Study selection and characteristics
The selection process of the literature is detailed in Figure 1. A total of the 35 articles were obtained by filtering from the four databases mentioned above. After completing the elimination of duplicates and irrelevant study topics, the titles, abstracts and full texts of these studies were read further and 11 articles were finally included in our Meta-analysis[16-20,22-26]. All included studies were from China and Contain a total of 1,227 patients with sample sizes of 35 to 215, and published between 2011 and 2022. A total of 10 types of cancer were involved: including breast cancer, colorectal cancer, esophageal squamous cell carcinoma, gastric cancer, retinoblastoma, lung squamous cell carcinoma, lung adenocarcinoma, ovarian cancer, hepatocellular carcinoma and melanoma. All included studies used Quantitative real-time polymerase chain reaction (Q-PCR) to detect HEIH expression, and were divided into high and low expression groups according to the difference in HEIH expression. None of the article had an Newcastle–Ottawa Scale (NOS). score below 6, and thus all the included article can be considered of high quality. The characteristics of all eligible articles are shown in Table 1.
Association between HEIH expression and prognosis
To assess the value of HEIH expression in the prognosis of cancer patients, a cumulative meta-analysis was used to performed. A total of 10 articles containing 990 patients were included. No statistically significant heterogeneity was found in these studies (I2 = 2%, P = 0.42), and therefore, a fixed effects models were used. The pooled HR was 2.03 (95% CI: 1.74-2.38, P < 0.00001) (Fig 2), indicating that the OS of the HEIH high expression group was significantly worse than that of the HEIH low expression group. This suggests that high expression of HEIH is promising as a marker for predicting poorer overall survival in cancer patients.
In addition, subgroup analysis was used to further validate the prognostic value of HEIH based on the classification of tumor type, tumor size, cut-off value, and sample type, and the results showed that high expression of HEIH in each subgroup still predicted poor prognosis (Table 2 and Fig S1).
Associations between HEIH expression and clinicopathological parameters
The OR and 95% CI were used to assess the relationship between HEIH expression and clinicopathological features as shown in Figure 3-4 and Table 3. Patients with high HEIH expression tended to have larger tumor size (large vs. small OR =2.65, 95% CI: 1.52–4.65, P =0.0006), more susceptible to lymph node metastasis (LM) (yes vs. no OR =2.07, 95% CI: 1.05–4.07, P <0.0001) and distant metastasis (DM) (yes vs. no OR =2.94, 95% CI: 1.75–4.96, P <0.0001), higher T stage (III/IV vs. I/II OR=4.76, 95%CI: 2.73–8.29, P <0.0001) and TNM stage (III/IV vs. I/II OR=2.41, 95% CI:1.54–3.77, P =0.0001). However, there was no significant correlation between HEIH expression and tumor differentiation (poor vs. well OR=1.54, 95% CI:1.00–2.37, P =0.05), age (old vs. young OR=1.13, 95% CI:1.87–1.47, P =0.35) and gender (male vs. female OR=1.18, 95% CI:0.89–1.56, P =0.26) (Fig S2). In addition, Funnel plots indicates publication bias did not affect the pooled results in the meta-analysis (Fig 5).
Publication bias and sensitivity analysis
The Begg and Egger test were used to evaluate the publication bias of the HEIH and OS, the result suggested that no significant publication bias was found (Egger test: P =0.204 and Begg test: P =0.436), and the results of Egger’s funnel plot and Begg’s funnel plot also support this conclusion (Fig 6a and 6b). In addition, sensitivity analysis showed that HR did not change significantly by removing any of the included articles, suggesting that the results are robust (Fig. 6c).