Clinical characteristics and transabdominal ultrasound features that associated with T3-T4 staging in gastric cancer: A retrospective study

Abstract

BACKGROUND

Transabdominal ultrasound after the oral administration of an echoic cellulose-based gastric ultrasound contrast agent has recently been suggested to be effective in T staging of gastric cancer (GC).

AIM

To explore the clinical characteristics and transabdominal features associated with T3-T4 staging of GC.

METHODS

In total, 113 patients who underwent transabdominal ultrasonography and had confirmed GC based on surgical pathology were included. Variables including clinical characteristics [age, body mass index, gender, clinical features, pathological type, histologic type, Lauren type, carcinoembryonic antigen (CEA), and CA19-9] and transabdominal ultrasound features (thickness and length of lesion, presence of angle sign, presence of ulcer, and lesion location) were recorded. Univariate and multivariable logistic regression analyses were performed to identify the factors associated with T3-T4 staging.

RESULTS

Ninety-two patients were T3-T4 staging and 21 were T1-T2 staging. Univariate analysis revealed that the thickness of gastric lesions (1.6 ± 0.6 cm vs 1.0 ± 0.4 cm, P < 0.001), length of gastric lesions (5.7 ± 2.2 cm vs 2.9 ± 1.0 cm, P < 0.001), presence of angle sign (92.4% vs 19.0%, P < 0.001), elevated CEA (36.9% vs 0%, P < 0.001) and elevated CA19-9 (27.2% vs 14.3%, P = 0.034) were statistically significant between the two groups. Multivariable logistic regression analysis revealed that the length of gastric lesions [odds ratio (OR) = 2.373, 95% confidence interval (95%CI): 1.281-4.396, P = 0.006] and presence of angle sign (OR = 31.083, 95%CI: 4.449-217.164, P < 0.001) were independent factors associated with T3-T4 staging. A receiver operating characteristic curve was plotted, and the area under the curve was 0.950 (95% CI: 0.906-0.994, P < 0.001).

CONCLUSION

Transabdominal ultrasound features, including an angle sign and lesion length, help identify T3-T4 staging.

Key Words: Tumor staging; Angle sign; Ultrasonography; Cellulose-based ultrasound contrast agent; Stomach neoplasms

Core Tip: The accuracy of gastric cancer (GC) staging is of great significance for the choice of surgical methods and patients’ prognosis. The length of gastric lesions and presence of angle sign were independent factors associated with T3-T4 staging. The combined use can improve diagnostic accuracy. Abdominal ultrasound provides a non-invasive, cost-effective, and reproducible imaging method for T staging of GC, suggesting its potential applicability in clinical settings.

INTRODUCTION

Gastric cancer (GC) is a common malignant tumor of the digestive tract that ranks fifth in cancer incidence globally and fourth among the causes of death[1]. Surgical resection is the primary treatment for GC. The accuracy of GC staging affects the choice of surgical method and plays an important role in patient prognosis. Whether patients with GC require splenic lymph node dissection remains controversial. According to the Chinese Society of Clinical Oncology (CSCO) guidelines, patients with GC with T1 and T2 staging do not need to undergo splenic lymph node dissection. GC patients with T3-T4 staging and those with primary tumors > 6 cm and lesions located on the greater curvature, splenic lymph node dissection is recommended because of the high risk of splenic hilar lymph node metastasis[2]. T3-T4 staging of GC can undergo omentum preservation or partial omentectomy during radical gastrectomy, which has a shorter surgical time, less bleeding, and a prognosis similar to that of total omentectomy[3,4]. When GC is T3-T4 staging, the tumor infiltrates and destroys the muscularis propria to the subserosa and serosa or invades adjacent tissues, and the prognosis is significantly worse than that of T1-T2 staging[5]. The five-year survival rate is 80%-90% for T1-T2 staging and 30%-60% for T3-T4 staging, which is related to the depth of tumor invasion[6]. Therefore, identifying T3-T4 staging accurately before surgery provides valuable information for patient management, thereby improving the postoperative quality of life and reducing mortality.

