Yttrium-90 Bremsstrahlung Imaging

Yttrium-90 imaging is probably not the first thing that comes to mind when scanning in Nuclear Medicine as an isotope, namely because it is outside of the realm of our usual Technetium-99m based radiopharmaceuticals. I have written about Y-90 and Theraspheres in the past blogs (parts I and II), but this time instead of discussing the treatment aspect, we will look at the imaging aspect of using Y-90 and its Bremmsstrahlung x-rays that it produces.

Imaging was performed on a patient who was diagnosed with hepatocellular carcinoma.  Early in the treatment planning with computed tomography (CT) and interventional arteriography, a large mass was localized in segment 4A/B in the liver.   Majority of the vascularity was provided by the left hepatic artery.  The middle hepatic artery and gastroduodenal arteries were then coiled embolized to limit perfusion.  This patient was on the third treatment cycle of Y-90 Therasphere.  The growth of the liver lesion was stabilized with the first two treatments but there were some suspicions of new metastases.

Fig. 1  Finding the common hepatic artery during the arteriogram.  Quite honestly, I can not figure how the interventional radiologists navigate through the arteries like that... There are no google maps for this!

Fig. 2  The right hepatic artery.  Notice the blush within the lesion as the contrast was infused.  I find this quite amazing to see in person.  In the end the interventional radiologists decide to use the left hepatic artery to infuse the Tc-99m MAA to determine the lung shunt fraction and in turn was used to infuse the Y-90. 

What we wanted to accomplish was to ensure that the Y-90 Theraspheres that we were infusing had truly localized within the liver segments that we wanted to treat.  In the past we had indirectly measured the activity using a dose rate meter (Bicron - mSv/hr) to examine the exposure rates to various parts of the chest and abdomen after the infusion.  The tricky part of the whole thing was that we had never scanned Bremmsstrahlung x-rays before.

Luckily there is an array of information on the internet, but the problem is trying to decipher all this information.  The Y-90 Bremsstrahlung spectrum looks very different, when low energy all purpose (LEAP), medium energy (ME) and high energy (HE) collimators are fitted onto the gamma camera.

Fig. 3  Top:  Represents the Bremsstrahlung spectra on a gamma camera without collimation.  Second:  Represents the spectra with a low energy all purpose (LEAP) collimator.  Third:  The spectra with a medium energy (ME) collimator.  Bottom:  The spectra with a high energy (HE) collimator.  Note 1:  If anyone who is reading this - can someone confirm what the lower peak is?  (It's a scatter peak of some sort - it may be a dumb question, but I need to ask it for my knowledge).  Note 2:  Diagram was taken from:  Planar Gamma Camera Imaging and Quantitation of Yttrium-90 Bremsstrahlung

Fig. 4  There is a characteristic x-ray photopeak between 75 - 79 keV with collimation.  This is the range where we decide to use our window for imaging.  However, note on the far right in Fig. 3 TOP, and in Fig. 4 another interesting peak occurs on around 180 - 210 keV.  This is probably the result of septal penetration due to some high energy Bremsstrahlung x-rays.

Furthermore, where do we centre the energy window and how big should the window be?  There is a characteristic x-ray peak, as indicated above, around 75 keV.  But what is interesting is that some protocols use a 79 keV peak with a 26% window, while others use a 90 keV peak and a window width of 15%. So what is the optimal imaging parameter in regards to the energy peak and energy window?  I am not really sure, and I think it really has to do with the testing that we need to do on our camera system (Seimens Symbia T-6) to figure this out in terms of sensitivity, resolution and target to background ratios.  However we did not have the luxury of time to test these parameters.  When we first started the trial, it was never our intention to image the Y-90 Therasphere patients, and this patient was a "one off".

However there is some agreement with respect to the collimation.  From what I have read between ME and HE collimators, generally speaking, most have used the ME's.  In our case, the administered dose was approximately 2.03 GBq and imaging was performed 3 days later, we opted for the ME's since there were not going to be any significant differences between sensitivity and resolution between ME and HE collimators for equivalent energy windows (ie. whether we were using a 26% or 15% window).  

So these are our results:

Fig. 5 Coronal fused section from the SPECT/CT (Symbia T-6) with localization in the segment 4 of the liver

Fig. 6  Sagittal fused section from the SPECT/CT (Symbia T-6)

Fig. 7  Transaxial fused section from the SPECT/CT (Symbia T-6)

The MIP after reconstruction looks like a "big blob".  It didn't look too bad with regards to the planar images which we also obtained.  We had captured them to quickly examine the target lesion to background liver fraction.

Fig. 8  Anterior image.  We imaged for 600 secs and obtained approximately 700K to 1700K total counts (posterior/anterior images respectively)

Fig. 9  A quick target lesion to liver background ratio was calculated to determine the amount that remained in the liver 3 days after the infusion

In the end we decided to use the MD Anderson's protocol.  It was a shot:

NM Planar & SPECT Y-90 Bremsstrahlung Imaging:

– 79keV/26% window, MELP collimation, 128x128 matrix, 4.8 mm2 pixels, 128 

views/360°, 28 s/view, non-circular step-shoot

So we did all this, BUT some might say... what about imaging the patient on the PET/CT unit? Well I wish we could, as you know, Y-90 is a beta emitter and this is certainly within the realm of possibilities, and with it, a huge array of literature on this as well.  We thought about it, but we couldn't get imaging time on the PET/CT unit... so we opted for regular gamma camera imaging.  QED.