Scholarlink

Purpose <p>This study aims to evaluate the predictive value of radiomics features extracted from [<sup>99m</sup>Tc]Tc-Octreotide SPECT images for predicting response to peptide receptor radionuclide therapy (PRRT) in neuroendocrine tumor (NET) patients, with particular emphasis on the discordance between tumor grade and Ki-67 proliferation index.</p> Methods <p>This retrospective study included 93 NET patients who underwent [<sup>99m</sup>Tc]Tc-Octreotide SPECT/CT before PRRT. Tumor grade and Ki-67 index were assessed via histopathology. Radiomics features were extracted using Pyradiomics. Patients were stratified into concordant (grade matches Ki-67) comprising 82 concordant and 11 discordant patients. A random forest classifier was developed using selected features to predict treatment response, validated by a 10-fold cross-validation scheme.</p> Results <p>Radiomics-based models showed promising results, with the random forest model achieving AUCs of 0.8981 (training/validation) and 0.89 (test). The Ki-67 index demonstrated a stronger correlation with treatment response than tumor grade, especially among discordant cases.</p> Conclusion <p>Radiomics analysis of [<sup>99m</sup>Tc]Tc-Octreotide SPECT images offers a non-invasive tool for predicting PRRT outcomes in NET. Discordance between Ki-67 and tumor grade impacts prognostication, and radiomics may aid in resolving diagnostic ambiguity in such cases.</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

[99mTc]Tc-Octreotide SPECT-based radiomics analysis for predicting response in patients with neuroendocrine tumors undergoing [177Lu]Lu-DOTA-TATE PRRT: a multicenter study on Ki-67 index and grade discordance

  • 期刊论文

  • 发表日期：

    2025年10月8日

查看全文
  • Mehdi Ghazizadeh,
  • Mahasti Amoui,
  • Mohammad Reza Deevband,
  • Kamran Aryana,
  • Abolhasan Divband,
  • Farivash Karamian,
  • Alireza Montazerabadi,
  • Zahra Mahboubi-Fooladi,
  • Sara Harsini,
  • Emran Askari

摘要

Purpose

This study aims to evaluate the predictive value of radiomics features extracted from [99mTc]Tc-Octreotide SPECT images for predicting response to peptide receptor radionuclide therapy (PRRT) in neuroendocrine tumor (NET) patients, with particular emphasis on the discordance between tumor grade and Ki-67 proliferation index.

Methods

This retrospective study included 93 NET patients who underwent [99mTc]Tc-Octreotide SPECT/CT before PRRT. Tumor grade and Ki-67 index were assessed via histopathology. Radiomics features were extracted using Pyradiomics. Patients were stratified into concordant (grade matches Ki-67) comprising 82 concordant and 11 discordant patients. A random forest classifier was developed using selected features to predict treatment response, validated by a 10-fold cross-validation scheme.

Results

Radiomics-based models showed promising results, with the random forest model achieving AUCs of 0.8981 (training/validation) and 0.89 (test). The Ki-67 index demonstrated a stronger correlation with treatment response than tumor grade, especially among discordant cases.

Conclusion

Radiomics analysis of [99mTc]Tc-Octreotide SPECT images offers a non-invasive tool for predicting PRRT outcomes in NET. Discordance between Ki-67 and tumor grade impacts prognostication, and radiomics may aid in resolving diagnostic ambiguity in such cases.