Preoperative evaluation for GC staging included endoscopic ultrasonography (EUS), computed tomography (CT), magnetic resonance imaging (MRI), and transabdominal ultrasonography. EUS can determine the depth of invasion of GC, but it also has some shortcomings, such as more contraindications and patient unwillingness to undergo EUS. In addition, as EUS is invasive, it is unsuitable for patients with upper gastrointestinal stenosis[7]. The value of CT in determining the T staging of GC is controversial because it is more difficult to diagnose T3 staging if the imaging of the intestinal serosa is irregular[8]. MRI is expensive and time-consuming; thus, it is not recommended as the first-line method for T staging[9]. Transabdominal ultrasonography displays the hierarchical structure and degree of gastric wall invasion. In addition, ultrasound has the advantage of being inexpensive, radiation-free, real-time, and repeatable. A GC lesion with high T staging could display an angle sign on ultrasound, which refers to the outer edge becoming blurry, uneven, angular, or serrated in the thickened muscularis propria resulting from cancer infiltration and invasion[10].

Some studies have found that the angle sign can help distinguish T3-T4 from T1-T2 staging[10]; however, few studies have comprehensively explored the value of the angle sign in combination with clinical characteristics and ultrasound features for T staging. Therefore, this study aimed to explore the clinical characteristics and transabdominal features associated with T3-T4 staging.

MATERIALS AND METHODS

Patients

Between January 2015 and April 2023, consecutive patients who underwent transabdominal ultrasound within two weeks before surgery were pathologically diagnosed with GC at our hospital. The inclusion criteria for this study were patients diagnosed with GC by surgical pathology, who underwent surgical treatment within one week after transabdominal ultrasound, had not undergone any radiation or chemotherapy before ultrasound, and had complete information. The exclusion criteria included patients with severe heart and lung disease or other organ failure who could not undergo surgical treatment, gastric stump cancer, or benign pathological results. In total, 113 patients were included in the final study. This retrospective study was approved by the Ethics Committee of the First People’s Hospital of Foshan (Guangdong Province, China), and the need for informed consent was waived. A flowchart of the patient inclusion process is shown in Figure 1.

Figure 1 Flow chart of patient inclusion.

Data collection

The following data were collected: (1) Clinical information: Age, gender, body mass index (BMI), clinical features, pathological type, histological type, Lauren classification, carcinoembryonic antigen (CEA), and CA19-9; and (2) Transabdominal ultrasound features: Location, hierarchical structure of the gastric wall, thickest part and length of the gastric lesion, presence of an ulcer, and presence of the angle sign. T-staging was performed according to the TNM staging criteria for GC in the seventh and eighth editions of the American Joint Commission on Cancer[11,12].

Transabdominal ultrasound examination

Transabdominal ultrasonography was performed by experienced investigators (Jian GL and Liu QX). Sonography was carried out using Esoate Mylab Twice or SIEMENS Acuson S2000, with a 1.0-8.0 MHz low-frequency convex array probe and a 3.0-11.0 MHz intermediate-frequency linear array probe.

A commercially available oral contrast agent (50 g per package; Tianxia; East Asia Medical Products Co., Ltd., Huzhou, Zhejiang, China) was mixed with 500 mL of boiling water to create a uniform thin paste. It was administered orally after cooling to a suitable temperature to aid stomach distension. This cellulose-based agent had a mildly sweet taste and was generally well-received by patients. Its acoustic properties were similar to those of the liver tissue, and the filled stomach appeared homogeneous, with a mid-to-high level of echogenicity. Antispasmodics were not administered.

The entire stomach was scanned in five steps. The first step was to scan the cardia. Second, the gastric fundus was scanned. Third, the stomach was scanned in continuous short-axis views, including the fundus, body, angle, and antrum. Fourth, the stomach was scanned in continuous coronal sections, including the fundus, body, incisura angularis, and antrum in the coronal plane. The fifth step was to scan the antrum and pylorus[13,14].

Ultrasound image analysis

The thickness and length of the lesions, presence of angle signs, presence of ulcers, and lesion location were analyzed using ultrasonography. The angle sign refers to a hypoechoic muscularis propria with a rough outer edge, fuzzy angulation, or serrated structure[6]. If any of the aforementioned conditions were present, the angle sign was identified. However, if the outer edge of the muscularis propria was smooth, no sign of an angle was observed (Figure 2). Two physicians (Jian GL and Liu QX) with more than 10 years of experience in transabdominal ultrasonography performed the image analysis without knowing the pathological results. They reached a consensus to determine whether the lesion had an angle sign.

Figure 2 Imaging features of lesions with and without angle sign on transabdominal ultrasound. A: T4a staging of gastric cancer (GC) lesion with an angle sign, which showed that hypoechoic intrinsic muscular layer protruded outward to form an angle (red arrow); B: T2 staging of GC lesion without an angle sign.

