Abstract
179
Objectives Amino acid PET with 11C-methionine (MET) is well established in the diagnostic work-up of malignant brain tumors. Analysis of MET-PET data using tumor-to-normal ratios (TNR) has high sensitivity and specificity for diagnosing glioma, is useful for grading glioma, and can accurately define the tumor boundary. However, as TNR reflects only a small volume of the tumor, this parameter does not make use of volume-related information and is sensitive to image noise and the image reconstruction method employed. Recently, in the investigation of malignancies other than brain tumor, so-called texture features from 18F-fluorodeoxyglucose (FDG) PET have been extensively studied. Texture analysis can evaluate intratumoral metabolic heterogeneity, which may vary in relation to the malignant potential of glioma. We investigated whether measurement of MET uptake heterogeneity in patients with glioma using textural analysis could provide additional information on tumor grade compared to conventional TNR.
Methods A total of 46 brain tumor patients underwent MET PET with a Siemens Biograph 64 PET/CT scanner. The clinical characteristics of patients are summarized in Table 1. The TNR was defined as SUVmax of the lesion divided by the reference brain tissue (=contralateral frontal lobe). The volume of interest for textural analysis was delineated using a threshold of 1.3 times the reference. In addition to histogram analysis, 4 kinds of texture matrices (co-occurrence matrix, gray-level run length matrix, gray-level zone length matrix, and neighborhood gray-level different matrix) were generated to calculate a total of 36 texture parameters (Figure 1). We evaluated the relationship of these parameters with histopathologic grading and the presence of oligodendroglial components (OC).
Results Of the 36 texture features, 12 parameters reached significant difference between low-grade (grade II) vs. high grade gliomas (grade III and IV, Table 2), whereas only 6 texture features showed significant difference between grade III vs. IV gliomas (Table 3). The conventional TNR significantly differed between low-grade vs. high-grade gliomas (1.88±0.29 vs. 3.53±0.16, P<0.05), but did not differ between grade III vs. IV gliomas (3.43±0.26 vs. 3.61±0.23, P=NS). Furthermore, among the glioblastoma patients, SZLGE was significantly higher in glioblastoma without OC than in glioblastoma with OC (0.0670±0.0161 vs. 0.0523± 0.0067, P<0.05). The sensitivity, specificity, and accuracy were 100 %, 64.3 %, and 81.0 %, respectively.
Conclusions Both the texture features and TNR differences were more remarkable between grade II vs. III than between grade III vs. IV, suggesting that both the heterogeneity and the intensity of amino acid metabolism may change as the tumor acquires malignancies, in particular in the step of grade II to III. Plus, glioblastoma with vs. without OC could be distinguished using texture analysis, by which the texture analysis may have potential to predict patient prognosis because glioblastomas with OC have better prognosis than glioblastomas without OC.
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.