Related Experiment Video
Updated: Jun 24, 2026
Predictive Immune Modeling of Solid Tumors
Published on: February 25, 2020
Genomic determinants and an exploratory prognostic model for immunotherapy outcomes in recurrent or metastatic
Lingling Gu1, Biqing Zhu2, Cuicui Liu3
1Department of Medical Image Center, Jiangsu Cancer Hospital, Jiangsu Institute of Cancer Research, The Affiliated Cancer Hospital of Nanjing Medical University, Nanjing 210000, China.
Background:
Immune checkpoint inhibitors (ICIs) improve outcomes in recurrent or metastatic cervical cancer, but responses are heterogeneous and biomarkers beyond programmed death-ligand 1 (PD-L1) are limited.
Materials And Methods:
Targeted sequencing of 437 cancer-related genes was performed in 42 patients receiving ICI-based therapy. Genomic correlates of progression-free survival (PFS) were evaluated, and an exploratory prognostic model was developed using least absolute shrinkage and selection operator regression and multivariable Cox regression. The Cancer Genome Atlas (TCGA) cohort was used for transcriptomic analysis.
Results:
The cohort included 42 patients, with a median age of 51 years; 85.7% were human papillomavirus-positive, 90.5% had squamous cell carcinoma, and 54.8% had a programmed death-ligand 1 combined positive score (PD-L1 CPS) ≥5. Frequent alterations included PIK3CA (57.1%), FBXW7 (21.4%), TERT (21.4%), BAP1 (19.0%), and EP300 (16.7%). Tumor mutational burden-associated mutations in PIK3CA, EP300, CREBBP, and TERT were associated with prolonged PFS (P < .001), whereas PD-L1-associated mutations showed numerically longer PFS (P = .170). EP300 (P = .007), PIK3CA (P < .01), and homologous recombination repair pathway alterations (P = .016) were associated with favorable PFS, whereas KEAP1 (P < .01) and TP53 (P = .010) mutations were associated with shorter PFS. A five-feature genomic model stratified patients into high- and low-risk groups (P < .001, 1-year AUC = .889). In the exploratory cohort, high-risk patients showed numerically shorter PFS (P = .060). In TCGA samples, high-risk tumors showed enrichment of angiogenesis, epithelial-mesenchymal transition, hypoxia, inflammatory response, and tumor necrosis factor-α/nuclear factor-κB signaling.
Conclusion:
Specific genomic alterations may help stratify immunotherapy outcomes in recurrent or metastatic cervical cancer. The proposed genomic risk model remains exploratory and requires validation in larger, independent cervical cancer cohorts.