An accurate diagnosis of diffuse gastric wall thickening is challenging for endoscopists since appropriate biopsy sampling is challenging.¹ Causes of diffuse gastric wall thickening include rugal hypertrophic gastritis (HG), amyloidosis involving the stomach, Zollinger-Ellison syndrome, lymphoma, and gastric cancer.^2-4 HG is a benign disease mimicking the morphology of diffuse infiltrative cancer such as Borrmann type 4 advanced gastric cancer (AGC B-4). Since the prognosis of advanced gastric cancer is poor, their timely and accurate diagnosis is of the utmost importance.^5,6

Although endoscopic findings, including rigidity, luminal distensibility, and the presence of mucosal break or ulcer may aid in differentiation,^7,8 endoscopic ultrasonography (EUS) can potentially assist in evaluating the gastric wall.^9-15 In AGC B-4, obtaining adequate tissue with forceps biopsy for pathologic diagnosis is challenging because of the deeply penetrative characteristics of diffuse infiltrative cancer beneath the mucosa.^1,16 Accordingly, various diagnostic methods are utilized, including strip biopsy, unroofing biopsy, EUS-guided fine needle aspiration/biopsy (EUS-FNA/B), or ultimately, surgery.¹⁰

An endoscopic biopsy can generally confirm the diagnosis without these modalities, while typical EUS findings can help differentiate the two diseases. However, in the case of early-stage AGC B-4, in which no malignant cells are found in repeated endoscopic biopsies or in which cancer invasion is limited to the muscularis propria layer without significant thickening, an early diagnosis might be delayed even if typical findings are seen in the endoscopy, which may also affect the prognosis of the patient. Therefore, we aimed to analyze the endoscopic and endosonographic features of HG and AGC B-4 and identify adequate diagnostic methods for the pathologic diagnosis of AGC B-4.

1. Patients

We retrospectively investigated patients who underwent EUS for the differential diagnosis of thickened gastric wall between January 2000 and December 2021 at Asan Medical Center. Medical history, clinical symptoms, laboratory findings, Helicobacter pylori infection status, endoscopic findings, pathologic results, EUS findings, and final diagnosis were reviewed from electronic medical records. Patients diagnosed with HG or AGC B-4 were selected, while those with other diagnoses were excluded. All diagnoses of AGC B-4 were pathologically confirmed. Most patients with HG were followed up at the hospital and confirmed to be stable without disease progression. Biopsy report showed chronic active gastritis with foveolar epithelial hyperplasia in patients with HG. Patients with AGC B-4 were classified according to the preservation of the 5-layered gastric wall structure on EUS and the presence of ulceration on endoscopy, respectively. This study was approved by the Institutional Review Board of Asan Medical Center (IRB number: 2021-1434). Informed consent was waived.

2. Endoscopic examination

All patients underwent esophagogastroduodenoscopy for evaluation of disease involvement. The presence of diffuse wall thickening was evaluated. Antral wall thickening was endoscopically defined when the background antral mucosa looks thickened or edematous, leading to a narrowed or constricted appearance of the lesion, which was not attributed to benign ulcers or deformation (Fig. 1). In addition, if there was superficial or deep ulceration between or on the thickened wall, it was defined as an ulcer. Ulcers unrelated to disease involvement, such as peptic ulcer, were excluded. In almost all patients, diffuse wall thickening could be seen on the official computed tomography reading (99.1%).

Figure 1. Endoscopic images of (A) thickened folds in a patient with hypertrophic gastritis, (B) thickened folds in a patient with Borrmann type 4 advanced gastric cancer, (C) a rare case of antral involvement in a patient with hypertrophic gastritis, and (D) antral involvement in a patient with Borrmann type 4 advanced gastric cancer.

