Glioma Grading via Analysis of Digital Pathology Images Using Machine Learning

Abstract

Cancer pathology reflects disease progression (or regression) and associated molecular characteristics, and provides rich phenotypic information that is predictive of cancer grade and has potential implications in treatment planning and prognosis. According to the remarkable performance of computational approaches in the digital pathology domain, we hypothesized that machine learning can help to distinguish low-grade gliomas (LGG) from high-grade gliomas (HGG) by exploiting the rich phenotypic information that reflects the microvascular proliferation level, mitotic activity, presence of necrosis, and nuclear atypia present in digital pathology images. A set of 735 whole-slide digital pathology images of glioma patients (median age: 49.65 years, male: 427, female: 308, median survival: 761.26 days) were obtained from TCGA. Sub-images that contained a viable tumor area, showing sufficient histologic characteristics, and that did not have any staining artifact were extracted. Several clinical measures and imaging features, including conventional (intensity, morphology) and advanced textures features (gray-level co-occurrence matrix and gray-level run-length matrix), extracted from the sub-images were further used for training the support vector machine model with linear configuration. We sought to evaluate the combined effect of conventional imaging, clinical, and texture features by assessing the predictive value of each feature type and their combinations through a predictive classifier. The texture features were successfully validated on the glioma patients in 10-fold cross-validation (accuracy = 75.12%, AUC = 0.652). The addition of texture features to clinical and conventional imaging features improved grade prediction compared to the models trained on clinical and conventional imaging features alone (p = 0.045 and p = 0.032 for conventional imaging features and texture features, respectively). The integration of imaging, texture, and clinical features yielded a significant improvement in accuracy, supporting the synergistic value of these features in the predictive model. The findings suggest that the texture features, when combined with conventional imaging and clinical markers, may provide an objective, accurate, and integrated prediction of glioma grades. The proposed digital pathology imaging-based marker may help to (i) stratify patients into clinical trials, (ii) select patients for targeted therapies, and (iii) personalize treatment planning on an individual person basis.

Radiomics at a Glance: A Few Lessons Learned from Learning Approaches

“…Finally, it is worth mentioning that recent studies reported ML-driven radiopathomics applications (e.g., SVM, logistic regression) for prognosis of glioblastoma [ 98 ] and grading of glioma [ 99 ], prediction of pathologic response in the locally advanced rectal cancer (LARC) [ 100 ], diagnosis of lung nodule subtypes [ 101 ], and detection of high-grade prostate cancer tumors missed by radiologists [ 102 ]. Model performance ranged from 0.8 to 0.9 in the accuracy, sensitivity, specificity, and AUC.…”

Malignant tumor purity reveals the correlation between CD3E and low grade glioma microenvironment

“…LGG is a common brain tumor, and the prognosis of patients is often poor [32]. However, whether it is surgery, radiation therapy or chemotherapy (usually using temozolomide), can't improve the prognosis and survival of patients [7,33,34]. The reasons for the lack of progress include the growth of invasive tumors in basic organs, which limits the utility of local therapies, and the protection of tumor cells by the blood-brain barrier, their inherent resistance to induced cell death, and the lack of dependence on a single, can targeted carcinogenic pathways [35].…”

“…Nevertheless, high recurrence and malignancy rate of LGG still bring great pain to patients [5,6]. Investigations on approaches to maintain the quality of life of LGG patients while prolonging the overall survival (OS) has become a common concern for clinicians and researchers [7].…”

Machine Learning Applied to Diagnosis of Human Diseases: A Systematic Review

“…Metabolic diseases [74,79] Clustering K-means Clustering [87] Clustering DBSCAN [171] Regression Random Forest [100] Classification SVM [106,109] Classification ID3 [115,116,118,120,122] Classification KNN [135] Classification Naïve Bayes [137,143] Classification Bayesian Networks [145] Regression Linear regression Cancer [75,81] Clustering K-means Clustering [84,86] Clustering DBSCAN [24] Clustering SNF [25] Clustering PINS [26] Clustering CIMLR [95,172] Classification SVM [108] Classification ID3 [130] Classification Naïve Bayes [136] Classification Bayesian Networks [148,173] Regression Linear regression [146,174] Regression Logistic regression [157] Classification Neural Networks + KNN [156] Classification Neural Networks + SVM [160] Classification Neural Networks + ID3 [161] Classification KNN [175] Classification DT [176] Classification DL…”