Intragastric Pressure Required to Achieve Strong Extension of the Gastric Wall for Invasion Depth Assessment in Early Gastric Cancer: A Single‐Center Prospective Observational Study

ABSTRACT

Background and Aims

Previously, we developed a simple and accurate method to measure intragastric pressure (IP) using upper gastrointestinal endoscopy. This study prospectively characterized the IP at the first point of visually defined strong extension of the gastric wall and evaluated the diagnostic performance of the non‐extension sign (NES) for assessing invasion depth in early gastric cancer (EGC).

Methods

This was a single‐center prospective study. Consecutive patients with EGC scheduled for upper gastrointestinal endoscopy were enrolled. The patients underwent continued air insufflation until strong extension of the gastric wall occurred, at which point the IP was measured. Endoscopic assessment of invasion depth was then performed using the NES.

Results

Forty‐six patients were enrolled; however, three were excluded, resulting in 43 patients included in the analysis. IP at the first point of strong extension of the gastric wall ranged from 5 to 20 mmHg; therefore, strong extension was achieved at an IP of 20 mmHg or less in all 43 patients. For the endoscopic diagnosis of the depth of invasion in EGC using the NES, this yielded a sensitivity of 80% (95% confidence interval (CI), 37.6%–96.4%), a specificity of 100% (95% CI, 90.8%–100%), and a diagnostic accuracy of 98% (95% CI, 87.9–99.6). No procedure‐related adverse events were observed.

Conclusions

IP at the first point of visually defined strong extension of the gastric wall ranged from 5 to 20 mmHg among technically evaluable patients. NES assessment showed high specificity and diagnostic accuracy, with no adverse events observed.

Trial Registration: UMIN 000046904

1. Introduction

In previous in vitro experiments, we developed a simple and accurate method for intragastric pressure (IP) measurement using an endoscope [1]. This method eliminates the need for a probe by connecting a measurement device directly to the forceps channel. Specifically, 20 mL of tap water was injected into the forceps channel for irrigation and subsequently aspirated for 10 s. This approach yielded accurate and reproducible pressure measurements with minimal complexity. We then investigated the IP at which strong extension of the gastric wall was first achieved. The mean IP ± standard deviation (SD) (range) at the time of strong gastric wall extension was 14.7 ± 3.6 mmHg (10–23). Strong extension occurred in 96% of patients (49/51) when the IP did not exceed 20 mmHg [1]. These results suggest that strong gastric wall extension can be achieved in nearly all patients when the IP remains at or below 20 mmHg during upper gastrointestinal endoscopy. However, the IP required to achieve strong extension of the gastric wall varied among patients. Moreover, it remained unclear whether a similar variation in IP would be observed in patients with early gastric cancer (EGC). For clinical application, utility must be assessed; therefore, accurate assessment of invasion depth is essential for determining the appropriateness of endoscopic treatment for EGC [2, 3]. We previously reported the non‐extension sign (NES) as a valuable marker for endoscopic diagnosis of submucosal invasive cancer, which requires strong extension of the gastric wall for evaluation [4, 5]. However, our previous evaluation of the diagnostic performance of the NES was retrospective, and IP was not prospectively measured or standardized during NES assessment. Therefore, the IP required to establish appropriate observation conditions for evaluating the NES in patients with EGC remains unclear. Thus, in this study, we aimed to investigate the IP for assessing invasion depth in EGC using the NES.

2. Methods

2.1. Patients

This single‐center prospective study was approved by the Institutional Review Board of Fukuoka University Chikushi Hospital (C22‐01‐002) and was registered with the clinical trials registry (UMIN 000046904). All participating patients were informed about the study and provided written informed consent. Consecutive patients with biopsy‐confirmed type 0 EGC [6] who were scheduled for pretherapeutic upper gastrointestinal endoscopy at the Department of Gastroenterology, Fukuoka University Chikushi Hospital from March 2022 to March 2025 were enrolled. The clinical indication for endoscopy was detailed assessment of lesion characteristics, horizontal extent, and invasion depth to guide selection of endoscopic or surgical treatment. Only those who met all inclusion criteria and none of the exclusion criteria were included.

