Abstract
Background. Postoperative monitoring of midline shift (MLS) and intracranial complications is critical for patients after decompressive craniectomy (DC). Routine computed tomography (CT) monitoring is associated with the risks of intrahospital transport of critically ill patients. Bedside transcranial sonography (TCS) through the trepanation window is a promising alternative; however, the limits of its absolute clinical agreement with CT for safe decision-making based on static thresholds (e.g., MLS > 5 mm) remain insufficiently studied. Purpose – to determine the diagnostic accuracy and absolute clinical agreement of bedside TCS compared to CT for quantifying MLS and detecting intracranial hematomas in patients following DC in the intensive care unit. Materials and methods. This prospective cohort observational study analyzed data from 25 patients with severe traumatic brain injury who underwent DC. Bedside TCS via the trepanation acoustic window was performed within 60 minutes of the refe-rence CT scan, maintaining blinding of the imaging results. The absolute agreement for MLS quantification was evaluated using Bland–Altman mathematical and graphical analysis, along with the intraclass correlation coefficient (ICC). Hematoma detection was assessed using the prevalence- and biasadjusted kappa index (PABAK). Results. The agreement evaluation for MLS measurements between CT and TCS demonstrated a moderate level of concordance (ICC = 0.736). The mean systematic bias was 0.84 mm (with TCS marginally underestimating the shift); however, the calculated 95% limits of agreement exhibited a substantial range (from –2.96 mm to 4.63 mm). The diagnostic accuracy of TCS for detecting intracerebral hematomas reached a sensitivity of 91.7% and a specificity of 92.3% (PABAK = 0.840). Furthermore, the method showed 100% accuracy in visualizing external ventricular drains and 100% specificity for detecting extra-axial hemorrhages. Conclusions. TCS is a highly effective, safe, and radiation-free bedside modality for identifying massive cerebral midline shifts, intracerebral hematomas, and verifying ventricular drain placement, entirely eliminating the risks associated with intra-hospital transport. Nevertheless, given the wide limits of agreement utilizing TCS for the absolute evaluation of static MLS thresholds is not recommended. The optimal clinical paradigm entails deploying TCS for dynamic trend monitoring relative to an individual’s baseline post-operative metrics.

This publication has 25 references indexed in Scilit: