Unsupervised machine learning using K-means identifies radiomic subgroups of pediatric low-grade gliomas that correlate with key molecular markers

Haldar, Debanjan; Kazerooni, Anahita Fathi; Arif, Sherjeel; Familiar, Ariana; Madhogarhia, Rachel; Khalili, Nastaran; Bagheri, Sina; Anderson, Hannah; Shaikh, Ibraheem S; Mahtabfar, Aria; Kim, Meen Chul; Tu, Wenxin; Ware, Jefferey; Vossough, Arastoo; Davatzikos, Christos; Storm, Phillip B.; Resnick, Adam; Nabavizadeh, Ali

doi:10.1016/j.neo.2022.100869

Cited by 12 publications

(4 citation statements)

References 21 publications

(26 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Radiomics has the ability to capture the breadth of heterogeneity in each tumor to aide in clinical treatment. Unsupervised techniques using radiomic feature clusterings in pediatric low grade gliomas correlate with molecular alterations [12]. In our study, we extracted interpretable tumor characteristics derived from shape, size, and texture features.…”

Section: Discussionmentioning

confidence: 99%

A machine learning approach to prediction of HER2/PR/ER status in metastatic breast cancer to the brain from magnetic resonance imaging.

Sabal,

Venteicher,

Taha

2024

Preprint

View full text Add to dashboard Cite

Introduction Breast cancer brain metastases (BCBM) are a clinical challenge, with 15–25% incidence among patients with metastatic breast cancer. Prediction of receptor status in BCBM is crucial for personalized treatment strategies. This study addresses the limitations of invasive biopsies and explores the use of machine learning techniques to predict BCBM receptor status based on primary breast cancer histology. Methods 1135 lesions from 196 scans and 173 unique patients were analyzed. Genetic information was obtained using next-generation sequencing or immunohistochemistry. We employed machine learning algorithms to predict receptor status from radiomic features extracted from T1-weighted post-contrast MRI images. Results Random Forest classifier demonstrated superior performance in predicting HER2 and ER status. The 'Minimum' feature from radiomic analysis was the most significant in determining mutation status. Unsupervised analysis showed distinct clustering for certain genetic combinations. Conclusion Machine learning models, particularly the Random Forest classifier, can effectively predict HER2 and ER receptor status in BCBM from MRI radiomic features. This approach could offer a pathway toward personalized therapy and potentially improved patient outcomes. This study is limited by known receptor discordance between primary breast lesions and their associated brain metastases. Further validation across diverse populations and multicenter studies is necessary.

show abstract

Section: Discussionmentioning

confidence: 99%

A machine learning approach to prediction of HER2/PR/ER status in metastatic breast cancer to the brain from magnetic resonance imaging.

Sabal,

Venteicher,

Taha

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…This may enable more confidence for empiric treatment with targeted therapies if tissue diagnosis is infeasible. pLGG mutational classification has been previously attempted in a few studies, most with manual segmentation-derived and/or pre-engineered radiomics (35)(36)(37)(38), which are known to fail when applied to the external dataset. Radiomic features have been extracted from MRI images and fitted to classifiers models like XGboost and SVM (17,35,36).…”

Section: Discussionmentioning

confidence: 99%

Noninvasive molecular subtyping of pediatric low-grade glioma with self-supervised transfer learning

