scholarly journals Oligodendroglial tumor

2020 ◽  
Author(s):  
2009 ◽  
Vol 95 (3) ◽  
pp. 343-354 ◽  
Author(s):  
Rubén Ferrer-Luna ◽  
Manuel Mata ◽  
Lina Núñez ◽  
Jorge Calvar ◽  
Francisco Dasí ◽  
...  

2017 ◽  
Vol 133 (1) ◽  
pp. 173-181 ◽  
Author(s):  
Michael G. Brandel ◽  
Ali A. Alattar ◽  
Brian R. Hirshman ◽  
Xuezhi Dong ◽  
Kate T. Carroll ◽  
...  

2008 ◽  
Vol 22 (S1) ◽  
Author(s):  
Carrie A. Mohila ◽  
Nathan E. Simmons ◽  
Camilo E. Fadul ◽  
Louise P. Meyer ◽  
Alan C. Hartford ◽  
...  

2017 ◽  
Vol 35 (15_suppl) ◽  
pp. 2043-2043
Author(s):  
Jingwei Wei ◽  
Jie Tian

2043 Background: Oligodendroglial tumor (OT) is one of the main types of gliomas, which is incurable in current situation. As a tumor-specific genetic alteration of OTs, loss of heterozygosity (LOH) on chromosome 1p/19q indicates a preferable response to chemo/radiotherapy and a better survial, which could be used as a key molecular signature for personalized treatment decision making. However, 1p/19q LOH status is now commonly obtained by fluorescence in situ hybridization after the tumor resection, which is highly likely to lead functin deficit with tumor locating on eloquent area. Thus, a noninvasive predicition on 1p/19q LOH is required. Here, we used a "Radiomics" method to achieve the prediction based on magnetic resonance imaging in this study. Methods: A cohort of 113 OT patients was collected from Beijing Tiantan Hospital. 593 three-dimensional imaging features were extracted on T2-weighted images including textural and non-textural features. We used "minimum redundancy maximum relevance" and "iterative backward elemination and forward inclusion" algorithms to pick up the most effective features with a p-value < 0.05. With the selected features as the input, support vector machine algorithm was adopted to predict the 1p/19q LOH status with 10-fold cross validation. Comparisons were made between the traditional clinical predictors and the established model. Results: The prediction accuracy for 1p/19q LOH turned out to be 86.6%. The top three features contributing most to the prediction were respectively: GlobalUniformity, GaborBankD4GLRLMLGRE and GaborBankD4GLRLMSRLGE. The predictive performance of the radiomics model was proved to be far more valid than the clinical predictors (indistict tumor border and heterogeneous signal intensity) with the higher area under curve (AUC). Compared with the best single feature (GlobalUniformity, AUC: 0.749), this combined-feature model has the best diagnostic performance with an AUC of 0.898. Conclusions: This study reveals the intrinsic association between the imaging features and 1p/19q LOH status, meanwhile, realizes the high precision prediction, providing reliable basis for the pre-operative treatment regime.


2014 ◽  
Vol 16 (suppl 5) ◽  
pp. v163-v163
Author(s):  
T. Matsutani ◽  
S. Hirono ◽  
N. Shinozaki ◽  
Y. Iwadate ◽  
N. Saeki

Neurology ◽  
2006 ◽  
Vol 66 (2) ◽  
pp. 247-249 ◽  
Author(s):  
A. U. Ty ◽  
S. J. See ◽  
J. P. Rao ◽  
J.B.K. Khoo ◽  
M. C. Wong

2012 ◽  
Vol 34 (7) ◽  
pp. 1326-1333 ◽  
Author(s):  
S. Fellah ◽  
D. Caudal ◽  
A.M. De Paula ◽  
P. Dory-Lautrec ◽  
D. Figarella-Branger ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document