Differentiating intraparenchymal hemorrhage from contrast extravasation on post-procedural noncontrast CT scan in acute ischemic stroke patients undergoing endovascular treatment

Introduction: This study aimed to identify the imaging characteristics that can help differentiate intraparenchymal hemorrhage from benign contrast extravasation on post-procedural noncontrast CT scan in acute ischemic stroke patients after endovascular treatment.

Results: Of the 135 patients studied, 74 (55 %) patients had hyperdense lesion(s) on immediate post-procedural CT scan. Follow-up scans confirmed the diagnosis of intraparenchymal hemorrhage in 20 of these 74 patients. A receiver operating characteristic analysis showed that the average attenuation of the most hyperdense lesion can differentiate intraparenchymal hemorrhage from contrast extravasation with an area under the curve of 0.78 (p = 0.001). An average attenuation of <50 Hounsfield units (HU) in the most visually hyperattenuating hyperdense lesion had 100 % specificity and 56 % sensitivity for identification of contrast extravasations. Petechial hyperdensity was seen in 46/54 (85 %) patients with contrast extravasation versus 9/20 (45 %) patients with intraparenchymal hemorrhage on the immediate post-procedural CT scan (p < 0.001).

Methods: We reviewed the clinical and imaging records of all acute ischemic stroke patients who underwent endovascular treatment in two hospitals over a 3.5-year period. The immediate post-procedural CT scan was evaluated for the presence of hyperdense lesion(s). The average attenuation of the lesion(s) was measured. Intraparenchymal hemorrhage was defined as a persistent hyperdensity visualized on follow-up CT scan, 24 h or greater after the procedure.

Conclusion: An average attenuation <50 HU of the most hyperattenuating hyperdense parenchymal lesion on immediate post-procedural CT scan was very specific for differentiating contrast extravasation from intraparenchymal hemorrhage in acute ischemic stroke patients after endovascular treatment.