Keywords:
Glioma, ITSS, rCBV, Perfusion MRI, Susceptibility
Financial Support And Sponsorship: Nil.
Conflicts of Interest:
There are no conflicts of interest.
BACKGROUND: Perfusion MRI is considered one of the best available methods in grading of gliomas and studies have shown it to be superior to conventional MRI sequences in this regard. Due to its high sensitivity to tumor microvasculature and hemorrhagic products, susceptibility-weighted imaging (SWI) is a promising imaging tool for evaluating angiogenesis, which in turn can help differentiate low grade from high grade gliomas.
AIMS: This study aims to compare the accuracy of SWI with Perfusion MRI in the grading of gliomas.
METHODS: In this prospective study on 40 patients with gliomas of different grades diagnosed based conventional MRI sequences (confirmed later on biopsy), SWI and perfusion MRI images were acquired by a 1.5 T MRI machine. rCBV (Relative cerebral blood volume) values were obtained by Perfusion MRI and ITSS (Intratumoral susceptibility signal) magnitude denoting the number of linear or dot like hypointensities within the tumor, was assessed from SWI images. Receiver operating characteristic (ROC) curve analysis was used to determine the individual diagnostic performances of rCBV and ITSS in grading of the gliomas. Reliability of both diagnostic tests was assessed by kappa statistic along with its 95% CI (confidence interval) and p Value.
RESULTS: High ITSS magnitude on SWI (5 and above) had sensitivity of 100% and specificity of 91.7% in predicting High grade gliomas. Positive and negative predictive values were 96.6% and 100% respectively. The overall diagnostic accuracy was 97.5% and the Kappa statistic was 0.94. Similarly a high rCBV value (Above 1.75) on perfusion MRI had sensitivity of 85.7% and specificity of 91.7% in predicting High grade gliomas. Positive and negative predictive values were 96% and 73.3% respectively. The overall diagnostic accuracy was 87.5% and the Kappa statistic was 0.72. p values for all the above results were <0.05.
CONCLUSIONS: The findings in our study show that non-invasive grading of gliomas based on their ITSS magnitude on SWI has a diagnostic accuracy comparable to rCBV values obtained by Perfusion MRI. These findings imply that visual assessment of the degree of intratumoral susceptibility signals in a glioma can predict its grade similar to Perfusion MRI, which has generally been considered the gold standard noninvasive method of grading gliomas.
INTRODUCTION:
Magnetic resonance imaging (MRI) is the gold standard in non-invasive diagnosis and grading of gliomas. Conventional MRI techniques may provide less than optimal information by assessing only the anatomic characteristics of the tumor. New complementary MRI techniques have been developed that provide quantitative parameters which provide functional or physiologic data, thereby improving their accuracy in grading these tumors[1].
Perfusion MRI is one such technique that has become a powerful tool by giving insight into angiogenesis, a critical component of gliomas of higher grade, thereby potentially helping to differentiate low grade from high grade gliomas. Studies have shown that it improves the accuracy in grading of gliomas and is generally considered superior to conventional MRI in this regard[2-4].
Due to its high sensitivity to tumor microvasculature and hemorrhagic products, susceptibility-weighted imaging (SWI) has recently been included as a potential imaging tool for evaluating angiogenesis and thereby help in grading gliomas. SWI has potential advantages in that it is a non-contrast technique and does not have some of the technical limitations and pitfalls in diagnostic interpretation present in perfusion MRI[5]. This study aims to compare the accuracy of SWI compared to perfusion MRI in grading gliomas.
MATERIALS AND METHODS:
Data collection:
A total of 40 patients with subsequently histopathologically proven gliomas of different grades were analysed from June 2017 to May 2018 by MRI. Perfusion MRI and SWI images were used to obtain rCBV and ITSS values respectively from each of the gliomas.
Imaging analysis:
All patients with suspected gliomas based on conventional MR imaging features were graded based on Perfusion MRI as low grade or high grade ( using an rCBV value of 1.75 as the cut off for differentiation, based on previous valid studies)[2,4] . The rCBV value was calculated from the site of maximal blood volume in the lesion, compared with a contralateral area of normal white matter. Quantitative assessment of intratumoral susceptibility signals (ITSS) was performed in the selected imaging slice that visually showed the highest frequency of ITSS within the tumor. This did not include hypointense lines and dots seen on the other conventional sequences, but only those hypointense foci seen only on the SWI sequence. Results were then be compared with the histopathological grade of the tumor obtained by surgery or stereotactic biopsy.
