Radiomics for glioblastoma survival analysis in pre-operative MRI: exploring feature robustness, class boundaries, and machine learning techniques

Suter, Yannick; Knecht, Urspeter; Alao, Mariana; Valenzuela, Waldo; Hewer, Ekkehard; Schucht, Philippe; Wiest, Roland; Reyes, Mauricio (2020). Radiomics for glioblastoma survival analysis in pre-operative MRI: exploring feature robustness, class boundaries, and machine learning techniques. Cancer imaging : the official publication of the International Cancer Imaging Society, 20(55), pp. 1-13. BioMed Central 10.1186/s40644-020-00329-8

[img]
Preview
Text
Suter2020_Article_RadiomicsForGlioblastomaSurviv.pdf - Published Version
Available under License Creative Commons: Attribution (CC-BY).

Download (2MB) | Preview

Background:
This study aims to identify robust radiomic features for Magnetic Resonance Imaging (MRI), assess feature selection and machine learning methods for overall survival classification of Glioblastoma multiforme patients, and to robustify models trained on single-center data when applied to multi-center data.

Methods:
Tumor regions were automatically segmented on MRI data, and 8327 radiomic features extracted from these regions. Single-center data was perturbed to assess radiomic feature robustness, with over 16 million tests of typical perturbations. Robust features were selected based on the Intraclass Correlation Coefficient to measure agreement across perturbations. Feature selectors and machine learning methods were compared to classify overall survival. Models trained on single-center data (63 patients) were tested on multi-center data (76 patients). Priors using feature robustness and clinical knowledge were evaluated.

Results:
We observed a very large performance drop when applying models trained on single-center on unseen multi-center data, e.g. a decrease of the area under the receiver operating curve (AUC) of 0.56 for the overall survival classification boundary at 1 year. By using robust features alongside priors for two overall survival classes, the AUC drop could be reduced by 21.2%. In contrast, sensitivity was 12.19% lower when applying a prior.

Conclusions:
Our experiments show that it is possible to attain improved levels of robustness and accuracy when models need to be applied to unseen multi-center data. The performance on multi-center data of models trained on single-center data can be increased by using robust features and introducing prior knowledge. For successful model robustification, tailoring perturbations for robustness testing to the target dataset is key.

Item Type:

Journal Article (Original Article)

Division/Institute:

10 Strategic Research Centers > ARTORG Center for Biomedical Engineering Research > ARTORG Center - Musculoskeletal Biomechanics
04 Faculty of Medicine > Service Sector > Institute of Pathology
04 Faculty of Medicine > Department of Radiology, Neuroradiology and Nuclear Medicine (DRNN) > Institute of Diagnostic and Interventional Neuroradiology
04 Faculty of Medicine > Department of Head Organs and Neurology (DKNS) > Clinic of Neurosurgery
10 Strategic Research Centers > ARTORG Center for Biomedical Engineering Research

Graduate School:

Graduate School for Cellular and Biomedical Sciences (GCB)

UniBE Contributor:

Suter, Yannick Raphael; Knecht, Urspeter; Alao, Mariana; Valenzuela, Waldo Enrique; Hewer, Ekkehard; Schucht, Philippe; Wiest, Roland and Reyes, Mauricio

Subjects:

500 Science > 570 Life sciences; biology
600 Technology > 610 Medicine & health
000 Computer science, knowledge & systems
600 Technology > 620 Engineering

ISSN:

1470-7330

Publisher:

BioMed Central

Funders:

[UNSPECIFIED] Krebsliga Schweiz

Language:

English

Submitter:

Yannick Raphael Suter

Date Deposited:

13 Aug 2020 14:59

Last Modified:

13 Aug 2020 14:59

Publisher DOI:

10.1186/s40644-020-00329-8

PubMed ID:

32758279

Uncontrolled Keywords:

Glioblastoma multiforme, MRI radiomics, Overall survival classification, Multi-center, Robustness

BORIS DOI:

10.7892/boris.145825

URI:

https://boris.unibe.ch/id/eprint/145825

Actions (login required)

Edit item Edit item
Provide Feedback