Scholarlink

Background <p>Carotid ultrasound is largely used to assess arterial aging. Radiomics (quantitative imaging feature analysis) applied on ultrasound may allow characterizing wall ultrastructure and arterial ageing. However, to date its application to the intima–media (IM) complex is unexplored. The aim of this study is to investigate the feasibility of radiomics-based B-mode and RF feature extraction and selection process.</p> Methods <p>Radio-frequency signals from 200 individuals (age 50–75y, 40% hypertensives, 20% diabetic) were used to extract RF and B-mode features. Feature stability across three frames from the same clip and five ROI sizes for each frame were evaluated by intraclass correlation analysis (threshold &gt; 0.50). Lasso L1 regression for chronological age prediction on ICC &gt; 0.50 features were used to analyze reproducibility and variation across frames and ROI sizes; 80/20 train-test split was used, with performance evaluated by MSE and <i>R</i><sup>2</sup>.</p> Results <p>Radiomic feature extraction was feasible in 190 among 200 individuals. 48 features showed an ICC &gt; 0.50. Feature selection for chronological age prediction showed consistent <i>R</i><sup>2</sup> (0.09–0.14 train, 0.05–0.13 test) and MSE (32.3–34.3 train, 30.7–42.3 test) across frames. Fixed ROI sizes of 1.0mm and 1.2mm had comparable performance to an ROI size manually tailored to wall thickness: <i>R</i><sup>2</sup> (0.07–0.15 train, 0.07–0.17 test) and MSE (33.15–36.22 train, 33.07–34.83 test), consistently selecting substantially similar three to six features.</p> Conclusion <p>While feature extraction and selection process are largely reproducible across frames, ROI size proved critical and thus must be carefully chosen. Further studies are required to demonstrate the validity of this approach for carotid wall characterization.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Radiomics Feature Extraction from Ultrasound B-Mode Images and Radio-Frequency Signals of the Carotid Arterial Wall: A Feasibility Study

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年3月31日

查看全文
  • Maryam Jadoon,
  • Federica Poli,
  • Pierre Boutouyrie,
  • Hakim Khettab,
  • Elisabetta Bianchini,
  • Francesco Faita,
  • Xavier Jouven,
  • Jean Philippe Empana,
  • Rosa Maria Bruno

摘要

Background

Carotid ultrasound is largely used to assess arterial aging. Radiomics (quantitative imaging feature analysis) applied on ultrasound may allow characterizing wall ultrastructure and arterial ageing. However, to date its application to the intima–media (IM) complex is unexplored. The aim of this study is to investigate the feasibility of radiomics-based B-mode and RF feature extraction and selection process.

Methods

Radio-frequency signals from 200 individuals (age 50–75y, 40% hypertensives, 20% diabetic) were used to extract RF and B-mode features. Feature stability across three frames from the same clip and five ROI sizes for each frame were evaluated by intraclass correlation analysis (threshold > 0.50). Lasso L1 regression for chronological age prediction on ICC > 0.50 features were used to analyze reproducibility and variation across frames and ROI sizes; 80/20 train-test split was used, with performance evaluated by MSE and R2.

Results

Radiomic feature extraction was feasible in 190 among 200 individuals. 48 features showed an ICC > 0.50. Feature selection for chronological age prediction showed consistent R2 (0.09–0.14 train, 0.05–0.13 test) and MSE (32.3–34.3 train, 30.7–42.3 test) across frames. Fixed ROI sizes of 1.0mm and 1.2mm had comparable performance to an ROI size manually tailored to wall thickness: R2 (0.07–0.15 train, 0.07–0.17 test) and MSE (33.15–36.22 train, 33.07–34.83 test), consistently selecting substantially similar three to six features.

Conclusion

While feature extraction and selection process are largely reproducible across frames, ROI size proved critical and thus must be carefully chosen. Further studies are required to demonstrate the validity of this approach for carotid wall characterization.