Semiautomated Volumetric Hemorrhage Quantification Using High-Frequency Ultrasound Following Spinal Cord Injury

Intraparenchymal hemorrhage (IPH) plays a significant role in the pathophysiology of traumatic spinal cord injury (SCI). Once IPH occurs following the initial mechanical trauma, the blood itself can trigger a secondary injury cascade that can worsen damage to surrounding neural tissue and lead to additional neurological deficits. Assessing IPH and its progression in the human setting is extremely challenging without performing serial imaging studies. This highlights the rationale for monitoring IPH in pre-clinical settings, where early changes can be studied and tracked more effectively. The aims of this study were to (1) characterize IPH progression during the first 7 h following SCI using high-frequency ultrasound (US) imaging in a porcine model and (2) develop and validate a semiautomated method for quantifying IPH using US. Seventeen female Yucatan miniature pigs were used in this study. Each animal underwent a weight-drop contusion-compression SCI followed by serial intraoperative US scanning during the first 7 h post-SCI, and then a final scan 7 days later. A semiautomated segmentation approach was developed and used to quantify early IPH progression, validated by the presence of red blood cells through hematoxylin and eosin staining. To assess the accuracy and reliability of the semiautomated method, we conducted inter-rater and intra-rater reliability assessments. IPH was consistently observed across all examined time points post-SCI. The expansion of IPH, in terms of both volume and length, occurred immediately following the injury, specifically at 0.5 h post-SCI. This expansion continued over the subsequent hours up to the 7-h time point, although at a slower rate. During this period, the expansion was primarily observed axially in the dorsal and ventral directions rather than along the rostral-caudal axis of the cord. The semiautomated quantification approach demonstrated excellent inter-rater and intra-rater reliability for IPH measurements, achieving greater consistency than traditional manual segmentation. In conclusion, we observed distinct regional patterns of IPH expansion over time, particularly in the dorsal and ventral areas during the early hours post-SCI, with no evident plateauing at 7 h post-injury (HPI). The implementation of a semiautomated quantification method marks a significant advancement in IPH assessment, enhancing measurement accuracy on US and indicating a move toward more objective IPH evaluations in SCI research.