Subclassification scheme for adenocarcinomas of the esophagogastric junction and prognostic analysis based on clinicopathological features

Abstract

BACKGROUND

Adenocarcinoma of the esophagogastric junction (AEG) has distinct malignant features compared with other esophageal and gastric cancers. Its management is controversial and largely influenced by tumor location and esophageal involvement. Hence, understanding the clinicopathological characteristics and prognosis of AEG is essential for optimizing treatment strategies.

AIM

To evaluate the prognosis and clinicopathological features of patients with AEG, providing insights for management strategies.

METHODS

This retrospective study analyzed cases with AEG admitted between January 2016 and December 2017. Patients meeting the inclusion criteria were categorized into three groups: Type E [tumors whose center was located within 5 cm above the esophagogastric junction (EGJ)]; Type Eg (tumors whose center was situated within 2 cm below the EGJ), with a 2-cm esophageal invasion; Type Ge (tumors whose center was situated within 2 cm below the EGJ, with an esophageal invasion of < 2 cm, based on tumor location and esophageal involvement. Then, clinicopathological characteristics and survival outcomes of the groups were compared to evaluate the predictive value of the American Joint Committee on Cancer/International Alliance against Cancer 8^th edition gastric cancer and esophageal adenocarcinoma staging systems. Statistical analysis included survival analysis and Cox regression to assess prognostic factors.

RESULTS

Totally, 153 patients with AEG were included (median follow up: 41.1 months; 22, 31, and 100 patients from type E, Eg, and Ge, respectively), with significant differences in maximum tumor length, esophageal involvement length, tumor type, pathology, differentiation, depth of invasion, and lymph node metastasis between the groups (P < 0.05). Lymph node metastasis rates at stations 1, 2, 3, and 7 were lower in type E than in Eg and Ge (P < 0.05). Survival rates for type E (45.5%) were significantly lower than those for Eg (48.4%) and Ge (73.0%) (P = 0.001). Type E tumors, vascular infiltration, T3-T4 invasion depth, and lymph node metastasis, were identified as independent prognostic factors (P < 0.05). The gastric cancer staging system outperformed the esophageal adenocarcinoma system for type Ge tumors.

CONCLUSION

Clinicopathological characteristics and prognoses varied between the AEG groups, with type E demonstrating distinct features. The gastric cancer staging system more accurately predicted type Ge AEG prognosis, guiding clinical decision-making.

Key Words: Adenocarcinoma of esophagogastric junction; Siewert classification; Survival rate; Prognosis; Risk factors

Core Tip: This study introduces a novel subclassification for adenocarcinoma of the esophagogastric junction (AEG) based on the tumor center location and esophageal invasion. By distinguishing between types E, Eg, and Ge AEGs, the study identified significant differences in clinicopathological features and prognoses. Type E tumors were associated with poorer outcomes, and the American Joint Committee on Cancer/International Alliance against Cancer gastric cancer staging system more accurately predicted the prognosis of type Ge AEGs than did the esophageal adenocarcinoma system. This subclassification therefore improves prognosis prediction and may guide tailored clinical management for patients with AEG.

INTRODUCTION

Despite the annual declined in the global incidence of gastric adenocarcinoma in recent years, a notable increase has been observed in the global incidence of adenocarcinomas of the esophagogastric junction (AEGs)[1,2]. Hence, the development of an accurate classification and staging system is crucial for establishing effective treatment strategies and assessing the prognosis of patients with AEG. However, given the lack of a clear definition for the esophagogastric junction (EGJ) and AEG, considerable controversy has surrounded the diagnosis, classification, and treatment of AEGs[3-6].

Currently, two primary classification methods have been utilized for tumors at the EGJ: The Siewert classification and the Nishi classification. The Siewert classification, initially introduced by Siewert et al[7] in 1987, categorizes adenocarcinomas into three types based on whether the tumor centers are located within a 5-cm range both proximal and distal to the EGJ, spanning or contacting the EGJ[7]. The Nishi classification, proposed by the Japanese scholar Nishi M in 1973, categorizes malignant tumors with a 4-cm diameter whose center is located around 2 cm from the EGJ into five types based on whether they involve or extend beyond the EGJ[8]. While the Nishi method is widely used in Japan, the Siewert classification remains common globally[9-11].

Another area of controversy is the staging system for AEGs. The 2016 International Alliance against Cancer (UICC)/American Joint Committee on Cancer (AJCC) has recommended the use of the gastric cancer staging system for Siewert type III AEGs for and the esophageal cancer staging system for type I and type II AEGs[12,13]. Recent studies suggest that the esophageal cancer staging system is universally applicable for tumors located above the EGJ. However, controversy persists among scholars from both Eastern and Western countries regarding the most appropriate staging system for AEG tumors located within 2 cm below the EGJ[14-17].

