Radiomics model building from multiparametric MRI to predict Ki-67 expression in patients with primary central nervous system lymphomas: a multicenter study DOI Creative Commons

Yelong Shen,

Siyu Wu,

Yanan Wu

et al.

BMC Medical Imaging, Journal Year: 2025, Volume and Issue: 25(1)

Published: Feb. 17, 2025

To examine the correlation of apparent diffusion coefficient (ADC), weighted imaging (DWI), and T1 contrast enhanced (T1-CE) with Ki-67 in primary central nervous system lymphomas (PCNSL). And to assess diagnostic performance MRI radiomics-based machine-learning algorithms differentiating high proliferation low groups PCNSL. 83 patients PCNSL were included this retrospective study. ADC, DWI T1-CE sequences collected their was examined using Spearman's analysis. The Kaplan-Meier method log-rank test used compare survival rates groups. radiomics features extracted respectively, screened by machine learning algorithm statistical method. Radiomics models seven different sequence permutations constructed. area under receiver operating characteristic curve (ROC AUC) evaluate predictive all models. DeLong utilized differences Relative mean (rADCmean) (ρ=-0.354, p = 0.019), relative (rDWImean) (b 1000) (ρ 0.273, 0.013) enhancement (rT1-CEmean) 0.385, 0.001) significantly correlated Ki-67. Interobserver agreements between two radiologists almost perfect for parameters (rADCmean ICC 0.978, 95%CI 0.966–0.986; rDWImean 0.931, 95% CI 0.895–0.955; rT1-CEmean 0.969, 0.953–0.980). PFS (p 0.016) OS 0.014) statistically significant. best prediction model our study a combination DWI, achieving highest AUC 0.869, while second ranked ADC an 0.828. rDWImean, rADCmean based on combined is promising distinguish from

Language: Английский

Treatment Response and Prognosis Evaluation in High‐Grade Glioma: An Imaging Review Based on MRI DOI
Qing Zhou, Caiqiang Xue,

Xiaoai Ke

et al.

Journal of Magnetic Resonance Imaging, Journal Year: 2022, Volume and Issue: 56(2), P. 325 - 340

Published: Feb. 7, 2022

In recent years, the development of advanced magnetic resonance imaging (MRI) technology and machine learning (ML) have created new tools for evaluating treatment response prognosis patients with high‐grade gliomas (HGG); however, patient has not improved significantly. This is mainly due to heterogeneity between within HGG tumors, resulting in standard methods benefitting all patients. Moreover, survival only related tumor cells, but also noncancer cells microenvironment (TME). Therefore, during preoperative diagnosis follow‐up HGG, noninvasive markers are needed characterize intratumoral heterogeneity, then evaluate predict prognosis, timeously adjust strategies, achieve individualized treatment. this review, we summarize research progress conventional MRI, MRI technology, ML evaluation HGG. We further discuss significance TME patients, associate features TME, indirectly reflecting tumor, shifting strategies from alone systemic therapy which may be a major direction future. Level Evidence 5 Technical Efficacy Stage 4

Language: Английский

Citations

29
Time-to-event overall survival prediction in glioblastoma multiforme patients using magnetic resonance imaging radiomics DOI Creative Commons
Ghasem Hajianfar, Atlas Haddadi Avval, Seyyed Ali Hosseini

et al.

La radiologia medica, Journal Year: 2023, Volume and Issue: 128(12), P. 1521 - 1534

Published: Sept. 26, 2023

Abstract Purpose Glioblastoma Multiforme (GBM) represents the predominant aggressive primary tumor of brain with short overall survival (OS) time. We aim to assess potential radiomic features in predicting time-to-event OS patients GBM using machine learning (ML) algorithms. Materials and methods One hundred nineteen GBM, who had T1-weighted contrast-enhanced T2-FLAIR MRI sequences, along clinical data time, were enrolled. Image preprocessing included 64 bin discretization, Laplacian Gaussian (LOG) filters three Sigma values eight variations Wavelet Transform. Images then segmented, followed by extraction 1212 features. Seven feature selection (FS) six ML algorithms utilized. The combination preprocessing, FS, (12 × 7 6 = 504 models) was evaluated multivariate analysis. Results Our analysis showed that best prognostic FS/ML combinations are Mutual Information (MI)/Cox Boost, MI/Generalized Linear Model Boosting (GLMB) Network (GLMN), all which done via LOG (Sigma 1 mm) method (C-index 0.77). filter mm method, MI, GLMB GLMN achieved significantly higher C-indices than other (all p < 0.05, mean 0.65, 0.70, 0.64, respectively). Conclusion capable MRI-based radiomics variables might appear promising assisting clinicians prediction GBM. Further research is needed establish applicability management clinic.

