基于高频超声鉴别皮肤良恶性肿瘤的机器学习模型。

Yishuo Qin, Zhirou Zhang, Xiaomeng Qu, Weijie Liu, Yumei Yan, Yanli Huang
{"title":"基于高频超声鉴别皮肤良恶性肿瘤的机器学习模型。","authors":"Yishuo Qin, Zhirou Zhang, Xiaomeng Qu, Weijie Liu, Yumei Yan, Yanli Huang","doi":"10.11152/mu-4504","DOIUrl":null,"url":null,"abstract":"<p><strong>Aim: </strong>This study aims to explore the potential of machine learning as a non-invasive automated tool for skin tumor differentiation.</p><p><strong>Material and methods: </strong>Data were included from 156 lesions, collected retrospectively from September 2021 to February 2024. Univariate and multivariate analyses of traditional clinical features were performed to establish a logistic regression model. Ultrasound-based radiomics features are extracted from grayscale images after delineating regions of interest (ROIs). Independent samples t-tests, Mann-Whitney U tests, and Least Absolute Shrinkage and Selection Operator (LASSO) regression were employed to select ultrasound-based radiomics features. Subsequently, five machine learning methods were used to construct radiomics models based on the selected features. Model performance was evaluated using receiver operating characteristic (ROC) curves and the Delong test.</p><p><strong>Results: </strong>Age, poorly defined margins, and irregular shape were identified as independent risk factors for malignant skin tumors. The multilayer perception (MLP) model achieved the best performance, with area under the curve (AUC) values of 0.963 and 0.912, respectively. The results of DeLong's test revealed a statistically significant discrepancy in efficacy between the MLP and clinical models (Z=2.611, p=0.009).</p><p><strong>Conclusion: </strong>Machine learning based skin tumor models may serve as a potential non-invasive method to improve diagnostic efficiency.</p>","PeriodicalId":94138,"journal":{"name":"Medical ultrasonography","volume":" ","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"A machine learning model based on high-frequency ultrasound for differentiating benign and malignant skin tumors.\",\"authors\":\"Yishuo Qin, Zhirou Zhang, Xiaomeng Qu, Weijie Liu, Yumei Yan, Yanli Huang\",\"doi\":\"10.11152/mu-4504\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Aim: </strong>This study aims to explore the potential of machine learning as a non-invasive automated tool for skin tumor differentiation.</p><p><strong>Material and methods: </strong>Data were included from 156 lesions, collected retrospectively from September 2021 to February 2024. Univariate and multivariate analyses of traditional clinical features were performed to establish a logistic regression model. Ultrasound-based radiomics features are extracted from grayscale images after delineating regions of interest (ROIs). Independent samples t-tests, Mann-Whitney U tests, and Least Absolute Shrinkage and Selection Operator (LASSO) regression were employed to select ultrasound-based radiomics features. Subsequently, five machine learning methods were used to construct radiomics models based on the selected features. Model performance was evaluated using receiver operating characteristic (ROC) curves and the Delong test.</p><p><strong>Results: </strong>Age, poorly defined margins, and irregular shape were identified as independent risk factors for malignant skin tumors. The multilayer perception (MLP) model achieved the best performance, with area under the curve (AUC) values of 0.963 and 0.912, respectively. The results of DeLong's test revealed a statistically significant discrepancy in efficacy between the MLP and clinical models (Z=2.611, p=0.009).</p><p><strong>Conclusion: </strong>Machine learning based skin tumor models may serve as a potential non-invasive method to improve diagnostic efficiency.</p>\",\"PeriodicalId\":94138,\"journal\":{\"name\":\"Medical ultrasonography\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2025-04-11\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Medical ultrasonography\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.11152/mu-4504\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Medical ultrasonography","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.11152/mu-4504","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

目的:本研究旨在探索机器学习作为非侵入性皮肤肿瘤分化自动化工具的潜力。材料和方法:回顾性收集2021年9月至2024年2月156例病变的数据。对传统临床特征进行单因素和多因素分析,建立logistic回归模型。基于超声的放射组学特征是在描绘感兴趣区域(roi)后从灰度图像中提取出来的。采用独立样本t检验、Mann-Whitney U检验和最小绝对收缩和选择算子(LASSO)回归来选择基于超声的放射组学特征。随后,使用五种机器学习方法基于所选择的特征构建放射组学模型。采用受试者工作特征(ROC)曲线和Delong检验评价模型的性能。结果:年龄、边缘不清、形状不规则是恶性皮肤肿瘤的独立危险因素。多层感知(MLP)模型表现最佳,曲线下面积(AUC)分别为0.963和0.912。DeLong检验结果显示,MLP与临床模型的疗效差异有统计学意义(Z=2.611, p=0.009)。结论:基于机器学习的皮肤肿瘤模型可作为一种潜在的无创诊断方法,提高诊断效率。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
A machine learning model based on high-frequency ultrasound for differentiating benign and malignant skin tumors.

Aim: This study aims to explore the potential of machine learning as a non-invasive automated tool for skin tumor differentiation.

Material and methods: Data were included from 156 lesions, collected retrospectively from September 2021 to February 2024. Univariate and multivariate analyses of traditional clinical features were performed to establish a logistic regression model. Ultrasound-based radiomics features are extracted from grayscale images after delineating regions of interest (ROIs). Independent samples t-tests, Mann-Whitney U tests, and Least Absolute Shrinkage and Selection Operator (LASSO) regression were employed to select ultrasound-based radiomics features. Subsequently, five machine learning methods were used to construct radiomics models based on the selected features. Model performance was evaluated using receiver operating characteristic (ROC) curves and the Delong test.

Results: Age, poorly defined margins, and irregular shape were identified as independent risk factors for malignant skin tumors. The multilayer perception (MLP) model achieved the best performance, with area under the curve (AUC) values of 0.963 and 0.912, respectively. The results of DeLong's test revealed a statistically significant discrepancy in efficacy between the MLP and clinical models (Z=2.611, p=0.009).

Conclusion: Machine learning based skin tumor models may serve as a potential non-invasive method to improve diagnostic efficiency.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
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
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
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 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信