Objective Development and validation of a nomogram integrating clinical characteristics, Lung Molecular Graded Prognostic Assessment (Lung-molGPA) score, and MRI-based radiomics to predict intracranial progression-free survival (iPFS) in epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) patients with brain metastases (BM) receiving concurrent EGFR-tyrosine kinase inhibitor (EGFR-TKI) and cranial radiotherapy. Materials and methods This study enrolled 409 eligible patients from three major hospitals. Clinical data, Lung-molGPA scores, and pre-treatment MRI-derived radiomic features were analyzed. Cluster analysis was used to evaluate the discriminative ability and intergroup consistency of radiomic features. Patients were divided into training, internal validation, and two independent external validation cohorts. Radiomic features were selected using the least absolute shrinkage and selection operator (LASSO) algorithm, and nomograms were developed and validated to predict iPFS. Model performance was evaluated using receiver operating characteristic (ROC) curves, concordance indexes (C-indexes), and decision curve analysis (DCA). Results The median iPFS of the overall cohort was 15.07 months. Independent clinical predictors of iPFS encompassed BM volume, radiotherapy modality, and first-line treatment status. Cluster analysis revealed that radiomic features could be stably classified into two distinct groups with strong intergroup consistency. Twelve radiomic features were ultimately selected to develop the radiomic signature. The integrated nomogram demonstrated superior predictive performance over single-modality approaches, with area under the curve (AUC) values of 0.888 (training cohort), 0.897 (internal validation), and 0.883/0.903 in two external validation cohorts. Conclusion The integrated nomogram provides a robust tool for predicting iPFS in EGFR-mutant NSCLC patients with BM. Key Points <Emphasis Type="BoldItalic">Question</Emphasis> Currently, effective predictive tools for intracranial progression-free survival following combined epidermal growth factor receptor tyrosine kinase inhibitor and cranial radiotherapy treatment are lacking. <Emphasis Type="BoldItalic">Findings</Emphasis> The integrated nomogram provided superior prediction of intracranial progression-free survival and radiotherapy modality was identified as a key independent prognostic factor. <Emphasis Type="BoldItalic">Clinical relevance</Emphasis> This non-invasive model enhances individualized therapy planning and optimizes cranial radiotherapy approaches, advancing precision care for non-small cell lung cancer patients with brain metastases. Graphical Abstract

Integrated nomogram for predicting intracranial progression-free survival with EGFR-TKI combined with cranial radiotherapy: a multicenter study

期刊论文
发表日期：

2025年10月31日

查看全文

Haoran Qi,
Yichen Hou,
Zhonghang Zheng,
Qiang Qiao,
Zihao Wang,
Xiao Liu,
Xiaorong Sun,
Ligang Xing

摘要

Objective

Development and validation of a nomogram integrating clinical characteristics, Lung Molecular Graded Prognostic Assessment (Lung-molGPA) score, and MRI-based radiomics to predict intracranial progression-free survival (iPFS) in epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) patients with brain metastases (BM) receiving concurrent EGFR-tyrosine kinase inhibitor (EGFR-TKI) and cranial radiotherapy.

Materials and methods

This study enrolled 409 eligible patients from three major hospitals. Clinical data, Lung-molGPA scores, and pre-treatment MRI-derived radiomic features were analyzed. Cluster analysis was used to evaluate the discriminative ability and intergroup consistency of radiomic features. Patients were divided into training, internal validation, and two independent external validation cohorts. Radiomic features were selected using the least absolute shrinkage and selection operator (LASSO) algorithm, and nomograms were developed and validated to predict iPFS. Model performance was evaluated using receiver operating characteristic (ROC) curves, concordance indexes (C-indexes), and decision curve analysis (DCA).

Results

The median iPFS of the overall cohort was 15.07 months. Independent clinical predictors of iPFS encompassed BM volume, radiotherapy modality, and first-line treatment status. Cluster analysis revealed that radiomic features could be stably classified into two distinct groups with strong intergroup consistency. Twelve radiomic features were ultimately selected to develop the radiomic signature. The integrated nomogram demonstrated superior predictive performance over single-modality approaches, with area under the curve (AUC) values of 0.888 (training cohort), 0.897 (internal validation), and 0.883/0.903 in two external validation cohorts.

Conclusion

The integrated nomogram provides a robust tool for predicting iPFS in EGFR-mutant NSCLC patients with BM.

Key Points

Question Currently, effective predictive tools for intracranial progression-free survival following combined epidermal growth factor receptor tyrosine kinase inhibitor and cranial radiotherapy treatment are lacking.

Findings The integrated nomogram provided superior prediction of intracranial progression-free survival and radiotherapy modality was identified as a key independent prognostic factor.

Clinical relevance This non-invasive model enhances individualized therapy planning and optimizes cranial radiotherapy approaches, advancing precision care for non-small cell lung cancer patients with brain metastases.

Graphical Abstract