This terrific article by Chris Lee from arstechnica comes courtesy of Nick Giovinco.
Lihong Wang from Washington University presented some fantastic results on photoacoustic imaging in a plenary talk at the Physics of Quantum Electronicsconference, bedazzling the audience with beautiful image after beautiful image.
Normally, I don't appreciate advertising talks: you know, the ones where the talk covers a vast array of stuff, leaving only time to see the pretty pictures and not get much of the details. However, although this was such a talk, it was a great way for me to catch up with the state of the art in a subject where I actuallyshould know what the state of the art is.
One of the major difficulties in diagnosing diseases like cancer is our inability to see inside the body accurately and inexpensively. MRI is great, but it's expensive and has relatively low resolution; while X-ray imaging had higher resolution, but it comes with dangers. Good old ultrasound is the happy medium, providing pretty good value for money in terms of resolution, safety, and cost.
Missing from that list of clinical techniques, you might notice, is light. This is because light and tissue don't play nice together. Tissues scatter, absorb, and generally play havoc with light, limiting direct imaging to just a few millimeters in depth. Which is a shame, because light imaging has great resolution, uses relatively cheap equipment, and is pretty safe. Let's just say that it is worth a bit of investment to see whether we can do more with light.
Photoacoustic imaging represents a sort of “best of both” approach to imaging. The basic process is that a relatively powerful pulse of light is sent into some part of the body. Wherever it is absorbed—and only small amounts of light are absorbed in any particular place, so it is safe—it generates heat, as its target expands and generates a small acoustic pressure wave. By picking up these sound waves with microphones, an image of the absorbing structures can be built up.
That basic technique has been around for a while—in fact, its origins are in high sensitivity gas sensing. So, we know that high sensitivity is possible, but the imaging systems have been nothing to write home about. This is because the resolution has the intrinsic limitation given by the sound waves.
What Wang has shown is that, with a bit of imagination, a flexible attitude, and a huge amount of work, photoacoustic imaging can be made into an all-singing, all-dancing imaging modality. It's just that you can't necessarily get it to sing and dance at the same time. Let me explain.
In the original photoacoustic imaging technique, the image resolution was about the same as that of ultrasound. So you could use it for whole organ imaging and detect things like tumors, provided they were already a few millimeters large and had started developing their own blood supply—blood is generally where the light is absorbed, so what you actually get is a picture of the blood supply.
But once you have spotted a tissue mass and want to investigate further, you then have a problem. Or maybe not. By putting a photoacoustic imaging system into an endoscope, Wang has shown that you can get up close and personal with various body parts. And, because you are already close, you can be a bit more subtle about things. So, instead of flooding the tissue with light and getting image resolution on the order of the acoustic wavelength, you can focus the light and use the photoacoustic signal as a contrast mechanism for distinguishing absorption at the scale of the wavelength of light.
This has allowed Wang's team to go much further, imaging with a resolution that is capable of detecting capillaries—so basically they have cellular resolution. Because they are detecting absorption, they can distinguish between oxygenated and deoxygenated blood, which is nice. But where they really hit the jackpot is detecting the oxygen saturation of the blood. This is a really important measure, as it seems that changes to oxygen saturation in the microcirculation can be predictive of a patient going into shock. (That is not from Wang, but rather from my faulty memory—if I am wrong, please tell me in the comments.)
So far, so cool. But can we go further? Why, yes, says Wang—how about optically diffraction-limited images? In this case, detecting hemoglobin is not that interesting. Instead, you need to choose something that is in every cell to provide contrast. How about DNA or RNA? That works, but you need to use UV light, which means that it doesn't travel very far through tissue before it is absorbed. So this is useful for imaging excised samples without staining them. Indeed, that is exactly what Wang has shown: by choosing the correct wavelength of light, he can selectively image DNA and RNA and get resolutions that are sub-cellular.
That is what makes photoacoustic imaging so different from any of the standard medical imaging techniques. Using the same basic apparatus—sure, you might have to change a few bits and pieces along the way, but it is pretty close to the same—you get images that have detail resolution appropriate to a huge range of samples, from entire organs down to individual cells. No other technique can do that.