Meeting Banner
Abstract #3071

Grading of glioma using a machine learning framework based on optimized features obtained from quantitative DCE-MRI and SWI

Banasmita Kar1, Anirban Sengupta1, Rupsa Bhattacharjee1,2, Neha Vats1, Virendra Yadav1, Dinil Sasi1, Rakesh Kumar Gupta3, and Anup Singh1,4

1Center for Biomedical Engineering, Indian Institute of Technology, Delhi, NEW DELHI, India, 2Philips Health Systems, Philips India Limited, Gurugram, India, 3Department of Radiology, Fortis Memorial Research Institute, Gurugram, India, 4Department of Biomedical Engineering, All India Institute of Medical Science, New Delhi, India, New Delhi, India

Potential of quantitative dynamic-contrast-enhanced(DCE) MRI parameters in glioma is well reported. However, in some glioma cases, biological behavior of tumors overlap between grades, therefore, correct grading becomes necessary for true classification and further treatment planning. In such cases histopathological glioma grading doesn’t necessarily correlate with DCE-MRI parameters based grading. Objective of this study is to improve the accuracy of grading of glioma using multi-parametric analysis i.e. combining Intra-Tumoral-Susceptibility-Signal(ITSS) volume generated from SWI with DCE-MRI parameters. Using a supervised-machine-learning based approach glioma grading can be improved particularly for the cases where DCE-MRI parameters underperform or vice-versa.

This abstract and the presentation materials are available to members only; a login is required.

Join Here