Format

Send to

Choose Destination
J Neurooncol. 2016 May;127(3):515-24. doi: 10.1007/s11060-016-2055-7. Epub 2016 Jan 11.

Differentiation between treatment-related changes and progressive disease in patients with high grade brain tumors using support vector machine classification based on DCE MRI.

Author information

1
Functional Brain Center, The Wohl Institute for Advanced Imaging, Tel Aviv Sourasky Medical Center, 6 Weizman St., 64239, Tel Aviv, Israel.
2
Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.
3
Department of Chemical Physics, Weizmann Institute, Rehovot, Israel.
4
Faculty of Engineering, Tel Aviv University, Tel Aviv, Israel.
5
Neuro-Oncology Service, Tel Aviv Sourasky Medical Center, Tel Aviv, Israel.
6
Functional Brain Center, The Wohl Institute for Advanced Imaging, Tel Aviv Sourasky Medical Center, 6 Weizman St., 64239, Tel Aviv, Israel. dafnab@tlvmc.gov.il.
7
Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. dafnab@tlvmc.gov.il.
8
Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel. dafnab@tlvmc.gov.il.

Abstract

Differentiation between treatment-related changes and progressive disease (PD) remains a major clinical challenge in the follow-up of patients with high grade brain tumors. The aim of this study was to differentiate between treatment-related changes and PD using dynamic contrast enhanced (DCE) MRI. Twenty patients were scanned using conventional, DCE-MRI and MR spectroscopy (total of 44 MR scans). The enhanced lesion area was extracted using independent components analysis of the DCE data. Pharmacokinetic parameters were estimated from the DCE data based on the Extended-Tofts-Model. Voxel based classification for treatment-related changes versus PD was performed in a patient-wise leave-one-out manner, using a support vector machine classifier. DCE parameters, K (trans), v e, k ep and v p, significantly differentiated between the tissue types. Classification results were validated using spectroscopy data showing significantly higher choline/creatine values in the extracted PD component compared to areas with treatment-related changes and normal appearing white matter, and high correlation between choline/creatine values and the percentage of the identified PD component within the lesion area (r = 0.77, p < 0.001). On the training data the sensitivity and specificity were 98 and 97 %, respectively, for the treatment-related changes component and 97 and 98 % for the PD component. This study proposes a methodology based on DCE-MRI to differentiate lesion areas into treatment-related changes versus PD, prospectively in each scan. Results may have major clinical importance for pre-operative planning, guidance for targeting biopsy, and early prediction of radiological outcomes in patients with high grade brain tumors.

KEYWORDS:

DCE-MRI; Disease progression; Support vector machine; Treatment-related changes

PMID:
26754857
DOI:
10.1007/s11060-016-2055-7
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center