Radiomics-Based Machine Learning Classification for Glioma Grading Using Diffusion- and Perfusion-Weighted Magnetic Resonance Imaging

Hashido, Takashi; Saito, Shigeyoshi; Ishida, Takayuki

doi:10.1097/rct.0000000000001180

Cited by 5 publications

(3 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The highdimensional quantitative information of imaging histology extracted from DWI in this study is consistent with the main references for histological grading of tumor pathology, which include: cellular pleomorphism and nuclear division heterogeneity, increased cell density with signi cant proliferation, microvascular embolism, proliferation, and necrosis. Similar results were found in the research of Takeshi et al [30]which show that HGGs and LGGs can be distinguished using Radiomics-based machine learning classi ers that leverage the quantitative ADC and cerebral blood ow (CBF) maps.…”

Section: Discussionsupporting

confidence: 85%

The value of Clinical-and DWI-based Radiomics Nomogram to predict Pathologic Upgrading in Biopsy-Proven Endometrial Cancer

Yue

Wang

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: It is important for biopsy formal endometrial cancer patients, especially young patients of childbearing age to determine the preservation of fertility and predict pathological escalation. Purpose: This study's goal was to determine the viability and effectiveness of a non-invasive quantitative imaging evaluation model built using the Diffusion Weighted Image (DWI) technique and based on Radiomics signatures and clinical parameters Analysis to evaluate Endometrial Cancer (EC) with Biopsy-Proven Pathologic Upgrading. Method: From January 2018 to December 2021, a total of 76 patients with endometrial cancer who had undergone surgery for the disease were retrospectively recruited (training cohort, n = 53; validation cohort, n = 23). The diffusion-weighted image (DWI) served as the source for the Radiomics features. All images were imported into 3D-slicer for whole tumor Segmentation and were used for radiomics feature extraction. Radiomic features were selected in target tumor volumes to build Radscore using the least absolute shrinkage and selection operator (LASSO) and Cox regression analysis Logistic regression, Next building a combined model incorporating rad-scores and clinical risk factors, compared with Radscore model, the clinical model. The models were evaluated by the receiver operating characteristic curve, and calibration curve as well as verified the model in the verification group. Results: AUC for identifying non-pathologic upgrading and pathologic upgrading in the training cohort was 0.606 and in the validation cohort was 0.708, Three of the 107 texture feature were retrieved and 3 parameters were preserved to create the Radscore. With the incorporation of clinical risk factors, the nomogram's AUC for the training and validation cohorts were 0.870 and 0.808, respectively. Both values were significantly higher than the AUC of the clinical model in these cohorts (0.830 and 0.815). The nomogram's training cohort and validation cohort's sensitivity and specificity were 0.938, 0.730, 0.900, and 0.769, respectively. The calibration curves for the nomogram had a good agreement. Conclusions: The Nomogram based on the Radiomics-clinical model in predicting Pathologic Upgrading in Biopsy-Proven Endometrial Cancer with high discriminatory ability.

show abstract

Section: Discussionsupporting

confidence: 85%

The value of Clinical-and DWI-based Radiomics Nomogram to predict Pathologic Upgrading in Biopsy-Proven Endometrial Cancer

Yue

Wang

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Therefore, since the main oncological goal is to prevent MT, additional non-invasive metabolic information may be useful in order to predict when the LGG has a higher risk of degeneration. First, an increase of perfusion or diffusion value(s) obtained through sequential and multimodal MRI could be a predictor of changes in glioma behavior [ 54 ], possibly identifiable using new machine-learning classifiers [ 55 ], and might prompt earlier (re)treatment. In the same spirit, recent advances in PET scanning using tracers easily accessible in routine practice (such as F-DOPA) have enabled an increase in sensitivity for the detection of foci of MT within the LGG, before the onset of enhancement [ 56 ].…”

Section: Predicting Oncological Interindividual Variability and Its C...mentioning

confidence: 99%

A Personalized Longitudinal Strategy in Low-Grade Glioma Patients: Predicting Oncological and Neural Interindividual Variability and Its Changes over Years to Think One Step Ahead

