Fig 2: Correlation between GPNMB, EGFR, p-PI3K, and Ki-67. Pearson was used to analyze the correlation between (a) GPNMB and EGFR, (b) GPNMB and p-PI3K, (c) GPNMB and Ki-67, (d) EGFR and p-PI3K, (e) EGFR and Ki-67, (f) p-PI3K and Ki-67, (g–i) GPNMB and EGFR at different differentiation levels, (j–l) EGFR and p-PI3K at different differentiation levels, and (m, n) GPNMB and EGFR at different AJCC stages.

Fig 3: GPNMB mRNA expression was increased in ESCC and correlated with the PI3K pathway. (a) The GPNMB expression difference between 28 pairs of ESCC and paired paracancer tissues in the GEO database was analyzed by paired T-test. Results are presented as means ± SD; the box plot shows the median value (quartile) of the two groups. (b) Differences in GPNMB expression between ESCC (n = 81) and normal esophageal tissues (n = 10) in TCGA were analyzed by Wilcoxon rank-sum test. The results were counted as the median difference; the box plot shows the median value (quartile) of the two groups. (c) In the GEO (GSE161533) database, ESCC samples with the highest (n = 7) and lowest (n = 7) GPNMB expression were selected for differential gene analysis. Heat maps of PI3K pathway-related gene expression in GPNMB upregulated and downregulated samples. Red indicates upregulated genes and blue indicates downregulated genes.

Fig 4: Expression of four proteins at different AJCC stages. The Kruskal-Wallis rank-sum test was used to analyze the expression differences of (a) GPNMB, (b) EGFR, (c) p-PI3K, and (d) Ki-67 in different stages. The box plot shows the median value (quartile) of the two groups.

Fig 5: Kaplan–Meier survival curves for patients. The difference in prognosis between the high and low expression groups of (a) GPNMB, (b) p-PI3K, (c) Ki-67, and (d) EGFR. Log-rank test was used to compare the difference between two curves. The pale area around the curve was the 95% confidence interval.