Multi-parametric MRI phenotype with trustworthy machine learning for differentiating CNS demyelinating diseases.
Ontology highlight
ABSTRACT: Background
Misdiagnosis of multiple sclerosis (MS) and neuromyelitis optica (NMO) may delay the treatment, resulting in poor prognosis. However, the precise identification of these two diseases is still challenging in clinical practice. We aimed to evaluate the value of quantitative radiomic features extracted from the brain white matter lesions for differential diagnosis of MS and NMO.Methods
We recruited 116 CNS demyelinating patients including 78 MS, and 38 NMO. Three neuroradiologists performed visual differential diagnosis based on brain MRI for comparison purpose. A multi-level scheme was designed to harness the selection of discriminative and stable radiomics features extracted from brain while mater lesions in T1-MPRAGE, T2 sequences and clinical factors. Based on the imaging phenotype composed of the selected radiomic and clinical features, Multi-parametric Multivariate Random Forest (MM-RF) model was constructed and verified with both 10-fold cross-validation and independent testing. Result interpretation was provided to build trust in diagnostic decisions.Results
Eighty-six patients were randomly selected to form the training set while the rest 30 patients for independent testing. On the training set, our MM-RF model achieved accuracy 0.849 and AUC 0.826 in 10-fold cross-validation, which were significantly higher than clinical visual analysis (0.709 and 0.683, p < 0.05). In the independent testing, the MM-RF model achieved AUC 0.902, accuracy 0.871, sensitivity 0.873, specificity 0.869, respectively. Furthermore, age, sex and EDSS were found mildly correlated with the radiomic features (p of all < 0.05).Conclusions
Multi-parametric radiomic features have potential as practical quantitative imaging biomarkers for differentiating MS from NMO.
SUBMITTER: Huang J
PROVIDER: S-EPMC8419989 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:PurposeMachine learning (ML) can extract high-throughput features of images to predict disease. This study aimed to develop nomogram of multi-parametric MRI (mpMRI) ML model to predict the risk of breast cancer.MethodsThe mpMRI included non-enhanced and enhanced T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), apparent diffusion coefficient (ADC), K trans, K ep, V e, and V p. Regions of interest were annotated in an enhanced T1WI map and mapped to other maps in every slice. 1,132 features and top-10 principal components were extracted from every parameter map. Single-parametric and multi-parametric ML models were constructed via 10 rounds of five-fold cross-validation. The model with the highest area under the curve (AUC) was considered as the optimal model and validated by calibration curve and decision curve. Nomogram was built with the optimal ML model and patients' characteristics.ResultsThis study involved 144 malignant lesions and 66 benign lesions. The average age of patients with benign and malignant lesions was 42.5 years old and 50.8 years old, respectively, which were statistically different. The sixth and fourth principal components of K trans had more importance than others. The AUCs of K trans, K ep, V e and V p, non-enhanced T1WI, enhanced T1WI, T2WI, and ADC models were 0.86, 0.81, 0.81, 0.83, 0.79, 0.81, 0.84, and 0.83 respectively. The model with an AUC of 0.90 was considered as the optimal model which was validated by calibration curve and decision curve. Nomogram for the prediction of breast cancer was built with the optimal ML models and patient age.ConclusionNomogram could improve the ability of breast cancer prediction preoperatively.
| S-EPMC7952867 | biostudies-literature
Project description:Objectives: To develop and validate a predictive model for discriminating clinically significant prostate cancer (csPCa) from clinically insignificant prostate cancer (ciPCa). Methods: This retrospective study was performed with 159 consecutively enrolled pathologically confirmed PCa patients from two medical centers. The dataset was allocated to a training group (n = 54) and an internal validation group (n = 22) from one center along with an external independent validation group (n = 83) from another center. A total of 1,188 radiomic features were extracted from T2WI, diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC) images derived from DWI for each patient. Multivariable logistic regression analysis was performed to develop the model, incorporating the radiomic signature, ADC value, and independent clinical risk factors. This was presented using a radiomic nomogram. The receiver operating characteristic (ROC) curve was utilized to assess the predictive efficacy of the radiomic nomogram in both the training and validation groups. The decision curve analysis was used to evaluate which model achieved the most net benefit. Results: The radiomic signature, which was made up of 10 selected features, was significantly associated with csPCa (P < 0.001 for both training and internal validation groups). The area under the curve (AUC) values of discriminating csPCa for the radiomics signature were 0.95 (training group), 0.86 (internal validation group), and 0.81 (external validation group). Multivariate logistic analysis identified the radiomic signature and ADC value as independent parameters of predicting csPCa. Then, the combination nomogram incorporating the radiomic signature and ADC value demonstrated a favorable classification capability with the AUC of 0.95 (training group), 0.93 (internal validation group), and 0.84 (external validation group). Appreciable clinical utility of this model was illustrated using the decision curve analysis for the nomogram. Conclusions: The nomogram, incorporating radiomic signature and ADC value, provided an individualized, potential approach for discriminating csPCa from ciPCa.
