Noninvasive neurophysiological diagnostics of delayed cerebral ischemia in aneurysmal subarachnoid hemorrhage: a scoping review

BACKGROUND: Delayed cerebral ischemia (DCI) is a major complication following aneurysmal subarachnoid hemorrhage (aSAH), affecting outcomes. Given its multifactorial pathophysiology, including vasospasm, microcirculatory dysfunction, and neuroelectric disturbances, a range of diagnostic modalities has been studied. Neurophysiological techniques such as transcranial Doppler (TCD), electroencephalography (EEG), and near-infrared spectroscopy (NIRS) offer complementary insights. This scoping review provides an overview of these modalities, their analytical approaches, diagnostic performance, and temporal aspects of signal change in predicting DCI. METHODS: The review followed PRISMA-ScR guidelines. A systematic search of PubMed, EMBASE, and Web of Science identified studies published between 2004 and January 2025. Eligible studies reported sensitivity, specificity, and/or area under the ROC curve (AUC) for predicting DCI or related outcomes in adult aSAH patients. 46 studies were included: 26 on TCD, 14 on EEG, and 6 on NIRS. RESULTS: TCD was the most studied modality. Mean flow velocity (MFV) of the middle cerebral artery (MCA) was most frequently analyzed, with AUCs ranging from 0.59 to 0.81. Combining TCD with other variables, such as melatonin levels or qEEG features, improved diagnostic accuracy (AUC up to 0.96). NIRS studies used ROC-based rSO₂ cutoffs (65-70%) with AUCs up to 0.93. EEG studies using quantitative features, especially alpha/delta ratio, showed strong predictive value. Diagnostic performance should be interpreted in the context of heterogeneous outcome definitions across studies. CONCLUSION: TCD, EEG, and NIRS each contribute unique physiological data. Signal changes may precede DCI, suggesting a potential window for early intervention. Combined use may enhance early DCI detection. Future research should focus on multimodal integration, threshold standardization, and real-time predictive modeling, including artificial intelligence (AI) based approaches.