A quantitative SVM approach potentially improves the accuracy of magnetic resonance spectroscopy in the preoperative evaluation of the grades of diffuse gliomas

Application of Machine Learning for Classification of Brain Tumors: A Systematic Review and Meta-Analysis

BACKGROUND

Classifying brain tumors accurately is crucial for treatment and prognosis. Machine learning (ML) shows great promise in improving tumor classification accuracy. This study evaluates ML algorithms for differentiating various brain tumor types.

METHODS

A systematic review and meta-analysis were conducted, searching PubMed, Embase, and Web of Science up to March 14, 2023. Studies that only investigated image segmentation accuracy or brain tumor detection instead of classification were excluded. We extracted binary diagnostic accuracy data, constructing contingency tables to derive sensitivity and specificity.

RESULTS

Fifty-one studies were included. The pooled area under the curve for glioblastoma versus lymphoma and low-grade versus high-grade gliomas were 0.99 (95% confidence interval [CI]: 0.98-1.00) and 0.89, respectively. The pooled sensitivity and specificity for benign versus malignant tumors were 0.90 (95% CI: 0.85-0.93) and 0.93 (95% CI: 0.90-0.95), respectively. The pooled sensitivity and specificity for low-grade versus high-grade gliomas were 0.99 (95% CI: 0.97-1.00) and 0.94, (95% CI: 0.79-0.99), respectively. Primary versus metastatic tumor identification yields sensitivity and specificity of 0.89, (95% CI: 0.83-0.93) and 0.87 (95% CI: 0.82-0.91), correspondingly. The differentiation of gliomas from pituitary tumors yielded the highest results among primary brain tumor classifications: sensitivity of 0.99 (95% CI: 0.99-1.00) and specificity of 0.99 (95% CI: 0.98-1.00).

CONCLUSIONS

ML demonstrated excellent performance in classifying brain tumor images, with near-maximum area under the curves, sensitivity, and specificity.

Ensemble Learning for Breast Cancer Lesion Classification: A Pilot Validation Using Correlated Spectroscopic Imaging and Diffusion-Weighted Imaging

The main objective of this work was to evaluate the application of individual and ensemble machine learning models to classify malignant and benign breast masses using features from two-dimensional (2D) correlated spectroscopy spectra extracted from five-dimensional echo-planar correlated spectroscopic imaging (5D EP-COSI) and diffusion-weighted imaging (DWI). Twenty-four different metabolite and lipid ratios with respect to diagonal fat peaks (1.4 ppm, 5.4 ppm) from 2D spectra, and water and fat peaks (4.7 ppm, 1.4 ppm) from one-dimensional non-water-suppressed (NWS) spectra were used as the features. Additionally, water fraction, fat fraction and water-to-fat ratios from NWS spectra and apparent diffusion coefficients (ADC) from DWI were included. The nine most important features were identified using recursive feature elimination, sequential forward selection and correlation analysis. XGBoost (AUC: 93.0%, Accuracy: 85.7%, F1-score: 88.9%, Precision: 88.2%, Sensitivity: 90.4%, Specificity: 84.6%) and GradientBoost (AUC: 94.3%, Accuracy: 89.3%, F1-score: 90.7%, Precision: 87.9%, Sensitivity: 94.2%, Specificity: 83.4%) were the best-performing models. Conventional biomarkers like choline, myo-Inositol, and glycine were statistically significant predictors. Key features contributing to the classification were ADC, 2D diagonal peaks at 0.9 ppm, 2.1 ppm, 3.5 ppm, and 5.4 ppm, cross peaks between 1.4 and 0.9 ppm, 4.3 and 4.1 ppm, 2.3 and 1.6 ppm, and the triglyceryl-fat cross peak. The results highlight the contribution of the 2D spectral peaks to the model, and they demonstrate the potential of 5D EP-COSI for early breast cancer detection.

Differentiation of astrocytoma between grades II and III using a combination of methionine positron emission tomography and magnetic resonance spectroscopy

Objective

This study aimed to establish a method for differentiating between grades II and III astrocytomas using preoperative imaging.

Methods

We retrospectively analyzed astrocytic tumors, including 18 grade II astrocytomas (isocitrate dehydrogenase (IDH)-mutant: IDH-wildtype = 8:10) and 56 grade III anaplastic astrocytomas (37:19). We recorded the maximum methionine (MET) uptake ratios (tumor-to-normal: T/N) on positron emission tomography (PET) and three MRS peak ratios: choline (Cho)/creatine (Cr), N-acetyl aspartate (NAA)/Cr, and Cho/NAA, between June 2015 and June 2020. We then evaluated the cut-off values to differentiate between grades II and III. We compared the grading results between contrast enhancement effects on MR and combinational diagnostic methods (CDM) on a scatter chart using the cutoff values of the T/N ratio and MRS parameters.

