The construction of a novel prognostic prediction model for glioma based on GWAS-identified prognostic-related risk loci

Jie Wei; Yujie Li; Wenqian Zhou; Xiaoya Ma; Jie Hao; Ting Wen; Bin Li; Tianbo Jin; Mingjun Hu

doi:10.1515/med-2024-0895

The construction of a novel prognostic prediction model for glioma based on GWAS-identified prognostic-related risk loci

Open Med (Wars). 2024 Mar 15;19(1):20240895. doi: 10.1515/med-2024-0895. eCollection 2024.

Authors

Jie Wei^{1

2

3}, Yujie Li^{1

2

3}, Wenqian Zhou^{1

2

3}, Xiaoya Ma^{1

2

3}, Jie Hao^{1

2

3}, Ting Wen^{1

2

3}, Bin Li^{1

4}, Tianbo Jin^{1

2

3}, Mingjun Hu^{3

5

6

7}

Affiliations

¹ Key Laboratory of Resource Biology and Biotechnology in Western China, Ministry of Education, Northwest University, Xi'an 710069, Shaanxi, China.
² Shaanxi Provincial Key Laboratory of Biotechnology, Northwest University, Xi'an 710069, Shaanxi, China.
³ College of Life Science, Northwest University, Xi'an 710127, Shaanxi, China.
⁴ Biomedicine Key Laboratory of Shaanxi Province, Northwest University, Xi'an 710069, Shaanxi, China.
⁵ School of Medicine, Northwest University, Xi'an 710127, Shaanxi, China.
⁶ Department of Neurosurgery, Xi'an Chest Hospital, Xi'an 710100, Shaanxi, China.
⁷ Department of Neurosurgery, Xi'an Chang'an District Hospital, Xi'an 710118, Shaanxi, China.

Abstract

Backgrounds: Glioma is a highly malignant brain tumor with a grim prognosis. Genetic factors play a role in glioma development. While some susceptibility loci associated with glioma have been identified, the risk loci associated with prognosis have received less attention. This study aims to identify risk loci associated with glioma prognosis and establish a prognostic prediction model for glioma patients in the Chinese Han population.

Methods: A genome-wide association study (GWAS) was conducted to identify risk loci in 484 adult patients with glioma. Cox regression analysis was performed to assess the association between GWAS-risk loci and overall survival as well as progression-free survival in glioma. The prognostic model was constructed using LASSO Cox regression analysis and multivariate Cox regression analysis. The nomogram model was constructed based on the single nucleotide polymorphism (SNP) classifier and clinical indicators, enabling the prediction of survival rates at 1-year, 2-year, and 3-year intervals. Additionally, the receiver operator characteristic (ROC) curve was employed to evaluate the prediction value of the nomogram. Finally, functional enrichment and tumor-infiltrating immune analyses were conducted to examine the biological functions of the associated genes.

Results: Our study found suggestive evidence that a total of 57 SNPs were correlated with glioma prognosis (p < 5 × 10^-5). Subsequently, we identified 25 SNPs with the most significant impact on glioma prognosis and developed a prognostic model based on these SNPs. The 25 SNP-based classifier and clinical factors (including age, gender, surgery, and chemotherapy) were identified as independent prognostic risk factors. Subsequently, we constructed a prognostic nomogram based on independent prognostic factors to predict individualized survival. ROC analyses further showed that the prediction accuracy of the nomogram (AUC = 0.956) comprising the 25 SNP-based classifier and clinical factors was significantly superior to that of each individual variable.

Conclusion: We identified a SNP classifier and clinical indicators that can predict the prognosis of glioma patients and established a prognostic prediction model in the Chinese Han population. This study offers valuable insights for clinical practice, enabling improved evaluation of patients' prognosis and informing treatment options.

Keywords: GWAS; LASSO Cox regression; glioma prognosis; nomogram; prognostic prediction model; risk loci.