Statistical analysis

Statistical analyses were conducted using IBM SPSS Statistics software (version 19.0). Continuous data (age, BMI, thickness, and length of gastric lesion) were presented as mean± SD deviation and compared using the t-test for normally distributed data. Categorical data (gender, presence of angle sign, presence of ulcer, location, CEA, CA19-9, clinical features, histological type, Lauren classification) were presented as numbers and percentages and compared using the χ² test or Fisher’s exact test. A multivariable logistic regression model was constructed to identify the independent factors associated with T3-T4 staging. Variables with significant results in the univariate analysis were included. A receiver operating characteristic (ROC) curve was plotted, and the area under the ROC curve was calculated to determine the predictive value of the logistic regression model. P < 0.05 was considered significant.

RESULTS

General information

A total of 113 patients (71 male, 42 female; mean age, 57.3 ± 11.9 years) were included in this study. The average thickness and length of GC were 1.5 ± 0.6 cm (range 0.4-4.9 cm) and 5.2 ± 2.3 cm (range 0.4-10.8 cm), respectively. An angle sign was observed on ultrasound in 89 patients (78.7%), whereas 24 patients (21.3%) did not exhibit this sign. The patient characteristics and lesion profiles are summarized in Table 1.

Table 1 Patients characteristics and lesion profiles (n = 113).

Characteristics	N = 113
Age (years)	57.3 ± 11.9
BMI (kg/m2)	21.2 ± 3.2
Gender, n (%)
Male/female	71 (62.8)/42 (37.2)
Lesion thickness (cm)	1.5 ± 0.6
Lesion length (cm)	5.2 ± 2.3
Presence of angle sign, n (%)
Yes/no	89 (78.7)/24 (21.3)
Presence of ulcer, n (%)
Yes/no	63 (55.7)/50 (44.3)
Lesion location, n (%)
Gastric antrum	53 (46.9)
Gastric fundus	12 (10.6)
Gastric body	33 (29.3)
Gastric fundus + body	11 (9.7)
Gastric angle	4 (3.5)
Clinical feature, n (%)
Epigastrium discomfort or dull pain	36 (31.9)
Abdominal pain with black stool	18 (15.9)
Abdominal pain with belching	22 (19.5)
Abdominal distension with anorexia	17 (15.0)
Abdominal pain with vomiting	15 (13.3)
Dysphagia	5 (4.4)
CEA, n (%)
< 5 μg/L/≥ 5 μg/L	79 (69.9)/34 (30.1)
CA19-9, n (%)
< 37 kU/L/≥ 37 kU/L	85 (75.2)/28 (24.8)
Pathological type, n (%)
Adenocarcinoma	113 (100.0)
Histologic type, n (%)
Poor differentiated	71 (62.8)
Moderate differentiated	37 (32.8)
Well differentiated	5 (4.4)
Lauren type, n (%)
Diffuse type	70 (61.9)
Hybrid type	15 (13.3)
Intestinal type	28 (24.8)

BMI: Body mass index; CEA: Carcinoembryonic antigen.

Univariate analysis of influence factors associated with T3-T4 staging

The thickness of gastric lesion for T3-T4 staging was 1.6 ± 0.6 cm, significantly higher than that for T1-T2 staging (1.0 ± 0.4 cm). The difference in lesion thickness between the two groups was statistically significant (P < 0.001). The length of gastric lesions for T3-T4 staging were 5.7 ± 2.2 cm, and that for T1-T2 staging were 2.9 ± 1.0 cm. The difference in lesion length between the two groups was statistically significant (P < 0.001). In the group of T3-T4 staging, 85 cases (92.4%) had an angle sign, whereas in the group of T1-T2 staging, there were 4 cases (19.0%; P < 0.001). The percentage of patients with elevated CEA levels was significantly higher in the group of T3-T4 staging than in the group of T1-T2 staging (36.9% vs 0%, P < 0.001). The percentage of patients with elevated CA19-9 levels was significantly higher in the group of T3-T4 staging than in the group of T1-T2 staging (27.2% vs 14.3%, P = 0.034). There were no significant differences in age, BMI, gender, presence of ulcer, lesion location, clinical features, histologic type, and Lauren type between the group of T3-T4 staging and T1-T2 staging (all P > 0.05; Table 2).

Table 2 Univariate analysis of influence factors between pathological T1-T2 staging and T3-T4 staging.