Results

The IDH-mutant group showed significant differences in the Cho/NAA ratio between grades II and III using univariate analysis; however, multiple regression analysis results negated this. The IDH-wildtype group showed no significant differences between the groups. Contrast enhancement effects also showed no significant differences in IDH status. Accordingly, regardless of the IDH status, no statistically independent factors differentiated between grades II and III. However, CDMs showed higher sensitivity and negative predictive value in distinguishing them than MRI contrast examinations for both IDH statuses. We demonstrated a significantly higher diagnostic rate of grade III than of grade II with CDM, which was more striking in the IDH-mutant group than in the wild-type group.

Conclusions

CDM could be valuable in differentiating between grade II and III astrocytic tumors.

Applications of Artificial Intelligence Based on Medical Imaging in Glioma: Current State and Future Challenges

Glioma is one of the most fatal primary brain tumors, and it is well-known for its difficulty in diagnosis and management. Medical imaging techniques such as magnetic resonance imaging (MRI), positron emission tomography (PET), and spectral imaging can efficiently aid physicians in diagnosing, treating, and evaluating patients with gliomas. With the increasing clinical records and digital images, the application of artificial intelligence (AI) based on medical imaging has reduced the burden on physicians treating gliomas even further. This review will classify AI technologies and procedures used in medical imaging analysis. Additionally, we will discuss the applications of AI in glioma, including tumor segmentation and classification, prediction of genetic markers, and prediction of treatment response and prognosis, using MRI, PET, and spectral imaging. Despite the benefits of AI in clinical applications, several issues such as data management, incomprehension, safety, clinical efficacy evaluation, and ethical or legal considerations, remain to be solved. In the future, doctors and researchers should collaborate to solve these issues, with a particular emphasis on interdisciplinary teamwork.

Molecular imaging in oncology: Current impact and future directions

The authors define molecular imaging, according to the Society of Nuclear Medicine and Molecular Imaging, as the visualization, characterization, and measurement of biological processes at the molecular and cellular levels in humans and other living systems. Although practiced for many years clinically in nuclear medicine, expansion to other imaging modalities began roughly 25 years ago and has accelerated since. That acceleration derives from the continual appearance of new and highly relevant animal models of human disease, increasingly sensitive imaging devices, high-throughput methods to discover and optimize affinity agents to key cellular targets, new ways to manipulate genetic material, and expanded use of cloud computing. Greater interest by scientists in allied fields, such as chemistry, biomedical engineering, and immunology, as well as increased attention by the pharmaceutical industry, have likewise contributed to the boom in activity in recent years. Whereas researchers and clinicians have applied molecular imaging to a variety of physiologic processes and disease states, here, the authors focus on oncology, arguably where it has made its greatest impact. The main purpose of imaging in oncology is early detection to enable interception if not prevention of full-blown disease, such as the appearance of metastases. Because biochemical changes occur before changes in anatomy, molecular imaging-particularly when combined with liquid biopsy for screening purposes-promises especially early localization of disease for optimum management. Here, the authors introduce the ways and indications in which molecular imaging can be undertaken, the tools used and under development, and near-term challenges and opportunities in oncology.

Radiomics Analysis of Contrast-Enhanced CT for Hepatocellular Carcinoma Grading

Objectives

To investigate the value of contrast-enhanced computer tomography (CT)-based on radiomics in discriminating high-grade and low-grade hepatocellular carcinoma (HCC) before surgery.

Methods

The retrospective study including 161 consecutive subjects with HCC which was approved by the institutional review board, and the patients were divided into a training group (n = 112) and test group (n = 49) from January 2013 to January 2018. The least absolute shrinkage and selection operator (LASSO) was used to select the most valuable features to build a support vector machine (SVM) model. The performance of the predictive model was evaluated using the area under the curve (AUC), accuracy, sensitivity, and specificity.

Results

The SVM model showed an acceptable ability to differentiate high-grade from low-grade HCC, with an AUC of 0.904 in the training dataset and 0.937 in the test dataset, accuracy (92.2% versus 95.7%), sensitivity(82.5% versus 88.0%), and specificity (92.7% versus 95.8%), respectively.

Conclusion

The machine learning-based radiomics reflects a better evaluating performance in differentiating HCC between low-grade and high-grade, which may contribute to personalized treatment.

