Radiomic Based Machine Learning Performance for a Three Class Problem in Neuro-Oncology: Time to Test the Waters?

Sarv Priya; Yanan Liu; Caitlin Ward; Nam H Le; Neetu Soni; Ravishankar Pillenahalli Maheshwarappa; Varun Monga; Honghai Zhang; Milan Sonka; Girish Bathla

doi:10.3390/cancers13112568

Back

Radiomic Based Machine Learning Performance for a Three Class Problem in Neuro-Oncology: Time to Test the Waters?

Journal article

Open access

Peer reviewed

Radiomic Based Machine Learning Performance for a Three Class Problem in Neuro-Oncology: Time to Test the Waters?

Sarv Priya, Yanan Liu, Caitlin Ward, Nam H Le, Neetu Soni, Ravishankar Pillenahalli Maheshwarappa, Varun Monga, Honghai Zhang, Milan Sonka and Girish Bathla

Cancers, Vol.13(11), p.2568

05/24/2021

DOI: 10.3390/cancers13112568

PMID: 34073840

Files and links (1)

url

https://doi.org/10.3390/cancers13112568View

Published (Version of record) Open Access

Abstract

Prior radiomics studies have focused on two-class brain tumor classification, which limits generalizability. The performance of radiomics in differentiating the three most common malignant brain tumors (glioblastoma (GBM), primary central nervous system lymphoma (PCNSL), and metastatic disease) is assessed; factors affecting the model performance and usefulness of a single sequence versus multiparametric MRI (MP-MRI) remain largely unaddressed. This retrospective study included 253 patients (120 metastatic (lung and brain), 40 PCNSL, and 93 GBM). Radiomic features were extracted for whole a tumor mask (enhancing plus necrotic) and an edema mask (first pipeline), as well as for separate enhancing and necrotic and edema masks (second pipeline). Model performance was evaluated using MP-MRI, individual sequences, and the T1 contrast enhanced (T1-CE) sequence without the edema mask across 45 model/feature selection combinations. The second pipeline showed significantly high performance across all combinations (Brier score: 0.311–0.325). GBRM fit using the full feature set from the T1-CE sequence was the best model. The majority of the top models were built using a full feature set and inbuilt feature selection. No significant difference was seen between the top-performing models for MP-MRI (AUC 0.910) and T1-CE sequence with (AUC 0.908) and without edema masks (AUC 0.894). T1-CE is the single best sequence with comparable performance to that of multiparametric MRI (MP-MRI). Model performance varies based on tumor subregion and the combination of model/feature selection methods.

Machine Learning

CNS lymphoma

glioblastoma

metastases

MRI

radiomics

texture

Details

Title: Subtitle: Radiomic Based Machine Learning Performance for a Three Class Problem in Neuro-Oncology: Time to Test the Waters?
Creators: Sarv Priya - University of Iowa Hospitals and Clinics
Yanan Liu - University of Iowa
Caitlin Ward - University of Iowa
Nam H Le - University of Iowa
Neetu Soni - University of Iowa Hospitals and Clinics
Ravishankar Pillenahalli Maheshwarappa - University of Iowa Hospitals and Clinics
Varun Monga - University of Iowa Hospitals and Clinics
Honghai Zhang - University of Iowa
Milan Sonka - University of Iowa
Girish Bathla - University of Iowa Hospitals and Clinics
Resource Type: Journal article
Publication Details: Cancers, Vol.13(11), p.2568
DOI: 10.3390/cancers13112568
PMID: 34073840
NLM abbreviation: Cancers (Basel)
ISSN: 2072-6694
eISSN: 2072-6694
Publisher: MDPI
Language: English
Date published: 05/24/2021
Academic Unit: Roy J. Carver Department of Biomedical Engineering; Radiology; Electrical and Computer Engineering; Hematology, Oncology, and Blood & Marrow Transplantation; Radiation Oncology; Nursing; Fraternal Order of Eagles Diabetes Research Center; Injury Prevention Research Center; Internal Medicine; Ophthalmology and Visual Sciences
Record Identifier: 9984187241202771

Metrics

26 Record Views

22 Times Cited - Web of Science