CHI3L1 Expression is a Prognostic Marker for Patients with Diagnosed Solid Tumors: Evidence from a Systematic Review and Meta-Analysis

Background: Accumulating studies have demonstrated YKL-40 associated with the prognosis of several cancers and contributed to the tumor progression through promoting tumor angiogenesis, migration, invasion and metastasis. The objective of this meta-analysis was to investigatethe relationship between YKL-40 and prognosis inpatients with solid tumors and seek for a new prognostic biomarker. Method: Relevant studies were searched in the Medline (PubMed), Web of science, and Embase. Pooled HR of overall survival and disease-free survival (DFS) were calculated to evaluate the strengthof the association between YKL-40 and cancer prognosis by using Stata software 14.0. Results: In total, 30 studies comprising 5160 patients were considered eligible and enrolled into the nal meta-analysis. According to the meta-analysis results, higher expression of YKL-40 predicted poorer OS (pooled HR=1.85 CI%=1.58-2.18 (cid:0) P<0.001) and DFS (pooled HR =3.63 95% CI =2.63-5.01; P<0.001). Conclusion: The current evidence suggests that YKL-40 has a predictive effect onsurvival of cancer patients as indexed by DFS and OS. YKL-40 tissue expression is a valuable prognostic biomarker and may be a promising therapeutic target for solid tumors. DFS, disease-free survival; OS, overall survival; GBM,glioblastoma (cid:0) EOC, epithelial ovarian carcinoma; STS and LT, Soft-tissue sarcomas and lipomatous tumors; BC, breast cancer; IHC, immunohistochemistry; IRS, immunoreactivity score; NSCLC, non-small-cell lung cancer; OSSC, oral squamous cell carcinoma; R, reported; NR;, no reported; SC, survival curve; UCB, carcinoma of the bladder;PGC (cid:0) Primary gastric cancer (cid:0) HC (cid:0) Hepatocellular carcinoma; NOS, Newcastle-Ottawa Scale.


Background
Cancer, a worldwide medical problem, is expected to rank as the leading cause of mortality in the 21st century(1). Although we have made a rapid progression in early diagnosis and treatment, cancer progression and development still cannot be effectively controlled by traditional diagnostic and therapeutic methods 2 . Therefore, a reliable biomarker to predict survival is needed, which will be bene cial to improve cancer patients' clinical outcome.
Human CHI3L1 is one of the members of the glycol-hydrolase family 18 (3). Human CHI3L1 is a 40-kDa mammalian glycoprotein, and an essential glutamic acid was substituted by leucine in its domain 4 . It lacks chitinase/hydrolase activity and is also known as chitinase-3-like-1(CHI3L1), YKL-40 or Chondrex 5 . YKL-40 was rst detected by Johansen in human articular chondrocytes and synovial cell 6 . YKL-40, an extracellular protein, presents in the Golgi apparatus and endoplasmic reticulum(ER) 7 . Past research showed that YKL-40 was secreted from different type of normal cells (activated neutrophils 8 , monocytes 9 , macrophages 10 ) and cancer cell lines (osteosarcoma (MG-63) 11 , ovarian cancer lines (SW626,SW480) 12 , human adenocarcinoma cell lines 13 and brain tumor cell lines (U87, SNB-75) 14 . Past studies have demonstrated that YKL-40 contributed to tumor progression and development through promoting tumor angiogenesis, migration, invasion and metastasis 15-17 . Pelloski was the rst reporter and found high YKL-40 tissue expression associated with aggressive behavior in Glioblastoma(GBM) (18). Several follow-up studies have proved that elevated serum level of YKL-40 or overexpression of YKL-40 was associated with poor prognosis in different type of tumors, such as lung cancer, anal cancer, glioblastoma, breast cancer, urologic neoplasms, Thyroid carcinoma and so on. A systemic review which investigated the prognostic value of serum/plasma YKL-40 in cancer patients has been nished by Bian et al 19 . His review revealed that elevated YKL-40 level in serum/plasma was a useful prognostic biomarker and associated with poor survival in cancer patients.
However, till now, no systemic review is to analyze the prognostic value of CHI3L1 expression in patients with solid tumor. The present systematic review and meta-analysis was conducted to investigate the prognostic value of YKL-40 expression and explore that protein maybe a promising therapeutic target.

Methods
This systematic review was conducted and followed PRISMA guidelines (20).

Study Eligibility
Studies was enrolled into the nal meta-analysis according the following inclusion criteria: (1) Studies must have been published in English and as original articles; (2) YKL-40 tissue expression was detected by IHC or RT-PCR; (3) Hazard ratios (HRs) with 95% con dence intervals (CIs) which investigated the relationship between YKL-40 levels and patients' survival outcomes were reported or enable to extrapolate these data from the data presented. Studies was excluded according the following exclusion criteria: (1) Lack of key information or could not be accessed in its entirely; (2) Letters to the editor, case report, conference abstracts, comments, review, and systemic review articles.