Multi-Modal Imaging in Down's Syndrome: Maximizing Utility Through Innovative Neuroimaging Approaches

In recent decades, the field of neuroimaging has experienced a surge of popularity and innovation which has led to significant advancements in the understanding of neurological disease, if not immediate clinical translation. In the case of Down's syndrome, a complex interplay of neurodevelopmental and neurodegenerative processes occur as a result of the trisomy of chromosome 21. The substantial potential impact of improved clinical intervention and the limited research under-taken to date make it a prime candidate for longitudinal neuroimaging-based study. However, as with a multitude of other multifaceted brain-based disorders, singular utilization of lone modality imaging has limited interpretability and applicability. Indeed, a present challenge facing the neuroimaging community as a whole is the methodological integration of multi-modal imaging to enhance clinical understanding. This review therefore aims to assess the current literature in Down's syndrome utilizing a multi-modal approach with regards to improvement upon consideration of a single modality. Additionally, we discuss potential avenues of future research that may effectively combine structural, functional and molecular-based imaging techniques for the significant benefit of the understanding of Down's syndrome pathology.

Thin-Slice Magnetic Resonance Imaging-Based Radiomics Signature Predicts Chromosomal 1p/19q Co-deletion Status in Grade II and III Gliomas

Objective: Chromosomal 1p/19q co-deletion is recognized as a diagnostic, prognostic, and predictive biomarker in lower grade glioma (LGG). This study aims to construct a radiomics signature to non-invasively predict the 1p/19q co-deletion status in LGG.

Methods: Ninety-six patients with pathology-confirmed LGG were retrospectively included and randomly assigned into training (n = 78) and validation (n = 18) dataset. Three-dimensional contrast-enhanced T1 (3D-CE-T1)-weighted magnetic resonance (MR) images and T2-weighted MR images were acquired, and simulated-conventional contrast-enhanced T1 (SC-CE-T1)-weighted images were generated. One hundred and seven shape, first-order, and texture radiomics features were extracted from each imaging modality and selected using the least absolute shrinkage and selection operator on the training dataset. A 3D-radiomics signature based on 3D-CE-T1 and T2-weighted features and a simulated-conventional (SC) radiomics signature based on SC-CE-T1 and T2-weighted features were established using random forest. The radiomics signatures were validated independently and evaluated using receiver operating characteristic (ROC) curves. Tumors with IDH mutations were also separately assessed.

Results: Four radiomics features were selected to construct the 3D-radiomics signature and displayed accuracies of 0.897 and 0.833, areas under the ROC curves (AUCs) of 0.940 and 0.889 in the training and validation datasets, respectively. The SC-radiomics signature was constructed with 4 features, but the AUC values were lower than that of the 3D signature. In the IDH-mutated subgroup, the 3D-radiomics signature presented AUCs of 0.950–1.000.

Conclusions: The MRI-based radiomics signature can differentiate 1p/19q co-deletion status in LGG with or without predetermined IDH status. 3D-CE-T1-weighted radiomics features are more favorable than SC-CE-T1-weighted features in the establishment of radiomics signatures.

Broad Learning Enhanced 1H-MRS for Early Diagnosis of Neuropsychiatric Systemic Lupus Erythematosus

In this paper, we explore the potential of using the multivoxel proton magnetic resonance spectroscopy (¹H-MRS) to diagnose neuropsychiatric systemic lupus erythematosus (NPSLE) with the assistance of a support vector machine broad learning system (BL-SVM). We retrospectively analysed 23 confirmed patients and 16 healthy controls, who underwent a 3.0 T magnetic resonance imaging (MRI) sequence with multivoxel ¹H-MRS in our hospitals. One hundred and seventeen metabolic features were extracted from the multivoxel ¹H-MRS image. Thirty-three metabolic features selected by the Mann-Whitney U test were considered to have a statistically significant difference (p < 0.05). However, the best accuracy achieved by conventional statistical methods using these 33 metabolic features was only 77%. We turned to develop a support vector machine broad learning system (BL-SVM) to quantitatively analyse the metabolic features from ¹H-MRS. Although not all the individual features manifested statistics significantly, the BL-SVM could still learn to distinguish the NPSLE from the healthy controls. The area under the receiver operating characteristic curve (AUC), the sensitivity, and the specificity of our BL-SVM in predicting NPSLE were 95%, 95.8%, and 93%, respectively, by 3-fold cross-validation. We consequently conclude that the proposed system effectively and efficiently working on limited and noisy samples may brighten a noinvasive in vivo instrument for early diagnosis of NPSLE.