Q&A: Tackling the Complexity of MR Safety

MR imaging offers incredible diagnostic information and insights for radiologists. But, it’s not always available for all patients. Those with metallic implants are frequently turned away due to uncertainty about whether they can be safely imaged.

That situation could change in the coming years, however, based on a new smartphone app developed by Emanuel Kanal, MD, director of magnetic resonance services at the University of Pittsburgh. Using data submitted by both implant manufacturers and MR scanner manufacturers, Kanal has created an app that visually displays the energy fields and imaging gradients used in MR imaging and how they relate to specific implants placed into specific machines.

The goal, he said, is to give technologists, radiologists, and physicists more information about whether patients with implants can safely enter an MR scanner. Diagnostic Imaging spoke with him about his app, how it works, and the benefits he anticipates.

DI: What is the need for a tool like this?

Kanal: There’s a limitation for patients who have metal implants that impacts their ability to undergo MR imaging. If I want to take a chest X-ray of you, I’m going to shine the X-rays at your chest. What if it’s an ultrasound of the liver? I’ll grab a transducer and shine it on the liver. It’s imaging from a different source. It’s all quite intuitive — you take whatever energy source you have and you aim it at whatever you’re studying. You gather information from that energy that you’ve probed the body with, and you move on.

But, to do an MRI of the body — to get an MRI of the liver — you put your whole body into a powerful magnetic field. Then, there’s a second energy source called the imaging gradient, and you turn them on and off tens of thousands of times per second. And, that imaging gradient covers roughly half of your body. Then, there’s a third energy source — radio-frequency oscillating fields — that covers part of your body and shines radio waves at you. The spatial distribution of these radiowaves over the body of the patient is different once again from the static magnetic field or from the imaging gradient magnetic field. So, you have three different energy sources. After exposing the human body to three different energy fields with three markedly different spatial distributions, you produce an image corresponding to 5-mm slice of your liver. It makes no intuitive sense.

Then, if you add an implant in the shoulder into the mix, the question is raised as to whether you can safely scan the patient. If you’re taking a chest X-ray, you’re not hitting the implant because you’re imaging the chest. But, with an MRI, to get a 5-mm slice of your liver, your whole body is in the magnetic field. Plus, there are the other energies we discussed - and they’re all invisible.

So, No. 1, there’s a general understanding that to generate imaging using X-ray, you use X-rays, ultrasound, you use ultrasound. But, with MRI, there are multiple different energy sources distributed at varied locations across the patient. People don’t intuitively understand that.

No. 2, each and every energy source to which we expose the patient will have its own safety concerns. We know the safety issues associated with X-rays so we try to keep exposure to as low as reasonably possible. It’s the same with the three energy sources with MRI, and each of them have safety concerns that are unrelated to the others. So, how can we expect to know whether we can safely perform a lumbar MRI on a person with a shrapnel fragment in their neck?

DI: How big is the concern over and interest in MRI safety?

Kanal: I’ve been involved in MRI safety since 1984. I’ve given hundreds of, if not over a thousand, lectures on MR safety. In my interactions with the audience I recognize that this is not intuitive for most audiences. There’s a lot going on with MRI that’s not intuitive. Our techs are superb, but they don’t routinely or necessarily have any formal detailed training in MR safety or the theory of MR image production. There is no mandate that radiologists undergo any formal training on MR imaging physics or especially MRI safety issues related to the MR imaging process.

So, for the first time in June 2014, I put together a whole course dedicated to – and only to — MR safety. Back then, it was a 16-hour course, but it quickly went to 20 then 24 hours over three-to-four calendar days. I didn’t anticipate the demand. The need for MRI safety knowledge is massive and growing. Since 2014, roughly 4,000 people have taken the course. I’ve given the course in numerous locations throughout the United States, as well as internationally. Interest in the course continues to grow stronger as time passes, and we routinely have 200-plus attendees at these courses multiple times per year.

DI: How did you come to the idea of developing an MRI safety app?

