Browsing by Subject "IDH mutation"

Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    Predicting Isocitrate Dehydrogenase 1 (IDH1) Mutation in Patients with Gliomas using a Novel Deep Learning Network
    (2019) Xiao, Haonan

    IDH is a gene that heavily affects the treatment response of gliomas and is associated with patient’s prognosis. Convolutional neural networks (CNNs) showed great potential in predicting IDH mutations. However, these CNN models require time-intensive image-preprocessing before being used for predictions. There are two main purpose of this study. The first purpose is to investigate the feasibility of applying a novel convolutional neural network based on the Inception-ResNet to reduce image preprocessing steps and improve accuracy in IDH mutation prediction. The second purpose is to evaluate different data augmentation methods on brain studies.

    MR images of 103 patients were selected from The Cancer Imaging Archive (TCIA). Intensity normalization of every individual slice is the only image pre-processing step. The T1w post-contrast, FLAIR, and T2w images at the same slice location were grouped together and considered as one training sample. This give rise to 209 IDH-mutant samples from 42 patients and 356 IDH-wild-type samples from 61 patients that were randomly selected to become training, validation, and test sets. To avoid overfitting in the model performance, data augmentation methods were applied individually to both training and validation sets in each training. The augmentation methods included duplication, noise addition, rotation, translation, cropping and mirroring. Images from one sample were fed to different input channels of Inception-ResNet, and the predictions were based on the extracted features and the patient’s age at diagnosis. Prediction accuracy was used to assess the performance of different augmentation methods.

    With only intensity normalization, the proposed model using training sets augmented by rotation and noise addition achieved the IDH prediction accuracies of 91.8% and 91.8%, respectively. On the same training, validation and test sets, the proposed model trained on data augmented by duplication, cropping, translation, and mirroring gave accuracies of 81.6%, 79.6%, 83.7%, and 85.7%, respectively.

    This work investigated the feasibility of the application of the novel convolutional neural network based on the Inception-ResNet on IDH mutation prediction, and high accuracies can be achieved with only intensity normalization as image preprocessing. Among all data augmentation methods, noise addition and rotation shows better performance and might suggest potential value for other clinical applications using machine learning algorithms.

  • Thumbnail Image
    ItemOpen Access
    Radiolabeled inhibitors as probes for imaging mutant IDH1 expression in gliomas: Synthesis and preliminary evaluation of labeled butyl-phenyl sulfonamide analogs.
    (Eur J Med Chem, 2016-08-25) Chitneni, Satish K; Reitman, Zachary J; Gooden, David M; Yan, Hai; Zalutsky, Michael R
    INTRODUCTION: Malignant gliomas frequently harbor mutations in the isocitrate dehydrogenase 1 (IDH1) gene. Studies suggest that IDH mutation contributes to tumor pathogenesis through mechanisms that are mediated by the neomorphic metabolite of the mutant IDH1 enzyme, 2-hydroxyglutarate (2-HG). The aim of this work was to synthesize and evaluate radiolabeled compounds that bind to the mutant IDH1 enzyme with the goal of enabling noninvasive imaging of mutant IDH1 expression in gliomas by positron emission tomography (PET). METHODS: A small library of nonradioactive analogs were designed and synthesized based on the chemical structure of reported butyl-phenyl sulfonamide inhibitors of mutant IDH1. Enzyme inhibition assays were conducted using purified mutant IDH1 enzyme, IDH1-R132H, to determine the IC50 and the maximal inhibitory efficiency of the synthesized compounds. Selected compounds, 1 and 4, were labeled with radioiodine ((125)I) and/or (18)F using bromo- and phenol precursors, respectively. In vivo behavior of the labeled inhibitors was studied by conducting tissue distribution studies with [(125)I]1 in normal mice. Cell uptake studies were conducted using an isogenic astrocytoma cell line that carried a native IDH1-R132H mutation to evaluate the potential uptake of the labeled inhibitors in IDH1-mutated tumor cells. RESULTS: Enzyme inhibition assays showed good inhibitory potency for compounds that have iodine or a fluoroethoxy substituent at the ortho position of the phenyl ring in compounds 1 and 4 with IC50 values of 1.7 μM and 2.3 μM, respectively. Compounds 1 and 4 inhibited mutant IDH1 activity and decreased the production of 2-HG in an IDH1-mutated astrocytoma cell line. Radiolabeling of 1 and 4 was achieved with an average radiochemical yield of 56.6 ± 20.1% for [(125)I]1 (n = 4) and 67.5 ± 6.6% for [(18)F]4 (n = 3). [(125)I]1 exhibited favorable biodistribution characteristics in normal mice, with rapid clearance from the blood and elimination via the hepatobiliary system by 4 h after injection. The uptake of [(125)I]1 in tumor cells positive for IDH1-R132H was significantly higher compared to isogenic WT-IDH1 controls, with a maximal uptake ratio of 1.67 at 3 h post injection. Co-incubation of the labeled inhibitors with the corresponding nonradioactive analogs, and decreasing the normal concentrations of FBS (10%) in the incubation media substantially increased the uptake of the labeled inhibitors in both the IDH1-mutant and WT-IDH1 tumor cell lines, suggesting significant non-specific binding of the synthesized labeled butyl-phenyl sulfonamide inhibitors. CONCLUSIONS: These data demonstrate the feasibility of developing radiolabeled probes for the mutant IDH1 enzyme based on enzyme inhibitors. Further optimization of the labeled inhibitors by modifying the chemical structure to decrease the lipophilicity and to increase potency may yield compounds with improved characteristics as probes for imaging mutant IDH1 expression in tumors.