Full-course resection control strategy in glioma surgery using both intraoperative ultrasound and intraoperative MRI

Abstract

Background: Intraoperative ultrasound(iUS) and intraoperative MRI (iMRI) are effective ways to perform resection control during glioma surgery. However, most published studies employed only one modality. Few studies have used both during surgery. How to combine these two techniques reasonably, and what advantages they could have for glioma surgery are still open questions.

Methods: We retrospectively reviewed a series of consecutive patients who underwent initial surgical treatment of supratentorial gliomas in our center. We utilized a full-course resection control strategy to combine iUS and iMRI: IUS for pre-resection assessment and intermediate resection control; iMRI for final resection control. The basic patient characteristics, surgical results, iMRI/iUS findings, and their impacts on surgical procedures were evaluated and reported.

Results: A total of 40 patients were included. The extent of resection was 95.43 ± 10.37%, and the gross total resection rate was 72.5%. The median residual tumor size was 6.39 cm³ (range 1.06-16.23 cm3). 5% (2/40) of patients had permanent neurological deficits after surgery. 17.5% (7/40) of patients received further resection after the first iMRI scan, resulting in four (10%) more patients achieving gross total resection. The number of iMRI scans per patient was 1.18 ± 0.38. The surgical time was 4.5 ± 3.6 hours. The pre-resection iUS scan revealed that an average of 3.8 borders of the tumor were beside sulci in 75% (30/40) patients. Intermediate resection control was utilized in 67.5% (27/40) of patients. In 37.5% (15/40) of patients, the surgical procedures were changed intraoperatively based on the iUS findings. Compared with iMRI, the sensitivity and specificity of iUS for residual tumors were 46% and 96%, respectively.

Conclusion: The full-course resection control strategy by combining iUS and iMRI could be successfully implemented with good surgical results in initial glioma surgeries. This strategy might stabilize resection control quality and provide the surgeon with more intraoperative information to tailor the surgical strategy. Compared with iMRI-assisted glioma surgery, this strategy might improve efficiency by reducing the number of iMRI scans and shortening surgery time.

Keywords: glioma; intraoperative MRI; intraoperative ultrasound; navigation; neurosurgery.

Figure 1

iMRI and iUS systems used in this study. (A) the iMRI system integrated a movable 3.0T MRI scanner. With the shielded door open, the scanner could move into the operating room to perform an iMRI scan. (B) iUS (white circle) and the navigation system (green circle) were registered during surgery to produce 3D iUS and co-planar MRI images. (C) During surgery, the shielded door was closed to isolate the magnetic field. iUS (white circle) and navigation system (Green circle) were placed on the left side of the surgeon. (D) iUS was registered with the navigation system by attaching a rigid reference frame to the transducer.

Figure 2

Illustrative case 1. (A) The preoperative MRI images revealed a low-grade glioma, predominantly in the right temporal lobe. It was difficult to determine how the tumor is related to the Sylvian fissure, insular lobe, or basal ganglia based on MRI images. (B) Pre-resection iUS image showing the tumor’s hyperechoic signal. Sylvian fissure (white arrow) and insular cortex were clearly visible on the iUS images. The insular cortex was not infiltrated by the tumor. The Sylvian fissure marked the upper border of tumor (C) In another pre-resection iUS image, the posterior margin was marked by a sulcus (white arrow). (D) Following partial removal of the tumor, the temporal stem (white arrows) was visible on the iUS image. There was no sign that the tumor (black arrow) had grown into basal ganglia through the temporal stem. The Green arrow indicated the ventricle. (E) Intermediated iUS revealed that the tumor was further removed along the temporal stem (white arrow). The Green arrow indicated the ventricle. (F) In iMRI images, the tumor was successfully removed along the Sylvian fissure and temporal stem. The insular cortex and basal ganglia were preserved.

Figure 3

Illustrative case 2. (A) The pre-resection iUS image was registered with MRI using navigation. (B) Overlay of iUS image on coplanar MRI images. The black arrows indicated the border of the enhanced tumor on T1+C images. Compared to the enhanced tumor, the hyperechoic area was larger. (C) The first tissue sample was taken during resection of the enhanced tumor. On the co-planar 3D iUS, this part was hyperechoic. (D) Hematoxylin and eosin (H&E) stain (×40) showed marked hypercellularity, cellular and nuclear atypia, increased mitotic figures, and hypervascularity. (E) The second tissue sample was taken outside the enhanced lesion, where MRI showed an abnormally long T1 signal. Co-planar 3D iUS showed a similar signal at this site as the enhanced part. (F) H&E stain (×40) demonstrated the same characteristics as image D, revealing moderate infiltrative growth of tumor cells. (G) the third sample was taken after all tissue with abnormal MRI signals had been removed. At this site, the iUS were still hyperechoic but slightly lower than the previous site. (H) H&E stain (×40) demonstrated the microstructure of white matter, with mild gliosis and nuclear atypia. (I, J) 3D reconstruction based on iMRI images. The removed tissues with hyperechoic signals were segmented as green volumes, which measured 60.6cm³. As a comparison, the enhanced part was segmented as red volume, which was only 8.1cm^3..

Figure 4

Illustrative case 3. (A) On the preoperative MRI images, a tumor could be seen in the right frontal lobe. It was diffuse and lacked defined boundaries. (B) Coplanar MRI and iUS images showed the hyperechoic area was much smaller than the area with abnormal signals on flare images (blue line). Purple lines represented the pyramidal tract (PT) boundaries revealed by DTI. (C) Intermediate iUS images showing the resection close to the PT. In accordance with iUS, we purposefully increased the spatial and temporal intensity of mapping. The red dots indicate areas where we got positive reactions at the bottom of the resection cavity. The lowest current intensity was 3mA. According to the mapping result, we should stop resection. However, the iUS images revealed a tumor on the medial side of the PT (white arrow). (D, E) The residual tumors (black arrowhead) were then removed with the help of a small bur hole transducer. (F) The spatial relation between the positive mapping sites (red dots) and the pyramidal tract (color streamlines). By integrating iUS with navigation, we were able to position mapping points accurately and comprehensively.