Kanal: In November of 2016, the Journal of Magnetic Resonance Imaging published a consensus document that defines the organizational structure for MR safety supervision in any site in which humans undergo MR imaging, both for clinical and/or research environments. These include three individuals: a Magnetic Resonance Medical Director (MRMD, who should be a licensed physician if/when patients are concerned), a Magnetic Resonance Safety Officer (MRSO, who can be anyone, but is often an MR technologist), and a Magnetic Resonance Safety Expert (MRSE, who again can be anyone but is often a physicist).

One of the biggest concerns in safety is, “Can I safely scan a patient?” Because there are three different energy sources, the question is often confusing and the answers may not be obvious or clear to many MR practitioners. It could be that scanning a patient with a given implant or device may pose no safety problem for energy fields A and B, but doing so might result in patient burn or injury from energy field C. Or, the patient might face life-threatening issues because of fields A and C, but not B. To answer the question of whether a patient with a given implant, device, or foreign body can safely undergo a requested magnetic resonance examination, one must understand the basic imaging principles underlying the MR imaging process, the potential safety concerns associated with each of the forces being used, and how and where these forces are used and spatially distributed.

To assist in this process, I first generated a decision tree for assessing implant safety for MRI. Designed like a pilot’s checklist, it took the user through each energy field used in the MR imaging process and analyzed the study in question and addressed the individual patient and device about to be scanned. After assessing for safety concerns associated with each of the three energy fields, if no concerns were found it would be assumed that there would be minimal risk for proceeding with the requested MR imaging study. If risk was found to be associated with exposing that patient’s implant(s) to one or more of these energies, and the risks could not be mitigated by the user, it was assumed that there would be significant risk associated with MR scanning of that device patient. This decision tree was so popular that I turned it into an app called “Kanal’s MR Risk Assessment” that I developed in conjunction with iCat Solutions, Ltd. headed by George Michalopoulos. This app was released in late 2016 and proved quite popular, and even won first runner up for Radiology App of the Year in 2016 by auntminnie.com six weeks after it was first released. However, the “Kanal’s MR Risk Assessment” app requires MR safety expertise to answer the questions that it inquires of the user, it is essentially intended to be used by the MR Medical Director, MR Safety Officer, or MR Safety Expert. The user of the first app would need to know if the patient would be exposed to gradient energy, or if the implant would be found in the volume that undergoes RF irradiation.

DI: What about the next iteration of the app?

Kanal: This is the app “for the rest of us.” For reasons that will shortly become obvious, I named it, “MagnetVision.” I wanted an app that helped answer whether there are significant risks of scanning a given patient for those involved in patient care and are not MRI safety experts. It doesn’t decrease the importance of having safety experts manning the sites – it just opens the understanding of the underlying safety issues to more MR practitioners.

I love math and physics, but I recognize that this is relatively unusual among radiologists. What radiologists DO collectively excel in is graphic pattern recognition. We’re very comfortable at interpreting and manipulating images. The whole idea of MagnetVision — the tag line, in fact — is “Turning the invisible – visible.” MagnetVision takes the energy fields used in the MR imaging process and displays them in a way that the end user can visually and graphically see where the fields are distributed and located relative to the patient, the implant, and the MR scanner and its components.

DI: Why is something like this important?

Kanal: Right now, there are implants that are approved by the FDA to undergo MR imaging at or up to certain conditions under which they were tested. The FDA allows manufacturers to market them as being safe in MR environments according to the tested and approved conditions of MR safety. One must consider the strength of the magnetic field and the magnetic spatial gradient, the speed at which imaging gradients are switched on and off, the radiofrequency transmitted energy strengths, and their durations. You look at these energy sources and get a sense for what’s safe under the thresholds of how the implants are tested, and you decide of whether you can scan a patient. MagnetVision provides graphic information as to how these energies work together to potentially affect – or not affect - the implant. It also provides clear visual feedback as to how close we are to reaching or exceeding those FDA approved MR energy field safety thresholds and whether – and how - we can actually safely scan that device patient.

