Neuroimaging selection and radiation-aware practice — formative SAQs

SAQ 1 (10 marks)

A 3-year-old boy is brought to the emergency department after falling off a playground slide. He cried immediately, did not lose consciousness, is alert and interactive with a Glasgow Coma Scale of 15, has not vomited, has no palpable skull fracture or signs of basilar skull fracture, no severe headache, and his parents feel he is acting normally. The fall was approximately one metre onto a rubber surface. The parent asks whether he needs a brain scan. ^{[1] [5]}

1. Apply the PECARN prediction rule and state the imaging disposition for this child, defending your reasoning. (4) ^[1]
2. The parent has read online that a CT scan will give her child cancer. Explain the attributable lifetime cancer risk of a single paediatric head CT in plain language, citing the evidence that underpins it. (3) ^{[2] [3] [4]}
3. Outline the four levers of the ALARA principle and give one practical example of each as it applies to this child and to children generally. (3) ^[5]

Model answer — SAQ 1

(1) PECARN application and disposition (4). This child is in the two-years-and-over PECARN arm. I work through the six predictors: altered mental status (absent — GCS 15, alert and interactive); signs of basilar skull fracture (absent — no haemotympanum, racoon eyes, Battle sign, or cerebrospinal-fluid leak); loss of consciousness (absent); severe mechanism of injury (absent — a one-metre fall onto a rubber surface is below the threshold of a fall of more than 1.5 metres or a motor-vehicle crash with ejection); severe headache (absent); and vomiting (absent). Meeting none of the predictors places him at very low risk of a clinically-important traumatic brain injury, with a negative predictive value well over 99 percent in the derivation cohort. CT is therefore not indicated. I observe him briefly in the department and discharge him with written head-injury advice, giving the explicit return precautions of persistent or repeated vomiting, worsening headache, increasing drowsiness, a seizure, or any focal neurology. I explain to the parent that the scan is more likely to cause harm than to find anything, and that the validated rule supports observation over imaging. ^[1]

(2) Attributable cancer risk in plain language (3). A single paediatric head CT delivers an effective dose of the order of 1 to 2 mSv, comparable to several months of natural background radiation (which is around 2 to 3 mSv per year in most regions). The attributable lifetime cancer risk is small and stochastic — meaning probabilistic, with no safe threshold assumed — and is of the order of one excess cancer per 10,000 scans in the cohort estimates. I frame this for the parent as a real but very small increase over the background lifetime cancer risk of roughly one in three, and I compare it with the everyday risks the family accepts. The evidence base is the lifetime-risk model of Brenner (2001), the UK CT-in-childhood retrospective cohort of Pearce (2012) which demonstrated a dose-response with cumulative brain and marrow dose, and the Dutch cohort of Meulepas (2019) with longer follow-up and refined dose estimates; these studies do not forbid CT, they justify minimising and justifying every scan. ^{[2] [3] [4]}

(3) The four levers of ALARA (3). First, justification: is any imaging needed, is it needed now, and what question does it answer — for this child, no imaging is needed because the question (is there intracranial injury) is answered by the validated PECARN rule and a well-looking child. Second, optimisation of the dose: when CT is the modality, use a paediatric protocol that scales the milliampere-seconds and kilovoltage peak to the child, limits the field, uses a single phase, and applies iterative reconstruction. Third, substitution of a non-ionising modality: where ultrasound or MRI answers the question, choose it — for example, a rapid MRI protocol for shunt surveillance, or cranial ultrasound in the neonate. Fourth, limitation of repeat scanning: cumulative dose accumulates across a lifetime, so each repeat request is justified on its own merits, and the shunted or oncology child is the population at greatest cumulative risk. ^[5]