Skip to main content
Fig. 2 | BMC Medicine

Fig. 2

From: Multimodal-based machine learning strategy for accurate and non-invasive prediction of intramedullary glioma grade and mutation status of molecular markers: a retrospective study

Fig. 2

Illustration of the study process. Stage I includes raw image acquisition (a), manual ROI segmentation (b), and auto ROI segmentation (c). Stage II includes feature extraction and selection. From both SAG and TRA images, radiomic features, including first-order statistical, shape, texture, and wavelet features, are extracted (d). All extracted features are screened out by ICC to select stable features (e). Informative features are then selected using LASSO (f). Stage III includes model construction and validation. Selected clinical and radiomic features are entered into the deep neural networks to predict the different tasks (g), and model performance is further tested in the external validation cohort (h). ATRX, alpha thalassemia/mental retardation syndrome X-linked; ICC, intraclass correlation coefficient; LASSO, least absolute shrinkage and selection operator; P53, tumor protein p53; ROI, region of interest; SAG, sagittal; TRA, transverse

Back to article page

BMC Medicine

ISSN: 1741-7015

Contact us