The strengths of the various energies used in the MR imaging process vary at different positions in the MR scanner and MR scan room and at different positions throughout the patient. How they vary can and do differ with different MR scanners and gradient coils and RF transmitting coils among varied scanners of the same or different MR manufacturers. All this information is built into the app, processed, calculated in real time for any selected or defined implant, MR scanner, coil, or patient, and rendered as a graphic color display – in real time.

DI: How does the app work?

Kanal: In this app, I don’t expect the user to have any special MRI safety expertise, but I do expect them to have expertise at their job. The app is designed specifically targeting MR technologists, radiologists, and physicists. That means that non-MR personnel would not be expected to understand how to use MagnetVision or how to interpret its output or reports, but any MR technologist or radiologist or physicist would be readily able to do so.

The app expects you to know how to position the patient — head first, feet first, supine, decubitus, angled, arms raised or lowered, etc. You need to know where you will center the study to be performed — the head, chest, knee, ankle, etc. These are all things that the MR technologist already knows and for which the MR technologist already has demonstrated expertise, and they can readily answer these questions. The app will ask about the patient’s implant. What kind is it, what does the labeling say? The technologist will either know that, or will be able to look at that up. For devices or objects for which there may be no FDA labeling, the app permits the user to define what strength fields they would be comfortable exposing the patient to – if any. The user can enter their own values if they so desire, and one can exercise their own expertise in MR safety – it’s just that the app doesn’t mandate that you do so or that you be an MR safety expert to use it.

There are four conceptual “rooms” within MagnetVision, and in each of them the technologist, radiologist, or physicist will recognize what is expected to transpire in each of them.

First, there’s the Waiting Room. Here the user “finds” or defines/creates their next patient. After entering the sex, height, and weight of the patient to be modeled, the app renders a photorealistic depiction of the body habitus of the patient just described. These can, of course, be fine-tuned by the user until the 3D avatar rendered by the app best approximates the appearance and body habitus of the patient about to be scanned.

The next room is the Implant Room. In this room the user selects from a list the implant that is found in the patient to be modeled. Should that particular implant or device or foreign body not be present in that list, the app provides the user with 3D tools that would enable them to generate just about any device or object of virtually any dimensions or shapes to try to best mimic the object in the patient being modeled. Thus, if your patient in real life has a stent or pacemaker or an artificial hip, etc., the user selects that implant from the list. The FDA-approved and/or user-defined conditions of MR safety are defined for each implant/device. If the device has FDA-approved MR safety labeling, the user can copy those values into the device specific conditions of MR safety in the app or, alternatively, the user can provide their own values or generate any what-if scenario they choose.

The patient is then brought to the Operating Room. Here the user implants within the avatar the selected device(s) in the same position and orientation as it might be found in the real patient being modeled.

The fourth – and most important — room is the MRI Room. It is in this room that the main teaching objectives of MagnetVision are accomplished. To understand what transpires in this room, let me first provide a bit of background information.

Every make and model MR scanner are different, with different and unique strengths and spatial distributions to the static magnetic fields, the static magnetic spatial gradients, the time varying imaging gradients, and the transmitted RF energies used in the MR imaging process. So, we went to the manufacturers of the MR scanners themselves, and they’ve been extremely kind and cooperative and have shared the data that describes the spatial distribution data for these various energies used in their MR scanners. The app displays that MR manufacturer provided data along with graphic depictions of the patient, their implant(s), the various energy fields, and the MR scanners and their various components.

The app will then permit the user to position the patient on the MR scanner just as they are about to do in real life on that same patient. They can then center the patient and advance that patient to scanning position at magnet isocenter just as they would do on that real patient being modeled. As the patient avatar, together with its implanted device(s), are advanced into the selected MR scanner, the app will calculate the strength of the various MR energy fields through which it will pass and will compare those values to the FDA and/or user defined maximum safe threshold values for those same energy fields labeled for that device/implant.

