Who is responsible for MRI safety screening before a scan?

Why is MRI safety a different problem from radiation safety?

Because there is nothing to shield and no dose to count. CT safety is about managing an exposure that stops when the beam stops. MRI has no ionizing radiation at all, but the static magnetic field runs continuously, day and night, and it does not care whether anyone is scanning. The hazards are mechanical, thermal and acoustic.

Start with the projectile effect, because it is the hazard that injures people fastest. A ferromagnetic object carried past the field boundary does not drift toward the bore. It accelerates, and it arrives with enough force to injure whoever is standing there and to pin a patient inside the tube. The objects that turn up in published incident reports are ordinary ones: oxygen cylinders, IV poles, floor buffers, mop buckets, scissors, a wheelchair, a fire extinguisher carried in by somebody answering an alarm. None of them were brought in by a careless person. They were brought in by someone doing their job in a hurry, usually during an emergency, and usually by staff who do not work in the department.

The same field exerts torque on ferromagnetic material already inside the patient, which is why a retained fragment matters even when it is small and even when it has been there for years. Radiofrequency energy is a separate mechanism. The transmitted pulses deposit energy in tissue as heat, and conductive material, wires, leads, looped cable and points where skin touches skin can concentrate it. Switched gradient fields produce the noise everyone associates with the exam and can stimulate peripheral nerves. A superconducting magnet also holds a cryogen supply. A quench boils that helium off in a rush, and the magnet is installed with a dedicated quench pipe whose job is to carry it outside the building. The magnet room becomes an oxygen displacement hazard when that path fails: a pipe that is blocked, damaged, undersized, or was never installed correctly in the first place. That failure case is the reason the room has oxygen monitoring, the reason the door is designed the way it is, and the reason the facility keeps a written quench plan at all.

Staff who trained around ionizing radiation carry an intuition that does not transfer here. Radiation safety rewards distance, time and shielding, and the risk falls to zero when the equipment is off. This equipment is never off. It holds its field overnight, through a service visit, through a power failure, and through the week the department closes for renovation. A team that runs MRI on the mental model it uses for CT will build the wrong habits, and in this room the habits are the entire control. That is why the safety program is written around who may enter rather than around how long anyone stays.

What is the zone model, and what does it actually control?

A layered access scheme. The outer area is open to anyone, the next is where patients are received and screened, the third is a controlled area that no unscreened person enters without supervision, and the fourth is the magnet room itself. The control is not the door. It is who holds the authority to open it.

The zones run from a freely accessible outer area through to the magnet room. Zone I is the part of the building anyone can walk into. Zone II is where patients are received, interviewed and screened, still outside the controlled area. Zone III is control in the real sense: physically restricted, supervised, and closed to anyone who has not been screened, because past that boundary an unscreened person or an unscreened object can cause an injury. Zone IV is the magnet room. The field boundary that a medical physicist surveys and marks tells you where the field is. The access control tells you who may cross it. Those are two different things and a facility needs both.

What defeats the model is almost never the design. It is the door wedged open, the badge issued to a contractor because it was easier than escorting him, the housekeeping cart that has always been parked there, the interpreter who walked in behind the patient. Stand at a Zone III corridor for an afternoon and count how many people cross it without anyone checking anything. That number is not zero anywhere. It is also why the screening station belongs in the path of travel rather than off to one side. A control people can walk around gets walked around.

The hardest case is the one where the model collides with a different emergency. A patient arrests inside the scanner. The team that runs to the room does not work in MRI, has not been screened, and arrives pushing a ferromagnetic oxygen cylinder and a steel crash cart. Every MRI unit needs a written answer to that scene, rehearsed with its own nurses and its own response team, and the answer is usually to get the patient out of Zone IV rather than to bring the response in. Rehearse it with the people who would actually come running, because the MRI staff are not the ones who need convincing.

Why is screening a documented process rather than a question at the door?

Because a patient's answer is only useful if somebody who can interpret it hears it. A patient who says they have a stent, a clip or a stimulator has given you a category, not a device. Turning that answer into a decision takes a record, a named person, and a second pass by staff trained in MR safety.

The screening form is the artifact most likely to produce false comfort. It gets handed to the patient at registration by someone whose training is in insurance verification, filled in standing up with a line forming behind, and filed. The patient ticks a box saying stent, or writes clip in head, or leaves the implant section blank because a hip does not feel like an implant to them. Nobody with MR training reads any of it until the patient is gowned and standing at the door with the next slot already booked. At that moment the pressure to proceed is at its highest and the time available to resolve an unknown is at its lowest. That is backwards, and it is the ordinary arrangement in a lot of departments.

The second reliable failure is the implant card. It names the manufacturer, the model and the conditions under which the device may go into the magnet, and it is usually in a drawer at home because nobody asked for it when the appointment was made. Two things follow, and neither is good. The scan gets canceled at the door and rebooked weeks out, or it proceeds on the strength of what the patient remembers being told on the day of surgery. Fix this at scheduling. The person booking the appointment does not need to know whether a device can be scanned. That person needs to know that certain answers trigger a request for the card, the model number, or a call to the service that placed the device, and that the request goes out the same day rather than on the morning of the exam.

Screening also has to work for people who cannot answer. A patient who is sedated, confused, aphasic, intubated or very young confirms nothing, and the relative standing beside them may know nothing about a surgery from twenty years ago. Departments handle this with a defined alternate pathway: chart and surgical history review, examination for scars and palpable hardware, review of prior imaging, and a written protocol for when radiographs are obtained to look for retained metal. That protocol belongs to the facility and its physicians. And screening reaches past the patient, because companions, interpreters, transport staff, biomedical engineers, cleaners, service technicians and anyone responding to an alarm all cross the same boundary.

