe you can’t see it on the camera but in the center of each channel is a little white dot. That’s your TSV, we’re using this through-silicon-via technology which allows you to do four side buttable detectors so you can build these up to any size you wish.
You see here we also have a 4 by 4 which is very small, 8 by 8 but we can also do a 16 by 16, we can do linear rays and there’s no reason why you couldn’t do other shapes as well if you so wished.
…you’re talking about a sensor that’s about an inch square roughly and it’s only got 64 pixels but it’s low resolution because it’s looking for extremely low energy volumes of particles coming back?
This one here’s an example where you actually have 1 by 1 millimeter, so you can see it’s the same form factor but a lot smaller channels that gives you a high resolution but what we had typically happen in the scanner was you’d have built these up into a big ring around the patient so you wouldn’t just have one of these, you’d have a whole ring and you’d have multiple banks of these detectors and then you’d couple a scintillator to this as well which then increase your resolution as well, so you’d have segmented scintillator crystals, it’s dual layer segmented to increase resolution further as well.
So when you see people being fed into in an MRI scanner it is a giant camera and the ring itself that goes all the way around is the sensor that gives you the pictures?
Yes. On a PET, you have a whole ring and you can see potentially here this is a schematic, yeah, the patient surrounded by a ring, so traditionally that would have been Hamamatsu photomultiplier tubes, now they know people are moving to the Hamamatsu silicon PM.
You mentioned automotive as well, do these types of sensors also apply… is the same R&D used in other areas as well?
Particularly in Europe the largest market for Hamamatsu is in the medical so we have a lot of business in dental, for example, dental x-ray, but really medical is a key application area but it varies from division to division and by country to country as well.
Certainly, the silicon PM which we’re showing today is a key technology for the future but some of the more traditional technologies that we’ve been manufacturing for 50 years are still used prevalently today and they enable things like they were going into systems for blood analysis, urine analysis, blood gas detectors.
So, the whole product range that we manufacture can be used in medical but it can also be used in other application areas completely separate to that as well, so this is why I was saying this can be used in medical, it can also be used in the automotive sector. They’re very different but they can they can both benefit from the technology.
Lastly, just talking again about the conference and our surroundings here, have you enjoyed the event?
It’s been brilliant. It’s the quality of the people that actually came and people who are doing the talks have been fantastic and particularly, as this is the first year, then I think it shows it’s going to be a good conference going forward as well, so we’re looking forward to attending again.