| S-EPMC7339043 | biostudies-literature
Project description:Current machine learning techniques provide the opportunity to develop noninvasive and automated glioma grading tools, by utilizing quantitative parameters derived from multi-modal magnetic resonance imaging (MRI) data. However, the efficacies of different machine learning methods in glioma grading have not been investigated.A comprehensive comparison of varied machine learning methods in differentiating low-grade gliomas (LGGs) and high-grade gliomas (HGGs) as well as WHO grade II, III and IV gliomas based on multi-parametric MRI images was proposed in the current study. The parametric histogram and image texture attributes of 120 glioma patients were extracted from the perfusion, diffusion and permeability parametric maps of preoperative MRI. Then, 25 commonly used machine learning classifiers combined with 8 independent attribute selection methods were applied and evaluated using leave-one-out cross validation (LOOCV) strategy. Besides, the influences of parameter selection on the classifying performances were investigated. We found that support vector machine (SVM) exhibited superior performance to other classifiers. By combining all tumor attributes with synthetic minority over-sampling technique (SMOTE), the highest classifying accuracy of 0.945 or 0.961 for LGG and HGG or grade II, III and IV gliomas was achieved. Application of Recursive Feature Elimination (RFE) attribute selection strategy further improved the classifying accuracies. Besides, the performances of LibSVM, SMO, IBk classifiers were influenced by some key parameters such as kernel type, c, gama, K, etc. SVM is a promising tool in developing automated preoperative glioma grading system, especially when being combined with RFE strategy. Model parameters should be considered in glioma grading model optimization.
| S-EPMC5564607 | biostudies-literature
Project description:Background Viral infections are a proposed possible cause of inflammatory central nervous system (CNS) demyelinating diseases, including multiple sclerosis (MS), neuromyelitis optica spectrum disorder (NMOSD), and myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD). During the past 2 years, CNS demyelinating events associated with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection have been reported, but causality is unclear. Objective To investigate the relationship between CNS demyelinating disease development and exacerbation with antecedent and/or concurrent SARS-CoV-2 infection. Methods A systematic literature review of all publications describing either a new diagnosis or relapse of CNS demyelinating diseases (MS, NMOSD, MOGAD) in association with SARS-CoV-2 infection was performed utilizing PRISMA guidelines. Descriptive statistics were used for data analysis, using a case analysis approach. Results Sixty-seven articles met the inclusion criteria for the study. Most of the reported cases of NMOSD (n = 13, 72.2% of reported cases) and MOGAD (n = 27, 96.5% of reported cases) were of new disease onset, presenting with typical clinical and radiographic features of these conditions, respectively. In contrast, reported MS cases varied amongst newly diagnosed cases (n = 10, 10.5% of reported cases), relapses (n = 63, 66.4%) and pseudo-relapses (n = 22, 23.2%). The median duration between COVID-19 infection and demyelinating event onset was 11.5 days (range 0–90 days) in NMOSD, 6 days (range−7 to +45 days) in MOGAD, and 13.5 days (range−21 to +180 days) in MS. Most cases received high-dose corticosteroids with a good clinical outcome. Conclusion Based upon available literature, the rate of CNS demyelinating events occurring in the setting of preceding or concurrent SARS-CoV-2 infection is relatively low considering the prevalence of SARS-CoV-2 infection. The clinical outcomes of new onset or relapsing MS, NMOSD, or MOGAD associated with antecedent or concurrent infection were mostly favorable. Larger prospective epidemiological studies are needed to better delineate the impact of COVID-19 on CNS demyelinating diseases.
