Neuro-POCUS for Detecting Cranial CT Pathology in Adult Emergency Department Patients: A Prospective Observational Study

Background: This study aimed to evaluate the diagnostic utility and clinical feasibility of a bedside Neuro-POCUS protocol—incorporating optic nerve sheath diameter (ONSD) assessment and transcranial Doppler-derived pulsatility index (TCCD PI)—for the early detection of intracranial hypertension in Emergency Department (ED) patients presenting with headache or mild traumatic brain injury (TBI) who underwent cranial computed tomography (CT). Methods: This single-center, prospective observational study was conducted at Baggiovara Hospital between February and May 2026. The cohort comprised 27 adult ED patients presenting with complaints of headache or mild TBI, all of whom subsequently underwent cranial CT and presented with a Glasgow Coma Scale score of 14 or 15. The Neuro-POCUS protocol included ocular sonography for ONSD measurement and papilledema assessment, alongside transcranial color-coded Doppler evaluation of the middle cerebral artery (MCA) pulsatility index. Cranial CT served as the diagnostic reference standard. The primary endpoint was defined as the detection of neuro-significant CT findings (intracranial haemorrhage, contusion, oedema, midline shift, hydrocephalus or mass lesion). Secondary endpoints included hospital admission rates, patient disposition, necessity for neurosurgical consultation, ED readmission, mortality, and total examination duration. Results: Nine patients (33 %) had neuro significant CT findings. Mean ONSD did not differ significantly between patients with and without CT pathology (3.82 ± 0.45 mm vs 4.04 ± 0.56 mm, p = 0.36); only one patient exceeded the 5 mm ONSD threshold. Conversely, PI-max was higher in patients with neuro significant CT (1.38 ± 0.61 vs 0.83 ± 0.25, p = 0.047). A logistic regression model demonstrated PI-max as an independent predictor of neuro significant CT (odds ratio ≈ 35, p = 0.044). Using PI > 1.2 as a cut off yielded 71 % sensitivity, 92 % specificity, positive predictive value of 83 %, and negative predictive value of 86 %. ONSD > 5 mm had negligible sensitivity. Neuro significant CT findings correlated strongly with hospital admission (8/9 patients) and neurosurgical consultation (7/9 patients). Median emergency department stay was approximately six hours. Conclusion: The TCCD PI demonstrated greater diagnostic potential compared to ONSD assessment in identifying patients with significant intracranial findings on cranial CT. Neuro-POCUS proved both rapid and clinically feasible within the ED environment. Nevertheless, these results necessitate cautious interpretation due to the limited sample size, incomplete data, and inherent technical challenges, particularly the prevalence of suboptimal acoustic windows. Ultimately, these findings advocate for the utility of Neuro-POCUS as a complementary bedside modality, rather than a replacement to be integrated alongside comprehensive clinical evaluation and standard neuroimaging protocols.