Survival prediction in glioblastoma on post-contrast magnetic resonance imaging using filtration based first-order texture analysis: Comparison of multiple machine learning models

Abstract

Objective Magnetic resonance texture analysis (MRTA) is a relatively new technique that can be a valuable addition to clinical and imaging parameters in predicting prognosis. In the present study, we investigated the efficacy of MRTA for glioblastoma survival using T1 contrast-enhanced (CE) images for texture analysis. Methods We evaluated the diagnostic performance of multiple machine learning models based on first-order histogram statistical parameters derived from T1-weighted CE images in the survival stratification of glioblastoma multiforme (GBM). Retrospective evaluation of 85 patients with GBM was performed. Thirty-six first-order texture parameters at six spatial scale filters (SSF) were extracted on the T1 CE axial images for the whole tumor using commercially available research software. Several machine learning classification models (in four broad categories: linear, penalized linear, non-linear, and ensemble classifiers) were evaluated to assess the survival prediction performance using optimal features. Principal component analysis was used prior to fitting the linear classifiers in order to reduce the dimensionality of the feature inputs. Fivefold cross-validation was used to partition the data iteratively into training and testing sets. The area under the receiver operating characteristic curve (AUC) was used to assess the diagnostic performance. Results The neural network model was the highest performing model with the highest observed AUC (0.811) and cross-validated AUC (0.71). The most important variable was the age at diagnosis, with mean and mean of positive pixels (MPP) for SSF = 0 being the second and third most important, followed by skewness for SSF = 0 and SSF = 4. Conclusions First-order texture features, when combined with age at presentation, show good accuracy in predicting GBM survival.

Predicting the Initial Treatment Response to Transarterial Chemoembolization in Intermediate-Stage Hepatocellular Carcinoma by the Integration of Radiomics and Deep Learning

“…However, the development of a precise model for predicting the initial response to TACE therapy is desired, and radiomics is a promising method that involves the extraction of several quantitative features from radiology images, which could be feasibly used (12,13). Previous studies have shown that conventional machine learning (cML) based on radiomics could be used to significantly predict clinical outcomes in cancers (14)(15)(16)(17)(18). In our previous studies, radiomics models could effectively predict microvascular invasion and progression-free survival before hepatectomy (19).…”

More than Meets the Eye: Using Textural Analysis and Artificial Intelligence as Decision Support Tools in Prostate Cancer Diagnosis—A Systematic Review

“…Amongst the most studied pathologies depicted through MRI protocols are the central nervous system tumors, neuroradiology comprising 19.71% of the international literature tackling radiomics-based tools, being the fastest developing research domain, with an annual growth rate of published papers of 316.02%, between 2013–2018 [ 51 ]. The greatest advantage that texture analysis brings in this context is the preoperative aggressiveness assessment of gliomas, differentiating low-grade gliomas from glioblastoma multiforme with an accuracy of 89% [ 52 ], having the potential of predicting the short (<12 months) or long-term survival rate (>24 months), based on the identified heterogeneity degree [ 53 ]. Finally, this emerging research area finds applicability in benign pathologies as well, being used for instance in differentiating endometriomas from hemorrhagic ovarian cysts, outperforming the classical pathognomonic “T2 dark spots” sign in terms of sensitivity (55.17%, versus 75%) and both “T2 shading” and “T2 dark spots” signs when it comes to specificity (35.71% and 64.29% versus 100%) [ 54 ].…”

Development and validation of a pyradiomics signature to predict initial treatment response and prognosis during transarterial chemoembolization in hepatocellular carcinoma

“…The radiomics algorithm offers an unprecedented opportunity to improve cancer decision-making in a low-cost and non-invasive manner. Previous studies have shown that radiomics models of radiology images are significantly associated with clinical outcomes in cancer patients (17)(18)(19)(20)(21). We previously found that a radiomics model based on CT images could precisely predict microvascular invasion in HCC patients and the machine learning algorithm could be used to predict clinic outcome in cancer (22,23).…”