3. Endo-ultrasonographic examination and review of EUS imaging

Experienced endoscopists (H.K.N., J.Y.A., J.H.L., K.W.J., D.H.K., K.D.C., H.J.S., G.H.L., and H.Y.J.) performed EUS with a radial/linear echoendoscope (Olympus GF series; Olympus Corp., Tokyo, Japan) or a mini-probe (UM-DP series; Olympus Optical, Tokyo, Japan) with a switchable ultrasound frequencies of 5, 7.5, 12, and 20 MHz. All patients were sedated using intravenous midazolam before the procedure, and the lumen of the stomach was filled with 300 to 600 mL of deaerated water. After the thickened gastric wall layer was endoscopically assessed, 5-gastric layers at the most thickened wall were evaluated by measuring the thickness of the proper muscle (PM) and total wall layers, and the location of the hypoechoic disruption of the 5-layered gastric wall structure was initially assessed by the endosonographers. Since the wall layers could have been influenced by conditions such as gastric distension, the examiner decided to take EUS measurements at the site where the wall thickening was most prominently observed during the procedure. Later, these findings were retrospectively reviewed again by an experienced endosonographer (D.H.K) with more than 15 years of experience using the images uploaded to the Picture Archive and Communication System. Total wall thickness was defined as thickness from the luminal to the extraluminal border. The maintained PM layer and subserosal layers were defined as a “preserved wall layer,” and disruption to the two layers was defined as a “destructed wall layer” (Fig. 2). The presence of ascites was also evaluated.

Figure 2. Endoscopic ultrasonographic images of patients diagnosed with hypertrophic gastritis (A), advanced gastric cancer with a preserved wall layer (B), and with a destroyed wall layer (C).

4. Method of biopsy

Endoscopic forceps biopsy, EUS-FNA/B, endoscopic mucosal resection (EMR) and unroofing biopsy, surgery, and other methods including ascites cytology and skin biopsy were performed for pathologic evaluation. EUS-FNA/B was conducted using a linear array echoendoscope at the thickened layers with 19- or 22-gauge needles. EMR or unroofing biopsy was performed using an electrocautery snare or knife for mucosal resection with or without submucosal injection of the saline–epinephrine solution. The success rate of each biopsy method, defined as the number of attempts divided by successful cases, was evaluated.

5. Statistical methods

All continuous variables are summarized using mean and standard deviation or median and interquartile range when they did not present with a normal distribution. Categorical variables are presented as percentages. The Student t-test was used to compare the thickness of each layer between the groups. Values of p<0.05 were considered significantly different. Receiver operating characteristic (ROC) curve analysis was used to evaluate the optimal PM layer thickness to predict advanced gastric cancer. We selected the significant factors such as thickened PM layer, sex, abdominal pain, weight loss, nausea or vomiting, antral wall thickening, and presence of ulceration and conducted a univariate analysis. Multivariable logistic analysis was performed in patients with AGC B-4 and a preserved wall layer and HG because measuring the PM layer in AGC B-4 with a destructed wall layer is insufficient. Statistical analyses were performed using SAS software (SAS Institute Inc., Cary, NC, USA).

1. Baseline characteristics

We identified 194 patients with thickened gastric wall who underwent EUS. Among them, we excluded those with normal-like rugae on endoscopy or EUS (n=8), early gastric cancer (n=2), other Borrmann type advanced gastric cancer (n=6), and lymphoma (n=13). The other Borrmann types were based on the postoperative pathologic report. A total of 50 and 115 patients were diagnosed with HG and AGC B-4, respectively. Patients with AGC B-4 were classified according to the presence of ulceration and preservation of the wall layer (Fig. 3).

Figure 3. Flowchart of patients with hypertrophied rugae who underwent endoscopic ultrasound.
EUS, endoscopic ultrasound; EGD, esophagogastroduodenoscopy; AGC, advanced gastric cancer; AGC B-4, Borrmann type 4 advanced gastric cancer.

The ratio of males to females in patients with AGC B-4 was 72:43. In contrast, female dominance was observed among those with HG (14:36). Clinical manifestations, including weight loss, nausea or vomiting, and dyspepsia were significantly more common in patients with AGC B-4 than in those with HG. Laboratory findings showed significantly lower hemoglobin and total albumin levels in patients with AGC B-4 than with HG (12.9 vs 14.6, p<0.001 and 3.9 vs 4.2, p=0.047) (Table 1).

Table 1 . Baseline Characteristics of Patients with AGC B-4 and Hypertrophic Gastritis.

