ARRT MRI common mistakes
ARRT MRI · Understand the traps and how to avoid them.
Classifying a contrast reaction by the wrong severity tier - for example labelling hives with mild hypotension as 'mild' and treating it with observation only, or calling a self-limited nausea episode 'severe'.
Candidates memorise symptom lists per category but never drill the boundary cases (urticaria plus mild hypotension is moderate, not mild), and the exam deliberately puts one overlapping symptom in each distractor.
What to do instead
Build a three-column table (mild / moderate / severe) and force each symptom into exactly one column by the rule 'airway, hypotension, or any organ-system involvement beyond skin = escalate'. Add a fourth column for the first-line drug and route at each tier (diphenhydramine for mild urticaria; epinephrine IM for moderate-to-severe), then recite it from memory twice a week.
Applying a single eGFR number as the gadolinium cutoff regardless of agent class, or stating that a low eGFR means the study must be cancelled outright.
Study guides quote 'eGFR 30' and 'eGFR 60' in different contexts (NSF risk vs. group assignment vs. linear vs. macrocyclic agents), and candidates collapse them into one memorised number.
What to do instead
Learn it as a decision tree, not a number: identify the agent class first (group I linear agents are the ones contraindicated in renal impairment), then apply the eGFR threshold, then state the alternative (macrocyclic agent at standard dose, or non-contrast study). Write the tree out from a blank page until the agent class, not the number, is your first question.
Mixing up which gradient axis performs which spatial-encoding job - for example stating the slice-select gradient encodes the phase direction, or that the frequency-encode gradient is applied between the RF pulse and the echo.
The three gradients are physically identical hardware; only timing and the accompanying RF or readout window define their role, and candidates learn the roles as a list rather than as a timeline.
What to do instead
Draw the pulse-sequence timeline (RF - slice select - phase encode - readout) for a spin-echo on every practice session and label each gradient lobe as it is drawn. When a question describes a gradient event, locate it on your drawn timeline before answering instead of matching keywords.
Confusing motion, pulsation, and wraparound/aliasing artifacts, or recommending a fix that addresses the wrong artifact (e.g., swapping phase and frequency direction to fix chemical shift).
Artifacts are taught as a gallery of pictures; when the exam gives a verbal description plus a proposed remedy, candidates pattern-match on the remedy word rather than on the artifact mechanism.
What to do instead
For each artifact, memorise the mechanism and the one parameter that changes it: phase-encode axis swap for wraparound, saturation bands or peripheral gating for flow/pulsation, increased bandwidth or fat saturation for chemical shift, shorter TE or metal removal for susceptibility. Quiz yourself mechanism-first ('what causes a ghost every N pixels?') rather than image-first.
Treating T1 and T2 as if both lengthen or shorten together, or asserting that fluid has a short T2 and a long T1 in the wrong combination.
The two time constants are memorised as a pair of numbers without the recovery-versus-decay mental model, so under time pressure the values swap.
What to do instead
Anchor on two reference tissues you can reason from: pure water/CSF has a very long T1 and a very long T2; fat has a short T1 and a short-ish T2. Then derive signal behaviour from TR (recovery, T1-weighting) and TE (decay, T2-weighting) rather than recalling numbers. Rebuild the long-T1/long-T2 table from these two anchors on scratch paper before the exam.
Miscomputing precession frequency - using the wrong constant, forgetting the field strength, or applying the Larmor equation to the wrong nucleus.
The gyromagnetic ratio (42.58 MHz/T for hydrogen) is quoted in several unit forms and the exam embeds it in multi-step field-strength questions where one arithmetic slip changes the answer.
What to do instead
Drill a fixed set of mental checkpoints: 1.5 T hydrogen = ~64 MHz, 3 T = ~128 MHz, and frequency scales linearly with field. Verify any computed answer against the nearest checkpoint (an answer wildly off 64/128 MHz at 1.5/3 T is arithmetically wrong before it is conceptually judged).
Confusing SNR effects with scan-time effects - for example claiming that increasing the number of phase-encoding steps shortens scan time, or that increasing receive bandwidth improves SNR.
Parameters are memorised as a list of 'what TR/TE/NEX/bandwidth do' without separating the two output axes the exam always tests: image quality and acquisition duration.
What to do instead
Keep a two-column ledger (SNR | scan time) and force every parameter into one or both columns: NEX and phase-encoding steps raise both; bandwidth raises SNR when lowered but does not change scan time; matrix and FOV affect SNR and resolution but not duration. Reproduce the ledger from memory before each practice set and check every parameter question against it.
Treating 'MR safe', 'MR conditional', and 'MR unsafe' as interchangeable labels, or answering a zone question with a person who is not permitted in that zone (e.g., allowing a non-screened family member into Zone IV).
The three device labels sound like degrees of the same thing, and the zone system is learned as four numbered rooms rather than as an access-control rule tied to screening status.
What to do instead
Learn the labels as definitions, not rankings: safe = no hazard in any MR environment; conditional = safe only under the stated conditions (field strength, SAR, orientation); unsafe = hazardous in all cases. For zones, memorise the gate rule - Zone IV is entered only by screened staff/patients and anyone who fails screening never reaches it - and answer zone questions by naming who is allowed, not which room is which.
Answering SAR questions as if SAR depended only on patient weight, or forgetting that SAR rises sharply with field strength and flip angle.
SAR limits are memorised as four numbers (whole body, head, torso, extremities) with no attached physics, so the scaling behaviour in the question stem is ignored.
What to do instead
Pair each limit with its scaling rule: SAR scales approximately with the square of field strength and with the square of flip angle, and falls with longer TR or reduced number of slices/RF pulses per unit time. Answer by identifying which variable the stem changed, then applying that dependence before quoting a limit.
Describing a quench response from memory of 'evacuate the room' without specifying oxygen monitoring, removal of the patient from the magnet room, and the fact that the cryogen gas displaces oxygen and is not visible.
Quench is a rare event so it is studied as a single sentence rather than as a sequence of actions, and the exam tests the sequence (and the reason for each step) rather than the headline.
What to do instead
Rehearse quench as an ordered checklist you can say aloud: activate alarm and remove the patient/staff from the magnet room, do not re-enter until oxygen level is confirmed adequate, ventilate, and account for the possibility of a helium release with no visible sign. Tie each step to the hazard it prevents (asphyxiation, not fire; no smoke to warn you) so the ordering is logical rather than memorised.