Scholarlink

<p>This study aimd to delineate the prognostic impact of Programmed death-ligand 1 (PD-L1) expression level in glioblastoma (GBM) patients, and investigate the value of using magnetic resonance imaging (MRI) histograms and Visually Accessible Rembrandt Images (VASARI) features to preoperatively predict the expression level of PD-L1. The clinical and imaging data of 124 patients with GBM at our institution between November 2018 and April 2023 were retrospectively analyzed. PD-L1 expression levels in tumor tissue samples were quantified using immunohistochemical staining. The optimal cutoff PD-L1 level was determined using the X-tile program through Kaplan-Meier survival analysis and log-rank test. The MRI histogram and VASARI features of the patients in the high and low PD-L1 expression groups were recorded. The predictive models for PD-L1 expression level were constructed using multivariable binary logistic regression, and a nomogram was generated. The GBM patients with high PD-L1 expression had unfavorable overall survival. The T1-weighted contrast-enhanced histogram features mean, 1st, 5th, 10th, 25th, 50th, and 75th percentiles and the VASARI feature F5 proportion enhancing were statistically significantly different between groups (all <i>p</i> &lt; 0.05). Multivariate logistic regression analysis showed that mean, 5th, 10th, and 50th percentiles, and F5 proportion enhancing were independent risk factors for predicting PD-L1 expression in GBM patients. The logistic regression model based on these 5 features showed a better predictive performance, and the area under the curve, accuracy, sensitivity, specificity were 0.795, 0.726, 0.887, and 0.621, respectively. The nomogram based on MRI histogram and VASARI features can show promise to non-invasively predict the level of PD-L1 expression in GBM and could be helpful in guiding immune checkpoint inhibitors therapy and predicting patient prognosis.</p><p><?noindent??><b>Clinical trial number</b> Not applicable.</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

MRI histogram and Visually AcceSAble Rembrandt Images features nomogram to predict PD-L1 levels in glioblastoma

  • 期刊论文

  • 发表日期：

    2025年4月21日

查看全文
  • Bin Zhang,
  • Qing Zhou,
  • Peng Zhang,
  • Caiqiang Xue,
  • Xiaoai Ke,
  • Yige Wang,
  • Tao Han,
  • Fengyu Zhou,
  • Wenjie Dong,
  • Liangna Deng,
  • Mengyuan Jing,
  • Yuting Zhang,
  • Xianwang Liu,
  • Junlin Zhou

摘要

This study aimd to delineate the prognostic impact of Programmed death-ligand 1 (PD-L1) expression level in glioblastoma (GBM) patients, and investigate the value of using magnetic resonance imaging (MRI) histograms and Visually Accessible Rembrandt Images (VASARI) features to preoperatively predict the expression level of PD-L1. The clinical and imaging data of 124 patients with GBM at our institution between November 2018 and April 2023 were retrospectively analyzed. PD-L1 expression levels in tumor tissue samples were quantified using immunohistochemical staining. The optimal cutoff PD-L1 level was determined using the X-tile program through Kaplan-Meier survival analysis and log-rank test. The MRI histogram and VASARI features of the patients in the high and low PD-L1 expression groups were recorded. The predictive models for PD-L1 expression level were constructed using multivariable binary logistic regression, and a nomogram was generated. The GBM patients with high PD-L1 expression had unfavorable overall survival. The T1-weighted contrast-enhanced histogram features mean, 1st, 5th, 10th, 25th, 50th, and 75th percentiles and the VASARI feature F5 proportion enhancing were statistically significantly different between groups (all p < 0.05). Multivariate logistic regression analysis showed that mean, 5th, 10th, and 50th percentiles, and F5 proportion enhancing were independent risk factors for predicting PD-L1 expression in GBM patients. The logistic regression model based on these 5 features showed a better predictive performance, and the area under the curve, accuracy, sensitivity, specificity were 0.795, 0.726, 0.887, and 0.621, respectively. The nomogram based on MRI histogram and VASARI features can show promise to non-invasively predict the level of PD-L1 expression in GBM and could be helpful in guiding immune checkpoint inhibitors therapy and predicting patient prognosis.

Clinical trial number Not applicable.