MR imaging centers may involve millions of dollars of investment - yet problems are often faced soon after going on line. This is not acceptable, writes Michael J. Lahita.

It never ceases to amaze me that someone would invest millions of dollars for an MR imaging center, only to be faced with image quality problems soon after going on line. I have had the 'honor' of being involved with radio frequency (RF) shielding systems for MRI applications since the early 80s. I realize this does not make me an expert by any means, but I have gained the ability when looking at problematic images to determine whether the fault lies with the 'system' or the RF shielded enclosure.

What I have found over the years is that over 90 per cent of all imaging problems are directly related to the RF enclosure's performance, or lack thereof. These problems may be caused by something as simple as the patient entry door being dirty, massive amounts of floor wax on the sill or missing contact points (finger stock). By the same token it could be something more serious, even if the enclosure is relatively new.

To demonstrate this point, I have elected to discuss actual conditions I have encountered in the past. Some of these cases were so ridiculous that I could not believe companies claiming to be experts in RF shielding design and construction were responsible. It is hard to understand how these companies could actually build or modify RF enclosures such that when all was said and done, their performance was so severely lacking.

The first case was truly amazing. The site, or shield, was an MR imaging center located in the North East that was having 'artifact' problems or ambient RF energy interference. (Artifacts are interference patterns caused by ambient RF energy at or close to the center frequency of an MR imaging system.) Prior to my arrival, a local 'RF shielding' company had tested the enclosure four different times at a cost of approximately $3,800 per test and repair action, for an approximate total of $15,200.

Unfortunately, the problem was never resolved. However, the client did get new finger stock applied to their door, the MRI connector or penetration panel was re-installed and numerous panel joints were fixed.

We were contracted to test and determine what the cause of the artifact was, and within an hour we had located and corrected the problem. A copper pipe that was installed, for a sprinkler system, through a brass threaded waveguide, was not electrically or mechanically bonded to the waveguide. (I was told at a later time that it was installed in this manner to maintain a single point ground.) The repair action consisted of $1 worth of bronze wool and about two minutes of labor.

The photos below display images taken at the imaging facility before and after the repairs. As can be seen, the 'artifact' in the first photo was significant.

Image quality before repairs

Image quality after repairs

In addition to the costs of the first testing company, our cost and the cost of down-time, this customer spent a lot of money and lost untold amounts of revenue for a very simple problem.

The second case was located in Northern Texas, in a brand new enclosure. In this case, the company that had installed the enclosure felt it was appropriate to allow the client to finish out the enclosure prior to performing a qualification test.

Upon completion of the finish out, it was found that a cell phone would operate without a problem and an FM radio played in the enclosure could receive almost every local station. As would be expected, this raised concerns amongst both the customer and the field installation technician of the magnet supplier. The installation company returned and reported that it was acceptable to be able to use a cell phone and play a radio in the enclosure, as both operate at higher frequency levels than the MRI. At this point the installation company was politely asked to leave the site.

What we found when we arrived at the site was astounding to say the least. The first problem we encountered was a quarter-inch gap that was approximately 54 inches in length, located above the doorway (see photos). Another area contributing to the RF leakage was the joining of the wall panels to the ceiling panels. Screws were used to connect the wall panels to the ceiling panel edges, causing the screws to become inadvertent radiators, or antennas. These items were corrected during the test and the enclosure met the client’s requirements as well as those of the MR imaging system supplier.

Above: 1/4 by 54 inch gap above doorway

Below: Wall and ceiling panels before repair

The last case I would like to describe was beyond anything I thought a reputable RF shielding company could allow. This particular imaging facility is located in the southern part of the country. It is a beautiful facility and has two active 1.5T imagers which are sited opposite each other.

One of the imagers was displaying very high noise levels that rendered certain scan procedures useless.  Upon my arrival, I had a brief discussion with the Head of Radiology regarding the history of the facility. I was told that one of the scanners had always provided substandard images from the day it went on line, less than 16 months prior. Evidently, the service technicians for the MRIs were convinced it was not a system problem but an RF shield problem, and would do nothing until the shield's performance was verified.