Characteristics	AGC B-4 (n=115)	Hypertrophic gastritis (n=50)	p-value
Age at diagnosis, yr	55.0 (44.5–64.0)	50.5 (43.0–60.0)	0.164
Sex, male/female	72/43	14/36	<0.001
Symptom
Pain	52 (45.2)	16 (32.0)	0.158
Weight loss	56 (48.7)	6 (12.0)	<0.001
Nausea or vomiting	23 (20.0)	3 (6.0)	0.042
Dyspepsia	45 (39.1)	6 (12.0)	0.001
Underlying disease
Hypertension	20 (17.4)	12 (24.0)	0.440
Diabetes mellitus	10 (8.7)	7 (14.0)	0.452
Chronic kidney disease	1 (0.9)	2 (4.0)	0.454
Liver cirrhosis	2 (1.7)	2 (4.0)	0.751
Cerebrovascular accident	1 (0.9)	0	1.000
Angina	4 (3.5)	0	0.433
Thyroid disease	2 (1.7)	0	0.870
Family history of gastric cancer	18 (15.7)	5 (10.0)	0.472
Helicobacter pylori status			<0.001
Infected	47 (40.9)	40 (80.0)
Non-infected	18 (15.7)	2 (4.0)
Previously treated	0	1 (2.0)
Unknown	50 (43.5)	7 (14.0)
Laboratory findings
Hemoglobin, g/dL	12.9 (12.1–14.3)	14.6 (13.4–16.2)	<0.001
Albumin, g/dL	3.9 (3.7–4.2)	4.2 (3.8–4.3)	0.047
Total protein, g/dL	7.0 (6.5–7.4)	7.1 (6.9–7.7)	0.021

Data are presented as median (interquartile range) or number (%)..

AGC B-4, Borrmann type 4 advanced gastric cancer..

2. Endoscopic findings

A significantly higher rate of antral wall thickening was associated with AGC B-4 than HG (39.1% vs 4.0%, p<0.001). There were also significantly more ulcerations in patients AGC B-4 than with HG (59.1% vs 4.0%, p<0.001) (Table 2).

Table 2 . Endoscopic Findings of Patients with AGC B-4 and Hypertrophic Gastritis.

Endoscopic findings	AGC B-4 (n=115)	Hypertrophic gastritis (n=50)	p-value
Location			<0.001
Antral wall thickening	45 (39.1)	2 (4.0)
Presence of ulcer			<0.001
Present	68 (59.1)	2 (4.0)
Absent	47 (40.9)	48 (96.0)

Data are presented as number (%)..

AGC B-4, Borrmann type 4 advanced gastric cancer..

3. Endosonographic findings

Patients with destructed layers were only identified in the AGC B-4 group (n=50). A mini-probe, linear probe, and radial probe were used for diagnosis. EUS revealed significant differences in total wall thickness between AGC B-4 with persevered and destructed wall layers and HG (median [interquartile range]: 9.6 mm [8.5–14.0] vs 14.3 mm [11.5–18.8] vs 9.9 mm [6.9–14.4), p<0.001). AGC B-4 with a preserved wall layer showed a significantly larger value in PM thickness than that of the HG (median [interquartile range]: 3.9 mm (2.9–4.8) vs 1.2 mm [0.9–1.7], p<0.001). In the presence of ascites, there were significantly more patients in AGC B-4 with a destructed wall layer than with a preserved wall layer (14.0% vs 1.5%, p=0.001). No ascites was observed in HG (Table 3).

Table 3 . Endoscopic Ultrasonographic Findings of Patients with AGC B-4 and Hypertrophic Gastritis.

	AGC B-4 with preserved layers (n=65)	AGC B-4 with destructed layers (n=50)	Hypertrophic gastritis (n=50)	p-value
EUS modalities
Mini-probe	36 (55.4)	16 (32.0)	19 (38.0)	0.030
Linear probe	7 (10.8)	3 (6.0)	2 (4.0)	0.351
Radial probe	41 (63.1)	48 (96.0)	40 (80.0)	<0.001
EUS findings
Wall thickness, mm	9.6 (8.5–14.0)	14.3 (11.5–18.8)	9.9 (6.9–14.4)	<0.001
PM thickness, mm	3.9 (2.9–4.8)	-	1.2 (0.9–1.7)	<0.001
Wall <2.39 mm	5 (7.7)	-	49 (98.0)
Wall ≥2.39 mm	60 (92.3)	-	1 (2.0)
Presence of ascites	1 (1.5)	7 (14.0)	0	0.001

Data are presented as number (%) or median (interquartile range)..

AGC B-4, Borrmann type 4 advanced gastric cancer; EUS, endoscopic ultrasound; PM, muscularis proper layer..

