Background
Cancer is a major public health problem worldwide [
1,
2]. Liver cancer is the most commonly diagnosed cancer and the leading cause of cancer-related death in China, followed by lung and stomach cancer [
3]. Strengthening the methods for early diagnoses of cancer, as well as improved treatments, will be of great importance to reduce the death rate [
4]. Although the diagnostic tool for early cancer detection could reduce the mortality, there have been still a lack of effective biomarkers that can be used for early diagnosis of cancer. Nowadays, the gold standard for cancer tissue detection is the histological evaluation of biopsy [
5]. Histological grade was evaluated following the National Comprehensive Cancer Network clinical practice guideline of oncology (V.1.2011). Though it is the most reliable way in cancer prediction with relatively high sensitivity and specificity, its usage is still limited in clinical for the pain of patients.
There are many clinical biomarkers, such as CEA and CA199, for tumor diagnosis, whereas their sensitivities and specificities are low [
6,
7]. Therefore, it is necessary for us to investigate novel effective biomarkers for cancer diagnosis.
Circular RNA (circRNA) is a class of non-coding RNA that was discovered in eukaryotes in 1979 [
8]. CircRNAs are generated from the back splicing of exons, introns, or both to form exonic or intronic circRNAs [
9,
10]. In human cells, circRNAs are usually composed of 1–5 exons, and the exon 2 is often the upstream “acceptor” exon [
11]. A genome-wide analysis found that at least 83% of the total circRNAs are overlapping with protein-coding regions [
12]. The copy number of circRNAs can be up to ten times greater than that of associated linear RNAs, suggesting that these circRNAs possess potential biological functions rather than accidental errors during splicing [
13].
Previously, circRNAs were found and considered to have no biological function [
14]. In the twenty-first century, with the development of RNA deep sequencing technology and bioinformatics analyses, the abundance and diversity of circRNAs were identified. Numerous of studies have confirmed the functions of circRNAs in tumor cell proliferation, migration and invasion, which may be drawn into multiple types of cancer, including colorectal cancer (CRC), hepatocellular carcinoma (HCC), breast cancer, gastric cancer (GC), and so on [
15‐
19].
Although their biological functions remain largely unknown, recent studies show that circRNAs have three main functions in mammalian cells [
11,
15,
20]. First, circRNAs can regulate transcription and RNA splicing. Second, circRNAs function as microRNA (miRNA) sponges. Third, they can be translated into protein driven by N
6-methyladenosine modification [
21]. Taking advantage of RNA sequencing (RNA-seq) technology, the expressions of circRNAs were found to be dysregulated in multiple types of cancer cell lines, tumor tissues, and even plasma samples from patients, which correlated with certain clinical outcomes, suggesting the potential roles of circRNAs in tumor progression. Considering their conserved sequences and stable structures, circRNAs may potentially serve as required novel biomarker for cancer [
22,
23].
In this article, we performed a meta-analysis to summarize the overall accuracy of circRNA in different types of cancers. The sensitivity and specificity of cirRNAs were evaluated to assess the feasibility as biomarkers of cancer diagnosis.
Discussion
In recent years, the detection of circRNAs expression in tumors was gradually recognized. However, there has been no meta-analysis on the expression of circRNAs in tumors. We collected published studies on the expression of circRNAs in various tumors, including 5 up-regulated circRNAs and 14 down-regulated circRNAs. There were four circRNAs including circZKSCAN1, hsa_circ_0004018, hsa_circ_0001649, and hsa_circ_0003570 down regulated and one circRNA (hsa_circ_0005075) up regulated in HCC, and nine circRNAs including hsa_circ_0000705, hsa_circ_0006633, hsa_circ_002059, hsa_circ_0000096, hsa_circ_00001649, hsa_circ_0000190, hsa_circ_0003159, hsa_circ_0014717, and hsa_circ_0001895 down regulated in gastric cancer, whereas three circRNAs including hsa_circ_006054, hsa_circ_100219, and hsa_circ_406697 were up regulated in breast cancer. Furthermore, circRNA hsa_circ_001988 down regulated in colorectal cancer and circRNA hsa_circ_0013958 up regulated in lung adenocarcinoma.
Only 5 studies noted that the patient accepted no adjunctive treatments before the surgery, involving chemotherapy, radiotherapy, and targeted therapy [
25‐
28], and other studies not mentioned. We excluded some studies which were based on the expression of circRNA in cell lines, peripheral blood, serum, or plasma [
29].
