Sunday, March 02, 2008

The First PET/CT Workstation Shootout:
Beam Me Up, Scotty! No Intelligent Life Here!

I've just returned from Las Vegas, having attended the Stanford University 2008 PET/CT Symposium. In my old age, I've decided that a good meeting teaches you some new things, but by and large reassures you that what you were already doing in day to day practice isn't too far below standard. This was a good meeting. There were numerous lectures by the luminaries of PET, as well as several case interpretation sessions, and what was (for me) the piece de resistance, the first-ever shootout between PET/CT workstations.

Of course, the second-most important part of the visit was a pilgrimage to the Star Trek Experience at the Las Vegas Hilton, which includes the Museum of the Future. The latter has changed a little, with the addition of props from the newest series, Enterprise, such as Captain Archer's uniform, communicators, and various other toys. My favorites, though, are still Dr. McCoy's medical kit and the tricorder patched into a network of vacuum tubes from the Original Series episode, "City on the Edge of Forever." The rides of the Experience have not changed at all, although I am sorry to report that the sound and video of Commander Riker on the Enterprise-D viewscreen in the Klingon Encounter have lost their synchronization. I took the "backstage" tour, and got to see how all of this works. We were all sworn to secrecy, but I can tell you that the riders don't actually go into space or the future. Darn.

There were a couple of photo ops that came with the tour package, so I took my turn in the command chair of the Enterprise-D. Sadly, the weapons were off-line, or some folks out there would be in real trouble.

The Borg are hanging around the Experience as well, and if you aren't careful, you might get assimilated, as I almost did:

Hmmmmmm...notice the green overtones. Could the BorG Eventually conquer Earth after all? Resistance is futile....

I also had the chance to see the comedian Carrot Top live at the Luxor. I caught his act here in the South at our State Fair, and he is absolutely hilarious. In Las Vegas, however, he can really let loose, and I haven't laughed so much in quite a while.

But on to the Shootout. This was a contest pitting four scanner manufacturers' workstations against each other. There were two players from GE, the Xeleris 2 Volumetrix, and the Advantage Windows Body Share 2. Siemens submitted the Syngo TrueD, and Philips brought its Brilliance platform. All were run by physicians experienced in their use, with the screens projected for the audience to see. All used dual-screens, although the Philips folks couldn't get the second monitor to project onto the overhead.

The audience was allowed to score each station's performance for three cases on the following criteria: ergonomics, speed, and display format, as well as an overall score. Sadly, these ratings were not collected, although the course director, Dr. Quon, told me that there might be an opportunity to enter them online at a later time. But perhaps more importantly, developers for GE and Siemens (I'm not sure about Philips) were present to see how their offspring fared (and perhaps to get a glimpse of the other guy's program.) I hope they paid close attention.

Each workstation was put through its paces with everyone watching. Here is a the script for one of the cases:

1. Open the PET/CT dated 5/2/07 and display all of the following simultaneously (or show as many as possible):

a. Coronal PET, PET/CT fusion, and CT

b. Transaxial PET, PET/CT fusion, and CT

c. Sagittal PET, PET/CT fusion, and CT


2. Threshold the MIP PET image appropriately.

3. Rotate the MIP image.

4. In the transaxial slices, start from the thymus and scroll inferiorly until reaching the stomach region.

5. Select and triangulate on the two lesions in the region of the fundus.

6. Enlarge the transaxial PET, PET/CT fusion, and CT images, and measure the maximum SUV of both lesions.

7. Window the PET/CT fusion image: go from all PET to all CT to demonstrate the presence of hypermetabolic gastric masses.

8. Scroll to the subtle focus in the periphery of the liver and measure the maximum SUV.

9. Window the CT image in an attempt to correlate the lesion.

10. Open the comparison PET/CT dated 9/5/07.

11. Select and triangulate on the lesion in the liver and display the focus on transaxial PET and PET/CT fusion.

12. Display the PET/CT from 5/2/07 and 9/5/07 simultaneously to compare the liver lesion side by side.

13. Window the 9/5/07 PET/CT fusion image to demonstrate the anatomical location of the liver focus.

14. Measure the maximum SUV of the liver focus on the 9/5/07 PET.

15. Window the 9/5/07 CT to best display the subtle liver lesion.

16. Scroll to the gastric lesions, window the CT and PET/CT fusion, and identify the site of the gastric lesions. Measure and compare the maximum SUV to the prior scan.

