Scholarlink

Background <p>Tumor-infiltrating lymphocytes (TILs) are prognostic biomarkers in breast cancer (BC), particularly in HER2-positive and triple-negative subtypes. Assessment follows the international guidelines, in which pathologists evaluate whole hematoxylin and eosin (H&amp;E)-stained slides while integrating representative regions of the invasive tumor. However, manual region selection can be labor-intensive, subjective, and may introduce variability, particularly across consecutive tissue sections. Automated region of interest (ROI) registration may mitigate this limitation, yet its impact on longitudinal TIL scoring has not been systematically evaluated. Here, we introduce three ROI registration strategies (direct, intermediate, and serial with/without quality control) and present an automated framework validated for consistent TIL scoring and clinical relevance in predicting relapse.</p> Methods <p>We analyzed 104 invasive BC cases, each with 12 consecutive H&amp;E slides. A pathologist annotated ROIs on both the first and twelfth slides. We registered these ROIs using the proposed strategies. We then evaluated them with performance metrics, including Intersection over Union (IoU), Dice Similarity Coefficient (DSC), failure rate, and execution time. Two pathologists scored TILs on manual and automated ROIs. We assessed longitudinal consistency between the first manual ROI and the twelfth slide’s corresponding ROIs. We also tested whether the association between TIL score and patient relapse outcomes was preserved.</p> Results <p>The direct registration strategy achieved the highest geometric accuracy (mean IoU&#xa0;=&#xa0;0.650, DSC&#xa0;=&#xa0;0.769), with &lt;&#xa0;1% failure rate and <InlineEquation ID="IEq1"><EquationSource Format="TEX">\(\sim \)</EquationSource></InlineEquation>5.8&#xa0;s execution time. TIL scores for both manual and automated ROIs closely matched (<i>P</i>&#xa0;=&#xa0;0.84), showing strong agreement (Spearman’s <InlineEquation ID="IEq2"><EquationSource Format="TEX">\(\rho \)</EquationSource></InlineEquation>&#xa0;=&#xa0;0.923, ICC&#xa0;=&#xa0;0.936, CCC&#xa0;=&#xa0;0.915, Cohen’s <InlineEquation ID="IEq3"><EquationSource Format="TEX">\(\kappa \)</EquationSource></InlineEquation>&#xa0;=&#xa0;0.786). Longitudinal analyses showed no significant variability across distant sections (<i>P</i>&#xa0;=&#xa0;0.79 and <i>P</i>&#xa0;=&#xa0;0.62). TIL concentrations differed significantly between patients with and without relapse, and this association was preserved on both the first and twelfth slides (<i>P</i>&#xa0;&lt;&#xa0;0.05).</p> Conclusions <p>We present the first automated framework for longitudinal ROI registration to support TIL scoring in BC. By reducing manual effort and variability, the framework supports scalable evaluation of immune biomarkers across tissue depth while preserving clinically relevant signals, supporting precision immuno-oncology applications.</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

Longitudinal evaluation of tumor-infiltrating lymphocyte scoring using automated region of interest registration in breast cancer

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月3日

查看全文
  • Alessio Fiorin,
  • Laia Adalid-Llansa,
  • Laia Reverté,
  • Esther Sauras-Colón,
  • Noèlia Gallardo-Borràs,
  • Hatem A. Rashwan,
  • Ramon Bosch-Príncep,
  • Alba Fischer-Carles,
  • Elena Goyda,
  • Marylène Lejeune,
  • Daniel Mata-Cano,
  • Domènec Puig,
  • Tábata Sánchez-Alcántara,
  • Mikel Relloso Ortiz de Uriarte,
  • Montserrat Llobera-Serentill,
  • José Antonio Izuel-Navarro,
  • Carlos López-Pablo

摘要

Background

Tumor-infiltrating lymphocytes (TILs) are prognostic biomarkers in breast cancer (BC), particularly in HER2-positive and triple-negative subtypes. Assessment follows the international guidelines, in which pathologists evaluate whole hematoxylin and eosin (H&E)-stained slides while integrating representative regions of the invasive tumor. However, manual region selection can be labor-intensive, subjective, and may introduce variability, particularly across consecutive tissue sections. Automated region of interest (ROI) registration may mitigate this limitation, yet its impact on longitudinal TIL scoring has not been systematically evaluated. Here, we introduce three ROI registration strategies (direct, intermediate, and serial with/without quality control) and present an automated framework validated for consistent TIL scoring and clinical relevance in predicting relapse.

Methods

We analyzed 104 invasive BC cases, each with 12 consecutive H&E slides. A pathologist annotated ROIs on both the first and twelfth slides. We registered these ROIs using the proposed strategies. We then evaluated them with performance metrics, including Intersection over Union (IoU), Dice Similarity Coefficient (DSC), failure rate, and execution time. Two pathologists scored TILs on manual and automated ROIs. We assessed longitudinal consistency between the first manual ROI and the twelfth slide’s corresponding ROIs. We also tested whether the association between TIL score and patient relapse outcomes was preserved.

Results

The direct registration strategy achieved the highest geometric accuracy (mean IoU = 0.650, DSC = 0.769), with < 1% failure rate and \(\sim \)5.8 s execution time. TIL scores for both manual and automated ROIs closely matched (P = 0.84), showing strong agreement (Spearman’s \(\rho \) = 0.923, ICC = 0.936, CCC = 0.915, Cohen’s \(\kappa \) = 0.786). Longitudinal analyses showed no significant variability across distant sections (P = 0.79 and P = 0.62). TIL concentrations differed significantly between patients with and without relapse, and this association was preserved on both the first and twelfth slides (P < 0.05).

Conclusions

We present the first automated framework for longitudinal ROI registration to support TIL scoring in BC. By reducing manual effort and variability, the framework supports scalable evaluation of immune biomarkers across tissue depth while preserving clinically relevant signals, supporting precision immuno-oncology applications.