Considering the variations in the etiology and common classification of AEG among Chinese and Western populations, accurate assessment of the clinicopathological features of AEG, including tumor location, upper margin, and lymph node metastasis, is crucial for exploring the most suitable classification and staging systems for individualized treatment strategies. This study aimed to provide valuable insights for the clinical management of patients with AEG by retrospectively analyzing the clinicopathological characteristics and prognosis of AEG tumors located within 2 cm of the EGJ.

MATERIALS AND METHODS

General information

A retrospective study was conducted using case data from AEG patients admitted to the Fourth Hospital of Hebei Medical University from January 2016 to December 2017. The inclusion criteria were as follows: (1) AEG radical resection (R0, no microscopic tumor) and AEG confirmed by postoperative pathological examination, with the tumor center located within 2 cm above or below the EGJ; (2) Primary treatment; (3) No history of malignancy; and (4) Complete study data. The exclusion criteria were as follows: (1) Nonadenocarcinoma or microscopic residual lesions after R0 resection; (2) Patients with distant metastasis; and (3) Patients with combined organ resection. Ultimately, 153 patients, comprising 126 males and 27 females with a mean age of 62.4 ± 8.5 years and a median follow up time of 41.1 months, were included. All case data were approved for use by the Ethics Office of the Fourth Hospital of Hebei Medical University (No. 2024KY198).

Classification scheme

This study defined the EGJ as the proximal part of the endoscopic gastric fold. The Nishi classification divides tumors into five types based on their relationship between the tumor center and the EGJ: E, EG, E = G, GE, and G. This classification is specifically applicable to regions within 2 cm of the EGJ and does not differentiate between pathological types (i.e., between adenocarcinoma or squamous carcinoma) in this area. Regarding the Siewert classification, type I tumors are those whose center is located within 1 cm to 5 cm of the EGJ, within the range of and involving the EGJ. Type II tumors are those whose center is positioned within 1 cm above the EGJ to 2 cm below the EGJ, also involving the EGJ. Type III tumors are those whose center is located 2 cm to 5 cm below the EGJ, still involving the EGJ. This study considered tumors whose center was located within or below the EGJ (as illustrated in Figure 1) and classified them into three types based on their distance from the EGJ: Type E (tumors whose center was located within 5 cm above the EGJ); Type Eg (tumors whose center was situated within 2 cm below the EGJ), with a 2-cm esophageal invasion; Type Ge (tumors whose center was situated within 2 cm below the EGJ, with an esophageal invasion of < 2 cm (Figure 1).

Figure 1 Tumor classification. In this study, adenocarcinoma of the esophagogastric junction was classified into three categories based on the location of the tumor center and the length of esophageal infiltration, within or 2 cm below esophagogastric junction (EGJ): Type E: The tumor center was situated within 5 cm above EGJ; Type Eg: The tumor center was situated within 2 cm below EGJ, with a 2 cm esophageal invasion; Type Ge: The tumor center was situated within 2 cm below EGJ, with an esophageal invasion of less than 2 cm. EGJ: Esophagogastric junction.

Staging instructions

Owing to the absence of a serosal structure in the esophagus, discrepancies in the T stages have been observed between esophageal adenocarcinoma and gastric cancer. Consequently, stages T4a and T4b, as per the gastric cancer staging system, were reclassified as T3 and T4a in accordance with the esophageal adenocarcinoma staging system. The staging for esophageal cancer and gastric cancer was determined following the AJCC/UICC 8^th edition tumor node metastasis (TNM) stage, adopting the TNM staging system for esophageal adenocarcinoma.

Treatment and follow up status

All patients underwent specific surgical treatment based on individual circumstances. This included preoperative or postoperative adjuvant chemotherapy and an adjuvant chemotherapy cycle. Postoperative follow up and telephone follow up were conducted, which concluded in August 2021. The endpoint for the follow up events was defined as either patient death or loss to follow up, and the total survival time (from the date of surgery to the date of death or last follow up) was recorded.

Observation indicators

Based on clinical data, surgical records, and follow up data, the following parameters were assessed. First, the general clinical characteristics of the three types of AEG patients were determined, including gender, age, maximum tumor length, length of esophageal invasion, tumor pathology type, nerve invasion, vascular invasion, tumor differentiation, tumor T stage, tumor N stage, tumor TNM stage, and lymph node metastasis rate (calculated by dividing the number of pathologically confirmed lymph node metastases by the total number of dissected lymph nodes, expressed as a percentage). Second, treatment specifics, encompassing the surgical approach and the number of lymph nodes dissected during the surgery, were evaluated. Third, survival statistics for the three groups were analyzed, accompanied by the creation of Kaplan-Meier survival curves to compare survival outcomes between the three groups. Fourth, factors affecting the prognosis of AEG patients were identified using univariate and multivariate analyses involving clinicopathological characteristics, treatment status, and other factors included in the univariate Kaplan-Meier analysis and multivariate Cox analysis model to elucidate the prognosis of AEG patients.

