Abstract #1847
Logistic regression of multiparametric MR for glioma grading
Lawrence Kenning 1 , Martin Lowry 2 , Martin Pickles 2 , Christopher Roland-Hill 3 , Shailendra Achawal 3 , Chittoor Rajaraman 3 , and Lindsay W Turnbull 2,3
1
Centre for Magnetic Resonance
Investigations, University of Hull, Hull, East Riding of
Yorkshire, United Kingdom,
2
Centre
for Magnetic Resonance Investigations, Hull York Medical
School at the University of Hull, Hull, East Riding of
Yorkshire, United Kingdom,
3
Hull
and East Yorkshire Hospitals NHS Trust, East Riding of
Yorkshire, United Kingdom
We investigated the role of functional MR parameters to
determine glioma grade using logistic regression models.
DTI, DCE and DSC were acquired from 55 glioma patients
and ADC, FA, q, RA, λL, λR, R1, Ktrans, ve, vb, rCBVGVF,
rCBVBOX, and K2 were calculated. Tumour volumes of
interest to sample the data were contoured using
morphological imaging. Using multi-parametric MR, 82.8%
of cases were correctly classified using a two part
logistic regression model decision tree. DTI and DCE
appeared to be the most useful for determining lesion
grade in this cohort of patients.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
