Siewert Classification of Gastroesophageal Junction Adenocarcinomas
The Siewert classification divides gastroesophageal junction (GEJ) adenocarcinomas into three anatomically distinct types based on tumor epicenter location, which directly determines the appropriate surgical approach: esophagectomy for Type I, extended total gastrectomy for Type III, and either approach for Type II depending on specific tumor characteristics. 1
Classification System (2000 Revision)
The currently used Siewert classification defines three tumor types based on the location of the tumor center relative to the anatomic GEJ 1:
Type I (Distal Esophageal Adenocarcinoma): Tumor center located 1-5 cm above the anatomic GEJ, typically arising from Barrett's esophagus 1
Type II (True Cardia Carcinoma): Tumor center within 1 cm above and 2 cm below the anatomic GEJ 1
Type III (Subcardial Gastric Carcinoma): Tumor center 2-5 cm below the anatomic GEJ, infiltrating the GEJ and distal esophagus from below 1
Biological and Prognostic Differences
These three types represent distinct biological entities with different clinical behaviors 2, 3:
Type I tumors are more frequently associated with Barrett's esophagus, show better histological differentiation, have lower rates of lymph node metastases, and demonstrate less aggressive features including reduced lymphatic, venous, and perineural invasion 4, 2
Type III tumors have significantly larger maximum tumor diameter, increased perineural invasion, greater vascular invasion, and worse overall survival (median 2.64 years vs. 4.96 years for Type I) 2
Median survival decreases progressively: Type I (4.96 years) > Type II (3.3 years) > Type III (2.64 years) 2
Lymphatic Drainage Patterns
The classification is critical because lymphatic spread differs markedly between types, necessitating different lymphadenectomy strategies 1, 3:
Type I: Lymphatic spread occurs cephalad to mediastinal nodes and caudally to the celiac axis 1
Type II and III: Metastases occur almost exclusively caudally to the celiac axis, splenic hilum, and para-aortic nodes 1, 3
Surgical Management Algorithm
For Type I tumors: Perform transthoracic esophagectomy with proximal gastric resection to address the cephalad lymphatic drainage pattern 1, 3
For Type III tumors: Perform transabdominal extended total gastrectomy, as esophagectomy offers no survival advantage and has higher postoperative mortality 3
For Type II tumors (the most controversial): Choose between transthoracic esophagogastrectomy or transabdominal extended total gastrectomy based on 5, 3:
- Preoperative T and N stage: Higher stages may favor more extensive resection
- Histological type: Diffuse-type tumors require longer margins that may not be achievable via gastrectomy alone 5
- Presence of Barrett's esophagus: If present, favor esophagectomy 5
- Tumor size and local invasion: Larger tumors may require esophagectomy for adequate margins
Evidence from a large single-center series of 1,002 patients demonstrates that extended gastrectomy achieves equivalent oncologic outcomes to esophagectomy for Type II tumors when R0 resection is achievable, with the advantage of lower perioperative complications and better postoperative quality of life 5, 3
Staging Implications
The AJCC staging system uses Siewert classification to determine which staging system applies 1:
Siewert Types I and II: Staged as esophageal adenocarcinoma 1
Siewert Type III: Staged using the gastric cancer staging system 1
This approach remains controversial and subject to debate, as the guidelines acknowledge that individualized therapeutic approaches may be preferred based on thorough preoperative staging, tumor location, nodal distribution, and specific requirements for local control 1
Critical Pitfalls to Avoid
Accurate anatomic localization is essential: The classification requires precise identification of the anatomic GEJ (approximately 40 cm from the incisors), which should be determined using endoscopy with both orthograde and retroflexed views, radiographic imaging, operative findings, and pathologic examination 6, 7
Complete tumor resection (R0) and lymph node status (pN0) are the dominant independent prognostic factors across all three types, regardless of surgical approach chosen 3
Do not assume histology or immunohistochemistry can reliably distinguish tumor types: Studies show that histologic patterns and immunophenotyping (CK7, CK20, CDX2, MUC markers) fail to reliably differentiate between proximal and distal tumors, making anatomic classification essential 4
The presence of Barrett's mucosa is more indicative of biological type than location alone and should influence surgical decision-making, particularly for Type II tumors 4