What do MR safe, MR conditional and MR unsafe mean?

Three labels from a standardized terminology. MR safe means the item poses no known hazard in any MR environment. MR unsafe means it must never enter. MR conditional means the item may be scanned only within the specific conditions its manufacturer documented, and those conditions are the whole substance of the label.

Conditional is where the work lives. The conditions can constrain field strength, sequence parameters, the coil in use, patient positioning, required monitoring, how the device is programmed before and after, and which anatomy may sit inside the bore. None of that survives on memory. The facility needs the manufacturer's documentation in hand, for that model. There is also a vocabulary problem that still causes accidents. The phrase MR compatible, or MRI compatible, has no defined meaning in the standard terminology. It was dropped because people used it for two opposite claims: that the device would keep working inside the scanner, and that the device was not dangerous inside it. If a vendor uses it in an email, ask for the label and the conditions.

Then there is everything with no label at all. Older implants with no paperwork, devices placed outside Puerto Rico, retained metallic fragments from an accident or from metalworking, and anything the patient cannot name. That is where the facility decides using its own protocol rather than a hurried internet search at the door. Say the obvious part out loud: metal by itself is not a reason to cancel. It is a reason to go find the paperwork, and the paperwork takes phone calls and records that nobody can produce in the five minutes before a slot. So the scheduling script needs those categories on it, alongside the screening form.

• Cardiac implantable electronic devices, including pacemakers, defibrillators and loop recorders.
• Neurostimulators, deep brain stimulators and spinal cord stimulators, along with their programming state.
• Cochlear implants and other implanted hearing devices.
• Programmable shunt valves, whose settings the facility protocol may require checking around the exam.
• Aneurysm clips, coils, filters, stents and other vascular hardware.
• Implanted drug pumps, tissue expanders and ports with magnetic components.
• Retained metallic fragments from injury or metalwork, and any device the patient cannot identify.

Who owns the decision that a patient can be scanned?

The performing facility, without exception. The people who can see the patient, hold the device documentation, control the room, adjust the protocol and manage whatever goes wrong are all standing in the same building. A remote radiologist is not one of them, and a remote interpretation service does not screen patients or clear a device.

The reason is physical rather than contractual. Screening happens in a room, with a person, before the scan. Interpretation happens afterward, from image data that already exists. By the time a remote radiologist opens the study, every decision screening governs has already been made and carried out. The reader cannot look at the patient's chest for a scar, cannot ask a follow-up question, cannot read the card in the patient's hand, cannot see what the technologist noticed on the way into the room. Those are not gaps in a service. They are properties of reading a study from somewhere else.

The question still gets asked. A facility calls to find out whether the reading service will sign off on scanning a patient with a conditional device, or whether a radiologist can confirm from the images that the implant was safe. The honest answer is no, and it should be said plainly rather than hedged, because a hedged answer gets written down as approval. Name the local role that owns the decision, put it in the operating procedure, and give it a backup. The ambiguous device does not turn up at ten in the morning with everyone available. It turns up at two in the morning, on a weekend, with the patient already on the table.

None of this leaves the radiologist out of it. Many facilities designate a physician responsible for MR safety inside their own organization, and that person may well be a radiologist. The distinction is the role, not the specialty. It is a local responsibility attached to the facility that performs the examination, and it does not travel down a network connection with the images. This website publishes no screening role, device clearance service or coverage commitment for DLA Imaging. Any division of responsibility would live in a signed facility agreement, not on a public page.

What does the remote reader need to know about the scan that was actually performed?

What was run, what was left out, and why. A study performed under a device's conditions is often not the standard protocol: sequences shortened, parameters constrained, positioning changed. If the reader does not know that, they will read a truncated study as though it were complete, or describe artifact as though it were pathology.

Device conditions change the protocol. When an examination is performed inside what the manufacturer documented, the technologist may have to shorten sequences, constrain parameters, change the coil or drop a series entirely. What arrives looks like an MRI of that body part and is not the study a reader would ordinarily expect. Without a note, the reader either reports a truncated examination as though it were complete or spends the read reconstructing what happened. The first is a safety problem. The second is wasted time, and it shows up as a slower report with no explanation attached.

Then there is artifact. Metal distorts the local field and produces signal loss, geometric distortion and bright rims that can swallow the anatomy nearest the device. A reader who knows what hardware is present reads that appearance as expected and says so. A reader who does not know hedges around it, or recommends a follow-up study that answers nothing. This information belongs in the technologist note and in the study record, where it travels attached to the images, rather than in a phone call that one person remembers. And when a device limited the examination, the report should say which part of it was limited.

• Which sequences were performed, and which were shortened or omitted, with the reason.
• Whether the examination was constrained by a device condition, and which anatomy that affected.
• Whether contrast was administered, and the technologist's note on how the patient tolerated the exam.
• Positioning, coil selection, and any series that had to be repeated.
• Whether the study ended early, and at what point.
• Whether prior MRI examinations exist and on what kind of system, since comparison across scanners changes appearance.

Sources and scope

• American College of Radiology, practice guidance on safe magnetic resonance imaging practices
• ASTM International, standard practice for marking medical devices and other items for safety in the magnetic resonance environment
• American Association of Physicists in Medicine, professional guidance on magnetic resonance imaging quality and safety
• International Society for Magnetic Resonance in Medicine, professional education on magnetic resonance safety
• Radiological Society of North America, professional education on magnetic resonance imaging and reporting