Statistical analysis:
Data was analyzed using SPSS 22.0. Comparison of continuous variables was assessed using independent sample t test. Susceptibility weighted imaging using an ITSS value and Perfusion MRI imaging using an rCBV value were compared with histopathology in distinguishing low grade (grade I and II) and high grade (Grade III and IV) gliomas . ROC curve analysis was used to define the cut off value for ITSS. The pre-fixed cut off value for rCBV based on previous studies as mentioned in the review of literature was 1.75[9]. Sensitivity, Specificity, Positive and Negative predictive values were used to interpret the accuracy of SWI and Perfusion MRI with respect to Histopathology. Reliability of both diagnostic tests was assessed by kappa statistic along with its 95% CI (confidence interval) and p Value. For all tests a p value of < 0.05 was considered as statistically significant.
RESULTS:
In the study of 40 patients, majority of the gliomas (57.5 %) were grade IV on histopathology (23 patients). Grade I gliomas accounted for 10% of the patients. Grade II and Grade III gliomas accounted for 20% and 12.5% of patients respectively. Grade I and II gliomas are classified as low grade gliomas and grade III and IV are classified as high grade gliomas by convention. By this classification, 70 % (28 cases) of the gliomas in this study were high grade and 30% (12 cases) were low grade.
The mean rCBV value (on perfusion MRI) of high grade gliomas was 3.61 ± 3.02 and low grade gliomas was 0.99 ± 0.78, and the mean difference (2.62) between the two groups was statistically significant (P value 0.006).
The rCBV value of the tumor had a good predictive validity in predicting histopathological high grade gliomas, as indicated by area under the curve of 0.921 (95% CI 0.812 to 1.00; p value <0.001).
Out of the histopathologically proved high grade gliomas, 85.7% were shown to be high grade on Perfusion MRI as well, using an rCBV cut off value of 1.75, giving it a sensitivity of 85.7%. Out of the histopathologically proved low grade gliomas, 91.7% of the lesions were low grade on Perfusion MRI as well, giving it a specificity of 91.7%. Positive predictive value was 96.0% and negative predictive
value was 73.3%; the overall diagnostic accuracy was 87.5%. The kappa statistic was 0.72 with a 95% CI of 0.5 to 0.95.
The mean ITSS magnitude on SWI of histopathological high grade gliomas was 16.25 ± 6.54 and low grade gliomas was 2.67 ± 2.84, and the mean difference (13.58) between the two groups was statistically significant (P value<0.001).
Based on the ROC curve analysis, an ITSS value of 5 was chosen as the cut off between high grade and low grade gliomas, as it was the point of maximum sensitivity and specificity.
In our study population, SWI showed better validity parameters than Perfusion MRI in predicting glioma grade, with the greatest difference being in the sensitivity and negative predictive value, in which SWI had values of 100% .The difference was however statistically insignificant with a p value of 0.1 and overlap in the 95 % confidence intervals of all the validity parameters between the two tests.
DISCUSSION
The present study was conducted with the objective of comparing the accuracy of Susceptibility-weighted Imaging (SWI) with Perfusion MRI in non-invasive grading of gliomas, using an optimal Intratumoral Susceptibility Signal (ITSS) cut-off value derived from the study.
The prospective study included 40 patients with histopathologically proved gliomas, of which 70% (28 in number) belonged to the high grade group and 30% (12 in number) belonged to the low grade group.
The accuracy of Perfusion MRI in predicting high grade gliomas was determined based on an rCBV threshold value of 1.75 (from previous studies); the accuracy of SWI was determined using an ITSS threshold value of 5, derived from the ROC curve analysis in this study.
The findings in our study show that grading of gliomas based on their ITSS magnitude on SWI has a diagnostic accuracy (sensitivity of 100 % and specificity of 91.7 %) comparable to that of Perfusion MRI based on their relative cerebral blood volume (rCBV) measurements (Sensitivity of 85.7% and specificity of 91.7%). Perfusion MRI has generally been considered the gold standard non-invasive method of grading gliomas and this study shows SWI to be an equally accurate alternative, the advantage being it does not require contrast administration and is free from software related technical pitfalls often present in perfusion MRI.