2.2. Inclusion Criteria

1. Eastern Cooperative Oncology Group performance status score of 0 or 1 [7].

2. Age at registration ≥ 20 years.

3. Written informed consent obtained for study participation.

4. Diagnosis of type 0 EGC [6] confirmed via endoscopy, classified as Vienna category 4 or 5 [8].

5. Tumor diameter ≥ 10 mm.

6. In cases of multiple lesions, only the largest lesion was considered for inclusion.

2.3. Exclusion Criteria

1. Presence of pedunculated gastric cancer [6].

2. History of gastric surgery or presence of a gastric tube.

3. Serious underlying diseases such as heart failure and ischemic stroke.

4. Psychiatric disorders or symptoms that could interfere with study participation.

5. Patients were also excluded if the study endoscopist considered additional gastric insufflation or prolongation of the procedure unsafe because of an unforeseen clinical condition, or if the patient could not cooperate sufficiently for reliable completion of the measurement protocol.

2.4. Premedication and Sedation

All patients received optimal preparation prior to the procedure. They consumed a mixture of mucolytic and defoaming agents dissolved in water 30 min before endoscopy [9]. The preparation formula consisted of 20 000 U of pronase (Kaken Pharmaceutical, Tokyo, Japan), 1 g of sodium bicarbonate, and 10 mL of dimethylpolysiloxane (20 mg/mL; Horii Pharmaceutical), all dissolved in 100 mL of water. Sedation with midazolam (5 mg/mL; Sandoz, Tokyo, Japan) was administered upon patient request. Endoscopic procedures were performed using an electronic endoscopy system (VISERA ELITE II and EVIS X1; Olympus Co., Tokyo, Japan) equipped with a high‐resolution upper gastrointestinal endoscope featuring an optical magnification function (GIF‐H260Z, GIF‐H290Z and GIF‐XZ1200; Olympus Co.). The GIF‐H260Z, GIF‐H290Z, and GIF‐XZ1200 had distal‐end outer diameters of 10.8, 9.9, and 9.9 mm, respectively; all had a 2.8‐mm working channel and a working length of 1030 mm. The endoscope model was not standardized.

2.5. Endoscopic Procedure for Measuring IP

After preparation and pretherapeutic endoscopic observation, the examination was performed according to the study protocol. The measurement method followed our previously reported technique [1]. Before intragastric pressure measurement, 20 mL of tap water was injected into the forceps channel (Figure 1a) and the forceps channel was then aspirated for 10 s to remove the water (Figure 1b). A pressure transducer (AP‐100; Asahi Biomed Co. Ltd., Tokyo, Japan) was connected directly to the forceps channel hole through the measurement tube (Figure 1c). During gradual insufflation, IP was recorded at the first point at which the gastric folds along the greater curvature of the gastric body or the gastric wall changed from an elevated state to flattening, which was the prespecified criterion for strong extension of the gastric wall. At that point, further insufflation was stopped, NES was assessed promptly, and excess intragastric gas was then suctioned. The assessment was made in real time by the study endoscopists. Interobserver agreement was not assessed. Strong extension of the gastric wall was defined as the flattening of the gastric folds or elevated gastric wall along the greater curvature following sufficient insufflation [1] (Figure 2).

FIGURE 1.

Method for measuring intragastric pressure in patients with early gastric cancer. (a) 20 mL of tap water was injected into the forceps channel for irrigation. (b) The endoscopist aspirated through the forceps channel for 10 s to remove the water. (c) A pressure measurement device was connected directly to the forceps channel.

FIGURE 2.

Strong extension of the gastric wall was defined as the flattening of the gastric folds or elevated gastric wall along the greater curvature following sufficient insufflation.