Language: Английский

Citations

21
Grading diffuse glioma based on 2021 WHO grade using self-attention-base deep learning architecture: variable Vision Transformer (vViT) DOI Creative Commons
Takuma Usuzaki, Kengo Takahashi, Ryusei Inamori

et al.

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 91, P. 106001 - 106001

Published: Feb. 6, 2024

To evaluate the diagnostic performance of self-attention-based model, termed variable Vision Transformer (vViT), in task predicting grade diffuse glioma based on 2021 World Health Organization (WHO) central nervous system (CNS) tumor classification. This cross-sectional study analyzed adult patients with histopathologically confirmed glioma, following WHO CNS We used age, sex, radiomic features, and four MRI sequences to predict gliomas. As binary classifications, we constructed three models: 2 vs. 3/4 (326 1575 1574 images), 3 2/4 (330 1726 4 2/3 (333 3292 images). a multiclass classification, model (334 2, 3, Metrics including accuracy area under curve receiver operating characteristic (AUC-ROC) were calculated. The highest AUC-ROC 0.84 (95% confidence interval; 0.75–0.93) classification (2 4) 0.94 (0.88–0.98) 2/3, respectively. Cohen's κ coefficient between ground truth predicted value was 0.54 obtained 4). vViT is competent multi-modal deep-learning that can gliomas which classified

Language: Английский

Citations

7
Artificial intelligence in paediatric radiology: Future opportunities DOI Open Access
Natasha Davendralingam, Neil J. Sebire, Owen J. Arthurs

et al.

British Journal of Radiology, Journal Year: 2020, Volume and Issue: 94(1117)

Published: Sept. 17, 2020

Artificial intelligence (AI) has received widespread and growing interest in healthcare, as a method to save time, cost improve efficiencies. The high-performance statistics diagnostic accuracies reported by using AI algorithms (with respect predefined reference standards), particularly from image pattern recognition studies, have resulted extensive applications proposed for clinical radiology, especially enhanced interpretation. Whilst certain sub-speciality areas such those relating cancer screening, wide-spread attention the media scientific community, children’s imaging been hitherto neglected. In this article, we discuss variety of possible ‘use cases’ paediatric radiology patient pathway perspective where either implemented or shown early-stage feasibility, while also taking inspiration adult literature propose potential future development. We aim demonstrate how ‘future, service’ could operate stimulate further discussion with avenues research.

Language: Английский

Citations

43
Deep Learning Aided Neuroimaging and Brain Regulation DOI Creative Commons
Mengze Xu,

Yuanyuan Ouyang,

Zhen Yuan

et al.

Sensors, Journal Year: 2023, Volume and Issue: 23(11), P. 4993 - 4993

Published: May 23, 2023

Currently, deep learning aided medical imaging is becoming the hot spot of AI frontier application and future development trend precision neuroscience. This review aimed to render comprehensive informative insights into recent progress its applications in for brain monitoring regulation. The article starts by providing an overview current methods imaging, highlighting their limitations introducing potential benefits using techniques overcome these limitations. Then, we further delve details learning, explaining basic concepts examples how it can be used imaging. One key strengths thorough discussion different types models that including convolutional neural networks (CNNs), recurrent (RNNs), generative adversarial network (GAN) assisted magnetic resonance (MRI), positron emission tomography (PET)/computed (CT), electroencephalography (EEG)/magnetoencephalography (MEG), optical other modalities. Overall, our on regulation provides a referrable glance intersection neuroimaging

Language: Английский

Citations

16
Advances on Liquid Biopsy Analysis for Glioma Diagnosis DOI Creative Commons
Panagiotis Skouras,

Mariam Markouli,

Theodosis Κalamatianos

et al.