Variable	T3-T4 staging (n = 92)	T1-T2 staging (n = 21)	t/χ2	P value
Age (years)	58.4 ± 12.2	57.7 ± 10.9	-0.250	0.803
BMI (kg/m2)	21.4 ± 3.2	20.1 ± 2.9	-1.862	0.065
Gender
Male/female	61/31	10/11	2.556	0.110
Lesion thickness (cm)	1.6 ± 0.6	1.0 ± 0.4	-4.022	< 0.001
Lesion length (cm)	5.7 ± 2.2	2.9 ± 1.0	-5.596	< 0.001
Presence of angle sign
Yes/no	85/7	4/17	50.684	< 0.001
Presence of ulcer
Yes/no	52/40	11/10	0.119	0.730
Lesion location
Gastric antrum	44	9	7.740	0.102
Gastric fundus	9	3
Gastric body	30	3
Gastric fundus + body	6	5
Gastric angle	3	1
CEA	58/34	21/0		< 0.001
< 5 μg/L/≥ 5 μg/L
CA19-9	67/25	18/3	0.911	0.034
< 37 kU/L/≥ 37 kU/L
Clinical feature
Epigastrium discomfort or dull pain	29	7		0.758
Abdominal pain with black stool	15	3
Abdominal pain with belching	17	5
Abdominal distension with anorexia	15	2
Abdominal pain with vomiting	13	2
Dysphagia	3	2
Histologic type
Poor differentiated	60	11	2.190	0.335
Moderate differentiated	29	8
Well differentiated	3	2
Lauren type
Diffuse type	60	10	2.579	0.275
Hybrid type	12	3
Intestinal type	20	8

BMI: Body mass index; CEA: Carcinoembryonic antigen.

Multivariate logistic regression analysis of influence factors associated with T3-T4 staging

Factors with statistical significance in univariate analysis were analyzed using multivariate logistic regression analysis. The results showed that the length of the gastric lesion [odds ratio (OR) = 2.373, 95% confidence interval (95%CI): 1.281-4.396, P = 0.006] and the presence of angle sign (OR = 31.083, 95%CI: 4.449-217.164, P < 0.001) were independent factors associated with T3-T4 staging (Table 3). The ROC curve (Figure 3) was plotted to evaluate the predictive value of the logistic regression model. The area under the ROC curve was 0.950 (P < 0.001, 95%CI: 0.906-0.994).

Figure 3 Receiver operating characteristic curve analysis of predictive value of the multivariate logistic regression model. The area under the curve was 0.950.

Table 3 Multivariate analysis of influence factors between pathological T1-T2 staging and T3-T4 staging.

	B	SE	P value	OR	95%CI
Lesion thickness (cm)	-	-	0.933	-	-
Lesion length (cm)	0.864	0.315	0.006	2.373	1.281-4.396
Angle sign (yes/no)	3.437	0.992	< 0.001	31.083	4.449-217.164
CA19-9 (< 37 kU/L/≥ 37 kU/L)	-	-	0.308	-	-
CEA (< 5 μg/L/≥ 5 μg/L)	-	-	0.997	-	-

OR: Odds ratio; 95%CI: 95% confidence interval.

DISCUSSION

In 2020, more than 1 million new cases and 768000 deaths from GC were reported worldwide. By 2040, these numbers will rise to 1.77 million new cases and 1.27 million deaths[15]. China accounts for approximately 350000 deaths from GC annually, representing 40% of the global total[1]. The CSCO guidelines[4] suggested different treatments for T1-T2 and T3-T4 staging of splenic lymph node dissection. National Comprehensive Cancer Network guidelines[16] recommend D2 dissection for T3-T4 staging to improve quality of life and reduce mortality, along with postoperative radiotherapy. Accurate preoperative T3-T4 staging identification provides reliable imaging for treatment planning and improves the postoperative quality of life. This study found that an angle sign and lesion length were independent risk factors for T3-T4 staging. These transabdominal ultrasound features help identify the T3-T4 staging of GC.

Transabdominal ultrasound after oral contrast agent administration is a new, noninvasive ultrasonic method that improves stomach imaging and enhances image quality and diagnostic accuracy. It offers high sensitivity (77.8%-100%) and specificity (94%-100%) for diagnosing various gastric lesions[17,18]. Transabdominal ultrasound can display a five-layer structure of the stomach wall[13], with three hyperechoic and two hypoechoic layers, namely the mucosal epithelium, deep mucosa, submucosa, muscularis propria, and serosa. The hypoechoic muscularis propria was more easily visualized when the stomach was filled. Abdominal ultrasound has a certain clinical value in determining T staging, with a diagnostic accuracy of 69.5%-82.9%[19-25].