2.6. Diagnostic Criterion for SM2: Definition of NES

According to the Japanese Classification of Gastric Carcinoma [2, 10], SM1 was defined as submucosal invasion of < 500 μm measured from the lower border of the muscularis mucosae, and SM2 was defined as submucosal invasion of ≥ 500 μm.

We used the NES, observed under a unified viewing method, as the sole endoscopic marker for diagnosing submucosal invasion classified as SM2 [4, 5]. This sign becomes visible only when the entire gastric wall is strongly distended through the insufflation of a large volume of air. Under such conditions, M‐SM1 cancers and the non‐cancerous mucosa surrounding an M‐SM1 cancer extend laterally and flatten (Figure 3). By contrast, SM2‐invasive areas fail to extend and instead form a trapezoidal elevation (Figure 4). Lesions exhibiting this characteristic were considered positive for NES. A negative NES indicated intramucosal (M) or superficial submucosal invasion (SM1), while a positive NES signified deep submucosal invasion (SM2).

FIGURE 3.

Negative non‐extension sign in SM1‐invasive early gastric cancer. (a) Weak gastric wall extension (intragastric pressure, 7 mmHg). The lesion and surrounding mucosa appear elevated (yellow arrow), mimicking deeply submucosal invasive cancer. (b) Moderate gastric wall extension (intragastric pressure, 11 mmHg). The tumor periphery tends to flatten (yellow arrow), although elevation remains evident. (c) Strong gastric wall extension (intragastric pressure, 18 mmHg). The lesion and surrounding mucosa are adequately extended (yellow arrow).

FIGURE 4.

Positive non‐extension sign in SM2 invasive early gastric cancer. With strong extension of the gastric wall through endoscopic insufflation of a large volume of air with an intragastric pressure of 17 mmHg, the SM2‐invasive area does not extend and forms a trapezoidal elevation, maintaining a distinct raised appearance (yellow arrows).

The sensitivity, specificity, and accuracy of NES for assessing the invasion depth in EGC were determined by comparing endoscopic findings with histopathological results obtained from surgical or endoscopic resection specimens.

2.7. Endpoints

The primary endpoint was the IP measurement at the first point during gradual insufflation at which the prespecified visual criterion for strong extension of the gastric wall was met. Secondary endpoints were the NES diagnostic performance compared with resection histopathology and procedure‐related adverse events.

2.8. Statistical Analysis

Continuous variables are presented as mean ± standard deviation and, where relevant, median and range; categorical variables are presented as number and percentage. The 95% confidence interval (CI) for mean IP was calculated using the distribution. Sensitivity, specificity, positive predictive value, negative predictive value, diagnostic accuracy, and technical success are presented with 95% CIs calculated using the Wilson score method. No between‐group hypothesis testing was performed. All statistical analyses were performed using EZR version 1.66 for Mac [11].

3. Results

Forty‐six consecutive patients were enrolled. In three patients, sufficient intragastric air could not be retained to achieve the protocol‐defined strong gastric wall extension; therefore, IP measurement and NES assessment could not be completed. These patients were considered technically unevaluable rather than excluded from the study population. The technical success rate was 93.5% (43/46; 95% CI, 82.5%–97.8%). Diagnostic performance was evaluated in the 43 technically evaluable patients (Table 1). Sedation was administered to all patients, and the protocol‐based examination was performed under sedation. Among them, 31 patients (72%) were men. The mean age ± SD was 73.2 ± 8.4 years, and the mean body mass index ± SD was 23.4 ± 2.7. Atrophic changes were observed in 41 patients (95%) [12, 13]. According to the Paris classification [6], 3 lesions were classified as type 0‐I, 13 as type 0‐IIa, 4 as type 0‐IIb, and 23 as type 0‐IIc. Lesion distribution by gastric location was as follows: upper third, 7 (16%); middle third, 22 (51%); and lower third, 14 (33%). Histological analysis revealed 39 differentiated adenocarcinomas and four undifferentiated adenocarcinomas.

TABLE 1.

Clinicopathological features of the 43 patients.

