{"title":"Radiomics models using machine learning algorithms to differentiate the primary focus of brain metastasis.","authors":"Yuping Xie, Xuanzi Li, Shuai Yang, Fujie Jia, Yuanyuan Han, Mingsheng Huang, Lei Chen, Wei Zou, Chuntao Deng, Zibin Liang","doi":"10.21037/tcr-24-1355","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Brain metastases are common brain tumors in adults. Brain metastases from different primary tumors have special magnetic resonance imaging (MRI) features. As a new technology that can extract and quantify medical image data, and with the rapid development of artificial intelligence, the machine learning model based on radiology has been successfully applied to the diagnosis and differentiation of tumors. This study aimed to develop radiomics models from post-contrast T1-weighted images using machine learning algorithms to differentiate lung cancer from breast cancer brain metastases.</p><p><strong>Methods: </strong>A retrospective analysis was conducted on 118 lung cancer brain metastases patients and 62 breast cancer brain metastases patients confirmed by surgery pathology or combined clinical and imaging diagnosis at The Fifth Affiliated Hospital of Sun Yat-sen University from August 2015 to September 2023. Patients were randomly divided into a training set (126 cases) and a validation set (54 cases) at a 7:3 ratio. Enhanced T1-weighted images of all patients were imported into ITK-SNAP software to manually delineate the region of interest (ROI). Radiomic features were extracted based on the ROI and feature selection was performed using the least absolute shrinkage and selection operator. Significant features were used to develop models using logistic regression (LR), support vector machine (SVM), K-nearest neighbors (KNN), multilayer perceptron (MLP), and light gradient boosting machine (LightGBM). The diagnostic performance of the models was assessed using the receiver operating characteristic (ROC) curve.</p><p><strong>Results: </strong>The LightGBM radiomics model exhibited the best diagnostic performance, with an area under the curve (AUC) of 0.875 [95% confidence interval (CI): 0.819-0.931] in the training set and 0.866 (95% CI: 0.740-0.993) in the validation set.</p><p><strong>Conclusions: </strong>The enhanced MRI radiomics model, especially the LightGBM model, can accurately predict the primary lesion types of brain metastases from lung cancer and breast cancer origins.</p>","PeriodicalId":23216,"journal":{"name":"Translational cancer research","volume":"14 2","pages":"731-742"},"PeriodicalIF":1.5000,"publicationDate":"2025-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11912090/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Translational cancer research","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.21037/tcr-24-1355","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/2/24 0:00:00","PubModel":"Epub","JCR":"Q4","JCRName":"ONCOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Background: Brain metastases are common brain tumors in adults. Brain metastases from different primary tumors have special magnetic resonance imaging (MRI) features. As a new technology that can extract and quantify medical image data, and with the rapid development of artificial intelligence, the machine learning model based on radiology has been successfully applied to the diagnosis and differentiation of tumors. This study aimed to develop radiomics models from post-contrast T1-weighted images using machine learning algorithms to differentiate lung cancer from breast cancer brain metastases.
Methods: A retrospective analysis was conducted on 118 lung cancer brain metastases patients and 62 breast cancer brain metastases patients confirmed by surgery pathology or combined clinical and imaging diagnosis at The Fifth Affiliated Hospital of Sun Yat-sen University from August 2015 to September 2023. Patients were randomly divided into a training set (126 cases) and a validation set (54 cases) at a 7:3 ratio. Enhanced T1-weighted images of all patients were imported into ITK-SNAP software to manually delineate the region of interest (ROI). Radiomic features were extracted based on the ROI and feature selection was performed using the least absolute shrinkage and selection operator. Significant features were used to develop models using logistic regression (LR), support vector machine (SVM), K-nearest neighbors (KNN), multilayer perceptron (MLP), and light gradient boosting machine (LightGBM). The diagnostic performance of the models was assessed using the receiver operating characteristic (ROC) curve.
Results: The LightGBM radiomics model exhibited the best diagnostic performance, with an area under the curve (AUC) of 0.875 [95% confidence interval (CI): 0.819-0.931] in the training set and 0.866 (95% CI: 0.740-0.993) in the validation set.
Conclusions: The enhanced MRI radiomics model, especially the LightGBM model, can accurately predict the primary lesion types of brain metastases from lung cancer and breast cancer origins.
期刊介绍:
Translational Cancer Research (Transl Cancer Res TCR; Print ISSN: 2218-676X; Online ISSN 2219-6803; http://tcr.amegroups.com/) is an Open Access, peer-reviewed journal, indexed in Science Citation Index Expanded (SCIE). TCR publishes laboratory studies of novel therapeutic interventions as well as clinical trials which evaluate new treatment paradigms for cancer; results of novel research investigations which bridge the laboratory and clinical settings including risk assessment, cellular and molecular characterization, prevention, detection, diagnosis and treatment of human cancers with the overall goal of improving the clinical care of cancer patients. The focus of TCR is original, peer-reviewed, science-based research that successfully advances clinical medicine toward the goal of improving patients'' quality of life. The editors and an international advisory group of scientists and clinician-scientists as well as other experts will hold TCR articles to the high-quality standards. We accept Original Articles as well as Review Articles, Editorials and Brief Articles.