Radiomics-Based Machine Learning in Differentiation Between Glioblastoma and Metastatic Brain Tumors

Abstract

Purpose: To investigative the diagnostic performance of radiomics-based machine learning in differentiating glioblastomas (GBM) from metastatic brain tumors (MBTs). Method: The current study involved 134 patients diagnosed and treated in our institution between April 2014 and December 2018. Radiomics features were extracted from contrast-enhanced T1 weighted imaging (T1C). Thirty diagnostic models were built based on five selection methods and six classification algorithms. The sensitivity, specificity, accuracy, and area under curve (AUC) of each model were calculated, and based on these the optimal model was chosen. Result : Two models represented promising diagnostic performance with AUC of 0.80. The first model was a combination of Distance Correlation as the selection method and Linear Discriminant Analysis (LDA) as the classification algorithm. In the training group, the sensitivity, specificity, accuracy, and AUC were 0.75, 0.85, 0.80, and 0.80, respectively; and in the testing group, the sensitivity, specificity, accuracy, and AUC of the model were 0.69, 0.86, 0.78, and 0.80, respectively. The second model was the Distance Correlation as the selection method and logistic regression (LR) as the classification algorithm, with sensitivity, specificity, accuracy, and AUC of 0.75, 0.85, 0.80, 0.80 in the training group and 0.69, 0.86, 0.78, 0.80 in the testing group. Conclusion: Radiomic-based machine learning has potential to be utilized in differentiating GBM from MBTs.

Comparison of Intraoperative Ultrasound B-Mode and Strain Elastography for the Differentiation of Glioblastomas From Solitary Brain Metastases. An Automated Deep Learning Approach for Image Analysis

“…Regarding the differentiation of GBM from SBM, we know that there are multiple radiological techniques available for pre-operative or non-invasive applications ( 5 , 8 – 11 , 13 , 15 – 17 , 55 ); besides, intraoperative histopathological techniques are currently the reference parameter for decision-making ( 56 ). Our study does not intend to make a comparison with the techniques mentioned above but to demonstrate, on the one hand, the high value that ultrasound and especially elastography owns in the study of brain tumors, and on the other hand, highlight that automatic image processing is a highly reliable technique.…”

Radiomics machine learning study with small sample size: single random training-test set split may result in unreliable results

“…Two classi cation tasks of different di culty levels were performed using radiomics features extracted from brain magnetic resonance imaging (MRI). The rst task was a "simple" task of differentiating between glioblastoma (GBM) and single metastasis; the accuracy of radiomics-based ML for this task has been reported to be up to 90% (12,13). In the current study, the rst task dataset consisted of 167 adult patients with pathologically con rmed single GBM (n=109) or brain metastasis (n=58) following brain MRI between January 2014 and December 2017.…”

Radioresistance in Glioblastoma and the Development of Radiosensitizers

“…Radiomics allows advanced non-invasive assessment of complex imaging features obtained by MRI that may serve as biomarkers [ 46 , 47 ] of disease aggressivity or response. Although these major advances in imaging techniques have substantially improved our ability to diagnose brain tumors, including GBM, overall survival and prognosis for patients with GBM continues to be poor, mostly due to inherent and developed resistance against standard-of-care therapy.…”