Mean age ± SD, (years)	73.2 ± 8.4
Sex
Male (%)	31 (73%)
Female (%)	12 (27%)
Mean body mass index ± SD, (kg m2)	23.4 ± 2.7
Macroscopic type
0‐Ia	3 (7%)
0‐IIa	13 (30%)
0‐IIb	4 (9%)
0‐IIc	23 (54%)
Location
Upper third	7 (16%)
Middle third	22 (51%)
Lower third	14 (33%)
Endoscopic gastric mucosal atrophy
Positive (%)	41 (95%)
Negative (%)	2 (5%)
Depth of invasion
Mucosa	33 (76%)
Submucosa 1 (1〜499 μm)	5 (7%)
Submucosa 2 (> 500 μm)	5 (7%)
Histological type
Differentiated	39 (91%)
Undifferentiated	4 (9%)

Abbreviation: SD, standard deviation.

As for the depth of invasion, 33 lesions were classified as intramucosal cancers (M), five lesions as superficial submucosal invasive cancers (SM1), and five lesions as deep submucosal invasive cancers (SM2).

3.1. Endpoints

All analyzable lesions showed strong extension of the gastric wall at an IP of 20 mmHg or less. The IP range was 5–20 mmHg (Figure 3). The mean ± SD IP at the first point of strong extension of the gastric wall was 14.2 ± 3.9 mmHg. The 95% CI was 13.0–15.4 mmHg, and the median value was 14.0 mmHg.

Real‐time assessment of the NES was performed by four expert endoscopists, each with more than 10 years of endoscopic experience and certification as a Board Certified Fellow of the Japan Gastroenterological Endoscopy Society (JGES). For the endoscopic diagnosis of the depth of invasion in EGC using the NES, the results were as follows: 4 true‐positive, 38 true‐negative, 0 false‐positive, and 1 false‐negative lesions (Table 2). Among the 43 technically evaluable patients, the sensitivity, specificity, and diagnostic accuracy were 80.0% (95% CI, 37.6%–96.4%), 100% (95% CI, 90.8%–100%), and 97.7% (95% CI, 87.9%–99.6%), respectively (Table 2). When technical feasibility and diagnostic correctness were considered together, 42 of the 46 enrolled patients (91.3%; 95% CI, 79.7%–96.6%) underwent both technically successful and histopathologically correct NES assessment. No procedure‐related adverse events were observed.

TABLE 2.

Diagnostic performance of conventional endoscopy for assessing depth of invasion in early gastric cancer.

		Histological diagnosis
		SM2	M‐SM1
Endoscopic diagnosis	SM2	4	0
	M‐SM1	1	38

Abbreviations: M‐SM1, mucosal cancer and superficial submucosal cancer (depth < 500 μm); SM2, deeper submucosal cancer (depth ≥ 500 μm).

4. Discussion

The mean ± SD IP at the first point of this strong extension was 14.2 ± 3.9 mmHg, with a 95% confidence interval of 13.0–15.4 mmHg and a median of 14.0 mmHg. Among the 43 technically evaluable patients, NES assessment yielded a sensitivity of 80.0% (95% CI, 37.6%–96.4%), specificity of 100% (95% CI, 90.8–%100%), and diagnostic accuracy of 97.7% (95% CI, 87.9%–99.6%). However, these estimates should be interpreted cautiously. Only five lesions were classified as SM2, resulting in substantial uncertainty in the estimate of sensitivity. In addition, three of the 46 enrolled patients could not retain sufficient intragastric air to achieve the protocol‐defined strong gastric wall extension and were therefore technically unevaluable. Because the diagnostic performance analysis was restricted to technically evaluable patients, the observed specificity and accuracy may overestimate performance in routine clinical practice. When technical feasibility and diagnostic correctness were considered together, 42 of 46 enrolled patients (91.3%) underwent both technically successful and histopathologically correct NES assessment. Therefore, the present findings should be regarded as preliminary evidence of the feasibility of NES assessment under objectively documented gastric distension rather than definitive validation of its diagnostic performance. The IP required to achieve strong gastric wall extension varied among patients (Figure 5), ranging from 5 to 20 mmHg. Therefore, the upper observed value of 20 mmHg should not be interpreted as an optimal pressure threshold for NES assessment. Rather, these findings indicate substantial interindividual variability in the pressure required to achieve the visually defined degree of gastric wall extension necessary for NES evaluation. The present study was not designed to determine whether IP monitoring improves NES assessment compared with conventional visual assessment without pressure measurement, because no such comparator group was included. Accordingly, our findings do not support routine IP monitoring or the use of a fixed pressure target in clinical practice. Instead, IP measurement should currently be regarded as a proof‐of‐concept approach for objectively characterizing and potentially standardizing gastric distension during NES assessment. Its clinical utility will require further evaluation in studies comparing prespecified pressure levels and NES assessment performed with and without IP monitoring.