Statistical analysis

Statistical analysis was conducted using statistical product and service solutions version 22.0. Data consistent with a normal distribution were described as mean ± SD, whereas those that did not follow a normal distribution were presented as median (lower and upper quartiles). For measurement data exhibiting normal distribution and homogeneity of variance, one-way analysis of variance was employed. For nonhomogeneous data, a nonparametric rank-sum test was utilized. Count data were reported as percentages (%) and analyzed using the χ² test or either the χ² test or Fisher exact probability method for those requiring continuity correction. The Kruskal-Wallis test was applied for three-grade data. Survival rates were assessed via Kaplan-Meier survival analysis, and comparisons of survival rates were made using the log-rank test. Prognostic factors were identified using multivariate Cox regression. To evaluate the homogeneity and discriminative power of the staging system, likelihood ratios were calculated based on the evaluation index of the relevant staging system. Higher likelihood ratios suggested improved homogeneity, indicating a smaller difference in the prognosis among patients at the same stage. The linear trend χ² test value was used to assess the discriminative capacity of the staging system, with higher values signifying greater discrimination and a more significant difference in prognosis among patients at different stages. The survival rate for patients at individual stages was calculated, and the log-likelihood estimate for each typing system was determined using the Cox proportional hazards model. Smaller log-likelihood estimates indicated that the model better predicted prognosis. In all analyses, a P value of < 0.05 indicated statistical significance.

RESULTS

Patient characteristics

This study included a total of 153 patients, comprising 126 males and 27 females, with a mean age of 62.4 ± 8.5 years and a mean maximum tumor of 44.2 mm ± 15.5 mm. Among the included patients, 86 (56.2%) and 67 (43.8%) had tumors measuring ≤ 40 mm and > 40 mm, respectively. The average length of esophageal invasion by the tumor was 13.4 mm ± 13.8 mm, with 101 (66.0%) and 52 patients (34.0%) having an invasion length of < 20 mm and ≥ 20 mm, respectively. According to the Siewert classification scheme, 11 and 142 patients were classified as Siewert type I and type II, respectively. Based on the Nishi classification, 7, 5, and 74 were type Eg, type E = G, and type GE, whereas 67 patients did not fit the Nishi classification, respectively. In accordance with the classification used in this study, 22, 31, and 100 patients were classified as type E, type Eg, and type Ge. Among the analyzed patients, 27 and 126 underwent transthoracic approach surgery and transabdominal hiatal approach surgery, respectively. Of these patients, 27, 34, and 92 had transthoracic esophagectomy plus proximal gastrectomy, transabdominal proximal total gastrectomy, and total gastrectomy, respectively. A total of 145 and 8 patients had adenocarcinoma and mixed adenocarcinoma, whereas 72 and 80 patients had well-moderately differentiated and moderately-poorly differentiated tumors, respectively. The tumor type was Lauren intestinal, diffuse, and mixed in 56, 44, and 51 patients, respectively. Endovascular infiltration and nerve invasion was observed in 25 (16.3%) and 80 patients (52.3%), respectively. Additionally, 48 and 105 patients had tumor stage T1-2 and T3-4, respectively. Lymph node metastases and distant metastases were identified in 80 and 6 patients, respectively. Based on the gastric cancer staging system, 24, 16, 8, 33, 30, 28, and 14 patients were classified as stage IA, IB, IIA, IIB, IIIA, IIIB, and IIIC, respectively. According to the esophageal cancer staging system, 5, 17, 16, 1, 4, 5, 49, and 56 patients were categorized as stage IA, IB, IC, IIA, IIB, IIIA, IIIB, and IIIC/IVA (Table 1).

Table 1 Patient characteristics, mean ± SD.

Characteristic	Variables	n	%
Age (year)	62.4 ± 8.5
	≤ 65	95	62.1
	> 65	58	37.9
Sex	Male	126	82.4
	Female	27	17.6
Maximal tumor size (mm)	44.2 ± 15.5
	≤ 40 mm	86	56.2
	> 40 mm	67	43.8
Esophageal invasion length (mm)	13.4 ± 13.8
	< 20 mm	101	66.0
	≥ 20 mm	52	34.0
Siewert classification	I	11	7.2
	II	142	92.8
Nishi classification	E = G	5	3.3
	GE	74	48.4
	EG	7	4.6
	Not applicable	67	43.8
This classification	E	22	14.4
	Eg	31	20.3
	Ge	100	65.4
Surgical approaches	Transthoracic approach	27	17.6
	Transabdominal hiatal approach	126	82.4
Extent of surgical resection	Subtotal esophagectomy with partial gastrectomy	27	17.6
	Proximal gastrectomy with partial esophagectomy	34	22.2
	Total gastrectomy with partial esophagectomy	92	60.1
Histologic type	Adenocarcinoma	145	94.8
	Mixed adenocarcinoma	8	5.2
Histological grade	Well-moderately differentiated	72	47.1
	Moderately-poorly differentiated	80	52.3
Lauren type	Intestinal type	56	36.6
	Diffuse type	44	28.8
	Mixed type	51	33.3
Preoperative chemotherapy	No	104	68.0
	Yes	49	32.0
Vessel invasion	No	128	83.7
	Yes	25	16.3
Perineural invasion	No	73	47.7
	Yes	80	52.3
Pathological depth of tumor invasion	T1-2	48	31.4
	T3-4	105	68.6
Lymph node metastasis	N0	73	47.7
	N1-3	80	52.3
Distant metastasis	pM0	147	96.1
	pM1	6	3.9
TNM-GC staging system	IA	24	15.7
	IB	16	10.5
	IIA	8	5.2
	IIB	33	21.6
	IIIA	30	19.6
	IIIB	28	18.3
	IIIC	8	5.2
	IV	6	3.9
TNM-EC staging system	IA	5	3.3
	IB	17	11.1
	IC	16	10.5
	IIA	1	0.7
	IIB	4	2.6
	IIIA	5	3.3
	IIIB	49	32.0
	IIIC/IVA	50	32.7
	IVB	6	3.9