In our study, there was a statistically significant difference between the mean ITSS magnitude of high grade and low grade gliomas, with a mean value of 16.25 ± 6.54 for high grade gliomas and 2.67± 2.84 for low grade gliomas. This result is similar to previous studies by Park et al5], Zhang et al[6]contrast enhanced T1WI (CE-T1WI, Li et al[7]9 of which were grades I-II, and 13 were grades III-IV. All examinations were performed on Signa DEx 3.0 T MRI scanner. Conventional imaging techniques (T1WI, T2WI, T2FLAIR, CE-T1WI and Wang et al8]and there was also a significant correlation between rCBV and tumor grade (r= 0.77; P < 0.001. For example, the study by Li et al7]9 of which were grades I-II, and 13 were grades III-IV. All examinations were performed on Signa DEx 3.0 T MRI scanner. Conventional imaging techniques (T1WI, T2WI, T2FLAIR, CE-T1WI in 22 gliomas showed an ITSS magnitude of 17.7 ± 12.71 in high grade gliomas and 7.9 ± 7.62 in low grade gliomas with a statistically significant difference.
In our study, using an ITSS magnitude of 5 as the optimal threshold value based on the ROC curve analysis, the sensitivity and specificity in predicting high grade gliomas was found to be 100% and 91.7% respectively. Positive and negative predictive values were 96.6% and 100% respectively. These results are similar to those of the previous study by Park et al[5], who also used the value 5 as their ITSS threshold in a study population of 41 gliomas (29 high grade and 12 low grade) and obtained a sensitivity of 85.2%, specificity of 92.9%, and positive and negative predictive values of 95.8% and 76.5% respectively.
Similar to perfusion MRI, the role of SWI in grading gliomas is also based on evaluating the process of neoangiogenesis that occurs more in high grade gliomas. However, unlike perfusion MRI, HR-SWI evaluates the anatomical aspect of the vascularity by assessing the number of tiny new vessels within the tumor, with their accompanying microhaemorrhages because of the leakiness associated with these immature new vessels. These are significantly higher in high grade gliomas than those of low grade.
CONCLUSION
ITSS quantification on the SWI sequence has a diagnostic performance comparable to Perfusion MRI in the grading of gliomas and is easier to perform and interpret with fewer technical limitations and pitfalls compared to Perfusion MRI, with an added advantage being non-requirement of contrast administration.
REFERENCES:
1. Floriano VH, Torres US, Spotti AR, Ferraz-Filho JRL, Tognola WA. The Role of Dynamic Susceptibility Contrast-Enhanced Perfusion MR Imaging in Differentiating between Infectious and Neoplastic Focal Brain Lesions: Results from a Cohort of 100 Consecutive Patients. PLoS ONE [Internet] 2013 [cited 2018 Jun 20];8(12). Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3855761/
2. Law M, Yang S, Wang H, Babb JS, Johnson G, Cha S, et al. Glioma Grading: Sensitivity, Specificity, and Predictive Values of Perfusion MR Imaging and Proton MR Spectroscopic Imaging Compared with Conventional MR Imaging. Am J Neuroradiol 2003;24(10):1989–98.
3. Direksunthorn T, Chawalparit O, Sangruchi T, Witthiwej T, Tritrakarn S-O, Piyapittayanan S, et al. Diagnostic performance of perfusion MRI in differentiating low-grade and high-grade gliomas: advanced MRI in glioma, A Siriraj project. J Med Assoc Thail Chotmaihet Thangphaet 2013;96(9):1183–90.
4. Al-Okaili RN, Krejza J, Wang S, Woo JH, Melhem ER. Advanced MR Imaging Techniques in the Diagnosis of Intraaxial Brain Tumors in Adults. RadioGraphics 2006;26(suppl_1):S173–89.
5. Park MJ, Kim HS, Jahng G-H, Ryu C-W, Park SM, Kim SY. Semiquantitative Assessment of Intratumoral Susceptibility Signals Using Non-Contrast-Enhanced High-Field High-Resolution Susceptibility-Weighted Imaging in Patients with Gliomas: Comparison with MR Perfusion Imaging. Am J Neuroradiol 2009;30(7):1402–8.
6. Zhang H, Tan Y, Wang X-C, Qing J-B, Wang L, Wu X-F, et al. Susceptibility-weighted imaging: the value in cerebral astrocytomas grading. Neurol India 2013;61(4):389–95.
7. Li C, Ai B, Li Y, Qi H, Wu L. Susceptibility-weighted imaging in grading brain astrocytomas. Eur J Radiol 2010;75(1):e81-85.
8. Wang X, Zhang H, Tan Y, Qin J, Wu X, Wang L, et al. Combined value of susceptibility-weighted and perfusion-weighted imaging in assessing who grade for brain astrocytomas. J Magn Reson Imaging JMRI 2014;39(6):1569–74.
