Automatisiere Auswertung der Hämatomgröße und der Schweregradbewertung nach experimenteller Subarachnoidalblutung: Integration der RGB-Bildbasierten Hämatomerkennung

Subarachnoid hemorrhage (SAH) presents high morbidity and mortality with limited treatment options, emphasizing the importance of experimental research using animal models. The widely employed middle cerebral artery perforation model mimics key aspects of human SAH but faces challenges in objectively assessing severity. This study focuses on developing an automated algorithm, employing RGB-image processing, to enhance the objective calculation of hematoma size and severity assessment in experimental SAH.

Following SAH induction, mouse brains (C57bl6, n = 15) were perfused, and high-resolution images were captured using a stereomicroscope. Automated analysis, previously validated,¹ utilized a custom JAVA code for ImageJ. Severity assessment employed a modified score by Sugawara et al.²

RGB color spaces were incorporated for further optimization, transforming the image into a three-dimensional matrix for efficient processing. The algorithm's effectiveness was evaluated based on R, G, and B values using bitwise and operations, calculating the ratio of selected bits to the whole brain surface area.

The established algorithm detected an average of 6.3% blood-covered area relative to the entire brain surface (mainly attributed to the central blood clot), while the optimized algorithm identified an average of 16.2% brain surface with blood (including correctly identified peripheral blood components at the skull base), resulting in approximately 10% more accurate quantification of hematoma portions on the skull base after the induction of SAH.

Overall, the extent of SAH is better captured with RGB-image-based blood clot detection, which offers an improved and objective approach to hematoma size calculation and severity assessment in experimental SAH. This researcher independent assessment method therefore helps to improve blood clot detection and severity assessment, thus enhancing clinical translatability of the model.

References

Lenz IJ, Plesnila N, Terpolilli NA. Role of endothelial nitric oxide synthase for early brain injury after subarachnoid hemorrhage in mice. J Cereb Blood Flow Metab. Jul 2021;41(7):1669-1681. doi:10.1177/0271678x20973787 Sugawara T, Ayer R, Jadhav V, Zhang JH. A new grading system evaluating bleeding scale in filament perforation subarachnoid hemorrhage rat model. J Neurosci Methods. Jan 30 2008;167(2):327-34. doi:10.1016/j.jneumeth.2007.08.004