| S-EPMC9530047 | biostudies-literature
Project description:Several laboratory and rotating frame quantitative MRI parameters were evaluated and compared for detection of changes in articular cartilage following selective enzymatic digestion. Bovine osteochondral specimens were subjected to 44?h incubation in control medium or in collagenase or chondroitinase ABC to induce superficial collagen or proteoglycan (glycosaminoglycan) alterations. The samples were scanned at 9.4?T for T1 , T1?Gd (dGEMRIC), T2 , adiabatic T1?? , adiabatic T2?? , continuous-wave T1?? , TRAFF2 , and T1?sat relaxation times and for magnetization transfer ratio (MTR). For reference, glycosaminoglycan content, collagen fibril orientation and biomechanical properties were determined. Changes primarily in the superficial cartilage were noted after enzymatic degradation. Most of the studied parameters were sensitive to the destruction of collagen network, whereas glycosaminoglycan depletion was detected only by native T1 and T1?Gd relaxation time constants throughout the tissue and by MTR superficially. T1 , adiabatic T1?? , adiabatic T2?? , continuous-wave T1?? , and T1?sat correlated significantly with the biomechanical properties while T1?Gd correlated with glycosaminoglycan staining. The findings indicated that most of the studied MRI parameters were sensitive to both glycosaminoglycan content and collagen network integrity, with changes due to enzymatic treatment detected primarily in the superficial tissue. Strong correlation of T1 , adiabatic T1? , adiabatic T2?? , continuous-wave T1?? , and T1?sat with the altered biomechanical properties, reflects that these parameters were sensitive to critical functional properties of cartilage. © 2015 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 34:1111-1120, 2016.
| S-EPMC4903086 | biostudies-literature
Project description:PurposeQuantitative multi-parametric MRI (mpMRI) methods may allow the assessment of renal injury and function in a sensitive and objective manner. This study aimed to evaluate an array of MRI methods that exploit endogenous contrasts including relaxation rates, pool size ratio (PSR) derived from quantitative magnetization transfer (qMT), chemical exchange saturation transfer (CEST), nuclear Overhauser enhancement (NOE), and apparent diffusion coefficient (ADC) for their sensitivity and specificity in detecting abnormal features associated with kidney disease in a murine model of unilateral ureter obstruction (UUO).MethodsMRI scans were performed in anesthetized C57BL/6N mice 1, 3, or 6 days after UUO at 7T. Paraffin tissue sections were stained with Masson trichrome following MRI.ResultsCompared to contralateral kidneys, the cortices of UUO kidneys showed decreases of relaxation rates R1 and R2 , PSR, NOE, and ADC. No significant changes in CEST effects were observed for the cortical region of UUO kidneys. The MRI parametric changes in renal cortex are related to tubular cell death, tubular atrophy, tubular dilation, urine retention, and interstitial fibrosis in the cortex of UUO kidneys.ConclusionMeasurements of multiple MRI parameters provide comprehensive information about the molecular and cellular changes produced by UUO. Magn Reson Med 79:2216-2227, 2018. © 2017 International Society for Magnetic Resonance in Medicine.
| S-EPMC5870889 | biostudies-literature
Project description:Myopathies often display a common set of complex pathologies that include muscle weakness, inflammation, compromised membrane integrity, fat deposition, and fibrosis. Multi-parametric, quantitative, non-invasive imaging approaches may be able to resolve these individual pathological components. The goal of this study was to use multi-parametric MRI to investigate inflammation as an isolated pathological feature. Proton relaxation, diffusion tensor imaging (DTI), quantitative magnetization transfer (qMT-MRI), and dynamic contrast enhanced (DCE-MRI) parameters were calculated from data acquired in a single imaging session conducted 6-8 hours following the injection of λ-carrageenan, a local inflammatory agent. T2 increased in the inflamed muscle and transitioned to bi-exponential behavior. In diffusion measurements, all three eigenvalues and the apparent diffusion coefficient increased, but λ3 had the largest relative change. Analysis of the qMT data revealed that the T1 of the free pool and the observed T1 both increased in the inflamed tissue, while the ratio of exchanging spins in the solid pool to those in the free water pool (the pool size ratio) significantly decreased. DCE-MRI data also supported observations of an increase in extracellular volume. These findings enriched the understanding of the relation between multiple quantitative MRI parameters and an isolated inflammatory pathology, and may potentially be employed for other single or complex myopathy models.