I was informed that the manufacturer of the shield had recently been contacted and had sent a tester to the facility to determine whether the problem was the RF shield. I was allowed to review the test data and noted that the enclosure did not even meet the RF shielding performance requirements of the MR at its center frequency or the required 85 dB to 100 dB at 100 MHz (85 dB for mobile applications, 100 dB for fixed sites). While reviewing the test data, I noticed that the comment was made, once again, that even though an FM radio could be played in the scan room, the FM frequencies are well above the magnet’s center frequency and should have no effect on the its performance. Go figure!

With this information in mind, testing and troubleshooting commenced. It was no surprise that the shielding effectiveness levels we recorded, at both 64 MHz and 100 MHz, were very similar to those recorded by the RF shield suppliers tester.

First of all we discovered the normal door finger adjustment problem and corrected it. When the 'fingers' are not positioned correctly on the door leaf, no contact is made between the finger stock and the conductive door jamb or door edge when the door is in the closed position. Additionally, loose bolts and screws were found between mating panel assembly edges.

Having taken care of the minor things, we proceeded to the penetration panel which can be a source of leakage if not installed properly, or if there is a 'bad' component on the panel. Sure enough, a few of the bolts that clamp the penetration panel to the shield face were loose. The bolts were tightened but there was still a very large amount of RF leakage. We next probed virtually every component on the penetration panel and found nothing.

The next area for troubleshooting was the incoming electrical power line filters - which are used to bring DC power for lighting and AC power for receptacles into the enclosure or scan room - and the 'data/communication or fire alarm' filters. We probed these filters using a small de-tuned antenna and found no problems. When the power line filters were probed, however, the received signal was off the screen of the spectrum analyzer.

Using the probing technique, we found that the incoming wires of the filters were conducting large amounts of RF energy. In this installation, the type of filter used had two access or cover panels. One access panel serves to cover the input to the filter and the second access panel covers the output of the filter. These cover plates are held in place with numerous machine screws. The output wires are fed through a tubular threaded nipple into the enclosure, and the nipple is secured to the filter housing and the shield with locknuts or flange nuts.

The first course of action was to ensure that the machine screws that held the access panels in place were tight. A few were found to be loose and a few were stripped. The next action was to tighten the lock nuts that secured the nipple to the filter body and the shield. These were re-tightened and probing continued.

To my dismay, the filters still leaked. The decision was made to replace the five existing power filters at a later date - that was acceptable to the imaging center and an electrician could be brought in to disconnect the filters.

When we returned a few weeks later, the filters had been disconnected and we started to remove them. When I pulled the first filter off, I was utterly amazed - no, dumbfounded - as to what I saw. Whoever had installed the filters originally, presumably the shield installer, evidently drilled the holes through the shield in the wrong location, or elevation. When such a mistake is made, standard practice and common sense dictate that the hole should be repaired. Sadly, whoever did this, just drilled new holes about 12 inches above the old ones. Rather than fixing or patching the wrong holes, the filter body was placed over the holes and nothing else. No patch, no repair, someone just covered the holes and hid them.

Before repairs, old filter with hole

After repairs, new filter and patch

Why this was never found during one of the two customary RF shielding effectiveness tests is beyond comprehension. It is customary and highly recommended to perform two tests. One is performed upon completion of the enclosure installation (prior to installing internal finishes) and a second, or final, test, just before ramping the magnet or bringing it to field.

What we have found recently is that many RF shielding companies are either eliminating one or both RF tests or allowing poorly trained individuals to perform the test. The reason for this is purely economical. By eliminating one or both tests, a price reduction of $4,000, minimum is realized on the price of a 'new' enclosure. Another way to reduce the price is to have the installer perform the test. This reduces the price by approximately $2,000, or even more. Bear in mind that most installers are mechanically inclined, not electronically. On the flip side, there are many very gifted installers who have a sound mechanical as well as electronic understanding.

Prior to issuing a purchase order or allowing a general contractor to pick the low bidder, it is wise to really research the company supplying your RF shielded enclosure. In these trying economic times, a company that claims to offer the best value for money may not be the wisest choice when quality is key. Remember, you get what you pay for and very rarely will you get anything more. Buyers, beware!

www.shieldingresources.com