The diagnostic value of circRNA as a biomarker for cancer was assessed according to published studies. Sensitivity and specificity are statistical measures of the value of diagnostic. The DOR and value of AUC were used to describe the characteristics of index test and its suitability as a diagnostic method. As for the overall circRNAs expressions from all cancers, the sensitivity, specificity, PLR, NLR, DOR with the corresponding 95%CIs and AUC values were 0.72 (95%CI:0.67–0.76), 0.74 (95%CI:0.69–0.78), 2.8 (95%CI:2.40–3.10), 0.38 (95%CI: 0.33–0.44), 7.00 (95%CI: 6.00–9.00), and 0.79, respectively. The results showed above indicated that circRNAs are suitable for use as diagnostic biomarkers for cancer. However, it should be noted that there was substantial heterogeneity in the pooled estimates. We performed the meta-regression based on the variables including cancer type, endogenous reference, and total sample sizes. The result showed that the heterogeneity may come from the tumor types, but not the endogenous reference and sample sizes.
Three groups of tumors had been evaluated repeatedly, which were divided into three subgroups: HCC, GC, and other tumors group. In HCC group, the expression of circRNAs showed satisfactory values of sensitivity and specificity (sensitivity: 0.73; specificity: 0.79), and its specificity values was higher than the sensitivity values. The specificity value (specificity: 0.72) of GC group was lower than HCC group. In other types of tumors group, the value of sensitivity and specificity (sensitivity: 0.68; specificity: 0.71) were disappointing, because of the low sensitivity. It is possible that the circRNAs might be suitable for use as diagnostic biomarkers of cancer. The value of sensitivity, specificity, and AUC of GC group were higher than other types of tumors group. Among all circRNAs investigated in this study, circRNA hsa_circ_0005075 showed effective diagnostic accuracy (AUC: 0.94; sensitivity: 0.833; specificity: 0.90) in HCC group, and circRNA hsa_circ_0000096 showed effective diagnostic accuracy (sensitivity: 0.88; specificity: 0.56; AUC: 0.82) in GC group. Other types of tumors group may also be available, but more researches might be needed to prove it.
CircRNAs are abundant in the brain and exosomes, with the exception of tumors. Their capability to transverse the blood–brain barrier makes them perfect candidates as potential diagnostic tools for central nervous system (CNS) disorders [
30]. A circRNA discovered in human named Cdr1as (antisense to the cerebellar degeneration-related protein 1 transcript), which also termed as ciRS-7 (circular RNA sponge for miR-7) aberrantly expressed in several cancers, including colorectal cancer [
31], hepatocellular carcinoma [
32], and gastric cancer [
20], but these study did not have completed data for a meta analysis.
circRNAs have multiple biological functions [
33]. (1) Transcription regulation. (2) Alternative splicing. (3) Regulation of parental gene translation. (4) Cell cycle regulation. (5) Protein sponge. (6) m6A-driven translation. (7) miRNA sponge. These suggested that the potential role of circRNAs in tumor progression. Detection of altered expression of circRNAs could have several advantages. First, the expressions of circRNAs were found to be dysregulated in multiple types of cancer cell lines, tumor tissues, and even plasma samples from patients. Second, circRNAs conserved sequences and stable structures. So circRNAs were deemed to be promising biomarkers for early diagnosis of cancer.
More importantly, the cancer profile in China is markedly different from those of developed countries. The diagnosis of the stomach and liver cancers in China comprise between one-third and one-half of the global incidence burden. The most common diagnostic in the United States is concentrated in prostate and breast cancers, however the expression of circRNAs in HCC and GC groups showed high values of sensitivity, specificity, and AUC. It indicates that the accurate diagnosis for the stomach and liver cancers is great social significance in China.
There are some deficiencies in the research. First, the samples are often lack of control of normal patients. Second, there is no complete data in some studies. We asked the author for data, but we did not get the detail data. Third, because the purposes of the study were different, some researches on the expression of circRNAs in cancer were excluded. We excluded some researches to ensure the accuracy of the analysis [
32,
34‐
37]. However, more study is needed.
Although technology for RNA detection have became mature, the detection reagents tend to be expensive. There are some promising methods were used in miRNA detection, including bioluminescence and high-throughput sequencing [
37,
38]. These detection technologies also can be used in circRNA detection, which might be improved before using in hospital. A quickly, accurately and cheaply detection method for circRNA may be used in hospitals in the future.