| S-EPMC4134016 | biostudies-other
Project description:INTRODUCTION:Magnetic resonance imaging (MRI) shows slight spatial variations in brain white matter (WM). We used quantitative multi-parametric MRI to evaluate in what respect these inhomogeneities could correspond to WM subtypes with specific characteristics and spatial distribution. MATERIALS AND METHODS:Twenty-six controls (12 women, 38 ±9 Y) took part in a 60-min session on a 3T scanner measuring 7 parameters: R1 and R2, diffusion tensor imaging which allowed to measure Axial and Radial Diffusivity (AD, RD), magnetization transfer imaging which enabled to compute the Macromolecular Proton Fraction (MPF), and a susceptibility-weighted sequence which permitted to quantify R2* and magnetic susceptibility (?m). Spatial independent component analysis was used to identify WM subtypes with specific combination of quantitative parameters values. RESULTS:Three subtypes could be identified. t-WM (track) mostly mapped on well-formed projection and commissural tracts and came with high AD values (all p < 10(-18)). The two other subtypes were located in subcortical WM and overlapped with association fibers: f-WM (frontal) was mostly anterior in the frontal lobe whereas c-WM (central) was underneath the central cortex. f-WM and c-WM had higher MPF values, indicating a higher myelin content (all p < 1.7 10(-6)). This was compatible with their larger ?m and R2, as iron is essentially stored in oligodendrocytes (all p < 0.01). Although R1 essentially showed the same, its higher value in t-WM relative to c-WM might be related to its higher cholesterol concentration. CONCLUSIONS:Thus, f- and c-WMs were less structured, but more myelinated and probably more metabolically active regarding to their iron content than WM related to fasciculi (t-WM). As known WM bundles passed though different WM subtypes, myelination might not be uniform along the axons but rather follow a spatially consistent regional variability. Future studies might examine the reproducibility of this decomposition and how development and pathology differently affect each subtype.
| S-EPMC6003690 | biostudies-literature
Project description:Cognitive impairment (CI) is prevalent in central nervous system demyelinating diseases, such as multiple sclerosis (MS) and neuromyelitis optica spectrum disorders (NMOSD). We developed a novel tablet-based modified digital Symbol Digit Modalities Test (MD-SDMT) with adjustable protocols that feature alternating symbol-digit combinations in each trial, lasting one or two minutes. We assessed 144 patients (99 with MS and 45 with NMOSD) using both MD-SDMT protocols and the traditional paper-based SDMT. We also gathered participants' feedback through a questionnaire regarding their preferences and perceived reliability. The results showed strong correlations between MD-SDMT and paper-based SDMT scores (Pearsons correlation: 0.88 for 2 min; 0.85 for 1 min, both p < 0.001). Among the 120 respondents, the majority preferred the digitalized SDMT (55% for the 2 min, 39% for the 1 min) over the paper-based version (6%), with the 2 min MD-SDMT reported as the most reliable test. Notably, patients with NMOSD and older individuals exhibited a preference for the paper-based test, as compared to those with MS and younger patients. In summary, even with short test durations, the digitalized SDMT effectively evaluates cognitive function in MS and NMOSD patients, and is generally preferred over the paper-based method, although preferences may vary with patient characteristics.
| S-EPMC11199480 | biostudies-literature
Project description:BackgroundDistinguishing between central nervous system lymphoma (CNSL) and CNS infectious and/or demyelinating diseases, although clinically important, is sometimes difficult even using imaging strategies and conventional cerebrospinal fluid (CSF) analyses. To determine whether detection of genetic mutations enables differentiation between these diseases and the early detection of CNSL, we performed mutational analysis using CSF liquid biopsy technique.MethodsIn this study, we extracted cell-free DNA from the CSF (CSF-cfDNA) of CNSL (N = 10), CNS infectious disease (N = 10), and demyelinating disease (N = 10) patients, and performed quantitative mutational analysis by droplet-digital PCR. Conventional analyses were also performed using peripheral blood and CSF to confirm the characteristics of each disease.ResultsBlood hemoglobin and albumin levels were significantly lower in CNSL than CNS infectious and demyelinating diseases, CSF cell counts were significantly higher in infectious diseases than CNSL and demyelinating diseases, and CSF-cfDNA concentrations were significantly higher in infectious diseases than CNSL and demyelinating diseases. Mutation analysis using CSF-cfDNA detected MYD88L265P and CD79Y196 mutations in 60% of CNSLs each, with either mutation detected in 80% of cases. Mutual existence of both mutations was identified in 40% of cases. These mutations were not detected in either infectious or demyelinating diseases, and the sensitivity and specificity of detecting either MYD88/CD79B mutations in CNSL were 80% and 100%, respectively. In the four cases biopsied, the median time from collecting CSF with the detected mutations to definitive diagnosis by conventional methods was 22.5 days (range, 18-93 days).ConclusionsThese results suggest that mutation analysis using CSF-cfDNA might be useful for differentiating CNSL from CNS infectious/demyelinating diseases and for early detection of CNSL, even in cases where brain biopsy is difficult to perform.
| S-EPMC10501233 | biostudies-literature