4. Diagnostic methods of AGC B-4

Patients with AGC B-4 with ulcers showed significantly higher success rates, especially in forceps biopsy, than those without ulcers (92.6% vs 42.6%, p<0.001). In AGC B-4 without ulcer, since forceps biopsy showed a lower success rate (42.6%), additional advanced methods other than forceps biopsy were required for the pathologic diagnosis, including EUS-FNA/B, EMR, unroofing biopsy, surgery, or extragastric pathologies such as ascites cytology or skin biopsy. In the case of EUS-FNA/B, it showed a 50.0% of success rate. EMR or unroofing biopsy showed 75.0% and 50.0% success rates in patients without and with an ulcer, respectively (Supplementary Table 1). Patients who ultimately underwent surgical treatment for diagnosis had a significantly lower proportion of ulcers compared to those who did not (0% vs 44.9%, p=0.021) (Supplementary Table 2).

5. ROC for prediction of advanced gastric cancer

The ROC curve according to the thickness of total wall layers and PM layer among patients with preserved wall layer was shown in Fig. 4. The area under the curves were 0.530 and 0.988, respectively. The cutoff value for the PM layer predicting AGC B-4 with preserved wall layer was 2.39 mm, and it showed the highest sensitivity and specificity with the values of 0.92 and 1.00, respectively.

Figure 4. (A) Receiver operating characteristics curves (ROCs) for the prediction of advanced gastric cancer based on the total wall thickness and proper muscle (PM) thickness on endoscopic ultrasound of with Borrmann type 4 advanced gastric cancer with a preserved wall layer or hypertrophic gastritis. (B) ROC for the prediction of advanced gastric cancer based on the PM thickness on endoscopic ultrasound of patients with advanced gastric cancer with a preserved wall layer or hypertrophic gastritis

6. Multivariable logistic analysis of risk factors for AGC B-4 with a preserved wall layer

In the univariable analysis, a thickened PM layer (≥2.39 mm), male sex, weight loss, antral wall thickening, and ulceration significantly increased the risk of AGC B-4 with a preserved wall layer. In the multivariable analysis, a thickened PM layer (odds ratio, 637.08; 95% confidence interval, 37.88 to 10,715; p<0.001) and the presence of ulceration (odds ratio, 48.62; 95% confidence interval, 2.61 to 906.81; p=0.009) were significant risk factors for AGC B-4 with a preserved wall layer (Table 4).

Table 4 . Multivariable Logistic Analysis of Risk Factors for Borrmann Type 4 Advanced Gastric Cancer with a Preserved Wall Layer.

Factor	Univariate analysis			Multivariable analysis
	OR (95% CI)	p-value		OR (95% CI)	p-value
Thickened pm layer (≥2.39 mm)	588.21 (66.47–5,201)	<0.001		637.08 (37.88–10,715)	<0.001
Sex (male)	5.39 (2.40–12.08)	<0.001
Abdominal pain	1.51 (0.70–3.27)	0.295
Weight loss	4.29 (1.59–11.56)	0.004
Nausea or vomiting	3.19 (0.84–12.13)	0.089
Antral wall thickening	12.28 (2.73–55.3)	0.001
Presence of ulcer	31.71 (7.10–141.74)	<0.001		48.62 (2.61–906.81)	0.009

OR, odds ratio; CI, confidence interval..

Gastric cancer is approximately twice as common in males than females and can cause various symptoms, including nausea, vomiting, and weight loss after progression.^17-20 In our study, patients with AGC B-4 showed male dominance and presented with symptoms such as weight loss, nausea, vomiting, or dyspepsia compared with those with HG. Hypoacid status caused by H. pylori infection is associated with HG, which showed a significantly higher rate of H. pylori than AGC B-4.²¹ Some patients in the AGC B-4 groups who underwent upfront surgery or were admitted to the oncology department for chemotherapy were not evaluated for H. pylori infection.

Akbas et al.²² previously reported that antral wall thickening on computed tomography (13.68 mm vs 9.22 mm, p<0.05), lower hemoglobin (10.78 g/dL vs 12.64 g/dL, p<0.05), and a lower albumin level (3.36±0.57 mg/dL vs 3.97±0.57 mg/dL, p<0.05) were more frequently observed in malignant gastric disease compared with benign gastric disease. In our study, Hb and albumin levels of AGC B-4 were significantly lower than that of HG.

A previous study demonstrated that patients with Menetrier disease had diffuse thickening of the gastric wall, often with antral sparing.²³ In our study, most cases of HG did not exhibit antral wall thickening, while AGC B-4 presented a significantly higher rate of this feature. Accordingly, if antral wall thickening was noted, it might suggest a higher possibility of malignant disease than benign disease.

