Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Nov 23:11:601425.
doi: 10.3389/fonc.2021.601425. eCollection 2021.

AI and High-Grade Glioma for Diagnosis and Outcome Prediction: Do All Machine Learning Models Perform Equally Well?

Luca Pasquini  1   2 , Antonio Napolitano  3 , Martina Lucignani  3 , Emanuela Tagliente  3 , Francesco Dellepiane  2 , Maria Camilla Rossi-Espagnet  2   4 , Matteo Ritrovato  5 , Antonello Vidiri  6 , Veronica Villani  7 , Giulio Ranazzi  8 , Antonella Stoppacciaro  8 , Andrea Romano  2 , Alberto Di Napoli  2   9 , Alessandro Bozzao  2
Affiliations

AI and High-Grade Glioma for Diagnosis and Outcome Prediction: Do All Machine Learning Models Perform Equally Well?

Luca Pasquini et al. Front Oncol. .

Abstract

Radiomic models outperform clinical data for outcome prediction in high-grade gliomas (HGG). However, lack of parameter standardization limits clinical applications. Many machine learning (ML) radiomic models employ single classifiers rather than ensemble learning, which is known to boost performance, and comparative analyses are lacking in the literature. We aimed to compare ML classifiers to predict clinically relevant tasks for HGG: overall survival (OS), isocitrate dehydrogenase (IDH) mutation, O-6-methylguanine-DNA-methyltransferase (MGMT) promoter methylation, epidermal growth factor receptor vIII (EGFR) amplification, and Ki-67 expression, based on radiomic features from conventional and advanced magnetic resonance imaging (MRI). Our objective was to identify the best algorithm for each task. One hundred fifty-six adult patients with pathologic diagnosis of HGG were included. Three tumoral regions were manually segmented: contrast-enhancing tumor, necrosis, and non-enhancing tumor. Radiomic features were extracted with a custom version of Pyradiomics and selected through Boruta algorithm. A Grid Search algorithm was applied when computing ten times K-fold cross-validation (K=10) to get the highest mean and lowest spread of accuracy. Model performance was assessed as AUC-ROC curve mean values with 95% confidence intervals (CI). Extreme Gradient Boosting (xGB) obtained highest accuracy for OS (74,5%), Adaboost (AB) for IDH mutation (87.5%), MGMT methylation (70,8%), Ki-67 expression (86%), and EGFR amplification (81%). Ensemble classifiers showed the best performance across tasks. High-scoring radiomic features shed light on possible correlations between MRI and tumor histology.

Keywords: genetics; glioblastoma; high-grade glioma (HGG); machine learning; radiomics; survival.

PubMed Disclaimer

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling editor declared a shared affiliation with several of the authors, LP, FD, MCR-E, GR, AS, AR, ADN, AB, at time of review.

Figures

Figure 1
Figure 1
Radiomic workflow followed in the present study.
Figure 2
Figure 2
Machine learning classifiers tested in the present study. Non-ensemble learners included KNeighbors, logistic regressor, and decision tree. Ensemble learners included boosting, stacking, and bagging classifiers.
Figure 3
Figure 3
Best ROC curves for Surv12 prediction: (A) AB classifier with ADC sequence on NET ROI; (B) xGB classifier with T2 sequence on NEC ROI; (C) xGB classifier with FLAIR sequence on NET ROI.
Figure 4
Figure 4
Best ROC curve for MGMT prediction: AB classifier with FLAIR sequence on CET ROI.
Figure 5
Figure 5
Best ROC curves for IDH prediction: (A) AB classifier with rCBV sequence on NET ROI; (B) AB classifier with T2 sequence on CET ROI; (C) AB classifier with T2 sequence on NEC ROI; (D) ST classifier with T1 sequence on NET ROI.
Figure 6
Figure 6
Best ROC curve for KI67 prediction: AB classifier with ADC sequence on CET ROI.
Figure 7
Figure 7
Best ROC curves for EGFR prediction: (A) AB classifier with rCBV sequence on CET ROI; (B) AB classifier with T2 sequence on CET ROI.
See this image and copyright information in PMC

References

    1. Tamimi AF, Juweid M. Epidemiology and Outcome of Glioblastoma. In: de Vleeschouwer S, editor. Glioblastoma. Brisbane (AU: ): Codon Publications; (2017). doi: 10.15586/codon.glioblastoma.2017.ch8 - DOI - PubMed
    1. Molinaro AM, Taylor JW, Wiencke JK, Wrensch MR. Genetic and Molecular Epidemiology of Adult Diffuse Glioma. Nat Rev Neurol (2019) 15:405–17. doi: 10.1038/s41582-019-0220-2 - DOI - PMC - PubMed
    1. Braunstein S, Raleigh D, Bindra R, Mueller S, Haas-Kogan D. Pediatric High-Grade Glioma: Current Molecular Landscape and Therapeutic Approaches. J Neuro Oncol (2017) 134:541–9. doi: 10.1007/s11060-017-2393-0 - DOI - PubMed
    1. Wang J, Bettegowda C. Genomic Discoveries in Adult Astrocytoma. Curr Opin Genet Dev (2015) 30:17–24. doi: 10.1016/j.gde.2014.12.002 - DOI - PMC - PubMed
    1. Louis DN, Wesseling P, Aldape K, Brat DJ, Capper D, Cree IA, et al. cIMPACT-NOW Update 6: New Entity and Diagnostic Principle Recommendations of the cIMPACT-Utrecht Meeting on Future CNS Tumor Classification and Grading. Brain Pathol (2020) 30:844–56. doi: 10.1111/bpa.12832 - DOI - PMC - PubMed