Scholarlink

Background <p>Accurate determination of H3 K27M mutation status is critical for prognosis evaluation and treatment planning in brainstem gliomas (BSGs). We developed a clinical-radiomics model integrating <sup>11</sup>C-methionine (<sup>11</sup>C-MET) positron emission tomography (PET), magnetic resonance imaging (MRI), and clinical features for non-invasive, preoperative prediction of H3 K27M mutation status in BSGs.</p> Methods <p>This retrospective study included 77 patients with newly diagnosed BSG who underwent preoperative <sup>11</sup>C-MET PET imaging and MRI. Among them, 53 had a histologically confirmed H3 K27M mutation. Patients were randomly divided into training (<i>n</i> = 55) and test (<i>n</i> = 22) cohorts. Clinical variables significantly associated with H3 K27M mutation status were used to construct a clinical model using the ‌extreme gradient boosting ‌(XGBoost) algorithm. Radiomics features were extracted separately from PET and MRI images to construct individual and combined models. A clinical-radiomics integrated model was then developed by combining clinical features and the radiomics model with the best predictive performance. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA).</p> Results <p>The clinical model achieved AUCs of 0.795 and 0.790 in the training and test cohorts, respectively. The combined PET-MRI radiomics model outperformed individual PET or MRI model, with AUCs of 0.876 and 0.848 in the training and test cohorts, respectively. The clinical-radiomics integrated model achieved the highest performance with AUCs of 0.886 in the training cohort and 0.876 in the test cohort. The calibration curve showed good agreement between predicted and actual outcomes, and DCA demonstrated a superior net clinical benefit of the integrated model.</p> Conclusions <p>Radiomics analysis based on <sup>11</sup>C-MET PET and MRI can effectively predict H3 K27M mutation status in BSGs. Incorporating clinical features further enhances predictive performance, supporting its use in non-invasive, preoperative molecular diagnosis.</p> Clinical trial number <p>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

Integrating clinical features with 11C-methionine PET-MRI radiomics for non-invasive prediction of H3 K27M mutation in brainstem gliomas

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月4日

查看全文
  • Shu Zhang,
  • Xiaotong Li,
  • Yanmei Wang,
  • Yuliang Wu,
  • Lin Ai,
  • Liwei Zhang,
  • Xiaobin Zhao

摘要

Background

Accurate determination of H3 K27M mutation status is critical for prognosis evaluation and treatment planning in brainstem gliomas (BSGs). We developed a clinical-radiomics model integrating 11C-methionine (11C-MET) positron emission tomography (PET), magnetic resonance imaging (MRI), and clinical features for non-invasive, preoperative prediction of H3 K27M mutation status in BSGs.

Methods

This retrospective study included 77 patients with newly diagnosed BSG who underwent preoperative 11C-MET PET imaging and MRI. Among them, 53 had a histologically confirmed H3 K27M mutation. Patients were randomly divided into training (n = 55) and test (n = 22) cohorts. Clinical variables significantly associated with H3 K27M mutation status were used to construct a clinical model using the ‌extreme gradient boosting ‌(XGBoost) algorithm. Radiomics features were extracted separately from PET and MRI images to construct individual and combined models. A clinical-radiomics integrated model was then developed by combining clinical features and the radiomics model with the best predictive performance. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA).

Results

The clinical model achieved AUCs of 0.795 and 0.790 in the training and test cohorts, respectively. The combined PET-MRI radiomics model outperformed individual PET or MRI model, with AUCs of 0.876 and 0.848 in the training and test cohorts, respectively. The clinical-radiomics integrated model achieved the highest performance with AUCs of 0.886 in the training cohort and 0.876 in the test cohort. The calibration curve showed good agreement between predicted and actual outcomes, and DCA demonstrated a superior net clinical benefit of the integrated model.

Conclusions

Radiomics analysis based on 11C-MET PET and MRI can effectively predict H3 K27M mutation status in BSGs. Incorporating clinical features further enhances predictive performance, supporting its use in non-invasive, preoperative molecular diagnosis.

Clinical trial number

Not applicable.