TNM: Tumor node metastasis; GC: Gastric cancer; EC: Esophageal cancer.

A comparison of the clinicopathological characteristics among patients with type E, type Eg, and type Ge AEGs is presented in Table 2. The average tumor length differed significantly between the three groups, with type Eg AEGs having the largest average tumor length (58.7 mm ± 12.6 mm), followed by type E (53.0 mm ± 18.2 mm), and type Ge (37.9 mm ± 11.2 mm) with the smallest average tumor length. The mean tumor esophageal infiltration length in type E AEG was 35.9 mm ± 13.0 mm, which was significantly greater than that type Eg (23.9 mm ± 6.0 mm) and type Ge (5.1 mm ± 5.4 mm) (P < 0.05). Significant differences in various aspects were observed between the three types of AEGs, including surgical approach, surgical resection range, tumor pathology type, tumor differentiation degree, depth of tumor invasion, lymph node metastasis, and tumor stage. Accordingly, 19 patients with type E AEGs underwent transthoracic surgery (86.4%), whereas 3 patients underwent proximal gastrectomy with partial esophagectomy through the transabdominal hiatal approach. Moreover, 8 patients with type Eg AEGs underwent transthoracic surgery (25.8%), whereas 23 patients underwent transabdominal hiatal surgery (74.2%), among whom 5 underwent proximal gastrectomy with partial esophagectomy and 18 underwent total gastrectomy with partial esophagectomy. All 100 patients with type Ge AEGs underwent transabdominal hiatal surgery, among whom 26 underwent proximal gastrectomy with partial esophagectomy and 74 underwent total gastrectomy with partial esophagectomy. Among our patients, 4 with type E (18.2%), 1 with type Eg (3.2%), and 3 with type Ge (3.0%) AEGs had mixed adenocarcinoma. Among type E, type Eg, and type Ge tumors, 12 (54.5%), 23 (74.2%), and 45 (45.0%) were moderately-poorly differentiated. In terms of tumor stage, a portion of type E (27.3%), type Eg (12.9%), and type Ge (38.0%) tumors were classified as T1-2. Lymph node metastases were observed in 59.1%, 24.0%, and 43.0% of patient with type E, of type Eg, and of type Ge AEGs, respectively. According to the gastric cancer staging system, 50.0%, 25.8%, and 62.0% of type E, type Eg, and type Ge tumors were classified as stage I-II, respectively. In contrast, according to the esophageal adenocarcinoma staging system, 9.1%, 12.9%, and 27.0% of type E, type Eg, and type Ge tumors were categorized as stage I–II, respectively. Notably, patients with type Ge AEGs tended to have earlier-stage tumors than did those with type E and type Eg AEGs (Table 2).

Table 2 Comparison of clinicopathological characteristics, mean ± SD.

Characteristic	Variables	Type E	Type Eg	Type Ge	P value
Age (year)		61.5 ± 11.7	61.0 ± 9.7	63.0 ± 7.2	0.417
	≤ 65	13	20	62
	> 65	9	11	38	0.922
Sex	Male	18	26	82	0.969
	Female	4	5	18
Tumor size (mm)		53.0 ± 18.2	58.7 ± 12.6	37.9 ± 11.2	0.000a
	≤ 40 mm	7	4	75	0.000a
	> 40 mm	15	27	25
Esophageal invasion length (mm)		35.9 ± 13.0	23.9 ± 6.0	5.1 ± 5.4	0.000a
	< 20 mm	1	0	100	0.000a
	≥ 20 mm	21	31	0
Surgical approaches	Transthoracic approach	19	8	0	0.000a
	Transabdominal hiatal approach	3	23	100
Extent of surgical resection	Subtotal esophagectomy with partial gastrectomy	19	8	0	0.000a
	Proximal gastrectomy with partial esophagectomy	3	5	26
	Total gastrectomy with partial esophagectomy	0	18	74
Histologic type	Adenocarcinoma	18	30	97	0.013a
	Mixed adenocarcinoma	4	1	3
Histological grade	Well-moderately differentiated	9	8	55	0.016a
	Moderately-poorly differentiated	12	23	45
Lauren type	Intestinal type	7	6	43	0.150
	Diffuse type	7	13	24
	Mixed type	7	12	32
Preoperative chemotherapy	No	20	19	65	0.042a
	Yes	2	12	35
Vessel invasion	No	18	24	86	0.512
	Yes	4	7	14
Perineural invasion	No	10	14	49	0.908
	Yes	12	17	51
Pathological depth of tumor invasion	T1-2	6	4	38	0.028a
	T3-4	16	27	62
Lymph node metastasis	N0	9	7	57	0.003a
	N1-3	13	24	43
Distant metastasis	M0	21	29	97	0.679
	M1	1	2	3
	I-II	11	8	62	0.002a
	III-IV	11	23	38
TNM-GC staging system	I-II	11	8	62	0.002a
	III-IV	11	23	38
TNM-EC staging system	I-II	2	4	27	0.003a
	III-IV	20	27	63

