Radiology EDU has performed applications for dozens of sites around the country. As I go through the muscoloskeletal protocols for these institutions I can't help notice a trend: STIR imaging is being pushed out in favor of standard chemical saturation sequences!
In reality I can appreciate the switch. Chemical saturation reigns supreme from a time efficiency standpoint. In a radiology world where you have to scan twice the patients for half the money you used to collect, this makes sense. Considering that most magnets have a field strength of 1.0 Tesla and above, the potential for a homogeneous saturation is fairly high.
With this being said, every day we see sites that have poor fat saturation time and time again, resulting in repeat studies and inaccurate diagnosis. Potential is great... but you have to succeed for it to be worth while.
The following images are typical Fat Sat failures we see every day.
As you can see, it's not just one body part that is routinely affected. Standard chemical saturation can be a beautiful thing, but it has certain requirements that must be met to turn out well: A homogeneous magnet shim and a lack of anything metal that can affect the scanners ability to produce a clean fat peak for which to suppress.
Other factors that can affect Fat Sat include the temperature of the magnet, patient body size, and different coil configurations. Even having the patient out of coil isocenter can produce an uneven saturation.
On top of all that, the technologist has to be able to recognize a sub optimal Fat Sat, for whatever the reason may be, in order to remedy.
For these reasons Radiology EDU recommends including a STIR sequence, in one plane, for all MSK studies. If they are performed with the right protocol selections, they can be produced in a reasonable time frame, with a similar (if not identical) appearance and weighting.
Consider the 2 images below; the one to the left is a coronal knee with a standard chemical Fat Sat. It took 3 minutes to complete. The image on the right is a STIR sequence that took 4 minutes to complete. Both were performed with roughly a 5,000TR and a 45TE. The STIR has a TI time of 140ms.
In both of these sequences the anatomy is well seen with the appropriate contrast resolution. The STIR reveals the meniscal tears and edema just as well as the chemical saturation. The time difference is minimal, especially considering it may save a repeat from poor saturation or even worse, a even a missed diagnosis.
Unfortunately STIR sequences get a bad rap in MRI land. Radiologists and technologists complain that this method produces images that are too dark. This can be easily remedied by understanding the mechanisms of their contrast: TE and TI.
Time to Echo is a significant factor in the contrast resolution of STIR sequences. Traditional T2 TE values (between 70- 100ms) make the image appear very dark. Per this reason, 30- 60ms echo times are most appropriate for STIR sequences, especially in MSK imaging.
Inversion time is, by far, the most significant factor in controlling the contrast of any STIR sequence. Traditional textbooks TI times for 1.5T and 3T magnets are 160 and 220ms respectively. With this being said, these times produce very dark images. We recommend TI times between 140- 150ms for a 1.5T magnet and 190- 200 for a 3T magnet. These times still suppress fat but also yield a contrast similar to, if not identical to, standard chemical saturation.
The following are examples of four ankle sequences that are identical apart from the TI time. As you can see, 140 and 150ms properly suppress fat while yielding a contrast resolution that is most consistent with a standard chemical saturation sequence. 130ms falls just short of fully suppressing fat, while 160ms is so dark that is obscures normal anatomy.
In summary, our routine fat saturation sequences will still be the backbone of musculoskeletal MRI... but there's still a place for the tested and true STIR. Understanding the need for consistent saturation, as well as how to optimize these sequences, is paramount to developing a protocol which is both diagnostic and efficient.