Duffau

2022

JPM

View full text Add to dashboard Cite

Diffuse low-grade glioma (LGG) is a rare cerebral cancer, mostly involving young adults with an active life at diagnosis. If left untreated, LGG widely invades the brain and becomes malignant, generating neurological worsening and ultimately death. Early and repeat treatments for this incurable tumor, including maximal connectome-based surgical resection(s) in awake patients, enable postponement of malignant transformation while preserving quality of life owing to constant neural network reconfiguration. Due to considerable interindividual variability in terms of LGG course and consecutive cerebral reorganization, a multistage longitudinal strategy should be tailored accordingly in each patient. It is crucial to predict how the glioma will progress (changes in growth rate and pattern of migration, genetic mutation, etc.) and how the brain will adapt (changes in patterns of spatiotemporal redistribution, possible functional consequences such as epilepsy or cognitive decline, etc.). The goal is to anticipate therapeutic management, remaining one step ahead in order to select the optimal (re-)treatment(s) (some of them possibly kept in reserve), at the appropriate time(s) in the evolution of this chronic disease, before malignization and clinical worsening. Here, predictive tumoral and non-tumoral factors, and their ever-changing interactions, are reviewed to guide individual decisions in advance based on patient-specific markers, for the treatment of LGG.

show abstract

“…In recent years, for example, a machine learning predictive model has been developed for the diagnosis of brain tumors based on routine blood tests 2 . Such algorithms can also be used for an automated detection and segmentation of meningiomas 3 and a preoperative classification of WHO grade of meningiomas and gliomas 4 – 6 .…”

Section: Introductionmentioning

confidence: 99%

Testing the applicability and performance of Auto ML for potential applications in diagnostic neuroradiology

Musigmann

Akkurt

Krähling

et al. 2022

Sci Rep

View full text Add to dashboard Cite

To investigate the applicability and performance of automated machine learning (AutoML) for potential applications in diagnostic neuroradiology. In the medical sector, there is a rapidly growing demand for machine learning methods, but only a limited number of corresponding experts. The comparatively simple handling of AutoML should enable even non-experts to develop adequate machine learning models with manageable effort. We aim to investigate the feasibility as well as the advantages and disadvantages of developing AutoML models compared to developing conventional machine learning models. We discuss the results in relation to a concrete example of a medical prediction application. In this retrospective IRB-approved study, a cohort of 107 patients who underwent gross total meningioma resection and a second cohort of 31 patients who underwent subtotal resection were included. Image segmentation of the contrast enhancing parts of the tumor was performed semi-automatically using the open-source software platform 3D Slicer. A total of 107 radiomic features were extracted by hand-delineated regions of interest from the pre-treatment MRI images of each patient. Within the AutoML approach, 20 different machine learning algorithms were trained and tested simultaneously. For comparison, a neural network and different conventional machine learning algorithms were trained and tested. With respect to the exemplary medical prediction application used in this study to evaluate the performance of Auto ML, namely the pre-treatment prediction of the achievable resection status of meningioma, AutoML achieved remarkable performance nearly equivalent to that of a feed-forward neural network with a single hidden layer. However, in the clinical case study considered here, logistic regression outperformed the AutoML algorithm. Using independent test data, we observed the following classification results (AutoML/neural network/logistic regression): mean area under the curve = 0.849/0.879/0.900, mean accuracy = 0.821/0.839/0.881, mean kappa = 0.465/0.491/0.644, mean sensitivity = 0.578/0.577/0.692 and mean specificity = 0.891/0.914/0.936. The results obtained with AutoML are therefore very promising. However, the AutoML models in our study did not yet show the corresponding performance of the best models obtained with conventional machine learning methods. While AutoML may facilitate and simplify the task of training and testing machine learning algorithms as applied in the field of neuroradiology and medical imaging, a considerable amount of expert knowledge may still be needed to develop models with the highest possible discriminatory power for diagnostic neuroradiology.

show abstract

Radiomics-Based Machine Learning Classification for Glioma Grading Using Diffusion- and Perfusion-Weighted Magnetic Resonance Imaging

Abstract: Supplemental digital content is available in the text.

Cited by 5 publications

References 39 publications

The value of Clinical-and DWI-based Radiomics Nomogram to predict Pathologic Upgrading in Biopsy-Proven Endometrial Cancer

The value of Clinical-and DWI-based Radiomics Nomogram to predict Pathologic Upgrading in Biopsy-Proven Endometrial Cancer

A Personalized Longitudinal Strategy in Low-Grade Glioma Patients: Predicting Oncological and Neural Interindividual Variability and Its Changes over Years to Think One Step Ahead

Testing the applicability and performance of Auto ML for potential applications in diagnostic neuroradiology

Contact Info

Product

Resources

About