Angiogenesis-related lncRNAs index: A predictor for CESC prognosis, immunotherapy efficacy, and chemosensitivity

Abstract

Cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) is a common gynecologic tumor and patients with advanced and recurrent disease usually have a poor clinical outcome. Angiogenesis is involved in the biological processes of tumors and can promote tumor growth and invasion. In this paper, we created a signature for predicting prognosis based on angiogenesis-related lncRNAs (ARLs). This provides a prospective direction for enhancing the efficacy of immunotherapy in CESC patients. We screened seven OS-related ARLs by univariate and multivariate regression analyses and Lasso analysis and developed a prognostic signature at the same time. Then, we performed an internal validation in the TCGA-CESC cohort to increase the precision of the study. In addition, we performed a series of analyses based on ARLs, including immune cell infiltration, immune function, immune checkpoint, tumor mutation load, and drug sensitivity analysis. Our created signature based on ARLs can effectively predict the prognosis of CESC patients. To strengthen the prediction accuracy of the signature, we built a nomogram by combining signature and clinical features.

Keywords: Angiogenesis; CESC; Chemotherapy; Immunotherapy; LncRNAs; Prognostic signature.

Figure 1

A flowchart outlining the study's primary ideas.

Figure 2

Screening of candidate ARLs. (A) Sankey diagram showing the co-expression relationship between ARLs and ARGs. (B) Prognosis of ARLs evaluated using univariate Cox regression analysis. (C) Adjusted parameter selection in Lasso analysis by tenfold cross-validation. (D) Lasso coefficient graph. (E) Heat map presenting the correlation of candidate ARLs and ARGs.

Figure 3

Verification of the accuracy of the ARLs signature in predicting prognosis. Distribution of risk scores between high and low risk groups in (A) TCGA all, (B) TCGA train, (C) TCGA test cohorts. Survival status between low and high risk groups in the (D) TCGA all, (E) TCGA train, (F) TCGA test cohorts. Survival Curve between low and high risk groups in the (G) TCGA all, (H)TCGA train, (I) TCGA test cohorts. Time-dependent ROC curves are demonstrated in (J) for all patient groups and (K) for the training group (L) for the test group. (M) Multi-index ROC analysis.

Figure 4

Candidate ARLs can better differentiate between high-risk and low-risk cohorts. PCA plots of the all genes (A), ARGs (B), ARLs (C), and candidate ARLs (D).

Figure 5

Association between risk scores and clinicopathological characteristics. (A) Heat map illustrating the proportion of risk scores and clinicopathological features in the TCGE-CESC cohort. Differences in the number of patients with different clinical characteristics in the high- and low-risk groups. These clinical characteristics include (B) age, (C) grade, (D) T stage, (E) M stage, and (F) N stage.

Figure 6

Prognostic efficacy of the ARL risk model for overall survival of different subtypes in the TCGA-PAAD cohort, (A) age≤65, (B) age>65, (C) grades I-II, (D) grades III-IV, (E) T1-T2, (F) T3-T4, (G) N0, (H) N1 and (I) female, respectively.

Figure 7

A nomogram constructed by combining risk scores and clinical characteristics. (A) Univariate and (B) Multivariate analysis of risk scores and multiple clinical characteristics. (C) Nomogram predicting overall survival of CESC patients at 1,3 and 5 years. (D) The calibration curve of the created nomogram.

Figure 8

Functional enrichment analysis of ARLs in the TCGA-CESC cohort. (A) GO enrichment analysis. (B) GSVA analysis between the high-risk and low-risk group with Kyoto Encyclopedia of Genes and Genomes (KEGG). GSVA analysis between the (C) low-risk and (D) high-risk group.

Figure 9

Risk model for ARLs predicts TME and immune cell infiltration. (A) Immune cell bubble map. (B) Discrepancies in immune cell infiltration between high- and low-risk cohorts. (C) Differences in immune checkpoints between high and low risk groups. (D) Immune function ssGSEA scores in the high and low-risk cohorts. (E) Variations in TIDE scores in high and low risk cohorts.

Figure 10

The waterfall plot shows the top 15 genes in the (A) high and (B) low risk groups in terms of mutation frequency. (C) TMB in high and low risk groups. (D) Survival curves of the high and low TMB groups. (E) Survival curves of the high-TMB+high-risk group, high-TMB+low-risk group, low TMB+high-risk group and low-TMB+low-risk group.

Figure 11

Drug sensitivity analysis in high and low-risk groups. (A) Afatinib, (B) AZD4547, (C) EPZ004777, (D) Erlotinib, (E) Gefitinib, (F) GNE-317, (G) Ibrutinib, (H) Lapatinib, (I) Sapitinib, (J) Trametinib, (K) Ulixertinib, and (L) VSP34_8731.