Stacking Ensemble Learning-Based [18F]FDG PET Radiomics for Outcome Prediction in Diffuse Large B-Cell Lymphoma.

IF 9.1 1区 医学 Q1 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING Journal of Nuclear Medicine Pub Date : 2023-10-01 Epub Date: 2023-07-27 DOI:10.2967/jnumed.122.265244
Shuilin Zhao, Jing Wang, Chentao Jin, Xiang Zhang, Chenxi Xue, Rui Zhou, Yan Zhong, Yuwei Liu, Xuexin He, Youyou Zhou, Caiyun Xu, Lixia Zhang, Wenbin Qian, Hong Zhang, Xiaohui Zhang, Mei Tian
{"title":"Stacking Ensemble Learning-Based [<sup>18</sup>F]FDG PET Radiomics for Outcome Prediction in Diffuse Large B-Cell Lymphoma.","authors":"Shuilin Zhao,&nbsp;Jing Wang,&nbsp;Chentao Jin,&nbsp;Xiang Zhang,&nbsp;Chenxi Xue,&nbsp;Rui Zhou,&nbsp;Yan Zhong,&nbsp;Yuwei Liu,&nbsp;Xuexin He,&nbsp;Youyou Zhou,&nbsp;Caiyun Xu,&nbsp;Lixia Zhang,&nbsp;Wenbin Qian,&nbsp;Hong Zhang,&nbsp;Xiaohui Zhang,&nbsp;Mei Tian","doi":"10.2967/jnumed.122.265244","DOIUrl":null,"url":null,"abstract":"<p><p>This study aimed to develop an analytic approach based on [<sup>18</sup>F]FDG PET radiomics using stacking ensemble learning to improve the outcome prediction in diffuse large B-cell lymphoma (DLBCL). <b>Methods:</b> In total, 240 DLBCL patients from 2 medical centers were divided into the training set (<i>n</i> = 141), internal testing set (<i>n</i> = 61), and external testing set (<i>n</i> = 38). Radiomics features were extracted from pretreatment [<sup>18</sup>F]FDG PET scans at the patient level using 4 semiautomatic segmentation methods (SUV threshold of 2.5, SUV threshold of 4.0 [SUV4.0], 41% of SUV<sub>max</sub>, and SUV threshold of mean liver uptake [PERCIST]). All extracted features were harmonized with the ComBat method. The intraclass correlation coefficient was used to evaluate the reliability of radiomics features extracted by different segmentation methods. Features from the most reliable segmentation method were selected by Pearson correlation coefficient analysis and the LASSO (least absolute shrinkage and selection operator) algorithm. A stacking ensemble learning approach was applied to build radiomics-only and combined clinical-radiomics models for prediction of 2-y progression-free survival and overall survival based on 4 machine learning classifiers (support vector machine, random forests, gradient boosting decision tree, and adaptive boosting). Confusion matrix, receiver-operating-characteristic curve analysis, and survival analysis were used to evaluate the model performance. <b>Results:</b> Among 4 semiautomatic segmentation methods, SUV4.0 segmentation yielded the highest interobserver reliability, with 830 (66.7%) selected radiomics features. The combined model constructed by the stacking method achieved the best discrimination performance. For progression-free survival prediction in the external testing set, the areas under the receiver-operating-characteristic curve and accuracy of the stacking-based combined model were 0.771 and 0.789, respectively. For overall survival prediction, the stacking-based combined model achieved an area under the curve of 0.725 and an accuracy of 0.763 in the external testing set. The combined model also demonstrated a more distinct risk stratification than the International Prognostic Index in all sets (log-rank test, all <i>P</i> < 0.05). <b>Conclusion:</b> The combined model that incorporates [<sup>18</sup>F]FDG PET radiomics and clinical characteristics based on stacking ensemble learning could enable improved risk stratification in DLBCL.</p>","PeriodicalId":16758,"journal":{"name":"Journal of Nuclear Medicine","volume":null,"pages":null},"PeriodicalIF":9.1000,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Nuclear Medicine","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.2967/jnumed.122.265244","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2023/7/27 0:00:00","PubModel":"Epub","JCR":"Q1","JCRName":"RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING","Score":null,"Total":0}
引用次数: 0

