Patients and study design
The inclusion criteria for the study was the presence of medial sphenoid wing meningiomas based on radiologic, intraoperative and pathologic findings from 2014 to 2019. Meningiomas originating in the tuberculum sellae, orbital roof, cavernous sinus, or middle or lateral aspects of the sphenoid wing were excluded. Each patient included needed a preoperative computed tomography (CT), magnetic resonance imaging (MRI) with and without contrast and magnetic resonance angiography (MRA) / computed tomography angiography (CTA). The study was approved by the local ethics committee as a retrospective study (Institutional Review Board No. 2019117). As this article is a retrospective study, ethics committees have been granted exemption of patient informed consent.
Image processing
The Digital Imaging and Communications in Medicine (DICOM) data format of images was imported from our institution’s Picture Archiving and Communication System (PACS) (Neusoft Corp., Shenyang, China) into Mimics Medical 15.0 (Materialise Company, Leuven, Belgium). Firstly, we created a mask of the imaging data to extract the targeted structure by determining the region of interest and adjusting the thresholding. Skull was delineated from CT. Blood vessels were delineated from MRA/ CTA. Tumor was delineated from the T1-weighted of MRI with contrast. Then, edits were done on the masks and the unclear margins were outlined manually. Finally, all the targeted structures were integrated into the same space and manually adjusted to the exact place to generate the 3D multimodality fusion imaging data. Tumors were set as transparent to assess the arteries inside. Tumor location, volume, CS involvement, vascular encasement, and bone invasion were evaluated. All the images were evaluated by a neuroradiologist and a young neurosurgeon, without knowledge of any patients’ clinical data.
The volume of the tumor was calculated after we established the 3D model of the tumor by Mimics software. Volume < 20cm3 was categorized into Grade 1, volume between 20cm3 and 50cm3 was categorized into Grade 2 and volume > 50cm3 was categorized into Grade 3. Grade of arterial encasement was determined as follows: involvement of any artery was given one point, and an additional point was given for each of the following features as shown in Figure 1: completely encircling (360°) the artery, involving more than one artery, and narrowing the lumen of the artery. The grade ranged from 0 to 4. CS involvement was defined as infiltration of the tumor into the cavernous sinus as shown in Figure 2. Bone invasion was observed by the 3D reconstruction of the skull.
Microsurgical Technique
Three neurosurgeons performed these surgeries and each neurosurgeon performed 27(58.7%), 11(23.9%) and 8 (17.4%) cases, respectively. For microsurgical tumor removal, pterional approach was performed on 45 patients (98%) and lateral supraorbital approach was performed on 1 patient (2%). A high-speed pneumatic drill was used to remove any thickened bony prominence of the lesser and greater sphenoid wings. After opening the dura, the operating microscope was used, and the tumor was approached through the sylvian fissure. The basal dural attachment of the tumor on the sphenoid ridge was coagulated to decrease the tumor’s blood supply and detach it from the dura. The capsule of the tumor was then opened, and the tumor was resected piecemeal. After resection of the tumor, the dura attachment was coagulated.
Extent of tumor resection
The extent of tumor resection was categorized according to the Simpson grade of resection[23]: Grade I, total tumor resection with excision of its dural attachment, and any abnormal bone; Grade II, total tumor resection and coagulation of dural attachments; Grade III, gross total tumor resection without resection or coagulation of its dural attachment, or alternatively, of its extradural extensions, e.g., an invaded sinus or hyperostotic bone. Grade IV, a partial removal, leaving intradural tumor in situ. Grade V, a simple decompression. The Simpson grade of our cases was determined by the surgical records, videos, and postoperative imaging studies.
Postoperative Symptoms and Follow-up
All the symptoms observed after surgery were recorded. To evaluate the outcome of surgical treatment, patients were assigned to a 4-category postoperative evaluation system used by Honig et al[11]: grade 0 represents a worse outcome with deterioration of symptoms; grade 1 is a fair outcome with unchanged symptoms; grade 2 is a good outcome with improvement of symptoms. Patients with grade 3 show an excellent outcome with complete regression of preoperative tumor-related symptoms.
Statistical Analyses
The variables included the volumes of the tumor, CS involvement, arterial involvement, and bone invasion. Kruskal-Wallis test for ordinal variables such as tumor volume grade and Pearson's chi-square test for nominal variables such as bone invasion were performed to compare between four different resection grade groups. Descriptive continuous data were reported as mean ± standard deviation (SD), categorical data were reported as percentages. The adjusted odds ratios (OR) were obtained by a logistic regression model including the variables showing significant differences in
univariate analysis (P < .05). The correlation between the variables and resection grade was detected respectively by Spearman’s correlation analysis. The predictive capacity of our scoring system was evaluated by discriminant analysis. The results were evaluated at 95% confidence interval (95% CI) and P < .05 was considered statistically significant. Statistical analyses were performed using SPSS version 25.0 software (IBM Corp., Armonk, New York, USA).