Time-resolved ADC analysis differentiates stable vs. progressive brain injury in post-cardiac arrest patients.

Abstract

To investigate whether serial MRI-based apparent diffusion coefficient (ADC) analysis can evaluate the time-dependent trajectories of Hypoxic Ischemic Brain Injury (HIBI) and predict outcomes in out-of-hospital cardiac arrest (OHCA) survivors. This retrospective cohort study included adult comatose OHCA survivors who underwent brain MRI within 6 h (ultra-early) and at 72-96 h (subacute) post-return of spontaneous circulation (ROSC). Quantitative voxel-based ADC metrics were extracted across thresholds from 200 to 1200 × 10^-6 mm²/s. ADC-R(x) was defined as the cumulative volume of voxels with ADC ≤ x. The primary outcome was poor outcome (CPC 3-5) at 6 months. Serum neuron-specific enolase (NSE) levels were measured serially and correlated with ADC values. Among 122 patients, 61 had poor outcomes. Subacute MRI showed stronger group separation and higher prognostic accuracy than ultra-early MRI, with the area under the curve peaking at 0.91 for ADC-R(400) and sensitivity improving by +0.53 at ADC-R(420). Patients with good outcomes demonstrated a rightward shift in ADC distributions and increased mid-to-high range ADC values, suggesting partial diffusion normalisation. Conversely, poor outcome patients showed progressive accumulation of low-ADC voxels (280-600 × 10^-6 mm²/s), indicating irreversible injury. ADC-R(x) correlated strongly with NSE in the poor outcome group at 6 h (ρ = 0.65), with overall cohort correlation improving by 72-96 h (ρ = 0.50). Serial ADC analysis reveals dynamic and diverging HIBI patterns in OHCA survivors. Subacute MRI more accurately reflects progressive HIBI and improves prognostic performance, supporting its use in neuroprognostication beyond 72 h post-ROSC.