Machine learning-based models for prediction of survival in medulloblastoma: a systematic review and meta-analysis

Machine learning models in the prediction of chronic or shunt-dependent hydrocephalus following subarachnoid hemorrhage: A systematic review and meta-analysis

PurposeChronic or shunt-dependent hydrocephalus is a frequent consequence of subarachnoid hemorrhage (SAH) with an unclear pathophysiology, making treatment challenging. Despite favorable outcomes following cerebrospinal fluid (CSF) diversion, high-risk surgical interventions remain necessary in some cases. Accurate prediction of chronic or shunt-dependent hydrocephalus in SAH patients can play an important role in their management. This systematic review and meta-analysis assessed the predictive performance of machine learning (ML) models in forecasting chronic or shunt-dependent hydrocephalus following SAH.MethodsA systematic search of PubMed, Embase, Scopus, and Web of Science was conducted. ML or deep learning (DL)-based models that predicted chronic or shunt-dependent hydrocephalus following SAH were included. To avoid bias, only the data of the best-performance model, which was defined by the highest area under the curve (AUC) of the models, were extracted. The pooled AUC, accuracy (ACC), sensitivity, specificity, and diagnostic odds ratio (DOR) were calculated using the R program.ResultsSix studies with 2096 individuals were included. The AUC, ACC, sensitivity, and specificity ranged from 0.8 to 0.92, 0.72 to 0.9, 0.73 to 0.85, and 0.7 to 0.92. The meta-analysis showed a pooled AUC of 0.83 (95%CI: 0.81-0.84) and ACC of 0.79 (95%CI: 0.66-0.91). The meta-analysis revealed a pooled sensitivity of 0.8 (95%CI: 0.73-0.85), specificity of 0.79 (95%CI: 0.68-0.86), and DOR of 12.13 (95%CI: 8.2-17.96) for predictive performance of these models.ConclusionML-based models showed encouraging predictive performance in forecasting chronic or shunt-dependent hydrocephalus following SAH.

Integrative Machine Learning Framework for Enhanced Subgroup Classification in Medulloblastoma

BACKGROUND

Medulloblastoma is the most common malignant brain tumor in children, classified into four primary molecular subgroups: WNT, SHH, Group 3, and Group 4, each exhibiting significant molecular heterogeneity and varied survival outcomes. Accurate classification of these subgroups is crucial for optimizing treatments and improving patient outcomes. DNA methylation profiling is a promising approach for subgroup classification; however, its application is still evolving, with ongoing efforts to improve accessibility and develop more accurate classification methods.

OBJECTIVES

This study aims to develop a supervised machine learning-based framework using Illumina 450K methylation data to classify medulloblastoma into seven molecular subgroups: WNT, SHH-Infant, SHH-Child, Group3-LowRisk, Group3-HighRisk, Group4-LowRisk, and Group4-HighRisk, incorporating age and risk factors for enhanced subgroup differentiation.

METHODS

The proposed model leverages six metagenes, capturing the underlying patterns of the top 10,000 probes with the highest variances from Illumina 450K data, thus enhancing methylation data representation while reducing computational demands.

RESULTS

Among the models evaluated, the SVM achieved the highest performance, with a mean balanced accuracy 98% and a macro-averaged AUC of 0.99 in an independent validation. This suggests that the model effectively captures the relevant methylation patterns for medulloblastoma subgroup classification.

CONCLUSIONS

The developed SVM-based model provides a robust framework for accurate classification of medulloblastoma subgroups using DNA methylation data. Integrating this model into clinical decision making could enhance subgroup-directed therapies and improve patient outcomes.

Deep Learning-Based Models for Ventricular Segmentation in Hydrocephalus: A Systematic Review and Meta-Analysis

BACKGROUND

Ventricular segmentation is a critical step in neuroimaging data evaluation, particularly in hydrocephalus. Current methods are mainly based on 2-dimensional measurements and ratios. Traditional manual and semiautomatic ventricular segmentation are time-consuming, operator-based, and lack flexibility in handling numerous radiological features. Recently, deep learning (DL) models have been developed to perform ventricular segmentation and have shown promising outcomes. The objective of the current study was to evaluate the performance of DL-based models in ventricular segmentation in the hydrocephalus setting.

METHODS

On December 5, 2024, a systematic search was conducted using an individualized search query in 4 electronic databases: PubMed, Embase, Scopus, and Web of Science. Studies that reported the mean dice similarity coefficient (DSC) of DL-based models in ventricular segmentation in patients with hydrocephalus were included. The mean DSC for the best-performance model was extracted.

