Prognostic and predictive significance of circulating biomarkers in patients with advanced upper gastrointestinal cancer undergoing systemic chemotherapy

Abstract

Objective: The prognosis of patients with advanced cancers of the upper gastrointestinal (UGI) tract is poor. Systemic chemotherapy forms the basis for their treatment, with limited efficacy. Biomarkers have been introduced into clinical practice for cancer management. This study aimed to investigate the predictive and prognostic values of circulating biomarkers in patients with advanced esophageal and gastric cancers receiving chemotherapy.

Design: Overall, 92 patients with advanced esophageal squamous cell carcinoma (ESCC; n = 38) and gastric adenocarcinoma (GAC; n = 54) were enrolled. We analyzed the association of circulating lymphocyte subsets, inflammatory markers, and blood cell counts with treatment efficacy and patient survival.

Results: Significant differences were identified in peripheral blood parameters between the groups with different clinicopathological features. Hemoglobin (Hb, p = 0.014), eosinophil counts (p = 0.028), CD4⁺CD28⁺T/CD4⁺T percentage (p = 0.049), CD8⁺CD38⁺T/CD8⁺T percentage (p = 0.044), memory CD4⁺T (p = 0.007), and CD4⁺CD28⁺T (p = 0.007) were determined as predictors for achieving non-PD (progression disease) in the ESCC cohort. High levels of eosinophils (p = 0.030) and memory CD4⁺T cells (p = 0.026) and high eosinophil-to-lymphocyte ratio (ELR, p = 0.013) were predictors of non-PD in patients with GAC. The combined detection models exhibited good ability to distinguish between partial response (PR)/non-PR and PD/non-PD in patients with ESCC and GAC, respectively. Using the multivariate Cox model, the Eastern Cooperative Oncology Group (ECOG) score status (hazard ratio [HR]: 4.818, 95% confidence intervals [CI]: 2.076-11.184, p < 0.001) and eosinophil count (HR: 0.276, 95% CI: 0.120-0.636, p = 0.003) were independent prognostic factors of progression-free survival (PFS) in patients with ESCC. Metastatic sites (HR: 2.092, 95% CI: 1.307-3.351, p = 0.002) and eosinophil-to-lymphocyte ratio (ELR; HR: 0.379, 95% CI: 0.161-0.893, p = 0.027) were independent prognostic factors for overall survival (OS) in patients with ESCC. Differentiation (HR: 0.041, 95% CI: 0.200-0.803, p = 0.010), memory CD4⁺T (HR: 0.304, 95% CI: 0.137-0.675, p = 0.003), NK cells (HR: 2.302, 95% CI: 1.044-3.953, p = 0.037), and C-reactive protein-to-lymphocyte ratio (CLR; HR: 2.070, 95% CI: 1.024-4.186, p = 0.043) were independent prognostic factors for PFS in patients with GAC. Total lymphocyte counts (HR: 0.260, 95% CI: 0.086-0.783, p = 0.017), CD8⁺T (HR: 0.405, 95% CI: 0.165-0.997, p = 0.049), NK cells (HR: 3.395, 95% CI: 1.592-7.238, p = 0.002), and monocyte-to-lymphocyte ratio (MLR; HR: 3.076, 95% CI: 1.488-6.360, p = 0.002) were identified as independent prognostic factors associated with OS of GAC.

Conclusion: Lymphocyte subsets, blood cell counts, and inflammatory parameters may predict the chemotherapeutic response and prognosis in ESCC and GAC. A combination of these markers can be used to stratify patients into risk groups, which could improve treatment strategies.

Keywords: biomarker; chemotherapy; esophageal cancer; gastric cancer; prognosis.

Figure 1

Scatter plots demonstrating the comparison of the counts of lymphocytes between groups stratified by differentiation (A), staging (B), the number of metastasis sites (C), and ECOG status (D). Comparison of the percentage of CD8⁺CD38⁺T/CD8⁺T among groups stratified by ages (E). The levels of B cells (F), CD4⁺T (G), memory CD4⁺T (H), and CD4⁺CD28⁺T cells (I) are compared between groups with different stages. The levels of B cells (J), CD4⁺T (K), NK cells (L), memory CD4⁺T (M), and CD4⁺CD28⁺T cells (N) are compared among groups stratified by the number of metastatic sites. B cells (O) are compared and demonstrated in scatter plots between groups with different ECOG status.

Figure 2

Scatter plots demonstrating the comparison of the counts of eosinophil (A), CD8⁺T (B), CD8⁺CD38⁺T (C), and ELR (D) between groups stratified by efficacy (PR vs. non-PR) in patients with ESCC. The levels of WBC (E), PLT (F), neutrophils (G), monocytes (H), and CD8⁺CD28⁺T cells (I) are compared between groups with different efficacies (PR vs. non-PR) in patients with GAC. ESCC, esophageal squamous carcinoma; GAC, gastric adenocarcinoma.

Figure 3

Receiver operating characteristic (ROC) curves for predicting non-PD in patients with ESCC by hemoglobin (Hb) (A), eosinophil (B), CD4⁺CD28⁺T/CD4⁺T percentage (C), CD8⁺CD38⁺T/CD8⁺T percentage (D), memory CD4⁺T (E), and CD4⁺CD28⁺T (F). ROC curves demonstrating the ability to discriminate patients with non-PD and those with PD in patients with GAC according to values of eosinophil (G), memory CD4⁺T (H), and ELR (I). ROC curves for predictive models showing the predicting potential of PR vs. non-PR (J) and non-PD vs. PD (K) for patients with ESCC, and the predicting ability of PR vs. non-PR (L) and non-PD vs. PD (M) for patients with GAC. ESCC, esophageal squamous carcinoma; GAC, gastric adenocarcinoma.

Figure 4

Results of univariable analysis for progression-free survival (PFS) and overall survival (OS) for patients with ESCC (A, D) and AGC (G, J), respectively. PFS curves by Kaplan–Meier analyses for patients with ESCC by ECOG (B) and eosinophil (C). OS curves of patients with ESCC by metastatic sites (E) and ELR (F). PFS Kaplan–Meier curves of patients with GAC by memory CD4⁺T (H), NK cells (I), and CLR (K).