Scholarlink

Background <p>Tumor regression grade (TRG) and ypTNM are primarily employed to evaluate the efficacy of Neoadjuvant chemotherapy (NAC) in gastric cancer (GC) patients, however, have limited prognostic value. In this study, we established a clinical-radiomic fusion model, without TRG information, for better prognosis assessment of patients following NAC.</p> Methods <p>A retrospective multicenter study comprising 875 GC patients from three centers was conducted. Cox hazard regression model was used for variable screening and risk weight assignment. Lasso regression was applied for dimensionality reduction and screening of radiomic features. Models were constructed for better prognosis assessment, and were verified in external cohorts.</p> Results <p>Survival analysis showed that dynamic T/N staging changes after NAC could effectively distinguish patients based on prognosis. Moreover, the Clinical SCORE model based on the dynamic T/N staging changes and other clinicopathological data had also been found in internal and external validations to be capable of effectively stratifying patients' risks. For CT images, the identified radiomics features were employed to establish the CT SCORE model, which was subsequently integrated with the Clinical SCORE model to construct the Final SCORE model for prognostic evaluation. In the training and validation cohorts, the prognostic discrimination performance of the Final SCORE model exceeded that of TRG and ypTNM. Furthermore, the final model might also be helpful for the screening of the population benefiting from postoperative adjuvant therapy.</p> Conclusion <p>The developed clinical-radiomic Final SCORE model showed superior prognostic assessment performance than TRG and ypTNM for prognostic assessment of GC patients following NAC.</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

Clinical-radiomic prognostic model integrating staging before and after neoadjuvant chemotherapy in gastric cancer: a multicenter retrospective study

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Yizhou Wei,
  • Siwei Pan,
  • Yahan Tong,
  • Guoliang Zheng,
  • Mengxuan Cao,
  • Yanqiang Zhang,
  • Ruolan Zhang,
  • Weiwei Zhu,
  • Qing Yang,
  • Ke Shen,
  • Mengya Zhou,
  • Ruixin Xu,
  • Jintao He,
  • Jiancheng Sun,
  • Zhiyuan Xu,
  • Xiangdong Cheng,
  • Can Hu

摘要

Background

Tumor regression grade (TRG) and ypTNM are primarily employed to evaluate the efficacy of Neoadjuvant chemotherapy (NAC) in gastric cancer (GC) patients, however, have limited prognostic value. In this study, we established a clinical-radiomic fusion model, without TRG information, for better prognosis assessment of patients following NAC.

Methods

A retrospective multicenter study comprising 875 GC patients from three centers was conducted. Cox hazard regression model was used for variable screening and risk weight assignment. Lasso regression was applied for dimensionality reduction and screening of radiomic features. Models were constructed for better prognosis assessment, and were verified in external cohorts.

Results

Survival analysis showed that dynamic T/N staging changes after NAC could effectively distinguish patients based on prognosis. Moreover, the Clinical SCORE model based on the dynamic T/N staging changes and other clinicopathological data had also been found in internal and external validations to be capable of effectively stratifying patients' risks. For CT images, the identified radiomics features were employed to establish the CT SCORE model, which was subsequently integrated with the Clinical SCORE model to construct the Final SCORE model for prognostic evaluation. In the training and validation cohorts, the prognostic discrimination performance of the Final SCORE model exceeded that of TRG and ypTNM. Furthermore, the final model might also be helpful for the screening of the population benefiting from postoperative adjuvant therapy.

Conclusion

The developed clinical-radiomic Final SCORE model showed superior prognostic assessment performance than TRG and ypTNM for prognostic assessment of GC patients following NAC.