Imaging Pitfalls

For Medical Students

What It Shows

Imaging goes wrong in four patterned ways: it finds something irrelevant, it misses something present, it is read without the clinical question, and it is technically inadequate for the question asked. 5.1.2 covers that comparison.

The first is the commonest and the least recognised, because a report naming a finding feels like an answer. 5.6.2 covers meniscal signal grading.

Knowing the patterns is what lets a clinician read a report rather than accept it. 5.20.8 covers common reporting errors.

Where It Misleads

Prevalent findings are present in asymptomatic knees and accumulate with age, so the older the patient the more the scan offers and the less of it means anything. 5.6.2 covers meniscal signal grading.

Technically inadequate studies answer a different question, and the inadequacy is often invisible on the report. 5.5.1 covers the standard protocol.

Artefact mimics pathology in nameable ways, and each modality has its own set. 5.2.19 covers artefacts.

Timing changes what is there: an effusion resolves, a bone bruise fades, and a healing structure looks worse before it looks better. 5.13.1 covers bone marrow oedema.

A report written without the clinical question answers the question the radiologist inferred. 5.20.1 covers structured reporting.

MRCS Level

How Good It Is

Recognising the patterns is what makes imaging usable, and it is a clinical skill rather than a radiological one.

Its best contribution is downgrading a finding that does not fit the clinical picture, which is the judgement nothing else in the assessment makes. 5.1.2 covers that comparison.

It is undermined by the authority of a written report, which is harder to argue with than an examination finding.

No figures appear for any accuracy comparison on this page, under decision 028, and no prevalence figure appears either, for the same reason.

Where It Misleads

In a viva this is the critical appraisal question and the expected answer names the patterns rather than listing examples.

The expected points are the prevalent finding, the technical inadequacy, the artefact, the timing and the missing clinical question.

Say that a degenerate meniscal finding in an older knee is prevalent rather than causal, and say what that means for the plan. 5.6.2 covers meniscal signal grading.

Say that the magic angle effect is a specific and nameable artefact rather than a general warning. 5.2.20 covers the magic angle effect.

Say that you would read the images rather than the report alone where the decision depends on them. 5.20.8 covers common reporting errors.

FRCS Level

How Good It Is

At decision level the pitfalls are what stand between a report and an operation.

A finding that does not match the examination is downgraded explicitly and in writing, because an unrecorded downgrade will be reversed by the next clinician. 4.1.8 covers examination documentation.

A technically inadequate study is repeated or worked around, and deciding which needs knowing what the protocol could not show. 5.5.1 covers the standard protocol.

An incidental finding needs a decision, and deciding to do nothing is a decision that belongs in the record. 5.1.5 covers cost effectiveness.

A finding that would change the operation is checked on the images rather than taken from the report. 5.20.8 covers common reporting errors.

Where It Misleads

The commonest error is operating on a report rather than on a patient. 5.1.2 covers that comparison.

A prevalent finding is treated because it is the only abnormality named. 5.6.2 covers meniscal signal grading.

An artefact is treated as a lesion, which the magic angle effect causes more often than any other. 5.2.20 covers the magic angle effect.

A previous study on an unknown protocol is compared with a current one as if they were a series. 5.5.1 covers the standard protocol.

The report's uncertainty is dropped on the way into the plan, so a possible becomes a definite between the scan and the clinic. 5.20.1 covers structured reporting.

Fellowship Level

Where It Misleads

At fellowship level the pitfalls are structural rather than individual, and that changes what can be done about them.

The prevalence problem has no clinical solution, because the finding is real and its relevance is a judgement every time. 5.6.2 covers meniscal signal grading.

Protocol variation between centres is invisible on reports, so comparison across studies is unreliable in a way nobody flags. 5.5.1 covers the standard protocol.

Advanced sequences introduce their own artefacts, so the techniques that answer the harder questions bring new pitfalls rather than fewer. 5.18.1 covers quantitative imaging.

Automated and assisted reading changes the error pattern rather than removing it, and the new pattern is less familiar than the old one. 5.18.8 covers artificial intelligence in imaging.

The pitfalls recur across this whole section, which is why they have a chapter rather than a paragraph on each page. 5.2.20 covers the magic angle effect.