Positional effect of preoperative neuronavigational magnetic resonance image on accuracy of posterior fossa lesion localization

Restricted access

OBJECTIVE

The aim of this study was to analyze the positional effect of MRI on the accuracy of neuronavigational localization for posterior fossa (PF) lesions when the operation is performed with the patient in the prone position.

METHODS

Ten patients with PF tumors requiring surgery in the prone position were prospectively enrolled in the study. All patients underwent preoperative navigational MRI in both the supine and prone positions in a single session. Using simultaneous intraoperative registration of the supine and prone navigational MR images, the authors investigated the images’ accuracy, spatial deformity, and source of errors for PF lesions. Accuracy was determined in terms of differences in the ability of the supine and prone MR images to localize 64 test points in the PF by using a neuronavigation system. Spatial deformities were analyzed and visualized by in-house–developed software with a 3D reconstruction function and spatial calculation of the MRI data. To identify the source of differences, the authors investigated the accuracy of fiducial point localization in the supine and prone MR images after taking the surface anatomy and age factors into consideration.

RESULTS

Neuronavigational localization performed using prone MRI was more accurate for PF lesions than routine supine MRI prior to prone position surgery. Prone MRI more accurately localized 93.8% of the tested PF areas than supine MRI. The spatial deformities in the neuronavigation system calculated using the supine MRI tended to move in the posterior-superior direction from the actual anatomical landmarks. The average distance of the spatial differences between the prone and supine MR images was 6.3 mm. The spatial difference had a tendency to increase close to the midline. An older age (> 60 years) and fiducial markers adjacent to the cervical muscles were considered to contribute significantly to the source of differences in the positional effect of neuronavigation (p < 0.001 and p = 0.01, respectively).

CONCLUSIONS

This study demonstrated the superior accuracy of neuronavigational localization with prone-position MRI during prone-position surgery for PF lesions. The authors recommend that the scan position of the neuronavigational MRI be matched with the surgical position for more precise localization.

ABBREVIATIONS FM = fiducial marker; MRI-NVS = MRI-based neuronavigation system; PF = posterior fossa.

Downloadable materials

  • Supplementary Materials (PDF 572 KB)

Article Information

Correspondence Chul-Kee Park: Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Republic of Korea. nsckpark@snu.ac.kr.

INCLUDE WHEN CITING Published online July 19, 2019; DOI: 10.3171/2019.4.JNS1989.

Y.S.D., Y.J.K., and K.G.K. contributed equally to this work.

Disclosures The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Schematic presentation of the flow of this study. Copyright Mi-Jin Jung, Biomedical Illustration & Design Company. Published with permission. Figure is available in color online only.

  • View in gallery

    A: The posterior fossa localization process during the surgery. After localization of the same test points, we confirmed the accuracies of the prone and supine MR images. The prone MRI showed better accuracy (yellow arrows). B: Posterior fossa localization was performed at the 64 test points, and the accuracies of the points were compared between the 2 MR images. Red dots—accurate in the prone MR image (n = 60, 93.8%); yellow dots—accurate in the supine MR image (n = 4, 6.2%). lat = lateral area; mid = midline area. Figure is available in color online only.

  • View in gallery

    A: The brains from the prone and supine MR images were reconstructed and synthesized 3-dimensionally by using BrainVAAS. Red: the brain from the prone MRI; light green: the brain from the supine MRI. B: The figure shows the actual distortion between the 2 brains based on panel A. The brain from the supine MRI is located in the posterior and superior direction compared to the brain from the prone MRI. C: Positional differences of the 2 specified coordinates in the brain on the prone and supine MR images, which shows the magnitude and direction of the vector between the 2 coordinates. For the coordinates specified in the prone MRI in the 3D space, all coordinates specified in the supine MRI are located in the posterior and the superior direction. The average distance of distortion between the prone and supine MR images was 6.3 mm. D: The X vector values between 2 coordinates in the 3D space were all 0, whereas both the Y and the Z vector values were positive. Copyright Mi-Jin Jung, Biomedical Illustration & Design Company (panels B and C). Published with permission. Figure is available in color online only.

  • View in gallery

    A: The brain from the prone and supine MR images was synthesized based on the FMs and sectioned by the axial, coronal, and sagittal planes in units with a 1-mm thickness. Gray area: distorted area of the brain on supine MRI compared to that of the brain on prone MRI; white area: distorted area of the brain on prone MRI compared to that of the brain on supine MRI. B: The degree of distortion of each cross-section (the ratio of the distorted area of the cross-section [the gray and white areas in panel A]/the total area of the prone MRI of the cross-section) was calculated and plotted in association with the brain region. Axial plane: the ratio of distortion is large on the caudal side. Coronal plane: the ratio of distortion is large on the posterior side. Sagittal plane: the ratio of distortion is large on the midline area. Copyright Mi-Jin Jung, Biomedical Illustration & Design Company (panel A). Published with permission. Figure is available in color online only.

  • View in gallery

    Box-and-whisker plots. A: Sixteen points were specified based on the anatomical landmarks of the prone and supine MR images, which were synthesized based on the FMs. After localizing the points on the 2 MR images, the distance between 2 points was measured. The locations in the supratentorial area had significantly greater distances than the locations in the infratentorial area (p < 0.001). B: The figure shows the distances between the prone and supine MR images for each of the 11 anatomical landmarks. The structures located in the supratentorial area have a larger distance value than the structures located in the infratentorial area. ACA = anterior cerebral artery; BA = basilar artery; FH = frontal horn; IT = infratentorial; ON = optic nerve; ST = supratentorial; Tent = tentorium; Tonsil = cerebellar tonsil; TSJ = transverse-sigmoid sinus junction; VA = vertebral artery; 4th V = fourth ventricle; 7/8 N = seventh/eighth cranial nerve complex.

  • View in gallery

    A: Sixteen points were specified based on the anatomical landmarks of the prone and supine MR images, which were synthesized based on the FMs. After localizing the points on the 2 MR images, the distance between 2 points was measured. Patients > 60 years have significantly greater distances than those of patients ≤ 60 years (p < 0.001). B: All 17 points (100%) on the scalp adjacent to the craniocervical muscles were more accurate in the prone MRI, 26 (72.2%) of the 36 parietal and occipital areas showed higher accuracy in the prone MRI, and 10 points (27.8%) showed higher accuracy in the supine MRI; these differences were statistically significant (p = 0.01).

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 428 428 75
Full Text Views 66 66 11
PDF Downloads 36 36 2
EPUB Downloads 0 0 0

PubMed

Google Scholar