Molecular and metabolic pattern classification for detection of brain glioma progression.

OBJECTIVES: The ability to differentiate between brain tumor progression and radiation therapy induced necrosis is critical for appropriate patient management. In order to improve the differential diagnosis, we combined fluorine-18 2-fluoro-deoxyglucose positron emission tomography ((18)F-FDG PET), proton magnetic resonance spectroscopy ((1)H MRS) and histological data to develop a multi-parametric machine-learning model. METHODS: We enrolled twelve post-therapy patients with grade 2 and 3 gliomas that were suspicious of tumor progression. All patients underwent (18)F-FDG PET and (1)H MRS. Maximal standardized uptake value (SUVmax) of the tumors and reference regions were obtained. Multiple 2D maps of choline (Cho), creatine (Cr), and N-acetylaspartate (NAA) of the tumors were generated. A support vector machine (SVM) learning model was established to take imaging biomarkers and histological data as input vectors. A combination of clinical follow-up and multiple sequential MRI studies served as the basis for assessing the clinical outcome. All vector combinations were evaluated for diagnostic accuracy and cross validation. The optimal cutoff value of individual parameters was calculated using Receiver operating characteristic (ROC) plots. RESULTS: The SVM and ROC analyses both demonstrated that SUVmax of the lesion was the most significant single diagnostic parameter (75% accuracy) followed by Cho concentration (67% accuracy). SVM analysis of all paired parameters showed SUVmax and Cho concentration in combination could achieve 83% accuracy. SUVmax of the lesion paired with SUVmax of the white matter as well as the tumor Cho paired with the tumor Cr both showed 83% accuracy. These were the most significant paired diagnostic parameters of either modality. Combining all four parameters did not improve the results. However, addition of two more parameters, Cho and Cr of brain parenchyma contralateral to the tumor, increased the accuracy to 92%. CONCLUSION: This study suggests that SVM models may improve detection of glioma progression more accurately than single parametric imaging methods. RESEARCH SUPPORT: National Cancer Institute, Cancer Center Support Grant Supplement Award, Imaging Response Assessment Teams.

Combined imaging biomarkers for therapy evaluation in glioblastoma multiforme: correlating sodium MRI and F-18 FLT PET on a voxel-wise basis.

We evaluate novel magnetic resonance imaging (MRI) and positron emission tomography (PET) quantitative imaging biomarkers and associated multimodality, serial-time-point analysis methodologies, with the ultimate aim of providing clinically feasible, predictive measures for early assessment of response to cancer therapy. A focus of this work is method development and an investigation of the relationship between the information content of the two modalities. Imaging studies were conducted on subjects who were enrolled in glioblastoma multiforme (GBM) therapeutic clinical trials. Data were acquired, analyzed and displayed using methods that could be adapted for clinical use. Subjects underwent dynamic [(18)F]fluorothymidine (F-18 FLT) PET, sodium ((23)Na) MRI and 3-T structural MRI scans at baseline (before initiation of therapy), at an early time point after beginning therapy and at a late follow-up time point after therapy. Sodium MRI and F-18 FLT PET images were registered to the structural MRI. F-18 FLT PET tracer distribution volumes and sodium MRI concentrations were calculated on a voxel-wise basis to address the heterogeneity of tumor physiology. Changes in, and differences between, these quantities as a function of scan timing were tracked. While both modalities independently show a change in tissue status as a function of scan time point, results illustrate that the two modalities may provide complementary information regarding tumor progression and response. Additionally, tumor status changes were found to vary in different regions of tumor. The degree to which these methods are useful for GBM therapy response assessment and particularly for differentiating true progression from pseudoprogression requires additional patient data and correlation of these imaging biomarker changes with clinical outcome.

Comparison of proton magnetic resonance spectroscopy with fluorine-18 2-fluoro-deoxyglucose positron emission tomography for assessment of brain tumor progression.

OBJECTIVES: We investigated the accuracy of high-field proton magnetic resonance spectroscopy ((1) H MRS) and fluorine-18 2-fluoro-deoxyglucose positron emission tomography ((18) F-FDG-PET) for diagnosis of glioma progression following tumor resection, stereotactic radiation, and chemotherapy. METHODS: Twelve post-therapy patients with histology proven gliomas (six grade II and six grade III) presented with magnetic resonance imaging (MRI) and clinical symptoms suggestive but not conclusive of progression were entered into the study. (1) H MRS data were acquired and 3-dimensional volumetric maps of choline (Cho) over creatine (Cr) were generated. Intensity of (18) F-FDG uptake was evaluated on a semiquantitative scale. RESULTS: The accuracy of (1) H MRS and (18) F-FDG-PET imaging for diagnosis of glioma progression was 75% and 83%, respectively. Classifying the tumors by grade improved accuracy of (18) F-FDG-PET to 100% in high-grade gliomas and accuracy of (1) H MRS to 80% in low-grade tumors. Spearman's analysis demonstrated a trend between (18) F-FDG uptake and tumor grading (ρ= .612, P-value = .272). The results of (18) F-FDG-PET and (1) H MRS were concordant in 75% (9/12) of cases. CONCLUSION: The combination of (1) H MRS data and (18) F-FDG-PET imaging can enhance detection of glioma progression. (1) H MRS imaging was more accurate in low-grade gliomas and (18) F-FDG-PET provided better accuracy in high-grade gliomas.

PET imaging of brain macrophages using the peripheral benzodiazepine receptor in a macaque model of neuroAIDS.

HIV infection in humans and simian immunodeficiency virus (SIV) infection in macaques result in encephalitis in approximately one-quarter of infected individuals and is characterized by infiltration of the brain with infected and activated macrophages. 1-(2-chlorphenyl)-N-methyl-N-(1-methylpropyl)-3-isoquinoline-carboxamide (PK11195) is a ligand specific for the peripheral benzodiazepine receptor abundant on macrophages and is expressed in low levels in the noninfected brain. We hypothesized that positron-emission tomography (PET) with the carbon-11-labeled, R-enantiomer form of PK11195 ([(11)C](R)-PK11195) could image brain macrophages and hence the development of encephalitis in vivo. [(11)C](R)-PK11195 binding was assessed in the brain using PET in 11 SIV infected macaques, six of which showed increased binding in vivo. Postmortem examination of the brain in these six macaques demonstrated encephalitis, while macaques that did not show an increase in [(11)C](R)-PK11195 binding did not develop SIV encephalitis. Brain tissue from SIV encephalitic macaques also showed increased [(3)H](R)-PK11195 binding compared with binding in nonencephalitic macaques. Increased PK11195 binding in vivo and in postmortem brain tissue correlated with abundance of macrophages but not astrocytes. Our results suggest that PET [(11)C](R)-PK11195 imaging can detect the presence of macrophages in SIV encephalitis in vivo and may be useful to predict the development of HIV encephalitis and in studies of the pathogenesis and treatment of HIV dementia.