Executive Summary. An MRI scanner is the strongest magnetic environment most people will ever enter. At 1.5 and 3 tesla, the static field exerts translational forces and torques on any ferromagnetic object, radiofrequency pulses can induce current in conductive jewelry, and even non-magnetic metals distort the image through susceptibility effects. The clinical physics explains why radiology departments insist on jewelry removal and which materials are genuinely safer.
The Static Field: Force and Torque
The static field B0 of an MRI scanner is measured in tesla. One tesla is roughly 20,000 times the Earth's surface field, and clinical scanners run at 1.5 T or 3 T, with research systems reaching 7 T. Two mechanical effects act on jewelry: translational force and torque.
Translational force pulls ferromagnetic objects toward the strongest field region, which in a superconducting magnet is the bore center. This is the projectile hazard. A loose steel barbell or ring can accelerate across the room or across the patient. The force scales with the field gradient, and the gradient is steepest at the bore entrance, which is why the danger zone starts before the patient is fully inside.
Torque rotates a magnetized object to align with the field. A ferromagnetic curved barbell or captive ring will try to twist in place. For a healed piercing this is uncomfortable but survivable; for a fresh fistula, the twisting can tear tissue and disrupt healing. Dedini et al. demonstrated at 1.5, 3, and 7 T that attraction forces scale dramatically with field strength, which is why behavior that is acceptable at low field can be dangerous at 3 T.
RF Heating: The Jewelry-as-Antenna Problem
MRI uses radiofrequency pulses at the Larmor frequency, roughly 64 MHz at 1.5 T and 128 MHz at 3 T. Conductive structures inside the bore can couple to these pulses and act as antennas. When a conductor forms a closed loop, a circular barbell, a ring, or a clicker, the RF field induces circulating current. Current density concentrates at small cross-sections such as the bead seat or the closure, and resistive losses appear as heat.
Winter et al. reviewed MRI-related heating of implants and devices and found that heating depends on geometry and position more than on material conductivity alone. The dominant variable is how well the conductor couples to the RF field, which is why a thin titanium ring can heat more than a thick steel bar if the ring forms a resonant loop. Görgülü et al. measured meaningful heating of small metallic orthodontic brackets at 3 T, and piercing jewelry is the same class of object, a small conductive loop at the body surface. Most scans are short enough that the temperature rise stays subclinical, but a localized burn at the piercing site is a documented risk, especially for jewelry that forms a complete conductive loop.
Susceptibility Artifacts: Why Non-Magnetic Is Not Invisible
Even non-ferromagnetic metals distort the magnetic field because their magnetic susceptibility differs from tissue. The local inhomogeneity dephases nearby protons and produces a signal void, the familiar black crescent around metal in an MRI image. Titanium has low susceptibility and produces a modest artifact. Stainless steel produces a larger one. Gold and silver, though non-magnetic, are highly conductive and distort the local field enough to matter.
The artifact matters even when the jewelry sits outside the primary region of interest. A tragus piercing during brain imaging or a navel barbell during lumbar spine imaging can degrade diagnostic quality enough to force a repeat scan. Li et al. reported tragus piercing as a risk factor during MRI precisely because of these interactions.
Material-by-Material Risk Assessment
| Material | Ferromagnetic? | RF loop risk | Artifact | MRI guidance |
|---|---|---|---|---|
| 316L stainless steel | Mostly no (austenitic) | Moderate | Large | Remove |
| Titanium (ASTM F136) | No | Low | Small | Preferred retainer |
| Niobium | No | Low | Small | Preferred retainer |
| Gold / silver | No | Moderate | Moderate | Remove |
| Magnetic jewelry (NdFeB) | Strongly yes | High | Severe | Never enter bore |
| Glass / PTFE retainers | No | None | None | Acceptable |
Clinical Takeaways
1. Remove all piercing jewelry before an MRI unless the imaging region is remote from the jewelry and the piece is verified non-ferromagnetic.
2. For piercings that cannot be removed, such as healing or tight-fitting jewelry, use glass or PTFE retainers, which are non-conductive and artifact-free.
3. Never scan with magnetic jewelry. Neodymium magnets are strongly ferromagnetic and experience both projectile force and torque at 3 T.
4. Titanium and niobium are the best metal choices for MRI-adjacent wear, but they still produce artifacts in nearby anatomy.
5. Tell the technologist about every piercing, including nipple and genital jewelry, because these are the ones most often missed in screening.
6. A healed piercing can close within hours without jewelry, so plan removal with a taper or retainer instead of gambling on the scan.
