We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Deep radiomic features from MRI scans predict survival outcome of recurrent glioblastoma

Abstract: This paper proposes to use deep radiomic features (DRFs) from a convolutional neural network (CNN) to model fine-grained texture signatures in the radiomic analysis of recurrent glioblastoma (rGBM). We use DRFs to predict survival of rGBM patients with preoperative T1-weighted post-contrast MR images (n=100). DRFs are extracted from regions of interest labelled by a radiation oncologist and used to compare between short-term and long-term survival patient groups. Random forest (RF) classification is employed to predict survival outcome (i.e., short or long survival), as well as to identify highly group-informative descriptors. Classification using DRFs results in an area under the ROC curve (AUC) of 89.15% (p<0.01) in predicting rGBM patient survival, compared to 78.07% (p<0.01) when using standard radiomic features (SRF). These results indicate the potential of DRFs as a prognostic marker for patients with rGBM.
Comments: Accepted in MICCAI RNO workshop
Subjects: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV); Quantitative Methods (q-bio.QM)
Cite as: arXiv:1911.06687 [cs.CV]
  (or arXiv:1911.06687v1 [cs.CV] for this version)

Submission history

From: Ahmad Chaddad [view email]
[v1] Fri, 15 Nov 2019 15:18:38 GMT (1412kb)

Link back to: arXiv, form interface, contact.