Iâ€™ve been experimenting with upgrading my hardware and software ever since we got a 16 slice CT machine. Before that, whatever software I was using (eFilm, Osirix) seemed to do ok with the images. But the advent of the CT study with 1000, 2000, or 3000 images really challenged the average mac or PC. There was a lot of hang time, crashing, image lag, and other issues that seemed to be factors. But what is the optimal hardware and software to use? Who has time to research these questions? Has someone discovered a better way? I created a survey of the ACVR membership to find out.
34 ACVR members answered the survey, for which I thank them profusely for their time and effort. Itâ€™s something that affects all of us who view images in a busy clinic or teleradiology setting, and our education has not prepared us to know what hardware and software combinations work best. I collected all the data so that we can make some informed decisions and recommendations for ourselves, and to our colleagues in IT.
What software are people using?
The two main software players are Osirix and eFilm. A few more people used Osirix than eFilm (16 vs 13 of 34). 5 people used other systems (Coactiv, Thinking Systems, ClearCanvas, Genesis) as alternatives to our well known software. I asked about RAM (random access memory) and software version, because Iâ€™ve found myself that this makes a big difference in software performance. The majority of eFilm users had version 3.3 with 4 having version 3.2 or 3.1. One person had version 3.4 which is the human rated workstation. Most Osirix users had the latest version (5.x) with others ranging a few versions behind. A few people had older versions due to the need to upgrade the OSX operating system in order to use the newer Osirix versions. Seven of 16 people had Osirix 64 bit installed. One person indicated they were using a 64 bit Windows machine with eFilm.
How much memory do the workstations need?
RAM of the machines used as workstations is important because it allows large volumes of data to be handled at the same time. RAM ranged from 2 GB to 16 GB. The Windows users tended to have 8 MB or less, and the Mac users 4 MB or more. There was a trend to decreased RAM when using eFilm software. There was a slight trend toward low RAM being associated with crashing though it was not significant (p=0.19). RAM was not associated with the overall user satisfaction of the software. With Osirix, higher RAM and 64 bit capability was a trend in the ability to have multiple series open at the same time (p=0.07, p=0.12). Crashing occurred with CT images >1000 across a range of RAM (3.12-16) in both systems, with most falling in the 4 MB (3/8) and 8 MB (3/8) categories. From these results, I think that increasing your RAM to more than 8 MB will increase the performance of your system.
Problems with software performance
Crashing was a significant issue with eFilm as opposed to Osirix and other viewers (p=0.007). Crashing also occurred slightly more frequently on systems not using 64 bit architecture, though it was close to 50% in both groups. Users reporting crashing with all CT studies of less than 1000 images had 3.12 and 4 MB RAM. EFilm users reported crashing with a mix of CT 1000 images (7), radiographs and other studies (6), and CT 1000 (3).
Image viewing when scrolling produced problems in 9/16 eFilm users and 2/13 Osirix users. The most frequent issue was image lag (9/16). This was not significantly associated with RAM. Skipped images (3/16) and pixellated images (1/16) were occasional issues with eFilm users. User satisfaction was slightly higher in Osirix users than eFilm users, however this was not significant. Osirix came out ahead with fewer performance problems in general. From the results above, increased RAM also helps software to perform its best.
Which viewer performs best?
From these results, I think that Osirix performs better as a viewer compared to eFilm. The reduced incidence of crashing, and fewer problems handling multiple series and studies, make it a better solution. Iâ€™ll say from my own experience that adding more RAM and purchasing the Osirix 64 bit option made a big difference in performance. Loading times were faster, all the images in a series could be loaded, and the 3D engine worked much faster.
Lack of 3D capability is also an issue for eFilm users. The function is there, however it doesnâ€™t seem to offer enough flexibility to get any real benefit from it (plus tendency to freeze or crash, in my experience). Iâ€™d be interested to hear if the 3.4 version has better capabilities. Iâ€™ve been told by IT personnel that eFilm is a 32 bit program (although it will run on a 64 bit machine), so it may be inherently limited in the volume of information it can handle.
Osirix came out far ahead of eFilm in terms of fewer issues, and better 3D capability. For many people, switching to a Mac is not practical, and a better Windows image viewer is definitely needed. I hope this information has highlighted some of the issues with the common viewers, and that we can all pitch in to discuss alternatives that will help people to interpret images without headaches caused by technology.