The angle sign on ultrasonography was defined as a rough outer edge, fuzzy angulation, or a serrated structure in the muscularis propria within a GC lesion. In a gross specimen with T4a staging pathology, we observed a serrated structure in the muscularis propria of the gastric wall. The pathological section showed that tumor cells in the T4a stage had penetrated the muscularis propria and invaded the serosa (Figure 4). Our study found that the angle sign accounted for 92.4% (85/92) in the group of T3-T4 staging and 7.6% (4/21) in the group of T1-T2 staging (P < 0.001), which is similar to values from previous reports[10]. Multivariate analysis showed that GC with the angle sign was 31.083 times more likely to be classified as T3-T4 staging than those without the angle sign, indicating that the angle sign has a high clinical application value in the preoperative T staging of GC. Among the pathological T3-T4 staging, angle signs were not found in 7.6% (7/92) of patients, which may be related to factors such as unclear angle signs of the tumor and unclear display of the lesion being obstructed by gas[26]. In the group of pathological T1-T2 staging, four cases were over staged owing to inflammation and exudation blurring the outer edge of the muscle layer[27].

Figure 4 The manifestations of T4a staging lesion on transabdominal ultrasound, gross specimen, and pathological section. A: An angle sign on ultrasound showed the significantly thickened gastric wall (red arrow) and serrated structure (red dotted line) in the gastric angle; B: Gross specimen showed the serrated structure (red dotted line) in the gastric angle; C: Pathological section (× 400) showed that tumor cells invaded the serosa, which was T4a staging.

The thickness and length of the gastric lesions were observed via transabdominal ultrasound. The stomach ultrasound report and data system[28] noted that the thickness of the gastric wall can be used as the main indicator for GC screening. Guan et al[29] proposed that a gastric wall thickness exceeds 9.5 mm, it can be used as the cutoff value for evaluating gastric lesions. This study found the thickness and length of gastric lesions in T3-T4 staging were significantly greater than those in T1-T2 staging (1.6 ± 0.6 cm vs 1.0 ± 0.4 cm, 5.7 ± 2.2 cm vs 2.9 ± 1.0 cm, all P < 0.001). The longer the lesion, the greater the possibility of T3-T4 staging pathologically. These results indicate that lesion size helps determine the T staging.

Currently, a CEA located on chromosome 19 is a commonly used tumor marker for diagnosing malignant tumors of the digestive tract. High CEA levels are closely related to tumor burden[30,31]. Previous studies have indicated that combining tumor markers (CEA, CA72-4, and CA19-9) could improve prognostic value. A survival prediction model established by combining tumor markers was accurate and effective[32]. Our study found that CEA ≥ 5 μg/L had statistical significance in both groups (P < 0.001). In the group of T3-T4 staging, the number of cases with CEA ≥ 5 μg/L was significantly higher than that in the group of T1-T2 staging [36.9% (34/92) vs 0% (0/21), P < 0.001]. Relevant reports also indicate that CEA plays an important role in diagnosing GC and makes sense for T staging and prognosis prediction in patients with GC[32]. The proportion of CA19-9 ≥ 37 kU/L in the group of T3-T4 staging was higher than that of T1-T2 staging [27.2% (25/92) vs 14.3% (3/21), P < 0.05], which was statistically significant between the two groups.

The limitations of this study include its retrospective design and selection bias owing to the inclusion of all patients undergoing surgery. Secondly, this study focused on patients with advanced GC. Due to inconspicuous symptoms and gastric wall thickening in early GC, the diagnosis was missed during ultrasonography, and such cases could not be included in the study. Moreover, the sample size of this study was relatively small, and larger samples are required to verify the results.

The thickness of gastric lesions, length of gastric lesions, presence of angle signs, elevated CEA, and elevated CA19-9 levels were significantly different between the T3-T4 staging group and the T1-T2 staging group. The length of the gastric lesions and angle sign were independent risk factors that could effectively predict T3-T4 staging for GC, improve diagnostic efficacy, and provide a basis for selecting clinical treatment plans.

CONCLUSION

Transabdominal ultrasound features, including angle signs and lesion length, help identify T3-T4 staging in GC. This study provides an imaging reference for the precise management of patients with GC.