Titanium is non-ferromagnetic and is generally considered MRI-safe for scanning, but it is not artifact-free. Commercially pure titanium and Ti-6Al-4V alloy have very low magnetic susceptibility, so the translational force and torque are negligible; however, the material will still produce a localized signal void due to susceptibility differences with tissue, which can degrade image quality if the piercing lies within the region of interest. A thin titanium ring that forms a closed conductive loop still carries an RF heating risk, as geometry and loop coupling are the dominant variables for heating, not material conductivity alone. Depending on the material, you risk projectile injury from the static field's translational force, torque-induced tissue tearing, or localized RF burns from induced currents. At 1.5 T, the force on a ferromagnetic steel barbell can exceed its weight several times over, and at 3 T that force scales up dramatically as demonstrated by Dedini et al., which is why a loose piece of steel can accelerate toward the bore center and become a projectile. Non-ferromagnetic but conductive metals like gold, silver, or copper rings can still couple to the RF field at 64 MHz (1.5 T) or 128 MHz (3 T) and heat enough to cause a burn, so radiology staff will typically halt the scan to have the jewelry removed. Gold is not ferromagnetic, so it will not be pulled toward the magnet, but it is highly conductive and creates meaningful susceptibility artifacts that can obscure the image. More importantly, a gold ring or clicker forms a closed conductive loop that can act as an antenna for RF pulses; Görgülü et al. measured significant heating in small metallic structures at 3 T, and gold's low resistance can lead to higher circulating current and localized thermal injury. Even a simple gold stud without a closed loop can heat at the contact point with skin, so gold jewelry should be removed before a scan. Possibly, but the bigger concern is that many "surgical steel" or "316L stainless steel" pieces are not purely austenitic and may contain ferromagnetic inclusions from manufacturing or cold working. That residual ferromagnetism is enough for the magnet's translational force to rotate the curved barbell or pull it forward, torquing the fistula and potentially tearing fresh tissue; the attraction scales with field strength, so a piece that feels safe at 1.5 T can become dangerous at 3 T. Because the force depends on the exact alloy composition and manufacturing process, radiology departments cannot certify it as safe without testing, which is why the rule is "when in doubt, remove it." Susceptibility artifacts are not confined to the immediate location of the metal; the local field inhomogeneity extends beyond the jewelry and dephases protons in surrounding tissue, creating a signal void that can appear as a black crescent over a wide field of view. Even a tragus piercing, as reported by Li et al., can degrade brain imaging quality, and a navel barbell can produce artifacts during lumbar spine scans because the distortion scale depends on the magnetic susceptibility difference, which for silver, gold, or stainless steel is large enough to affect adjacent slices. At 3 T, the artifact size increases with field strength, so a piercing that is "small" in size can still force a repeat scan if it overlaps the region of diagnostic interest.FAQ
Can I wear titanium piercing jewelry in an MRI?
What happens if I forget to remove my metal piercing before an MRI?
Is 14k gold piercing jewelry safe to wear during an MRI?
Will a surgical steel belly button ring set off the metal detector at an MRI facility?
Why does my MRI report mention an artifact from my nose piercing even though it was outside the scanning area?
6. MRI Safety by Material: Is Your Piercing Magnetic or Conductive?
People ask me two questions more than any others: are body piercings magnetic, and are piercings conductive? The honest answer depends entirely on the alloy. "Acier inoxydable", stainless steel, is a blanket term, not a specification. Most 316L surgical steel is austenitic and effectively non-ferromagnetic, which is why you can walk a basic steel ring past a fridge magnet without drama. But cheap jewellery and some "surgical steel" imports contain ferritic or martensitic phases, and those respond to a static field. That distinction matters, and magnetic response is only one of three hazards in the scanner.
Conductivity is the second. Even a metal that is not pulled by the magnet can still heat. At 1.5 and 3 T, radiofrequency pulses induce current in any conductive loop, a captive bead ring, a barbell with a snug ball, a nose hoop. The tighter the loop and the closer it sits to the coil, the higher the local specific absorption rate. This is why "non-magnetic" and "MRI-safe" are not synonyms, and why I never let a supplier conflate them. Titanium is non-ferromagnetic and comparatively low-risk for heating, but it still produces susceptibility artifacts that smear the image around the piercing site. Gold behaves similarly, and 14k or 18k alloys can contain enough ferromagnetic trace content to twist under torque.
My own position, after 25 years making this product: if you are heading into an MRI and cannot remove the jewellery, the retainer material is the decision that actually matters. BioFlex® is PP-R, a medical-grade plastic, not an elastomer, not PTFE, not silicone. It is non-metallic, so there is no ferromagnetic torque and no RF-induced current path. It does not generate the susceptibility artifacts metal does. That is the whole reason I developed it, and it is why the same material people wear for years in mouth, eyebrow, and cartilage piercings is also the sensible choice for imaging day.
Two practical rules before you scan. First, tell the radiographer exactly what you are wearing, where, and in what material, do not let them guess. Second, if the piercings are within the scan field, replace them with PP-R retainers and confirm retention before you enter the room, because a loose retainer is its own problem. For a material-by-material comparison before your appointment, work through our MRI retainer selection guide and check the imaging considerations for each option. And if you want the full materials breakdown behind these claims, start with BioFlex®.
