English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Preoperative magnetic resonance imaging (MRI) is a critical modality for the determination of glioblastoma (GBM) treatment strategy, as it is thought to reflect the biology of the tumor to some extent. The authors attempted to predict prognosis of newly diagnosed GBM (nGBM) using machine learning-based texture analysis of preoperative MRI in this study. METHOD A total of 160 nGBMs with determined overall survival were collected from Kansai Molecular Diagnosis Network for CNS tumors. Preoperative MRI scans (T1WI, T2WI, and Gd-T1WI) from all cases were semi-quantitatively analyzed leading to acquisition of 489 texture features as explanatory variables using Matlab-based in-house software. Dichotomous overall survival (OS) with a cutoff of 15 months was regarded as the response variable (short or long OS). Lasso regression was employed for feature selection to ensure robustness of the prediction model. One hundred patients were randomly assigned as training dataset (TR), followed by predictive model construction via 5-fold cross-validation. Subsequently, the constructed model was transferred to the remaining 60 patients, which was assigned as test dataset (TD). The survival distribution between populations with predicted short and long OS was compared using log-rank test. RESULTS Distributions of the analyzed data were as follows; 53 short OS cases in the TR (53.0%) and 27 cases in the TD (45.0%). As for the result of transfer analysis in TD, 38 cases out of 60 (63.3%) were predicted to be short OS (76.3 % of recall, 54.3% of precision, and 63.5% of F-measure). The population of predicted short OS significantly showed poorer prognosis (median OS 14.0 vs 19.1 months) (p=0.02, log-rank test). CONCLUSION Short OS was successfully identified from preoperative MRI with high recall rates with our algorithm. The presented result ensures the potential of machine learning-based texture analysis for prognostic stratification of nGBM.

NI-13 PREDICTION OF PROGNOSIS IN NEWLY DIAGNOSED GLIOBLASTOMA USING MACHINE LEARNING-BASED TEXTURE ANALYSIS OF PREOPERATIVE MRI