Tak

Zapaishchykova

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

PurposeTo develop and externally validate a scan-to-prediction deep-learning pipeline for noninvasive, MRI-based BRAF mutational status classification for pLGG.Materials and MethodsWe conducted a retrospective study of two pLGG datasets with linked genomic and diagnostic T2-weighted MRI of patients: Boston Children’s Hospital (development dataset, N=214), and Child Brain Tumor Network (CBTN) (external validation, N=112). We developed a deep learning pipeline to classify BRAF mutational status (V600E vs. fusion vs. wild-type) from whole-scan input via a two-stage process: 1) 3D tumor segmentation and extraction of axial tumor images, and 2) slice-wise, deep learning-based classification of mutational status. We investigated knowledge-transfer approaches to prevent model overfitting with a primary endpoint of the area under the receiver operating characteristic curve (AUC). To enhance model interpretability, we developed a novel metric, COMDist that quantifies the accuracy of model attention with respect to the tumor.ResultsA combination of transfer learning from a pretrained medical imaging-specific network and self-supervised label cross-training (TransferX) coupled with consensus logic yielded the highest AUC, taken as a weighted average across the three mutational classes, (0.82 [95% CI: 0.70-0.90], Accuracy: 77%) on internal validation and (0.73 [95% CI 0.68-0.88], Accuracy: 75%) on external validation. Training with TransferX also led to an AUC improvement of 17.7% and a COMDist Improvement of 6.42% over training from scratch on the development dataset.ConclusionTransfer learning and self-supervised cross-training improved classification performance and generalizability for noninvasive pLGG mutational status prediction in a limited data scenario.Summary StatementThe authors developed and externally validated an automated, scan-to-prediction deep learning pipeline that classifies BRAF Mutational status in pediatric low-grade gliomas directly from T2-Weighted MRI scans.Key ResultsAn innovative training approach combining self-supervision and transfer learning (“TransferX”) is developed to boost model performance in low data settings;TransferX enables the development of a scan-to-prediction pipeline for pediatric LGG mutational status (BRAF V600E, fusion, or wildtype) with high accuracy and mild performance degradation on external validation;An evaluation metric “COMDist” is proposed to increase interpretability and quantify the accuracy of the model’s attention around the tumor.

show abstract

“…Overall, 41/62 of the reviewed studies (66%) focused on predicting IDH mutation and 1p/19q codeletion status only, while 33 studies (53%) analyzed other molecular subgroups. These were TERT [9,37,[40][41][42][43][44][45][46], ATRX [8,[47][48][49][50][51], H3K27 [4,[15][16][17][18], MGMT [50,[52][53][54][55], P53 [8,16,51,53], CDKN2A/B [12,30,35,56], EGFR [36], chr7/10 [57] and BRAF alterations [3,[5][6][7]. The reported AUC values range from 0.6 to 0.98 for these predictions with an average of 0.82 to 0.9.…”

Section: Molecular Subgroupsmentioning

confidence: 99%

“…Most studies analyzed adult populations and high-grade gliomas, with only five studies (8%) analyzing pediatric populations [3][4][5][6][7], ten studies (16%) analyzing low-grade glioma [3,5,6,[8][9][10][11][12][13][14] and only four studies (6%) focusing on diffuse midline glioma [15][16][17][18]. Data on the analyzed patient population are shown in Table 2.…”

Section: Patient Populationmentioning

confidence: 99%

Novel Imaging Approaches for Glioma Classification in the Era of the World Health Organization 2021 Update: A Scoping Review

Richter,

Ernemann,

Bender

2024

Cancers

View full text Add to dashboard Cite

The 2021 WHO classification of CNS tumors is a challenge for neuroradiologists due to the central role of the molecular profile of tumors. The potential of novel data analysis tools in neuroimaging must be harnessed to maintain its role in predicting tumor subgroups. We performed a scoping review to determine current evidence and research gaps. A comprehensive literature search was conducted regarding glioma subgroups according to the 2021 WHO classification and the use of MRI, radiomics, machine learning, and deep learning algorithms. Sixty-two original articles were included and analyzed by extracting data on the study design and results. Only 8% of the studies included pediatric patients. Low-grade gliomas and diffuse midline gliomas were represented in one-third of the research papers. Public datasets were utilized in 22% of the studies. Conventional imaging sequences prevailed; data on functional MRI (DWI, PWI, CEST, etc.) are underrepresented. Multiparametric MRI yielded the best prediction results. IDH mutation and 1p/19q codeletion status prediction remain in focus with limited data on other molecular subgroups. Reported AUC values range from 0.6 to 0.98. Studies designed to assess generalizability are scarce. Performance is worse for smaller subgroups (e.g., 1p/19q codeleted or IDH1/2 mutated gliomas). More high-quality study designs with diversity in the analyzed population and techniques are needed.

show abstract

Unsupervised machine learning using K-means identifies radiomic subgroups of pediatric low-grade gliomas that correlate with key molecular markers

Cited by 12 publications

References 21 publications

A machine learning approach to prediction of HER2/PR/ER status in metastatic breast cancer to the brain from magnetic resonance imaging.

A machine learning approach to prediction of HER2/PR/ER status in metastatic breast cancer to the brain from magnetic resonance imaging.

Noninvasive molecular subtyping of pediatric low-grade glioma with self-supervised transfer learning

Novel Imaging Approaches for Glioma Classification in the Era of the World Health Organization 2021 Update: A Scoping Review

Contact Info

Product

Resources

About