The AJCC approach to staging of gastric adenocarcinoma is based on the TNM classification. Using the TNM classification, cancers are grouped from a stage of best prognosis (stage 0: Tis N0 M0) to poorest prognosis (stage IV: T4 N1 M1) (Table 22.2). The disease stage is stated according to actual information for each TNM factor. Both clinical TNM (cTNM) and pathological TNM (pTNM) staging schemes exist for tumors because histopathological confirmation of stage may not always be available when treatment decisions are being made, and the two types of staging may each predict outcome differently.

In the fifth revision of the AJCC classification of gastric adenocarcinoma, the classification of nodal metastases (N staging) was restructured to reflect the number of involved lymph nodes as opposed to their anatomical stations. The sixth revision did not bring about any major changes to the staging of gastric cancer. The current classification is reflected in Table 22.1, and the stage groupings are shown in Table 22.2.

Several groups, including the Japanese, have retrospectively evaluated their respective cohorts of gastric adenocarcinoma patients in the context of the nodal staging classifications brought about in the fifth edition (14,15). A multivariate analysis of various factors in gastric adenocarcinoma demonstrated that the depth of tumor invasion and the new nodal stage classification were the most significant prognostic factors (16). More recently, a study compared survival rates for 1,244 gastric cancer patients using the Japanese Classification system versus the AJCC/UICC system, and found that the AJCC/UICC system was more rational and homogenous than the Japanese system (17).

For many years, the staging of gastric carcinoma in Japan has followed the TNM model, similar to the AJCC system. The Japanese system, though, has several differences. In addition to the level of invasion of the primary tumor, it looks at the gross morphologic characteristics of the tumor. The Japanese system also puts greater emphasis on the location and groups of lymph nodes involved and separates the location of distant metastases into peritoneal, hepatic, and distant sites. These classifications were published for the first time in English in 1995 (18). The Japanese classification criteria are presented in Table 22.3, and the stage groupings are shown in Table 22.4.

Barium radiography and upper gastrointestinal endoscopy are both useful modalities in the diagnosis and assessment of gastric adenocarcinoma, but neither modality has been shown to accurately predict the stage of cancer so they both remain strictly diagnostic tools. Although barium radiography has an excellent record as an effective screening tool for high-risk mass population screening in Japan, endoscopy has become the procedure of choice in the United States and Japan for the
diagnostic evaluation of symptomatic patients. Endoscopy offers the advantage of biopsy at the time of the procedure, thus providing a definitive diagnosis. The location, extent, and gross morphologic characteristics of a tumor are readily identified by either method, but the depth of tumor invasion and the assessment of lymph nodes cannot be achieved by either technique. Therefore, neither procedure alone provides enough information to allow for complete staging of gastric carcinoma.

Traditionally, the major usefulness of CT scan imaging in staging gastric cancer is related to its ability to identify metastatic disease to the liver, peritoneum, intraabdominal or retroperitoneal lymph nodes, or other more distant locations. However, in prospective studies comparing the results of CT with intraoperative findings, the sensitivity of CT for staging gastric cancer has been called into question (19). In the staging of hepatic and peritoneal metastases, preoperative staging by CT has been suggested in one study to have an accuracy of 79% for hepatic disease and 81% for peritoneal disease; however, the sensitivity for each in the same study was quite low, 52% and 8%, respectively (20). In another prospective study of a series of 111 patients undergoing laparoscopic staging for gastric cancer, preoperative CT was shown to have missed as many as 37% of small metastatic lesions that were detected by laparoscopy or laparotomy (21). The majority of these lesions were small peritoneal implants, with only 3 of 38 representing liver metastases. This suggests that although CT provides important information regarding liver metastases in gastric carcinoma, it misses a substantial proportion of peritoneal disease. In the staging of the primary tumor and locoregional lymph node metastases, CT has performed even more poorly, with accuracy rates ranging from 25% to 43% for T staging and 33% to 51% for N staging (22,23).

Newer studies have focused on the role of spiral or thin-section CT in gastric cancer staging. Early investigations using this technology were disappointing, demonstrating results similar to traditional CT, with a sensitivity of 71% for peritoneal metastases and 57% for hepatic metastases in one study (24). The sensitivity for lymph node staging in that same study was 24% for N1 disease and 43% for N2 disease. Subsequent studies have shown more promise, with locoregional staging capabilities comparable to EUS (25,26). In one study of 51 patients, helical CT achieved correct T staging in 76% and correct
N staging in 70% of cases (26). As with EUS, the greatest difficulty in T staging with helical CT relates to differentiating between T2 and T3 lesions. Another study demonstrated that the ability to accurately predict T and N staging has improved with progressive CT technology and with physician experience in using that technology (27).