A few studies reported that the average gastric wall thickness in trans-abdominal ultrasonography was approximately 4 to 7 mm,^24,25 and Rapaccini et al.²⁶ suggested further evaluation for pathologic tests when the gastric wall thickness is greater than 7 mm on computed tomography. Tongdee et al.²⁷ proposed an antral thickness of 10 mm as a cutoff point criterion for differentiating malignancy and non-malignancy conditions and Lim et al.¹³ have shown that gastric wall thicknesses greater than 9.8 mm and thickened muscularis propria on EUS suggest the possibility of malignant disease. However, most patients had a wall thickness greater than 10 mm in our study. In addition, when plotting the ROC curve for total wall thickness and advanced gastric cancer, a value close to a straight line was observed (Fig. 4A), suggesting that total wall thickness alone could not predict malignancy. On the other hand, when the ROC curve for the PM layer thickness was drawn, the value of the area under the curve was 0.987 for predicting malignancy, and as a cutoff point, a PM layer of 2.39 mm showed the highest sensitivity and specificity, strongly suggesting malignancy (Fig. 4B). There were no cases of HG in the destructed PM layer, indicative of malignancy.

Since standard forceps biopsy in thickened gastric wall could frequently show a negative result for malignancy,^9,16 many studies have reported on the efficacy of EMR, unroofing biopsy, or EUS-FNA/B as a diagnostic method.¹⁰ In our study, we divided the approaches into patients with or without ulcers on endoscopy. In patients with ulcers, a simple forceps biopsy of the ulcer showed a 92.6% success rate, so advanced biopsy methods were not required. In contrast, the yield of forceps biopsy fell to 42.6% in patients without ulcers, so other diagnostic methods were required. The success rates of EUS-FNA/B and EMR and unroofing biopsy were 50.0% (5/10) and 70.0% (14/20), respectively. Although a study reported an EUS-guided biopsy accuracy rate of 38% to 98% in the thickened gastric wall,¹⁰ this result was likely due to the difficulty of targeting the PM layer, and the number of cases was too insufficient to analyze in our study.

There are a few limitations to this study. First, the EUS procedure was initially performed by multiple endoscopists and retrospectively reviewed later by one endosonographer. Even though the reviewing endosonographer possesses over 15 years of experience, interpreting and analyzing images captured by multiple practitioners can lead to some variability in assessment. However, the expertise of the reviewing endosonographer and the rigorous EUS process implemented should have minimized these discrepancies to some extent. Second, the PM layer thickness measured by EUS can vary depending on the stomach distension from water or air and the pressure applied by a balloon. These factors introduce potential variability, which should be considered when interpreting results based on the specified cutoff value. Therefore, given the inconsistency of the thickness of every wall layer, measurements should have been taken at multiple sites and compared using the average. However, although this was not done, we repeatedly observed multiple sites with long inspection times and tried to measure the thickness of the stomach wall at the site considered the thickest, where the pathological findings were the most prominent. Third, only AGC B-4 and HG were targeted in our study. Comparing the result with other malignant diseases in which the gastric wall is thickened, such as lymphoma or metastatic cancer, is warranted. Fourth, given this was a retrospective study in a single center with a small patient population, a multicenter prospective study is warranted.

In conclusion, sex, clinical symptoms, low hemoglobin, and albumin values significantly differed between HG and AGC B-4. AGC B-4 had a significantly higher rate of antral wall thickening and the presence of ulcers than HG. Destructed wall layers and 2.39 mm of PM thickness as a cutoff value are strongly indicative of AGC B-4. Forceps biopsy at the ulcer showed an excellent pathologic success rate in AGC B-4 with ulcer. Therefore, we propose EUS as a highly prognostic method for malignancy in patients with thickened PM layers of more than 2.39 mm or with an ulcer on endoscopy.

J.Y.A. is an editorial board member of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.

Study concept and design: J.Y.S., D.H.K. Data acquisition: J.Y.S. Data analysis and interpretation: J.Y.S., D.H.K. Drafting of the manuscript: J.Y.S. Critical revision of the manuscript for important intellectual content: D.H.K., J.Y.A., K.D.C., H.K.N., J.H.L., K.W.J., H.J.S., G.H.L., H.Y.J. Statistical analysis: J.Y.S., H.J.K. Approval of final manuscript: all authors.

Supplementary materials can be accessed at https://doi.org/10.5009/gnl230307.