Role of MR Morphology and Diffusion-Weighted Imaging in the Evaluation of Meningiomas: Radio-Pathologic Correlation

ABSTRACT Introduction Preoperative differentiation of benign, atypical and malignant meningiomas would significantly help in surgical planning and treatment. The aim of this study is to look at radio-morphologic behavior of various histopathological types and grades of meningiomas and their diffusion characteristics. Methods We performed an analytical cross-sectional study including all patients operated on for meningiomas at our hospital during January 2016 to July 2018. We studied 38 meningiomas in 38 patients aged 14 to 73 years old. All patients underwent MRI prior to surgery, including diffusion-weighted sequences, in a 1.5T scanner. Signal intensity in T2-weighted images, diffusion-weighted images (b=0, 90 and 1,000), and Apparent Diffusion Coefficient maps within the tumors and in the normal parietal white matter as a reference were evaluated. In the histological study, cellularity, proliferation index, histological grade, and cerebral invasion were evaluated. Results There was female predilection with male:female ratio of 1:2.4. Most meningiomas were supratentorial with most common origin being parafalcine and convexity. Of the 38 meningiomas, 31 were WHO grade I, 6 were WHO grade II (atypical) and one was WHO grade III (anaplastic). Among various tumors' behaviors, incomplete CSF cleft, pial invasion and parenchymal invasion were significantly high in high-grade tumors. Similarly, tumors showing pial invasion, breached tumor-brain interface, no capsular enhancement and parenchyma invasion showed significantly low NADC. Mean ADC value was 0.722±7.7x10−3 mm2/s (normalized ADC 0.9±0.1) in the atypical group and 0.876±24.56x10−3 mm2/s (normalized ADC 1.11±0.31) in the typical group. No statistically significant differences of ADC/NADC were found between histologic subtypes. Two subtypes of typical meningiomas, metaplastic and angiomatous meningioma had the highest values in the ADC maps. Conclusions MR morphology like pial invasion, breached tumors brain interface, parenchymal invasion can predict aggressiveness and atypical nature of meningiomas. Meningioma shows moderately restricted diffusion. The signal on the ADC map is associated with tumors cellularity and aggressiveness suggesting its usefulness for predicting the histological grade.


Meningiomas comprise about 14% to 20% of all
Free Full Text Articles are Available at www.jnma.com.npintracranial tumors. 1 Most meningiomas are benign and even asymptomatic.Atypical meningiomas make 7.2% and rarer malignant ones constitute 2.4% of all meningiomas. 2These have higher propensity of recurrence with aggressive growth pattern thereby increasing patient morbidity and mortality. 2I is modern day imaging modality of choice for meningioma.Preoperative differentiation of benign and atypical meningiomas significantly helps in surgical and treatment planning.This however is not reliably accomplished assessing the imaging features on routine MRI alone. 3Diffusion-weighted imaging (DWI) has been used in primary brain neoplasms.Correlations between apparent diffusion coefficient (ADC) values, tumors cellularity, and tumors grade have been made.5][6] However role of DWI in the diagnosis or prognosis of extra-axial neoplasms is unclear.
This study aims to examine various morphometric and signal characteristics of meningiomas on conventional sequences and correlate diffusion coefficient with histopathology.We hypothesized certain conventional MRI characters and diffusion constant may help distinguish benign and atypical meningiomas.