17. Display DICOM information and demographics.

This more or less simulates how one might approach a real case. It would be very tedious to describe the events in detail, and the winner is in the eye of the beholder. Each station had its pluses and minuses, and given the subjective nature of the criteria, I'll bet a $100 chip that everyone in the audience came away thinking that a different product was the winner. Frankly, I thought that all of the products were quite fast, although the Xeleris appeared to choke while loading the second and third cases. The AW and the Syngo TrueD both ran significantly faster than the versions I use. (I guess it's time to hit up the powers that be for some upgrades!) As far as Ergonomics and Display go, this is almost purely a subjective choice. I thought the TrueD was the winner overall, followed closely by the Philips Brilliance. The GE (fraternal) twins were powerful, but I don't think their controls are particularly intuitive. The other two were a little better on this. Creating a custom display by yourself was said to be easy on the Brilliance (although it wasn't demonstrated). I know that doing this on the AW is difficult, and it is pretty much impossible on the older Siemens eSoft. I don't know about the TrueD's capabilities here. I haven't a clue about the Xeleris.

The sad fact to me is that none of these stations are as easy to use as they should be. Of the four, I would prefer the Siemens TrueD, acknowledging the fact that it has some esoteric controls that one must master. But somebody needs to do better.

Dr. Quon does plan to include some of the other non-scanner vendors in next year's shootout, such as MIM, TeraRecon, Voxar, and so on. You have to ask why few, if any, customers buy, say, a TrueD for a Philips scanner, or an AW for a Siemens scanner. One usually just takes what comes with. But I think the best option for PET/CT viewing might well come from the PACS vendors. At RSNA, I saw some preliminary versions of PET/CT reading modules from Intelerad, Agfa, Dynamic Imaging (GE!), and Amicas. I know ScImage has one as well, and I'm sure there are others. From my brief peek, I think that at least some of these could do the same tasks as outlined in the shootout script, and do them in a more intuitive and rapid manner than we are seeing with the high-horsepower stations.

Eventually, there will probably be some narrowing of the spectrum between the PACS workstation and the high-level imaging workstation such as those above, with thin-clients filling in the gaps. But PET/CT doesn't really require quite the level of processing we see in some 3D renderings of the beating heart, and I suspect we will see more and more PACS programs that can handle it readily, and as well (or better) than the big ol' workstation.

In the meantime, Live Long and Prosper! (Which is difficult to do in Las Vegas for more than a few days).


Patrick said...

I have 2 TrueD/3D workstations, 2 Xeleris, 2 esoft.
For PET-CT reading, I only use now TrueD alongside 3D.
esoft and Xeleris belongs to the 20th century.
Did you try Osirix? I have to say I just purchased a MacBook Pro after more than 20 years of fidelity to DOS/windows. It has a PACS.

As an additionnal question: does WB diffusion MRI competes hard with PET-CT in your country at this moement? How do you see things going on this side?

Patrick, France

Dalai said...

I have yet to try the recent Osirix, as I haven't got access to a Mac. I'm sorely tempted by the Mac Book Air, however!

Diffusion MRI is making inroads, and in fact there was a talk about it at this PET meeting. I don't think it's there yet, but it will be a competitor in some areas, especially for liver lesions.

Anonymous said...

Have to give the Philips Radiologist some credit for being able to make all those dual monitored based protocols run smoothly on-the-fly when forced to work with a single monitor environment due to the projector KVM "malfunction".

Medical Imaging for IT Professionals said...

Ya' gotta' give OsiriX a try. It's well worth wrangling a Mac from a friend or coworker or enemy, for that matter. I've spent the last twenty years fiddling with PC based systems, Windows and Linux. I use them professionally... but deep in my little heart of hearts, I love the look and feel of the Mac systems. And... OsiriX gets better with each release.