The app will then leave colored spheres as a trail of “breadcrumbs” passed through by the implant on the way from initial position to scanning position as the patient avatar is advanced into the scanner. These “breadcrumb” spheres will change color from green to yellow to red as the displayed energies to which the implant is exposed passes through energy levels that are well below, or approaching, or exceed published/defined MR safety thresholds for that implant. Thus, if the app demonstrates that the device is anticipated to traverse energy levels that may be beyond the safety labeling for that device, it may be potentially unsafe to scan that patient/device in that MR scanner, and the technologist should bring this to the attention of the radiologist or MR Medical Director.

DI: This could change imaging for countless numbers of patients who have implants, but how do you intend for it to be used?

Kanal: MagnetVision is first and foremost an educational tool. A radiologist or technologist may not have a 3D understanding of the various energies used in generating MR studies. They likely have heard that there are magnetic fields and radio waves used in the MR imaging process, but they might not recognize where they are located within the MR scanner or how strong these fields may be at different locations within the scanner. The app will allow them to see the MR scanner being used and the complexity of the 3D spatial distribution of the relative strengths of the various energies that are about to be applied to this patient to generate the requested MR imaging examination. They can also see the specifications of the energies for each make and model based on the scanner being used, and how it might differ should they wish to proceed with the requested MR imaging examination on another MR scanner instead.

MagnetVision permits direct visualization of the spatial distributions and relative strengths in space of the various energy fields used in their MR scanners and allows them to more thoroughly understand and even visualize where, in space, these energies are being deposited relative to where in the scanner that patient and implant will be situated. This information might better prepare the user to make truly informed benefit risk decisions regarding scanning a patient with a particular device or foreign body which may be quite far or physically remote from, for example, imaging gradient magnetic fields.

In deciding, for example, whether to accept a pregnant patient for a head CT, we have to know what energy/energies is/are being used (X-rays), where they are being deposited (primarily the head), what the potential risks may be (carcinogenesis), and what may be done to decrease that risk (eg, shielding) before deciding whether to permit a given patient to undergo the requested head CT examination.

MagnetVision provides the knowledge base for the user to make similar informed decisions regarding MRI scan exposure, where there is more than just one energy field (and, therefore, risk) being utilized and where there are unique spatial distributions to where each of these energy fields are being deposited over the patient/implant. We’re applying the same common-sense information that we require to safely do our jobs to a field whose imaging physics is not governed by as common or intuitive reasoning as is, for example, CT.

DI: How does this app help?

Kanal: It reduces the consideration of most of the safety risks associated with MR scanning of various implants and devices into the reach of the average technologist, radiologist, and physicist. It can improve the understanding of the MR imaging process and its associated energy fields to those who operate all MR sites, including not just academic centers, but also private practices and/or outpatient imaging clinics. It presents these rather complex concepts in a manner that can be seen, understood, and applied by any MR practitioner. In this regard, it is truly the app “for the rest of us.”

DI: So, when you look forward, how do you see this tool potentially being able to change the knowledge field we have?

Kanal: By reducing these concepts to visual graphics to an audience which already has expertise and already excels at processing visual graphic information and pattern recognition, the app will provide much needed data in a manner that is readily comprehensible by the MR practitioners caring for our patients. It will, thus, provide for them an overall gestalt and a 3D understanding of where energies are found and what that could mean for an implant in my patients. This information is invaluable for assessing patient safety of requested MR examinations in device/implant patients.

As I like to stress in my MR safety courses, saying “no” is easy. It takes knowledge, understanding, and a willingness to clinically apply them in order to say “yes.” MagnetVision should help provide our installed MR base with significant advances in their knowledge and understanding of the underlying MR safety issues associated with MR imaging of implants. In that manner, it should help to solidify why certain patients perhaps should NOT be scanned if it might be unreasonably unsafe for them to undergo a requested MR study. Similarly, MagnetVision should help notably decrease the number of unnecessary MR scan cancellations for device patients who might actually be able to safely undergo a requested MR examination – and now the radiologist may clearly understand and literally “see” why and how to do so.

It is important for all to recognize that withholding a requested MR study can also be associated with its own patient risks. A patient might not get the diagnostic information they need or they could be sent for a more invasive study as a result of that study cancellation. To this end, the app provides the knowledge we need in order to make more knowledge-based risk-benefit decisions.