^aP < 0.05.

TNM: Tumor node metastasis; GC: Gastric cancer; EC: Esophageal cancer.

Among the 153 patients, 80 (52.3%) exhibited lymph node metastasis classified as N1-3, with an average of 37.0 ± 21.9 lymph nodes involved. Among the 22 patients with type E AEGs, 13 showed lymph node metastasis, with 11 of them having more than 15% lymph node involvement (84.6%). Among the 31 patients with type Eg AEGs, 24 had lymph node metastases, with 13 having > 15% lymph node involvement (54.2%). Among the 100 patients with type Ge AEGs, 43 exhibited lymph node metastasis, with 16 having metastasis exceeding 15% (37.2%). Regarding lymph node metastasis distribution between the three groups, patients with type E, type Eg, and type Ge AEGs presented with lymph node involvement in various regional groups. Specifically, 6 (27.3%), 19 (61.3%), and 19 (19.0%) patients with type E, type Eg, and type Ge AEGs had group 1 lymph node metastasis (right of the cardia). Notably, no group 2 lymph node metastasis (left of the cardia) was observed in patients with type E AEGs. However, 9 (29.0%) and 18 (18.0%) patients with type Eg and type Ge AEGs experienced group 2 lymph node metastasis, respectively. Moreover, 2 (9.1%), 15 (48.4%), and 26 (26.0%) patients with type E, type Eg, and type Ge AEGs showed lymph node metastasis in group 4 (gastric curvature), with a significant difference between these groups. Differences in the pattern of lymph node metastasis for group 7 (left gastric artery) and group 20 (esophageal hiatus) were also observed among the different types of AEGs (Table 3).

Table 3 Number of patients with positive nodes.

Variable	Variable	Type E	Type Eg	Type Ge	P value
Overall		13/22	24/31	43/100	0.003a
The rate of lymph node metastasis	≤ 15%	2	11	27	0.010a
	> 15%	11	13	16
Pathological depth of tumor invasion	T1-2	4	3	4	0.141
	T3-4	9	21	39
Pathological depth of tumor invasion	T1	0/1	0/1	1/23	0.956
	T2	4/5	1/3	3/15	0.051
	T3		2/3	2/4	0.659
	T4	9/16	21/24	37/58	0.058
Location of lymph node	No. 1	6	19	19	0.017a
	No. 2	0	9	18	0.018a
	No. 3	2	15	26	0.010a
	No. 4	1	6	9	0.443
	No. 5	0	5	4	0.134
	No. 6	0	1	2	0.735
	No. 7	12	20	27	0.047a
	No. 8	0	2	4	0.527
	No. 9	1	4	9	0.541
	No. 10	0	0	1	0.647
	No. 11	0	4	8	0.249
	No. 20	2	1	0	0.038a
	No. 21	1	0	0	0.074

^aP < 0.05.

Prognostic factors

In this study, the overall survival rate of patients with AEG was 64.1%. Among the 55 patients who succumbed to their disease, 49 died due to their tumor, 1 patient died due to postoperative complications, and 5 patients experienced cardiovascular and other unexpected causes of death. The mean survival time for patients with type E, type Eg, and type Ge AEGs was 42.1, 47.6, and 61.0 months, respectively. The cumulative survival rates for these types were 45.5%, 48.4%, and 73.0%, respectively, with significant differences between them (P = 0.001). The survival rate of patients with type Siewert I tumors was slightly lower (54.6%) than that of patients with type II tumors (64.8%), though the difference was not significant (P = 0.156). According to the Nishi classification, those with type Eg AEGs had a survival rate of 57.1%, which was slightly lower than that for patients with type E = G (60.0%) and type GE (75.7%). However, the difference in survival rates was not significant (P = 0.402). Univariate Cox analysis revealed that type E tumors, esophageal invasion length ≥ 2 cm, tumor length > 4 cm, mixed carcinoma, vascular invasion, nerve invasion, invasion depth T3-4, and lymph node metastasis (N1-3) were prognostic factors for patients with AEGs. Multivariate Cox analysis demonstrated that type E tumors remained an important independent prognostic factor for patients with AEGs. Patients with type E tumors had a significantly higher survival risk than did patients with type Eg [hazard ratio (HR) = 0.451; 95% confidence interval (CI): 0.207-0.984; P = 0.004] and Type Ge (HR = 0.300; 95%CI: 0.147-0.609; P = 0.001) tumors. Vascular infiltration (HR = 2.545; 95%CI: 1.399-4.631; P = 0.002), depth of infiltration T3-4 (HR = 2.945; 95%CI: 1.095-7.922; P = 0.032), and lymph node metastasis (HR = 3.643; 95%CI: 1.728-7.678; P = 0.001) were also identified as independent factors influencing the outcomes of patients with AEGs (Figure 2, Table 4 and Table 5).

