Pediatrics
Non-Accidental Trauma: A Constellation, Never One Finding
Fellowship-level neuroradiology teaching by E. Brooke Schrickel, MD. Open this topic in the interactive reading room →
Core clinical idea
No single finding is diagnostic of abuse — the determination is a constellation of imaging plus physical findings plus clinical history, made by the multidisciplinary team, not the scan. The radiologist's job is accurate, complete description (especially features the maltreatment team uses for timelines), recognition of the few more-specific patterns, and whole-neuraxis imaging driven by mechanism, without overclaiming what imaging can prove.
Bottom line
Describe accurately, image the whole neuraxis by mechanism, know the more-specific patterns — and let the constellation, not the scan, make the call.
Core workstation questions
- Are the subdurals bilateral/asymmetric, and are the interhemispheric/vertex and posterior-fossa regions and the spine covered?
- Can I describe density/distribution/membranes precisely without overclaiming an age (no comparison = limited dating; enhancing membranes suggest a chronic component)?
- Are any more-specific patterns present — parenchymal lacerations, spinal subdural — and have I described rather than 'diagnosed' them?
- Does the mechanism (shake vs impact) match where I'm finding injury, and have I obtained GRE/SWI for bridging-vein thrombosis?
- Is there an HII pattern signaling a life-threatening event to escalate?
- Have I framed findings as a constellation requiring clinical correlation, not a single diagnostic sign?
What changes reporting / management
- Describe subdural density, distribution (interhemispheric/vertex, posterior fossa), and membranes precisely; state that dating is limited and that without a comparison exam enhancing membranes are the most reliable sign of a chronic component.
- Image the whole neuraxis; support total-spine MRI when abuse is suspected, given spinal subdural hematomas as a more-specific marker.
- Use GRE/SWI to look for bridging-vein thrombosis (lollipop/tadpole sign) as support for a traumatic cause of a subdural.
- Escalate a hypoxic-ischemic injury pattern (diffuse supratentorial or parasagittal watershed) as it may be the only imaging sign of a life-threatening event.
- Frame all findings as a constellation requiring clinical correlation and team determination; do not assert abuse from a single imaging sign or convert incidence figures into a probability about the individual child.
Practical traps
- Dating a subdural by parenchymal-hematoma signal-evolution rules.
- Over-reading fracture timing — an absent scalp hematoma does not exclude an acute fracture, and calvarial fractures lack vigorous callus.
- Relying on diffuse axonal injury to distinguish abusive from accidental trauma in young infants.
- Treating any single finding (subdural, fracture, retinal hemorrhage) as diagnostic of abuse rather than supportive within the constellation.
- Under-imaging by ignoring mechanism (missing spine, posterior fossa, or GRE/SWI bridging-vein assessment).
Teaching pearls
- No single finding diagnoses abuse — it's the constellation of imaging, exam, and history, decided by the team.
- Don't date a subdural by parenchymal-blood rules; without a prior exam, enhancing membranes are the most reliable sign of a chronic component.
- Parenchymal lacerations and spinal subdurals are the more-specific patterns — describe them, image the whole spine.
- Infant contusions sit by the scalp hematoma/fracture; the basifrontal/temporal-pole pattern is the older-child story.
- An HII pattern may be the only imaging clue to a life-threatening event — escalate it.
- No fracture pattern is diagnostic; complex/diastatic/depressed fractures are concerning when the history doesn't fit.
Source lectures
- Non-Accidental Trauma
Educational material for radiology residents and neuroradiology fellows. Nothing here drives individual patient care, and it contains no patient data.