METHODS
Neuroradiology imaging and neuropathology database of all consecutive cases of meningioma that were admitted and underwent surgical resection at Upendra Devkota Memorial National Institute of Neurological and Allied Sciences from January 2016 to July 2018 were retrospectively reviewed.Atypical and malignant meningiomas were diagnosed based on the WHO classification of grades II and III tumors, respectively.The exclusion criteria were previous radiotherapy or radiosurgery, preoperative trans-arterial embolization, and incomplete or uninterpretable preoperative MRI studies.Institutional review board of the respective institute approved this study.A total of 38 patients were enrolled, including 31 (75%) with Grade I meningiomas and 7 (25%) with high-grade (Grade II or III) meningiomas.
Preoperative MRI was available for each patient and was performed using a 1.5-T MR unit (Magnetom, Essenza, Siemens).The MRI protocol was TR 3900 msec, TE 111 msec, matrix size 230×230, section thickness 5 mm, and intersection gap 0.21 mm with b-values of 0, 90 and 1000 s/mm 2 in 3 orthogonal directions.Routine images of the whole brain, including spin echo T1weighted images, spin echo T2-weighted images, and fluid-attenuated inversion recovery (FLAIR) images were obtained.Spin echo contrast-enhanced T1-weighted images were obtained in the coronal, sagittal, and axial planes after intravenous Gadolinium administration (0.1 mmol/kg body weight).Diffusion-weighted imaging (DWI) was acquired in the axial plane using a singleshot, spin echo, echo planar imaging sequence.
Signal intensities of the meningiomas on T1-and T2weighted imaging were recorded as hypointense, isointense, or hyperintense relative to the intensity of the gray matter.Meningiomas with distinct peritumoral rims and CSF clefts, which were hypointense on T1weighted imaging and hyperintense on T2-weighted imaging, were defined as clear tumors-brain interface.In contrast, tumors without distinct borders were defined as unclear tumors-brain interface.
The pattern of contrast enhancement after Gd administration was divided into homogeneous or heterogeneous.Intratumoral cystic change, defined as an area of hyperintensity on T2-weighted imaging and hypointensity on T1-weighted imaging without contrast enhancement, was regarded as heterogeneous enhancement in this study.Capsular enhancement was defined as the entire enhanced layer at the tumorsbrain interface and was categorized as positive or negative.The presence of brain edema was judged as a hyperintense extension adjacent to tumors on T2-weighted imaging and was judged as positive or negative.
The DWI was visually inspected and classified as hyperintense, isointense, or hypointense in comparison with normal white matter.According to the particular site of origin, the location of each intracranial meningioma was divided into convexity, tentorial and bony on which dura of origin is predominantly based.The image interpretation of each MRI feature was described and confirmed by 2 experienced radiologists.
All data were analyzed using the SPSS statistical program version 20.The association between radiological features of MRI along with patient age and sex and the histopathological grade of meningiomas were examined by univariate and multivariate analyses.Logistic regression was used to identify significant factors that were predictive of high-grade meningiomas.

RESULTS
Mean age of the patients was 43.5 years with age range of 14-73 years.Out of 38 total meningiomas, 11 (28.9%) were found in male and 27 (71.1%)were found in female.

.2) 0
There was no significant difference in nodularity, lobulations and tumors margin between low grade and high-grade tumors.Likewise, no significant difference in parenchymal edema was seen between these two groups.Capsular enhancement, flow voids, venous sinus involvement and tumors brain interface also showed no significant difference between low grade and high-grade tumors.However, there was significant difference in pial invasion and parenchymal invasion between these two groups with pial invasion and parenchymal invasion significantly higher in high-grade meningiomas.Incomplete CSF cleft also showed to be significantly high in high-grade meningioma (Table 2).Tumors with lower NADC values showed significantly more pial invasion, parenchymal invasion, breached tumors brain interface and breached capsular enhancement.Significant difference in NADC value was seen in various histologic subtypes of meningioma.However, low ADC value was seen in atypical meningioma (0.618x10 -3 mm 2 /s), anaplastic meningioma (0.645x10 -3 mm 2 /s) and meningothelial meningioma (0.727x10 -3 mm 2 /s) and highest ADC was seen in metaplastic meningioma.
No significant difference in tumors grade was seen between meningiomas arising in skull base and those arising at sites other than skull base.