FIGURE 5.

Histogram of measured intragastric pressure. All cases showed strong gastric wall extension at an intragastric pressure of 20 mmHg or less.

High specificity (100%) observed in this study suggests that establishing strong wall extension as a diagnostic criterion allows for precise identification of lesions suitable for endoscopic treatment, potentially avoiding unnecessary surgical interventions. IP requirements for adequate evaluation of invasion depth varied among cases (Figure 5). Notably, six of the 43 patients achieved strong gastric wall extension at an IP below 10 mmHg. These low IP values may reflect interindividual differences in gastric wall compliance, baseline gastric tone, or the amount of air required to achieve strong extension. However, repeated IP measurements were not performed in these patients. Therefore, although flattening of the gastric folds or gastric wall was visually confirmed at the time of measurement, the reliability of these low IP values could not be independently verified, and the possibility of measurement error related to residual fluid within the endoscope channel or other technical factors cannot be completely excluded. Future studies should include repeated measurements, particularly when unexpectedly low IP values are obtained, to confirm their reproducibility.

The endoscopic pressure study integrated system (EPSIS) and its simplified variants use pressure measurements obtained during gastric insufflation to assess gastroesophageal junction and anti‐reflux barrier function [14, 15, 16, 17, 18, 19]. The technical principle of simplified EPSIS is particularly similar to our method because both connect an external pressure‐measuring system to the endoscope working channel [17]. However, the clinical purpose and measurement endpoint are distinct. EPSIS primarily evaluates dynamic pressure behavior, including maximum IP, waveform pattern, pressure difference, and pressure gradient, as physiological markers of anti‐reflux barrier competency, whereas the present study measured IP at the first point of prespecified strong gastric wall extension to standardize the observation conditions for NES‐based invasion‐depth assessment in EGC. More recently, Tanaka et al. examined gastric fold stretching as a visual surrogate for IP without special equipment [19]. They reported that near‐complete flattening of the longitudinal folds or intervening mucosal ridges (Time range 3) was observed in 23 of 33 patients (70%) and occurred at a mean IP of 17.7 mmHg with low interindividual variability [19]. This visual concept is closely related to the strong‐extension criterion used in our study. However, Tanaka et al. evaluated folds along the greater curvature at the level of the cardia in a retroflex view and used near‐complete flattening as the visual endpoint, primarily to estimate anti‐reflux barrier competency. In contrast, our endpoint was the first point of sufficient flattening of the gastric folds or elevated gastric wall along the greater curvature of the gastric body for immediate NES assessment. In addition, all patients in the study by Tanaka et al. had non‐atrophic stomachs, whereas 95% of patients in our EGC cohort had endoscopic gastric atrophy. The procedural protocols also differed, including the insufflation gas and sedation. These differences in the visual endpoint, observation site, patient population, and procedural conditions may partly explain the difference between the mean IP of 17.7 mmHg reported by Tanaka et al. and the mean IP of 14.2 mmHg observed in the present study. Therefore, these pressure values should not be interpreted as directly interchangeable.

