Differentiation of multiple sclerosis lesions and low-grade brain tumors on MRS data: machine learning approaches

BIOM-62 SENSITIVE DETECTION AND DISCRIMINATION OF INTRACRANIAL TUMORS BY BLOOD

Neuro-Oncology ◽

10.1093/neuonc/noaa215.059 ◽

2020 ◽

Vol 22 (Supplement_2) ◽

pp. ii15-ii15

Author(s):

Farshad Nassiri ◽

Ankur Chakravarthy ◽

Shengrui Feng ◽

Roxana Shen ◽

Romina Nejad ◽

...

Keyword(s):

Dna Methylation ◽

Brain Tumors ◽

Model Performance ◽

High Sensitivity ◽

Circulating Tumor Dna ◽

Low Grade ◽

Learning Approaches ◽

Intracranial Tumors ◽

Cell Of Origin ◽

Tumor Dna

Abstract BACKGROUND The diagnosis of intracranial tumors relies on tissue specimens obtained by invasive surgery. Non-invasive diagnostic approaches, particularly for patients with brain tumours, provide an opportunity to avoid surgery and mitigate unnecessary risk to patients. We reasoned that DNA methylation profiles of circulating tumor DNA in blood can be used as a clinically useful biomarker for patients with brain tumors, given the specificity of DNA methylation profiles for cell-of-origin. METHODS We generated methylation profiles on the plasma of 608 patients with cancer (219 intracranial, 388 extracranial) and 60 healthy controls using a cell-free methylated DNA immunoprecipitation combined with deep sequencing (cfMeDIP-seq) approach. Using machine-learning approaches we generated and evaluated models to distinguish brain tumors from extracranial cancers that may metastasize to the brain, as well as additional models to discriminate common brain tumors included in the differential diagnosis of solitary extra-axial and intra-axial tumors. RESULTS We observed high sensitivity and discriminative capacity for our models to distinguish gliomas from other cancerous and healthy patients (AUC=0.99, 95%CI 0.96–1), with similar performance in IDH mutant and wildtype gliomas as well as in lower- and high-grade gliomas. Excluding non-malignant contributors to plasma methylation did not change model performance (AUC=0.982, 95%CI 0.93–1). Models generated to discriminate intracranial tumors from each other also demonstrated high accuracy for common extra-axial tumors (AUCmeningioma=0.89, 95%CI 0.80–0.97; AUChemangiopericytoma=0.95, 95%CI 0.73–1) as well as intra-axial tumors ranging from low-grade indolent glial-neuronal tumors (AUC 0.93, 95%CI 0.80 – 1) to diffuse intra-axial gliomas with distinct molecular composition (AUCIDH-mutant glioma = 0.82, 95%CI 0.66 -0.98; AUCIDH-wildtype-glioma = 0.71, 95%CI 0.53 – 0.9). Plasma cfMeDIP-seq signals originated from corresponding tumor tissue DNA methylation signals (r=0.37, p< 2.2e-16). CONCLUSIONS These results demonstrate the potential for cfMeDIP-seq profiles to not only detect circulating tumor DNA, but to accurately discriminate common intracranial tumors that share cell-of-origin lineages.

Download Full-text

Amide proton transfer weighted (APTw) imaging based radiomics allows for the differentiation of gliomas from metastases

Scientific Reports ◽

10.1038/s41598-021-85168-8 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Elisabeth Sartoretti ◽

Thomas Sartoretti ◽

Michael Wyss ◽

Carolin Reischauer ◽

Luuk van Smoorenburg ◽

...

Keyword(s):

Machine Learning ◽

Random Forest ◽

Brain Tumors ◽

Proton Transfer ◽

Multilayer Perceptron ◽

Random Forest Classifier ◽

Amide Proton ◽

Low Grade ◽

Who Grade ◽

Amide Proton Transfer

AbstractWe sought to evaluate the utility of radiomics for Amide Proton Transfer weighted (APTw) imaging by assessing its value in differentiating brain metastases from high- and low grade glial brain tumors. We retrospectively identified 48 treatment-naïve patients (10 WHO grade 2, 1 WHO grade 3, 10 WHO grade 4 primary glial brain tumors and 27 metastases) with either primary glial brain tumors or metastases who had undergone APTw MR imaging. After image analysis with radiomics feature extraction and post-processing, machine learning algorithms (multilayer perceptron machine learning algorithm; random forest classifier) with stratified tenfold cross validation were trained on features and were used to differentiate the brain neoplasms. The multilayer perceptron achieved an AUC of 0.836 (receiver operating characteristic curve) in differentiating primary glial brain tumors from metastases. The random forest classifier achieved an AUC of 0.868 in differentiating WHO grade 4 from WHO grade 2/3 primary glial brain tumors. For the differentiation of WHO grade 4 tumors from grade 2/3 tumors and metastases an average AUC of 0.797 was achieved. Our results indicate that the use of radiomics for APTw imaging is feasible and the differentiation of primary glial brain tumors from metastases is achievable with a high degree of accuracy.

Download Full-text

A Comparison of Machine Learning Approaches for Classifying Multiple Sclerosis Courses Using MRSI and Brain Segmentations

Artificial Neural Networks and Machine Learning – ICANN 2017 - Lecture Notes in Computer Science ◽

10.1007/978-3-319-68612-7_73 ◽

2017 ◽

pp. 643-651 ◽

Cited By ~ 1

Author(s):

Adrian Ion-Mărgineanu ◽

Gabriel Kocevar ◽

Claudio Stamile ◽

Diana M. Sima ◽

Françoise Durand-Dubief ◽

...

