Role of endoscopic ultrasound-guided tissue acquisition for the diagnosis of gastric wall thickening: a retrospective study with meta-analysis

Introduction

Abnormal gastric wall thickening poses a diagnostic challenge for the endoscopist. When the pathology is located in the subepithelial layers, with accompanying reactive fibrosis in the mucosal layer, endoscopic mucosal biopsies are often inadequate for diagnosis [1]. Ji et al showed a low diagnostic yield of 28% when deep mucosal biopsies using the bite-on-bite technique were taken for subepithelial gastric lesions [2]. Abnormal gastric wall thickening can be caused by a wide range of benign and malignant conditions, and expedient diagnosis is required to commence the appropriate treatment [3,4].

Various techniques have been described to obtain histology from the deeper layers of the stomach. Bite-on-bite mucosal biopsies are not highly targeted and may not sample deep enough to reach the abnormal area of interest [2]. Endoscopic submucosal dissection allows for deeper sampling, but the abnormal area may not be clearly visible on an endoscopic view, increasing the risk of unsuccessful sampling. Moreover, endoscopic submucosal dissection (ESD) is associated with adverse events (AE) such as bleeding and gastric wall perforation [5,6].

Endoscopic ultrasound (EUS)-guided fine-needle aspiration (FNA) or biopsy (FNB) is a promising solution to this diagnostic challenge. The lesion can be localized and targeted via EUS. Certain EUS features, such as a thickened muscularis propria layer, non-preserved gastric wall layers, and the presence of ascites and lymphadenopathy, are highly suggestive of malignancy [7,8]. EUS-guided tissue acquisition (TA) is also a reliable method of obtaining histology required for subsequent resection or systemic chemotherapy. There is a paucity of data on the diagnostic utility of EUS-TA for abnormal gastric wall thickening, with wide variation in the reported data on the outcomes.

Therefore, we conducted a retrospective analysis to determine the diagnostic outcomes of EUS-TA for abnormal gastric wall thickening. We also aimed to summarize the current evidence on the diagnostic outcomes of EUS-TA for abnormal gastric wall thickening by performing a systematic review and meta-analysis.

Patients and methods

Results

Present study

The analysis included 30 patients (18 male; median age 46, range 23-76 years). All patients had prior inconclusive endoscopic biopsies (median 2, range 1-4). Table 1 shows the baseline characteristics of each patient, along with the details of the lesion and procedure. Fig. 1 shows the details of one of the cases included in this study.

Table 1 Characteristics, procedural details, and outcomes of patients in the present study

Figure 1 (A) Endoscopy showing thickened gastric wall along the greater curvature. (B) Computed tomography showing thickened gastric wall (arrow) with loss of stratification. (C) Endoscopic ultrasound (EUS) demonstrating hypoechoic thickened gastric wall of 13 mm with loss of normal layered pattern. (D) Fine-needle aspiration under EUS guidance with 22-G needle (arrow). (E) Aspiration specimen in formalin. (F) Histopathology suggestive of poorly differentiated adenocarcinoma

The median thickness of the gastric wall, measured by EUS, was 16 (range 10-28) mm. The most common site of involvement was the gastric body, with or without the involvement of the antrum. An FNB needle was used in 93.3% of cases. The median number of passes was 2 (range 1-5). Additional samples were taken from involved perigastric lymph nodes in 14/30 (46.7%) patients. The perigastric lymph nodes were mostly hypoechoic, heterogeneous, oval or round-shaped, with sizes ranging from 12-30 mm. ROSE was available in 7/30 (23.3%) of cases. There were no reported intra- or postprocedural adverse events following the EUS-TA.

An adequate sample was obtained in 29/30 (96.6%) cases of gastric wall thickening. The EUS-FNA from gastric thickening was diagnostic in 27/30 (90%) of patients. Gastric adenocarcinoma was the most common diagnosis (13/30, 43.3%), followed by gastric lymphoma (6/30, 20%), signet ring cell carcinoma (4/30, 13.3%), metastatic disease (2/30, 6.7%), and tuberculosis (2/30, 6.7%). Of the 3 cases with an inconclusive diagnosis from pathological examination of the EUS-FNA/B sample, 1 was diagnosed as adenocarcinoma with an FNB sample from a metastatic lymph node. Two other patients with gastric thickening did not develop any progression over a 6-month follow up. Fig. 1 and 2 show the details of 2 cases of gastric thickening with negative endoscopic biopsy, 1 of which was diagnosed as gastric adenocarcinoma and another as gastric lymphoma.