DI: What do you see as the overall benefits of the app?

Kanal: The educational benefits of the app are, like the tag line says, it makes the invisible - visible. The magnetic fields, the energy sources, the sources of potential injury and harm about which many may have limited to no detailed knowledge or understanding – this app will provide a very graphic and visual display of where they are in space and what their strengths and relative strengths are and how they compare to the labeling of the implant/device. This will help the user make intelligent and knowledgeable scan/no scan decisions.

The app also shows them that you don’t have to be an expert to use spatial distribution information. They won’t have to look at graphs or perform esoteric calculations. And, the app will constantly update hardware information. It will upload directly through Apple’s App Store. The app can auto-update through the Apple Store as with any other Apple app and is planned to continually add other and new MR hardware and hardware components from these and other MR manufacturers over time.

DI: How is the app available?

Kanal: At present, the app is just completing beta testing. It is scheduled to be distributed starting late March/early April 2018. In order to limit my own potential legal liability as to how this app may be used or applied (or misused or misapplied), right now I plan to distribute it at no cost to anyone to whom I have personally taught what it is and is not, how to use it, and how not to use it, etc. But, since I am not able to distribute this app on the Apple site to some users at no cost and not distribute it to others, the plan is to charge a sufficiently high price for it on the Apple store so that no one actually purchases, and, then, distribute it no charge via promo codes to those whom I have taught how to correctly and properly use it.

DI: How is the app designed to merge with the electronic medical record?

Kanal: My first app (“Kanal’s MR Risk Assessment”) has a button to produce a PDF output, and this one does, as well. It generates reports and graphic capability which provides, at the push of a button, an exceedingly detailed report of the quantification of all energy fields experienced by the implant at all positions traversed by it and/or within the MR scanner relative to the published/defined labeled threshold values for each energy for each device in that patient. The report shows graphically and with text the outcome of the modeling and communicates whether any safety thresholds were approached or exceeded and if so, exactly where in 3D space this occurred – for each device, implant, or foreign body so modeled for each MR scanner for which this modeling had been performed. The data can be entirely HIPAA anonymized and all Protected Health Imformation (PHI) can be removed (or never entered) or, should the site so desire, they can use HIPPA-compliant methodology and hardware and save PDF reports to the patient’s electronic medical records. Data can be stored and anonymized and/or can be prepared as examples for educational purposes and teaching files.

DI: What feedback are you already getting as far as efficacy and efficiency are concerned?

Kanal: One of my past attendees actually wrote an article about the methodology employed by the first app (Kanal’s MR Risk Assessment”) for the AHRA’s RSNA edition last year. They reported Press Ganey score improvements from the patients, as well as, considerable throughput benefits and decreases in the number of patients canceled when applying the approach and standardized methodology I teach in my course and embodied in heap. I must admit that I was quite impressed with her results. I am personally most interested in the safety issues, but I had not anticipated such other benefits as patient throughput and patient satisfaction scores from considerable decreases in patient cancellations, etc. With increased MR safety knowledge, sites have repeatedly shown that they are able to confidently and safely approve scans on more patients and decrease unnecessary patient cancellations, and empty magnet time typically plummets.

With this app, I just began using it for teaching in my most recent MRMD/MRSO MR Safety Training Course that was held in Boston in October 2017. The feedback was so universally powerful, strongly positive, and supportive that I have accelerated the development cycle and release schedule for MagnetVision and its derivative future version is already well underway. Based on the feedback from the 200-plus course attendees, as well as, the more than 100 beta testers of this app, I can say that I am guardedly, but quite realistically, expecting that in the foreseeable short-term future MagnetVision and its derivative future versions will change how our industry practices and approaches the assessment of safety decisions relating to MR scanning of patients in whom there may be implants, devices, or foreign bodies.

Now that I am imminently ready to launch this app my primary objective is to ensure that it finds itself in the hands of all MR practitioners as rapidly as possible, and to ensure that MagnetVision is found and used in quite literally every MR scanner in the world to help improve MR safety practices for all device/implant/foreign body patients everywhere.