Figure 2 Comparison of the survival curves of adenocarcinoma of the esophagogastric junction. A: Comparison of the survival curves of type E, Eg, and Ge adenocarcinoma of the esophagogastric junction (AEG) type E survival 45.5%, type Eg survival 48.4%, and type Ge survival 73.0%, P = 0.001; B: Comparison of the survival curves of type Siewert I and Siewert II AEG type Siewert I survival 54.5% and Siewert II survival 64.8%, P = 0.156; C: Comparison of the survival curves of type Nishi EG, E = G, and GE AEG type EG survival 57.1%, type E = G survival 60.0%, and type GE survival 75.5%, P = 0.402. AEG: Adenocarcinoma of the esophagogastric junction.

Table 4 Univariate Cox proportional hazards analysis of overall survival.

Variable	Hazard ratio	95% confidence interval	P value
This study classification
Type E	1.0		0.002
Type Eg	0.774	0.366-1.640	0.504
Type Ge	0.331	0.167-0.656	0.002
Age
< 65 years	1.0
≥ 65 years	1.110	0.647-1.904	0.705
Sex
Male	1.0
Female	0.850	0.416-1.738	0.657
Tumor size
≤ 40 mm	1.0
> 40 mm	2.235	1.307-3.823	0.003
Esophageal invasion length
< 20 mm	1.0
≥ 20 mm	2.431	1.355-4.362	0.003
Histologic type
Adenocarcinoma	1.0
Mixed adenocarcinoma	2.775	1.102-6.988	0.030
Histological grade
Well-moderately differentiated	1.0
Moderately-poorly differentiated	1.707	0.993-2.936	0.053
Lauren type
Intestinal type	1.0		0.207
Diffuse type	1.776	0.922-3.418	0.086
Mixed type	1.538	0.798-2.964	1.198
Preoperative chemotherapy
No	1.0
Yes	1.484	0.865-2.547	0.152
Vessel invasion
No	1.0
Yes	3.688	2.069-6.573	0.000
Perineural invasion
No	1.0
Yes	2.523	1.420-4.484	0.002
Pathological depth of tumor invasion
T1-2	1.0
T3-4	6.159	2.452-15.471	0.000
Lymph node metastasis
L0	1.0
L1	5.821	2.928-11.570	0.000
Distant metastasis
M0	1.0
M1	2.931	1.053-8.159	0.040

Table 5 Multivariate Cox proportional hazards analysis of overall survival.

Variable	Hazard ratio	95% confidence interval	P value
This study classification
Type E	1.0		0.004
Type Eg	0.451	0.207-0.984	0.045
Type Ge	0.300	0.147-0.609	0.001
Vessel invasion
V0	1.0
V1	2.545	1.399-4.631	0.002
Perineural invasion
P0	1.0
P1	1.129	0.599-2.127	0.709
Pathological depth of tumor invasion
T1-2	1.0
T3-4	2.945	1.095-7.922	0.032
Lymph node metastasis
L0	1.0
L1	3.643	1.728-7.678	0.001

For patients with type E tumors, significant differences in survival curves were observed according to the staging system of gastric cancer and esophageal adenocarcinoma (P = 0.039 vs P = 0.011). Patients with type Eg AEGs also showed significant differences in survival curves according to the two staging systems (P = 0.000 vs P = 0.002). In contrast, patients with type Ge tumors showed significant differences according to stage when utilizing the gastric cancer staging system (P = 0.005) but not the esophageal adenocarcinoma staging system (P = 0.113) (Figure 3).

Figure 3 Comparison of survival curves for adenocarcinoma of the esophagogastric junction with different stages in the staging system of gastric cancer and esophageal adenocarcinoma. A and B: Comparison of survival curves for type E adenocarcinoma of the esophagogastric junction (AEG) with different stages in the staging system of gastric cancer and esophageal adenocarcinoma (P = 0.039 vs P = 0.011); C and D: Comparison of survival curves for type Eg AEG with different stages in gastric cancer and esophageal adenocarcinoma staging system (P = 0.000 vs P = 0.002); E and F: Comparison of survival curves of type Ge AEG with different stages in staging system of gastric cancer and esophageal adenocarcinoma (P = 0.005 vs P = 0.113). AEG: Adenocarcinoma of the esophagogastric junction.