Abstract

This study aimed to develop an analytic approach based on [18F]FDG PET radiomics using stacking ensemble learning to improve the outcome prediction in diffuse large B-cell lymphoma (DLBCL). Methods: In total, 240 DLBCL patients from 2 medical centers were divided into the training set (n = 141), internal testing set (n = 61), and external testing set (n = 38). Radiomics features were extracted from pretreatment [18F]FDG PET scans at the patient level using 4 semiautomatic segmentation methods (SUV threshold of 2.5, SUV threshold of 4.0 [SUV4.0], 41% of SUVmax, and SUV threshold of mean liver uptake [PERCIST]). All extracted features were harmonized with the ComBat method. The intraclass correlation coefficient was used to evaluate the reliability of radiomics features extracted by different segmentation methods. Features from the most reliable segmentation method were selected by Pearson correlation coefficient analysis and the LASSO (least absolute shrinkage and selection operator) algorithm. A stacking ensemble learning approach was applied to build radiomics-only and combined clinical-radiomics models for prediction of 2-y progression-free survival and overall survival based on 4 machine learning classifiers (support vector machine, random forests, gradient boosting decision tree, and adaptive boosting). Confusion matrix, receiver-operating-characteristic curve analysis, and survival analysis were used to evaluate the model performance. Results: Among 4 semiautomatic segmentation methods, SUV4.0 segmentation yielded the highest interobserver reliability, with 830 (66.7%) selected radiomics features. The combined model constructed by the stacking method achieved the best discrimination performance. For progression-free survival prediction in the external testing set, the areas under the receiver-operating-characteristic curve and accuracy of the stacking-based combined model were 0.771 and 0.789, respectively. For overall survival prediction, the stacking-based combined model achieved an area under the curve of 0.725 and an accuracy of 0.763 in the external testing set. The combined model also demonstrated a more distinct risk stratification than the International Prognostic Index in all sets (log-rank test, all P < 0.05). Conclusion: The combined model that incorporates [18F]FDG PET radiomics and clinical characteristics based on stacking ensemble learning could enable improved risk stratification in DLBCL.

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
基于堆叠集成学习[18F]FDG PET放射组学用于弥漫性大B细胞淋巴瘤的预后预测。
本研究旨在开发一种基于[18F]FDG PET放射组学的分析方法,使用堆叠集成学习来改进弥漫性大B细胞淋巴瘤(DLBCL)的预后预测。方法:将来自2个医疗中心的240名DLBCL患者分为训练集(n=141)、内部测试集(n=61)和外部测试集(n=38)。使用4种半自动分割方法(SUV阈值为2.5,SUV阈值为4.0[SUV4.0],SUVmax的41%,平均肝脏摄取的SUV阈值[PERIST])从患者水平的预处理[18F]FDG PET扫描中提取放射组学特征。所有提取的特征都与ComBat方法进行了协调。组内相关系数用于评估通过不同分割方法提取的放射组学特征的可靠性。通过Pearson相关系数分析和LASSO(最小绝对收缩和选择算子)算法从最可靠的分割方法中选择特征。基于4个机器学习分类器(支持向量机、随机森林、梯度增强决策树和自适应增强),应用堆叠集成学习方法构建仅用于放射组学和组合临床放射组学的模型,用于预测2-y无进展生存期和总生存期。使用混淆矩阵、受试者操作特征曲线分析和生存分析来评估模型性能。结果:在4种半自动分割方法中,SUV4.0分割产生了最高的观察者间可靠性,有830个(66.7%)选择了放射组学特征。采用叠加方法构建的组合模型具有最佳的识别性能。对于外部测试集中的无进展生存期预测,基于叠加的组合模型的受试者工作特性曲线下的面积和准确性分别为0.771和0.789。对于整体生存预测,基于堆叠的组合模型在外部测试集中实现了0.725的曲线下面积和0.763的精度。在所有集合中,联合模型也显示出比国际预后指数更明显的风险分层(log-rank检验,均P<0.05)。结论:结合[18F]FDG PET放射组学和基于堆叠集合学习的临床特征的联合模型可以改善DLBCL的风险分层。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Journal of Nuclear Medicine
Journal of Nuclear Medicine 医学-核医学
CiteScore
13.00
自引率
8.60%
发文量
340
审稿时长
1 months
期刊介绍: The Journal of Nuclear Medicine (JNM), self-published by the Society of Nuclear Medicine and Molecular Imaging (SNMMI), provides readers worldwide with clinical and basic science investigations, continuing education articles, reviews, employment opportunities, and updates on practice and research. In the 2022 Journal Citation Reports (released in June 2023), JNM ranked sixth in impact among 203 medical journals worldwide in the radiology, nuclear medicine, and medical imaging category.
期刊最新文献
C-X-C Motif Chemokine Receptor 4-Directed Scintigraphy of Multiple Myeloma Using [99mTc]Tc-PentixaTec. Debating the Future of Nuclear Medicine: The Greek Experience. Peptide Receptor Radionuclide Therapy Is Effective for Clinical Control of Symptomatic Metastatic Insulinoma: A Long-Term Retrospective Analysis Identification of (R)-[18F]YH134 for Monoacylglycerol Lipase Neuroimaging and Exploration of Its Use for Central Nervous System and Peripheral Drug Development Value of68Ga-FAPI-04 and18F-FDG PET/CT in Early Prediction of Pathologic Response to Neoadjuvant Chemotherapy in Locally Advanced Gastric Cancer
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1