Quality Assessment
Two reviewers extracted data independently from the eligible studies and based on a standardized form, and any disagreement was resolved by the third author. To assess the quality of each study, we use Newcastle-Ottawa Scale (NOS) on the guidelines of the Newcastle-Ottawa Quality Assessment Scale. The main standards were: selection (0-4 points), outcome assessment (0-3points), and comparability (0-2 points). Finally, we calculated the NOS score of each study and assess the methodological quality.

Statistical analysis
Page 4/28 The cutoff value of YKL-40 tissue expression was de ned and provided in each article. The pooled hazard rations (HRs) and 95% CI for two endpoints (OS and DFS) were used to evaluate the relationship between YKL-40 and cancer prognosis. Statistical heterogeneity among studies was assessed by using Cochran's Q test and the I 2 statistic. If a P value of ≥ 0.05 or I 2 ≤ 50%, indicating the statistically signi cant, a xedeffects model was used. A random-effects model was applied to merge the HR when a P value of < 0.05 or I 2 > 50%. The potential factors contributed to the heterogeneities were analyzed via sensitivity, subgroup analysis, and meta-regression analysis. Funnel plots, Begg's and Egger's tests were employed to estimate the publication bias. All statistical analyses were conducted using Stata SE 14.0.

Study selection
A total of 1449 articles were collected from databases. After screening the full the articles, reasons for exclusion included manuscripts that focus on YKL-40 in the serum or plasma(n = 41), the study was systemic review(n = 1), insu cient data(n = 4), no survival data(n = 3). Finally, 30 articles that included 5160 participants were considered eligible in nal meta-analysis. The ow chart of the study selection is shown in Fig. 1 studies had a score of 7, and 7 studies had a score of 6. The details of assessment results and the relationship between YKL-40 and other clinical features were presented in the Table 1.

Heterogeneity Analysis
Meta-regression analyses were conducted to explore the potential reasons of heterogeneity, and we use follow covariates: geographical location, tumor type, tumor source, YKL-40 detected method, HR acquisition method and number of cases. The results show in Table 3

Publication Bias And Sensitivity Analysis
To evaluate the con dence of this study, Begg's rank correlation and Egger's linear regression were used to assess publication bias. The shapes of the funnel plot for the OS is asymmetric (Figure S4), and the p value of Egger's test was < 0.01, but the p value of Begg's tests was 0.159. However, the funnel plot for the DFS did not show obvious evidence of asymmetry ( Figure S5); the p value of Begg's and Egger's test were 0.533 and 0.571, respectively. Sensitivity analysis were further conducted to assess the stability and credibility of the heterogeneity through omitting individual studies in the OS analysis. The results were shown in Figure S3 and no individual study obviously dominated the combined HR. Although the prognostic value of YKL-40 was con rmed in our study, some limitations still need to be acknowledged. First, the total sample size was relatively small and only 30 studies were included in this meta-analysis, so failed to detect the association between overexpression of YKL-40 and some clinicopathological parameters. Second, the method of detected YKL-40 and the cutoff value of YKL-40 expression were not uniform in all the studies. Third, of the 30 studies, 14 directly provided HRs, and individual HRs of the remaining studies were extracted from survival curve using the methods reported by Tierney et al (63), which inevitably produced small statistical errors. Finally, a subgroup analysis of study types found that there was a big difference among different subgroup analysis and overall results. Still, we cannot explain the high heterogeneity for the investigated outcomes through sensitivity and metaregression analyses.

Conclusion
In summary, our comprehensive analysis clearly demonstrated the relationship between high YKL-40 expression in solid tumor tissues and cancer patient's survival. Thus, we can conclude that YKL-40 is a reliable prognostic biomarker and may be a promising therapeutic target for solid tumors. Given the limitations of the current study, prospective clinical trials, multicenter, and higher-quality studies with a uni ed criterion for determining YKL-40 expression are necessary to con rm the results of this study.

Declarations
Ethics approval and consent to participate Ethical approval for this study was obtained from Ethical Review Committeefor Biomedical Research, Anhuimedical University. Thestudy was performed in accordance with the Declaration of Helsinki. Thestudy is a systematic review and meta-analysis and no patients involved.

Consent for publication
This study did not contain any individual person's data.

Availability of data and material
This study is a systematic review and meta-analysis, the data was extractedfrom published research. The data is available by contacting correspondingauthor or extracting from original published research.

Competing interests
The authors declare that they have no competing interests.    Summary estimates and 95% CIs for overall survival, for associations between YKL-40 and survival of patients with solid tumors. Weights are from random effects analysis. CI, con dence interval; HR, hazard ratio; W (random), Weights (random effects model).

Figure 3
Forest plot of studies evaluating the relationship between high YKL-40 expression and OS in patients with different cancers. GBM, glioblastoma; EOC, epithelial ovarian carcinoma; BC, breast cancer; OS, overall survival; CI, con dence interval; HR, hazard ratio.

Figure 4
Forest plot of studies evaluating the relationship between high YKL-40 expression and DFS in patients with solid tumor. CI, con dence interval; HR, hazard ratio.