For patients with type E tumors, the survival curves exhibited significant differences between the staging system of gastric cancer and esophageal adenocarcinoma (P = 0.039 vs P = 0.011). Patients with type Eg AEG also showed significant differences in survival curves between the two staging systems (P = 0.000 vs P = 0.002). In contrast, for patients with type Ge tumors, the survival curves of different stages were significantly different when utilizing the gastric cancer staging system (P = 0.005), but in the esophageal adenocarcinoma staging system, there was no significant difference in survival curves (P = 0.113) (Figure 3).

Staging systems

Several statistical values were assessed to comprehensively compare the staging systems for gastric cancer and esophageal adenocarcinoma in patients with AEG. For type E tumors, the gastric cancer staging system demonstrated a greater likelihood ratio (χ² value) than did the esophageal adenocarcinoma staging system (10.896 vs 6.615). Additionally, using a linear trend, we found that the esophageal adenocarcinoma staging system had greater χ² values than did gastric cancer staging system (5.178 vs 3.327) and that the log-likelihood estimate for the gastric cancer staging system was smaller than that for the esophageal adenocarcinoma staging system (37.018 vs 37.300). For type Eg tumors, the esophageal adenocarcinoma staging system had greater χ² values than did the gastric cancer staging system (18.760 vs 17.032). Using a linear trend, we found that the gastric cancer staging system showed greater χ² values than did the esophageal adenocarcinoma staging system (11.923 vs 9.899) and that the log-likelihood estimate for the gastric cancer staging system was smaller than that for the esophageal adenocarcinoma staging system (55.624 vs 62.748). For type Ge tumors, the gastric cancer staging system had greater χ² values and linear trend χ² values than did the esophageal adenocarcinoma staging system (13.514 vs 7.575; 2.873 vs 1.196). The log-likelihood estimate for the gastric cancer staging system was smaller than that for the esophageal adenocarcinoma staging system (181.106 vs 186.322) (Table 6).

Table 6 Comprehensive comparison of staging systems for gastric cancer and esophageal adenocarcinoma in adenocarcinoma of the esophagogastric junction patients.

Project	Likelihood ratio χ2 price		Linear trend χ2 price		The log-likelihood estimate
	Gastric cancer TNM staging system	Esophageal cancer TNM staging system	Gastric cancer TNM staging system	Esophageal cancer TNM staging system	Gastric cancer TNM staging system	Esophageal cancer TNM staging system
Type E	10.896	6.615	3.327	5.178	37.018	37.300
Type Eg	17.032	18.760	11.923	9.899	55.624	62.748
Type Ge	13.514	7.575	2.873	1.196	181.106	186.322

TNM: Tumor node metastasis.

DISCUSSION

Surgical resection has been the treatment of choice for patients with resectable AEGs. The primary goal of surgery is the complete removal of the primary tumor and its lymphatic drainage. An individualized therapeutic strategy, tailored to the tumor’s location, is essential for improving overall survival[18]. Therefore, accurate preoperative evaluation is a prerequisite for selecting the optimal therapeutic strategy.

The EGJ is the virtual anatomical junction where the tubular esophagus connects with the distal stomach. However, the definition of this region varies based on whether endoscopy, imaging, and histopathology is used. During endoscopy, the EGJ is identified based on the presence of the distal longitudinal palisade vessels in the lower esophagus or the proximal margin of the gastric mucosa. On imaging, the EGJ is defined as the narrowest part of the lower esophageal segment, whereas during histopathology, it is determined as the transition from squamous to gastric epithelium. Given that the location of the distal end of the palisade vessels and the lower esophagus can be influenced by the presence of tumors at the junction and that the anatomical junction does not precisely coincide with these definitions, we adopted the definition based on the location of the proximal gastric fold, which is consistent with the widely accepted World Health Organization definition. This approach can be considered more reasonable for our study[19-21].

The Nishi classification, also referred to as the Japanese classification, presents a significant limitation in that it uniformly categorizes malignant tumors near the EGJ without differentiating between squamous cell carcinoma and adenocarcinoma. Moreover, its applicability is confined to tumors located within 2 cm above and below the EGJ. In Japan, where the early detection rates for gastric cancer has been relatively high, EGJ tumors are often identified at earlier stages. Outside of Japan, however, particularly in China and Western countries, AEG tumors are frequently diagnosed at more advanced stages, with patients often presenting with tumor diameters exceeding 4 cm. Consequently, the Nishi classification exhibits certain limitations and has been predominantly utilized within Japan, with limited international applicability. However, the widely used Siewert classification also has certain limitations, such as its inability to account for the extent of proximal and distal tumor spread and limited guidance for determining the appropriate scope of surgical resection. For example, a large-diameter Siewert type II tumor located at or below the EGJ may exhibit more esophageal infiltration than would a small-diameter Siewert type I or II AEG tumor. Apart from tumor staging and Siewert classification, the distance between the proximal tumor and the EGJ is of particular significance in clinical practice[9]. This distance is crucial considering the close association between the length of tumor invasion and the rate of mediastinal lymph node metastasis and plays a pivotal role in determining the surgical approach, resection method, and reconstruction technique. Several studies focusing on various surgical methods for Siewert type II AEGs have emphasized the importance of esophagectomy length and the distance of the proximal tumor margin as critical evaluation criteria. Such studies have shown that failure to ensure a minimum length of 2 cm can significantly impact patients’ prognosis[22,23]. A multicenter study found that the rate of mediastinal lymph node metastasis was not significantly correlated with tumor type but rather with the length of tumor invasion into the esophagus, suggesting that mediastinal node clearance is necessary when the esophagus invasion exceeds 2 cm[24]. Some scholars have underscored the importance of cleaning the lymph nodes at stations 1, 2, 3, and 7 among patients with Siewert type II and III AEGs[25,26]. These conclusions are consistent with our research findings, which showed that the lymph nodes at stations 1, 2, 3, and 7 among patients with type Eg and type Ge AEGs have markedly increased metastasis rates.