In clinical practice, ensuring safety is particularly important when distending the gastric wall. Although no serious events were reported, even at a pressure of 20 mmHg across all patients, the risk of complications from prolonged high‐pressure exposure cannot be entirely excluded. Therefore, achieving strong gastric wall extension and completing the diagnostic assessment quickly at that pressure level remain essential. Additionally, in three cases (7%, 3/46), adequate air retention within the stomach was difficult, preventing assessment of invasion depth. In our previous study, three patients (5.6%, 3/54) were also excluded because of difficulty retaining air within the stomach [1]. Therefore, we consider that a certain proportion of patients may be unable to retain sufficient air in the stomach despite insufflation. We believe that the inability to retain intragastric air, for example because of belching, was attributable to anatomical or physiological factors rather than failure of the measurement procedure. Factors such as belching hinder air retention in some patients, rendering NES evaluation ineffective. In such cases, alternative modalities, such as endoscopic ultrasonography should be employed [20]. One false‐negative lesion was identified, an EGC with submucosal invasion of 1020 μm. The short invasion distance into the submucosa may have contributed to the absence of visible NES features in this case [4, 5].

Although the sedation protocol was not standardized, sedation was administered to all patients in the present study. Sedation may have altered lower esophageal sphincter tone and gas retention. Although the measurement reflected the actual IP at the time of observation, an effect of sedation on the ability to achieve or maintain strong extension cannot be excluded.

This study has several limitations. First, this was a single‐center study with a small sample, including only five SM2 cancers, resulting in a wide CI for sensitivity and substantial uncertainty in the diagnostic performance estimates. Furthermore, three enrolled patients were technically unevaluable because sufficient intragastric air could not be retained. Because sensitivity, specificity, and accuracy were calculated only among technically evaluable patients, exclusion of these technical failures may have resulted in overestimation of diagnostic performance. Therefore, the reported diagnostic performance should not be generalized to all patients undergoing NES assessment. Second, both the achievement of strong gastric wall extension and the presence or absence of the NES were determined visually in real time, and these two sequential judgments were made by the same expert endoscopist. No independent blinded assessment was performed, and interobserver and intraobserver reliability were not evaluated. Therefore, the initial judgment that sufficient gastric wall extension had been achieved may have influenced the subsequent NES assessment, and observer bias cannot be excluded. Furthermore, because all assessments were performed by expert endoscopists, the reproducibility of these visual criteria among different endoscopists, particularly those with less experience, remains unknown. Future studies should include independent assessment of recorded examinations by multiple blinded observers and repeated evaluations to determine the interobserver and intraobserver reliability of both strong gastric wall extension and NES assessment. Third, sedation and endoscope model were not standardized, and their effects on gas retention and measured IP cannot be excluded. The measurement method was previously validated in vitro only with the GIF‐H290Z [1]. Fourth, IP was measured only once in each patient, and within‐patient repeatability was not assessed. Therefore, the possibility of measurement error, particularly in the six patients with values below 10 mmHg, cannot be excluded. Finally, the sample was insufficient to establish safety or generalizability, and multicenter validation is required.

In the present study, the IP required to achieve strong extension of the gastric wall varied among patients; therefore, a universally optimal IP could not be established. Future multicenter prospective studies should standardize relevant procedural factors, including the insufflation gas, sedation protocol, and endoscope model; assess the reproducibility of IP measurements and interobserver agreement in the assessment of gastric wall extension; and enroll a larger number of patients with SM2 cancers. Such studies may facilitate the establishment of patient‐specific optimal IP values for achieving adequate gastric wall extension.

In conclusion, IP at the first point of visually defined strong extension of the gastric wall ranged from 5 to 20 mmHg among technically evaluable patients. NES assessment showed high specificity and diagnostic accuracy, and no procedure‐related adverse events were observed.

Funding

The authors have nothing to report.

Ethics Statement

Approval of the research protocol by an Institutional Reviewer Board. This study was approved by the Institutional Review Board of Fukuoka University Chikushi Hospital (C22‐01‐002).

Consent

All participating patients received detailed information about the study and provided written informed consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.