Keyword(s):

Machine Learning ◽

Multiple Sclerosis ◽

Learning Approaches

Download Full-text

Analysis of defective pathways and drug repositioning in Multiple Sclerosis via machine learning approaches

Computers in Biology and Medicine ◽

10.1016/j.compbiomed.2019.103492 ◽

2019 ◽

Vol 115 ◽

pp. 103492 ◽

Cited By ~ 3

Author(s):

Enrique J. deAndrés-Galiana ◽

Guillermina Bea ◽

Juan L. Fernández-Martínez ◽

Leo N. Saligan

Keyword(s):

Machine Learning ◽

Multiple Sclerosis ◽

Drug Repositioning ◽

Learning Approaches

Download Full-text

When tumefactive demyelination is truly a tumor: case report of a radiographic misdiagnosis

CNS Oncology ◽

10.2217/cns-2020-0028 ◽

2021 ◽

pp. CNS69

Author(s):

Breanna Taylor ◽

Mallika P Patel ◽

Katherine B Peters

Keyword(s):

Multiple Sclerosis ◽

Case Report ◽

Brain Tumors ◽

Disease Progression ◽

Symptom Management ◽

Low Grade ◽

Appropriate Therapy ◽

Well Differentiated ◽

Adult Male Patient ◽

Slow Growing

Oligodendrogliomas are slow-growing tumors that account for 15–20% of gliomas. This case report describes the case of an adult male patient diagnosed initially with tumefactive demyelination and multiple sclerosis, which was subsequently found to be a well-differentiated low-grade oligodendroglioma. This case emphasizes the importance of timely diagnosis in oligodendrogliomas and other brain tumors for the prompt initiation of appropriate therapy, to minimize the likelihood of disease progression, ensure symptom management and escalation of unnecessary treatments for multiple sclerosis.

Download Full-text

Ensemble learning predicts multiple sclerosis disease course in the SUMMIT study

npj Digital Medicine ◽

10.1038/s41746-020-00338-8 ◽

2020 ◽

Vol 3 (1) ◽

Author(s):

Yijun Zhao ◽

◽

Tong Wang ◽

Riley Bove ◽

Bruce Cree ◽

...

Keyword(s):

Machine Learning ◽

Multiple Sclerosis ◽

San Francisco ◽

Ensemble Learning ◽

Disease Onset ◽

Machine Learning Algorithms ◽

Machine Learning Techniques ◽

Learning Approaches ◽

Disease Course ◽

Learning Techniques

AbstractThe rate of disability accumulation varies across multiple sclerosis (MS) patients. Machine learning techniques may offer more powerful means to predict disease course in MS patients. In our study, 724 patients from the Comprehensive Longitudinal Investigation in MS at Brigham and Women’s Hospital (CLIMB study) and 400 patients from the EPIC dataset, University of California, San Francisco, were included in the analysis. The primary outcome was an increase in Expanded Disability Status Scale (EDSS) ≥ 1.5 (worsening) or not (non-worsening) at up to 5 years after the baseline visit. Classification models were built using the CLIMB dataset with patients’ clinical and MRI longitudinal observations in first 2 years, and further validated using the EPIC dataset. We compared the performance of three popular machine learning algorithms (SVM, Logistic Regression, and Random Forest) and three ensemble learning approaches (XGBoost, LightGBM, and a Meta-learner L). A “threshold” was established to trade-off the performance between the two classes. Predictive features were identified and compared among different models. Machine learning models achieved 0.79 and 0.83 AUC scores for the CLIMB and EPIC datasets, respectively, shortly after disease onset. Ensemble learning methods were more effective and robust compared to standalone algorithms. Two ensemble models, XGBoost and LightGBM were superior to the other four models evaluated in our study. Of variables evaluated, EDSS, Pyramidal Function, and Ambulatory Index were the top common predictors in forecasting the MS disease course. Machine learning techniques, in particular ensemble methods offer increased accuracy for the prediction of MS disease course.

Download Full-text

Diagnostic Dilemma: An Atypical Case of Astrocytoma in a Patient with Relapsing–Remitting Multiple Sclerosis

Neurology International ◽

10.3390/neurolint13020025 ◽

2021 ◽

Vol 13 (2) ◽

pp. 240-251

Author(s):

Chantal Kahovec ◽

Aman Saini ◽

Michael C. Levin

Keyword(s):

Multiple Sclerosis ◽

Magnetic Resonance ◽

Brain Tumors ◽

Resonance Spectroscopy ◽

Low Grade ◽

Relapsing Remitting ◽

Mri Scans ◽

Remitting Multiple Sclerosis ◽

Relapsing Remitting Multiple Sclerosis ◽

Progressive Course

Distinguishing between tumefactive demyelinating lesions (TDLs) and brain tumors in multiple sclerosis (MS) can be challenging. A progressive course is highly common with brain tumors in MS and no single neuroimaging technique is foolproof when distinguishing between the two. We report a case of a 41-year-old female with relapsing–remitting multiple sclerosis, who had a suspicious lesion within the left frontal hemisphere, without a progressive course. The patient experienced paresthesias primarily to her right hand but remained stable without any functional decline and new neurological symptoms over the four years she was followed. The lesion was followed with brain magnetic resonance imaging (MRI) scans, positron emission tomography–computed tomography scans, and magnetic resonance spectroscopy. Together, these scans favored the diagnosis of a TDL, but a low-grade tumor was difficult to rule out. Examination of serial brain MRI scans showed an enlarging lesion in the left middle frontal gyrus involving the deep white matter. Neurosurgery was consulted and an elective left frontal awake craniotomy was performed. Histopathology revealed a grade II astrocytoma. This case emphasizes the importance of thorough and continuous evaluation of atypical MRI lesions in MS and contributes important features to the literature for timely diagnosis and treatment of similar cases.

Download Full-text