RESULTS

Twenty-four studies with 2911 patients were included. The mean DSC ranged from 0.671 to 0.99 across the best-performance models. The meta-analysis revealed a pooled mean DSC of 0.89 (95% CI: 0.84-92). The subgroup analysis yielded a pooled mean DSC of 0.88 (95% CI: 0.80-0.96) for magnetic resonance imaging-based models, 0.91 (95% CI: 0.86-0.95) for computed tomography-based models, and 0.84 (95% CI: 0.81-0.87) for ultrasound-based best-performance DL-based models.

CONCLUSIONS

DL-based models have demonstrated favorable outcomes in ventricular segmentation in patients with hydrocephalus. Application of these models in clinical practice can optimize the treatment protocol and enhance the clinical outcomes of hydrocephalus patients.

Machine learning radiomics for H3K27M mutation prediction in gliomas: A systematic review and meta-analysis

PURPOSE

Noninvasive prediction and identification of the H3K27M mutation play an important role in optimizing therapeutic strategies and improving outcomes in gliomas. In this systematic review and meta-analysis, we aimed to evaluate the performance of machine learning (ML)-based models in predicting H3K27M mutation in gliomas.

METHODS

Literature records were retrieved on September 16th, 2024, in PubMed, Embase, Scopus, and Web of Science. Records were screened according to the eligibility criteria, and the data from the included studies were extracted. The meta-analysis, sensitivity analysis, and meta-regression were conducted using R software.

RESULTS

A total of 15 studies were included in our study. Our meta-analysis demonstrated a pooled AUC, sensitivity, and specificity of 0.87 (95% CI: 0.77-0.97), 92% (95% CI: 83%-96%), and 89% (95% CI: 86%-91%)), respectively. The subgroup meta-analysis revealed that despite the higher sensitivity of the deep learning (DL) models, the sensitivity is not superior to ML (P = 0.6). In contrast, the ML-based pooled specificity was significantly higher (P < 0.01). The meta-analysis revealed a 78.1 (95% CI: 33.3 - 183.5). The SROC curve indicated an AUC of 0.921, and the estimated sensitivity is 0.898 concurrent with the false positive rate of 0.126, which indicates high sensitivity with a low false positive rate.

CONCLUSION

Our systematic review and meta-analysis demonstrated that ML-based magnetic resonance imaging (MRI) radiomics models are associated with promising diagnostic performance in predicting H3K27M mutation in gliomas.

Prediction of facial nerve outcomes after surgery for vestibular schwannoma using machine learning-based models: a systematic review and meta-analysis

Postoperative facial nerve (FN) dysfunction is associated with a significant impact on the quality of life of patients and can result in psychological stress and disorders such as depression and social isolation. Preoperative prediction of FN outcomes can play a critical role in vestibular schwannomas (VSs) patient care. Several studies have developed machine learning (ML)-based models in predicting FN outcomes following resection of VS. This systematic review and meta-analysis aimed to evaluate the diagnostic accuracy of ML-based models in predicting FN outcomes following resection in the setting of VS. On December 12, 2024, the four electronic databases, Pubmed, Embase, Scopus, and Web of Science, were systematically searched. Studies that evaluated the performance outcomes of the ML-based predictive models were included. The pooled sensitivity, specificity, area under the curve (AUC), and diagnostic odds ratio (DOR) were calculated through the R program. Five studies with 807 individuals with VS, encompassing 35 models, were included. The meta-analysis showed a pooled sensitivity of 82% (95%CI: 76-87%), specificity of 79% (95%CI: 74-84%), and DOR of 12.94 (95%CI: 8.65-19.34) with an AUC of 0.841. The meta-analysis of the best performance model demonstrated a pooled sensitivity of 91% (95%CI: 80-96%), specificity of 87% (95%CI: 82-91%), and DOR of 46.84 (95%CI: 19.8-110.8). Additionally, the analysis demonstrated an AUC of 0.92, a sensitivity of 0.884, and a false positive rate of 0.136 for the best performance models. ML-based models possess promising diagnostic accuracy in predicting FN outcomes following resection.

Long Non-Coding RNAs in Malignant Human Brain Tumors: Driving Forces Behind Progression and Therapy

Long non-coding RNAs (lncRNAs) play a pivotal role in regulating gene expression and are critically involved in the progression of malignant brain tumors, including glioblastoma, medulloblastoma, and meningioma. These lncRNAs interact with microRNAs (miRNAs), proteins, and DNA, influencing key processes such as cell proliferation, migration, and invasion. This review highlights the multifaceted impact of lncRNA dysregulation on tumor progression and underscores their potential as therapeutic targets to enhance the efficacy of chemotherapy, radiotherapy, and immunotherapy. The insights provided offer new directions for advancing basic research and clinical applications in malignant brain tumors.