Study Selection and Characteristics
The process of literature search and selection was detailed in Figure 1. In total, 75 potentially relevant records were obtained. After excluding the duplicated and unqualified papers, 13 studies involving 914 patients with 8 different types of cancers were enrolled in this meta-analysis ultimately [5,10,13,14,17-25]. These included studies comprised renal cell carcinoma , glioma [14,18], hepatocellular carcinoma , colorectal cancer [13,20,21,24], ovarian clear cell carcinoma , gastric cancer , esophageal squamous cell carcinoma  and osteosarcoma [23,25].
The major characteristics of the eligible articles were summarized in Table 1. All included studies were conducted in China and published from 2017 to 2019. The sample size of the included studies ranged from 30 to 141. The expression level of lncRNA SNHG6 was detected by quantitative real-time polymerase chain reaction (qRT-PCR) in all studies and all patients of each study were divided into high and low groups based on the expression of SNHG6. Of the 13 studies, 6 studies recorded the HR and corresponding 95% CI for OS, and data on OS of the other 7 studies were extrapolated through Kaplan-Meier curves indirectly. Additionally, all included studies were considered high quality because of the NOS scores were more than 6 for each study.
Prognostic value of SNHG6 expression in solid cancers
The HR and 95%CI from 13 studies (including 914 patients) was combined to determine the association between lncRNA SNHG6 expression and OS. As shown in Figure 2, no obvious heterogeneity was observed among the studies (I2 = 0.00%, p =0.994). Therefore, a fixed-effect model was applied. The pooled HR was 2.14 (95%CI: 1.61~2.67, p < 0.001), indicating that patients with increased expression of lncRNA SNHG6 predicted a poor OS in 8 types of human cancers (Figure 2). Meanwhile, the independent prognostic value of SNHG6 expression was also assessed based on the multivariate analysis in 6 studies with 514 patients (Figure 3A). The pooled results revealed that SNHG6 expression was an independent prognostic factor for OS in cancer patients (HR = 2.21, 95%CI: 1.46-2.96, p < 0.001; I2 = 0.0%, p = 0.892) . Particularly, for colorectal cancer, the pooled HR was 2.62 with 95%CI (1.23-4.01) (Figure 3B). In addition, the prognostic value of SNHG6 expression for RFS was also assessed in 2 studies with 221 patients (Figure 3C). The pooled result indicated that increased SNHG6 expression was associated with a poor RFS in hepatocellular carcinoma and colorectal cancer (HR = 3.27, 95%CI: 1.42-5.12, p < 0.001; I2 = 0.0%, p = 0.93).
Furthermore, subgroup analysis of OS was also performed according to types of tumor, sample size, and survival analysis, as shown in Figure 3. Stratified analysis showed that SNHG6 overexpression could predict unfavourable OS in digestive system (HR = 2.5, 95%CI: 1.57-3.48, p < 0.001; I2 = 0.0%, p = 0.998), and other system (HR = 1.97, 95%CI: 1.33-2.61, p < 0.001; I2 = 0.0%, p < 0.874). And we also found that evaluated SNHG6 level significantly related to unfavorable OS in the studies with sample size <70 (HR = 2.70, 95%CI: 1.29-4.11, p < 0.001; I2 = 0.0%, p = 0.950), as well as those with sample size ≥ 70 (HR = 2.05, 95%CI: 1.48-2.62, p < 0.001; I2 = 0.0%, p = 0.970). Moreover, higher SNHG6 expression could predict poorer outcome in the studies carried out by univariate and multivariate (U/M) analysis (HR = 2.21, 95%CI: 1.46-2.96, p < 0.001; I2 = 0.0%, p = 0.892), as well as those without U/M analysis (HR = 2.07, 95%CI: 1.32-2.82, p < 0.001; I2 = 0.0%, p = 0.961).
Correlation between SNHG6 and clinicopathologic characteristics
A correlation between lncRNA SNHG6 expression and clinicopathological features were obtained from OR analysis. The combined results were shown in Table 2. The pooled results from 4 studies indicated that the high lncRNA SNHG6 expression was related to tumor invasion depth (OR = 1.76, 95%CI: 1.18-2.63, p = 0.006, I2 = 0.24%), lymph node metastasis (LNM) (OR = 1.60, 95%CI: 1.18-2.17, p = 0.002, I2 = 5.57%), distant metastasis (DM) (OR = 1.90, 95%CI: 1.37-2.64, p < 0.001, I2 = 0.73%) and advanced TNM stage (OR = 1.88, 95%CI: 1.36-2.60, p < 0.001, I2 = 1.3%). In addition, for colorectal cancer, the pooled results also suggested that the elevated SNHG6 expression was associated with LNM (OR = 1.80, 95%CI: 1.11-2.92; Figure 5A), DM (OR = 1.92, 95%CI: 1.15-3.20; Figure 5B), and TNM (OR = 1.82, 95%CI: 1.22-2.73; Figure 5C). Therefore, our meta-analysis suggested that lncRNA SNHG6 overexpression was associated with advanced clinicopathological characteristics.
Publication bias and sensitivity analysis
The publication bias was evaluated by Begg’s funnel plot and Egger’s linear regression tests in the present meta-analysis. Visual inspection of the funnel plot revealed the absence of asymmetry (Figure 6A), as well as Egger’s test showed probable evidence for publication bias in our meta-analysis (t = 7.12, p < 0.001). Therefore, we preformed trim and fill analysis with a fixed-effect model to assessed the impact of potential publication bias. The pooled analysis incorporation the hypothetical studies continued to show a significant association between SNHG6 expression with OS in human cancers (corrected HR = 2.07, 95%CI: 1.73-2.48, p < 0.001). As shown in Figure 6B, We also performed trim and fill analysis when evaluating the independent prognostic value of SNHG6 expression for OS in cancers because of the present of asymmetry of funnel plot and the result of Egger’s test (t = 8.52, p = 0.001). The pooled data also showed a relationship between SNHG6 overexpression with poor OS in human cancers (corrected HR = 2.16, 95%CI: 1.70-2.75, p < 0.001). Publication bias in the RFS groups was not analyzed due to the small number of studies.
The sensitivity analysis was carried out by removing each study in turn from the pooled analysis to examine the impact of the removed study on the overall HRs. As shown in Figure 6C-D, the pooled HR was not significantly changed when removing any of the included studies, suggesting the robustness of the results in the present research.