Biomedicines, Journal Year: 2023, Volume and Issue: 11(9), P. 2371 - 2371

Published: Aug. 24, 2023

Gliomas comprise the most frequent primary central nervous system (CNS) tumors, characterized by remarkable genetic and epigenetic heterogeneity, difficulty in monitoring, increased relapse mortality rates. Tissue biopsy is an established method of tumor cell collection analysis that enables diagnosis, classification different types, prediction prognosis upon confirmation tumor's location for surgical removal. However, it invasive often challenging procedure cannot be used patient screening, detection mutations, disease or resistance to therapy. To this end, minimally liquid has emerged, allowing effortless sampling enabling continuous monitoring. It considered a novel preferable way obtain faster data on potential risk, personalized prognosis, recurrence evaluation. The purpose review describe advances glioma diagnosis management, indicating several biomarkers can utilized analyze characteristics, such as cell-free DNA (cfDNA), RNA (cfRNA), circulating proteins, cells (CTCs), exosomes. further addresses benefit combining with radiogenomics facilitate early accurate diagnoses, enable precise prognostic assessments, real-time aiming towards more optimal treatment decisions.

Language: Английский

Citations

16
Machine-Learning and Radiomics-Based Preoperative Prediction of Ki-67 Expression in Glioma Using MRI Data DOI
Jiaying Ni, Hongjian Zhang, Qing Yang

et al.

Academic Radiology, Journal Year: 2024, Volume and Issue: 31(8), P. 3397 - 3405

Published: March 7, 2024

Language: Английский

Citations

5
Challenges and opportunities for artificial intelligence in oncological imaging DOI
Helen Cheung, Daniel L. Rubin

Clinical Radiology, Journal Year: 2021, Volume and Issue: 76(10), P. 728 - 736

Published: April 24, 2021

Language: Английский

Citations

28
Application of Artificial Intelligence Methods for Imaging of Spinal Metastasis DOI Open Access

Wilson Ong,

Lei Zhu, Wenqiao Zhang

et al.

Cancers, Journal Year: 2022, Volume and Issue: 14(16), P. 4025 - 4025

Published: Aug. 20, 2022

Spinal metastasis is the most common malignant disease of spine. Recently, major advances in machine learning and artificial intelligence technology have led to their increased use oncological imaging. The purpose this study review summarise present evidence for applications detection, classification management spinal metastasis, along with potential integration into clinical practice. A systematic, detailed search main electronic medical databases was undertaken concordance PRISMA guidelines. total 30 articles were retrieved from database reviewed. Key findings current AI compiled summarised. techniques include image processing, diagnosis, decision support, treatment assistance prognostic outcomes. In realm oncology, technologies achieved relatively good performance hold immense aid clinicians, including enhancing work efficiency reducing adverse events. Further research required validate tools facilitate routine

Language: Английский

Citations

22
Preoperative Diagnosis and Molecular Characterization of Gliomas With Liquid Biopsy and Radiogenomics DOI Creative Commons
Carmen Balañá, Sara Castañer,

Cristina Carrato

et al.

Frontiers in Neurology, Journal Year: 2022, Volume and Issue: 13

Published: May 26, 2022

Gliomas are a heterogenous group of central nervous system tumors with different outcomes and therapeutic needs. Glioblastoma, the most common subtype in adults, has very poor prognosis disabling consequences. The World Health Organization (WHO) classification specifies that typing grading gliomas should include molecular markers. characterization implications for prognosis, treatment planning, prediction response. At present, diagnosed via tumor resection or biopsy, which always invasive frequently risky methods. In recent years, however, substantial advances have been made developing methods through analysis products shed body fluids. Known as liquid biopsies, these analyses can potentially provide diagnostic prognostic information, guidance on choice treatment, real-time information status. addition, magnetic resonance imaging (MRI) is another good source data; radiomics radiogenomics link phenotypes to gene expression patterns insights biology underlying signatures. Machine deep learning computational techniques also use quantitative features non-invasively detect genetic mutations. key obtained biopsies be useful not only diagnosis but help predict response specific treatments guidelines personalized medicine. this article, we review available data using non-invasive biopsy MRI suggest tools could used future preoperative gliomas.

Language: Английский

Citations

21