Gastric cancer is an aggressive disease, which is the second leading cause of cancer-related death globally [
1]. It is well acknowledged that the pathogenesis of stomach cancer is a multi-step and sequential process beginning with chronic atrophic gastritis, going through intestinal metaplasia (IM), intraepithelial neoplasia and finally developing into adenocarcinoma [
2‐
5]. The IM is generally considered as a premalignant lesion contributing to the development of gastric tumor [
6,
7]. Current diagnosis of IM is based on pathological assessment of biopsy specimens with white-light endoscopy [
8]. This conventional method is time-consuming and inefficient. It fails to detect IM whose mucosal surface looks normal. Improved endoscopic techniques such as chromoendoscopy, magnifying endoscopy, narrow-band imaging (NBI) technique have been shown to improve detection and diagnosis of IM during endoscopy [
9,
10]. Above all techniques, however, are suboptimal for the detection of IM. None of them can distinguish the structure of individual cells or microstructures, and so pathologic assessment is still required [
11,
12]. Recently, a new endoscopic technique called Confocal Laser Endomicroscopy (CLE) is applied to the detection of many gastrointestinal diseases [
13‐
16]. It combines conventional white-light endoscopy with confocal laser microscopy and can be divided into two types:endoscope-based CLE (e-CLE) and probe-based CLE (p-CLE) [
17]. The greatest advantage of CLE is to simultaneously provide macroscopic and microscopic images of the gastrointestinal epithelium [
18]. The CLE has been reported to reveal high diagnostic value for digestive diseases [
13,
14,
19,
20]. However, a comprehensive systematic review of the diagnostic performance of CLE on IM has not been reported.
In this meta-analysis, the goal is to establish diagnostic accuracy of CLE in diagnosis of IM in the high-risk group of gastric cancer.