DISCUSSION
Predicting pre-surgical histopathological grade of meningioma is helpful in appropriate treatment plan.
The association between specific radiological features and aggressive biological behavior has been studied ][9][10] Our study attempts to incorporate the use of diffusion restriction and other radio-morphological findings in predicting grades of meningioma.As established by previous studies, in this study as well, there is female predilection of meningioma with male to female ratio of 1:2.4.The finding that age was a risk factor for high-grade meningioma is controversial.It has been reported that age is an independent variable in predicting tumors recurrence and degree of differentiation according to previous reports. 11However, this is not supported by our study result.Heterogeneous MRI enhancement after Gd injection is associated with uneven distribution of tumor cells or even ischemic necrosis, hemorrhage, cystic degeneration, accumulation of tumor cell secretion, and evidence of rapid tumor expansion, the biological features of malignant tumors. 2,12Several reports have stated that Grade II and III meningiomas have significantly more intratumoral cystic changes compared with Grade I meningiomas. 3,13In the present study, heterogeneous enhancement, as well as the presence of an intratumoral cyst, was an important factor predictive of high-grade meningioma, consistent Free Full Text Articles are Available at www.jnma.com.np with previous studies.
The interface between the tumor and the brain is determined by the expression of a peritumoral rim.
A clear peritumoral rim indicates the presence of a physiological barrier between the meningioma and brain parenchyma and an unclear peri-tumoral rim suggests tumor adhesion and invasion of the surrounding brain tissue, the logical feature of aggressive biological behavior. 14,15As in previous reports, an unclear tumorbrain interface was a significant indicative factor in predicting high-grade meningiomas in both univariate and multivariate analyses in our study.
Likewise, a positive capsular enhancement, defined as the enhanced layer at the tumor-brain interface, was another identified predictor in our study.Meningiomas with unclear tumors-brain interface had negative capsular enhancement or partial loss of capsular enhancement.This means that unclear tumor-brain interface is a negative confounder in determining the association between positive capsular enhancement and high-grade meningioma.This fact is supported by the findings of this study that absence or incomplete capsular enhancement, unclear tumor-brain interface, pial invasion and parenchymal invasion were significantly higher in high-grade meningioma.In addition to this, loss of CSF cleft also showed significant difference being higher in high-grade meningioma.
7][18] Nakano et al. reported that the invasive pattern of tumor-brain interface including irregular tumors margins, loss of the peri-tumoral rim, and hyperintensity of the tumor on T2-weighted imaging was associated with meningiomarelated brain edema.However, several studies showed no significant correlation between histological subtypes of meningiomas and peri-tumoral brain edema. 2,19In the present study as well, no statistically significant difference in brain edema was found between benign and high-grade meningiomas.
Some studies have reported that meningiomas with skull base locations were associated with a decreased risk of being high grade. 7,20We observed in our study that there was no significant difference in the grade between non-skull base meningiomas and skull base meningiomas.
We found that the mean ADC value and NADC ratio were lower (i.e.relatively restricted diffusion) in atypical/ malignant meningioma than in benign tumors (Figure 4).Two previous studies showed similar results. 1,8There are various explanations of this difference in literature like increased tumor cellularity, tumor matrices, fibrous or gliotic tissues, or a combination of these factors. 125][26][27][28][29][30] The absolute cut-off and reliability of ADC measurement is controversial, with different b values, areas of measurement (tumors peduncle, peripheral part of the tumor, and central region of the tumor), and methods of measurement (minimum ADC, mean ADC, maximum ADC, and normalized ADC) used in respective studies.
The measurement of absolute ADC values may vary across different scanners, and the NADC ratio may be a consistent parameter to use.The NADC ratio minimizes the differences in absolute ADC values that may be obtained with different DW imaging sequences and hardware configurations, thereby eliminating inter-image variability.This study also identified that meningiomas with lower ADC/NADC had significant association with breach in tumor brain interface, capsular enhancement, pial invasion, parenchymal invasion as well as heterogenous pattern of contrast enhancement.
In clinical practice, predicting meningiomas with a lower probability of advanced histopathological grade; selective resection balanced against the risk of a surgical procedure is recommended.Otherwise, more aggressive resection, and even dura substitution, should be considered for those with a higher probability of a high-grade meningioma.
Our study has some limitations.First, this is a retrospective study, and further prospective reports are needed to test the validity of our prediction model.Second, the patient population comes from a tertiary medical care center, and therefore the sample might not be representative of the entire population.Third, the description of imaging findings is somewhat subjective, with the possible existence of intra-observer and interobserver variability.Fourth and most importantly, the sampling bias of 7 patients with high-grade meningioma and 31 patients with benign meningioma would have influenced the probability calculation.

CONCLUSIONS
This study shows some important association between few MR morphologic characters with aggressiveness and grade of meningioma.High grade/atypical meningiomas had loss of capsular enhancement, breach in tumor brain interface, loss of CSF cleft and pial/parenchymal invasion.DWI findings also correlated with above mentioned MR parameters.Atypical/Grade II/III tumors had statistically significant lower ADC mean values than grade I meningiomas.
ADC mean can be helpful in preoperative distinction between benign and atypical/malignant meningioma.

Table 2 . Correlation of tumors grades with various tumors characters.
Free Full Text Articles are Available at www.jnma.com.np