Figure 2 (A) Endoscopy showing thickening of the antropyloric region with (B) endoscopic ultrasound-guided fine-needle aspiration (arrow) from gastric wall thickening with a maximum thickness of 16 mm (dotted line); (C) high-power microscopy (400×) showing sheets of large atypical lymphoid cells; and (D) immunohistochemistry (400×) showing diffusely CD20 positivity suggestive of diffuse large B-cell lymphoma

Systematic review and meta-analysis

Literature search and study characteristics

A total of 3215 records were retrieved using the above search strategy, of which 14 studies were included in the final analysis [14-26]. Fig. 3 shows the PRISMA flowchart for the study selection and inclusion process. The baseline characteristics of the included studies are summarized in Table 2. The majority of the studies were retrospective in nature, with numbers of patients ranging from 8-104. The thickness of the gastric wall varied from 5 mm to 30 mm. Two studies used FNA needles [16,18], 2 used trucut biopsy needles [14,15], 2 used FNB needles [21,25] and the rest used either FNA or FNB. The dry suction method was used in the majority of studies. The median number of passes was 2 in the majority of the studies. Supplementary Fig. 1 shows the study quality analysis using the QUADAS-2 tool. One study had a low risk of bias [22], 3 had an intermediate risk of bias [16,20,23], and the rest had a high risk of bias.

Figure 3 PRISMA flowchart for study identification, selection, and inclusion process

FNA, fine-needle aspiration

Table 2 Baseline characteristics of studies included in the present analysis

Sample adequacy

Data on sample adequacy were reported by 10 studies (n=195) [14,15,17,20-22,24-26]. The pooled rate of sample adequacy with EUS-TA from gastric wall thickenings was 94.1% (95% confidence interval [CI] 90.0-98.2; I²=28.5%) (Supplementary Fig. 2). Considering only malignant lesions (n=157), the sample adequacy rate was 92.4% (95%CI 87.4-97.4; I²=26.2%).

Sensitivity and specificity for the diagnosis of malignancy

Data from 9 studies were used to calculate the pooled sensitivity and specificity [14,15,18-20,22,23,25]. The pooled sensitivity and specificity for diagnosing malignant lesions with EUS-TA from gastric wall thickening were 94.8% (95%CI 91.3-97.2; I²=26.4%) and 100% (95%CI 93.6-100; I²=0.0%), respectively (Fig. 4). The Fagan nomogram showed that a positive result increased the pretest probability of malignancy from 50-99%, whereas a negative result decreased the pretest probability from 50% to 6% (Supplementary Fig. 3). The summary receiver operating characteristic curve showed an area under the curve of 1.00 (95%CI 0.99-1.00) (Supplementary Fig. 4).

Figure 4 Forest plot for pooled sensitivity and specificity with endoscopic ultrasound-guided transesophageal tissue acquisition from gastric wall thickening

CI, confidence interval

Diagnostic accuracy

All 14 studies (n=399) reported diagnostic accuracy with EUS-TA from gastric wall thickening [11-23]. The pooled diagnostic accuracy was 91.3% (95%CI 87.0-95.5; I²=63.0%), with moderate heterogeneity among the studies (Fig. 5). With respect only to malignant lesions, the data from 357 patients showed a pooled diagnostic accuracy of 89.5% (95%CI 84.5-94.5; I²=63.6%).

Figure 5 Forest plot for pooled diagnostic accuracy with endoscopic ultrasound-guided transesophageal tissue acquisition from gastric wall thickening

CI, confidence interval

AEs

None of the studies reported any AEs with EUS-TA from gastric wall thickening.

Publication bias and sensitivity analysis

Funnel-plot assessment showed the presence of publication bias for sample adequacy and accuracy (Supplementary Fig. 5). Deek’s plot showed significant publication bias for pooled sensitivity and specificity (Supplementary Fig. 6). Egger’s test showed the presence of a small-study effect for all the outcomes (Supplementary Table 3). There was no change in the overall effect with the leave-one-out analysis.

Discussion

Tissue diagnosis is essential for managing thickened gastric walls detected on endoscopy or cross-sectional imaging. EUS-TA is a valuable tool for the etiological diagnosis of thickened gastric walls in patients with prior negative endoscopic biopsies. The present multicenter study of 30 patients undergoing EUS-TA for evaluation of gastric wall thickening showed a sample adequacy of 96.6% and a diagnostic accuracy of 90%, without any procedure-related AE. The meta-analysis of data from 427 patients showed a pooled rate of sample adequacy and a diagnostic accuracy of 94.1% and 91.3%, respectively, for EUS-TA from gastric wall thickening. The pooled sensitivity and specificity for diagnosing malignant lesions with EUS-TA from gastric wall thickening were 94% and 99%, respectively.

Gastric wall thickening is due to a variety of conditions, both benign and malignant. Of concern are the malignant etiologies such as linitis plastica and lymphoma. EUS plays an important role in the management of early gastric cancer by determining the T stage and N stage, aiding in treatment by endoscopic therapy. In instances of diffuse wall thickening, a diagnostic EUS aids in identifying the layer of origin and the preservation of layers and extension through various layers [27]. As the standard biopsy forceps provide samples with only a mucosal layer, the yield is very low in patients with these malignancies, which usually present as wall thickening in the absence of mucosal abnormalities. Standard endoscopic biopsies have a low yield of 51.2% in diagnosing these malignancies of the infiltrative pattern [28]. This led to the European Society of Gastrointestinal Endoscopy’s guideline recommending EUS-guided sampling after the failure of standard biopsy techniques in patients with diffuse gastric wall thickening [29].