The current study defined AEG cases as those in which the tumor was located within 5 cm above and 2 cm below the EGJ and categorized them into the following three groups based on the location of the tumor’s center and 2 cm of esophageal invasion: Type E (esophageal type), type Eg (esophagogastric type with predominant esophageal involvement), and type Ge (gastroesophageal type with predominant gastric involvement). Specifically, type E tumors include those classified as type I and certain type II AEGs based on the Siewert classification, which may be closely associated with Barrett’s esophagus and the presence of esophageal adenocarcinoma extending into the EGJ. In contrast, type Eg and Ge tumors encompass other Siewert II AEGs, which might be more akin to “true cardia adenocarcinoma,” characterized by adenocarcinoma originating from the cardia and extending into the EGJ. Our study demonstrated that patients with type E AEGs had the poorest survival outcomes, with a significantly higher risk of mortality than did type Eg and Ge AEGs. This difference may be attributed to longer esophageal invasion lengths, a higher incidence of nodal metastasis, elevated rates of mediastinal lymph node involvement, an increased likelihood of vascular invasion, and the presence of mixed histological subtypes in type E cases, which have been identified as significant factors affecting the prognosis of AEG patients[27,28]. In contrast, patients with type Eg and Ge AEGs had comparable survival outcomes, which might be linked to the similar tumor characteristics and uniform surgical approaches. For these two AEG subtypes, opting for abdominal diaphragmatic esophageal aperture surgery appears to yield similar surgical results. Therefore, we conclude that the treatment approach for patients with type E tumors is appropriate and aligns with that for patients with type Eg and Ge tumors.

The 8^th edition of the AJCC/UICC TNM staging system for esophageal cancer and gastric cancer was used in the staging of AEGs. This suggests that for Siewert I AEGs, the esophageal cancer TNM staging system (TNM-EC) and corresponding treatment guidelines should be applied, whereas for Siewert III AEGs, the gastric cancer TNM staging system (TNM-GC) and related guidelines are more appropriate. Nevertheless, the classification and treatment of Siewert II AEGs remains controversial[29]. In fact, a retrospective study found that the 8^th TNM-GC scheme is superior to TNM-EC in predicting the prognosis of Siewert II AEGs[14,30]. In the current study, the prognosis of different AEG types was evaluated using both the TNM-EC and TNM-GC systems. Our findings revealed that for type E AEGs, the TNM-EC better predicted the survival risk of patients than did the TNM-GC, indicating greater homogeneity and discriminative power. This could be due to the similarities in esophageal invasion length, lymph node metastasis range, and histological type between type E AEGs and esophageal cancer. For type Ge AEG, the TNM-GC was more effective at distinguishing survival differences among patients with different tumor stages than did the TNM-EC, showing an increasing trend in the risk ratio and better separation of survival curves. This might be related to the similarities in gastric involvement, lymph node metastasis range, and histological type between type Ge AEGs and gastric cancer. In the case of type Eg AEGs, both systems effectively differentiated survival differences among stages, but the TNM-GC demonstrated better risk ratio trends, survival curve separation, homogeneity, and discriminative power than did the TNM-EC. Given the patients with type Eg AEGs exhibit mixed biological behavior of esophageal and gastric cancer, comprehensive prognostic evaluation is warranted. However, due to the detailed classification of AEG lymph node metastasis in the TNM-GC and its alignment with surgical lymph node dissection, the TNM-GC may better reflect the prognosis of patients with type Eg AEGs.

CONCLUSION

The present study retrospectively analyzed 153 patients with AEGs and proposed a novel and practical classification method for AEG by comparing the applicability of two commonly used staging systems for different AEG types. Although this research offers valuable insights for clinical diagnosis and treatment, it is subject to certain limitations. The study’s single-center and retrospective nature, along with its relatively small sample size and limited follow up duration, may introduce biases. Furthermore, the study did not thoroughly examine the impact of preoperative chemotherapy on the prognosis of AEGs. Moreover, the proposed AEG classification method has not been validated by other centers or in large samples, hightailing the need for further confirmation of its generalizability and stability. To validate the presented conclusions and methods, large-scale, prospective, multicenter, randomized controlled trials are essential.