Multiple features of EUS can give clues to an underlying malignant etiology. Ascites, loss of wall structure, impaired gastric distension, and presence of pathologic lymph nodes point towards malignancy [19]. Further, Thomas et al [5] reported a significant difference in the wall thickness between benign disease and malignant disease (10 mm vs. 15 mm; P=0.037). In a study by Liu et al [17], certain features such as enlarged or effaced rugal folds, loss of distensibility despite air insufflation, circumferentially infiltrating lesions associated with hyperemic mucosal change, and small erosions on white light endoscopy (WLE) were used to determine the presence and extent of gastric linitis plastica. Their study showed a higher positive rate with EUS-FNA (71.43%) than with WLE (47.37%). However, the difference did not reach statistical significance (P=0.12), perhaps because of the small number of patients.

In the present study, associated perigastric lymph nodes were seen in 46.7% of the cases, and EUS-FNA/B was carried out at these nodes as well as the gastric thickening. Of the 3 cases with an inconclusive diagnosis on the EUS-FNA/B sample, 1 was diagnosed as adenocarcinoma with an FNB sample from a metastatic lymph node. Hence, in the presence of associated perigastric lymphadenopathy, samples should be obtained from the gastric wall as well as from nodes, as this would improve the diagnostic accuracy.

In patients with suspected gastrointestinal lymphoma, auxiliary methods such as flow cytometry (FCM) and gene rearrangement provide additional information for diagnosis and management. Yu et al utilized a gastric thickening sample obtained by EUS-FNA for FCM and analysis of gene rearrangement [16]. Adequate material for FCM was obtained in 84.6% of cases (33/39), while only 50% (14/28) of the cases had adequate material for analyzing monoclonal gene rearrangement. An FNA needle was used for TA in this study. Whether an FNB needle can improve the adequacy of the sample for conducting auxiliary tests remains a topic for future research. FNB samples may also provide a further advantage in maintaining the tissue architecture for additional histopathological assessment. Recent studies have shown that genomic profiling of an EUS-FNB sample can identifying clinically applicable druggable mutations in various cancers [30]. Therefore, further studies on the role of EUS-FNB are required for patients with gastric thickening.

In the included studies, the median number of negative endoscopic biopsies prior to EUS-TA varied from 2 to 4. Though guidelines endorse EUS after the failure of sampling by standard techniques, the number of failed endoscopic biopsies [29] after which EUS-TA should be sought is not specified in the guidelines, and this delays the diagnosis. Furthermore, the high false negative rates with standard sampling methods in gastric wall thickening indicate the impending need to adopt EUS-guided tissue acquisition as the standard of care for these patients. Hence, the threshold for EUS-TA should be lower, and patients with gastric wall thickening should be scheduled for EUS-TA as the first-line modality for tissue acquisition or immediately after one report of a negative biopsy with forceps.

Multiple other techniques have been studied to improve tissue acquisition and diagnostic yield from gastric thickening. A new technique called after-EUS judgment (AEJ) biopsy was used by Liu et al [26]: after the target lesion in the gastric wall was identified using EUS, biopsies were performed using biopsy forceps under EUS guidance. For diffuse infiltrative lesions, EUS-FNA was performed. The positive rate of biopsy by WLE was 77.93%, whereas that of AEJ biopsy was 89.38%. Notably, for diffuse thickening of the gastric wall, the positive rate of EUS was 91.67%, while it was 0% for WLE [26]. Shan et al reported a technique similar to mucosal incision-assisted biopsy for subepithelial lesions [31], called via mucosa incision EUS-guided sampling [32]. After identification of the submucosal lesion by EUS, a small incision is made, through which biopsy forceps are inserted, and a biopsy is taken from the target lesion under EUS guidance. The authors reported no perforation or massive bleeding. However, an incision is always associated with a risk of bleeding, which may make this technique less useful. Further studies are required to study the role of these techniques in the diagnosis of gastric thickening.

To the best of our knowledge, this is the first meta-analysis on the diagnostic efficacy of EUS-TA from a thickened gastric wall. The lack of significant heterogeneity for the majority of outcomes adds to the strength of the study. Nevertheless, the present meta-analysis had multiple limitations. As most studies included in the meta-analyses were retrospective, it may be associated with selection bias. No data were available for comparison of the diagnostic outcomes of EUS-FNA and FNB. Further studies with large sample sizes are needed to determine the choice of needle and suction and also compare EUS-TA with other modalities, such as standard biopsy, mucosal incision-assisted biopsy and endoscopic submucosal sampling. Moreover, 3 studies with low diagnostic accuracy [16,17,24] did not mention the sensitivity, which led to a high pooled sensitivity in the analysis. Lastly, reference standards were not mentioned in the majority of studies.

To conclude, EUS-FNA/B from gastric wall thickening provides a safe and effective diagnostic modality after indeterminate endoscopic biopsy. EUS-TA also allows assessment of the depth of invasion and simultaneous sampling of perigastric lymph nodes, which may be helpful for staging and increasing the diagnostic accuracy. Further prospective studies are required to ascertain the optimal technique of EUS-TA to improve diagnostic accuracy, and whether EUS-TA should be advised early for gastric wall thickening rather than waiting for multiple negative endoscopic biopsies.