Gastric carcinoma is the most common cancer in the world after lung cancer and is a major cause of mortality and morbidity. Though a marked reduction has been observed in the incidence of gastric carcinoma in North America and Western Europe in the last 50 years, 5-year survival rates are less than 20%, as most patients present late and are unsuitable for curative, radical surgery. Gastric tumors are seen in the images below.
Adenocarcinomas account for approximately 95% of all malignant gastric neoplasms. The remaining 5% of tumors are lymphomas, leiomyosarcomas, carcinoids, or sarcomas.
Gastric adenocarcinomas are divided into 2 types. Type 1 adenocarcinomas are intestinal tumors and have well-formed glandular structures. This form of gastric carcinoma is more likely to involve the distal stomach and occur in patients with atrophic gastritis (seen in the image below). It has a strong environmental association.
A type 2 adenocarcinoma is a diffuse type with poorly cohesive cells, which tend to infiltrate the gastric wall. Tumors of this type may involve any part of the stomach, especially the cardia; they have a worse prognosis than type 1 tumors.
Prognosis and staging
The prognosis of gastric carcinoma is related to the stage of the disease at the time of diagnosis and to the histologic grade of the carcinoma. 
Pathologic staging of these neoplasms is based on tumor stage, nodal stage, and metastasis stage (TNM). The stages are as follows:
The T stage, representing the extent of penetration through the gastric wall, is categorized as follows:
Tis - Carcinoma in situ, intraepithelial tumor
T1 - Tumor extension to submucosa
T2 - Tumor extension to the muscularis propria or subserosa
T3 - Tumor penetration of the serosa
T4 - Tumor invasion of the adjacent organs
The N stage, representing the number and site of draining lymph nodes involved, is categorized as follows:
N0 - No lymph nodes involved
N1 - Metastases in 1-6 regional lymph nodes
N2 - Metastases in 7-15 regional lymph nodes
N3 - Metastases in >15 regional lymph nodes
The M stage, representing the presence of metastases, is categorized as follows:
M0 - No distant metastases
M1 - Distant metastases
Table: Staging and 5-Year survival rates (Open Table in a new window)
|Stage||TNM Stage||5-Year Survival|
|1||T1N0M0, T1N1M0, or T2N0M0||88%|
|2||T1N2M0, T2N1M0, or T3N0M0||65%|
|3a||T2N2M0, T3N1M0, or T4N0M0||35%|
|4||T4N1-3M0, TxN3M0, or TxNxM1*||5%|
|*Tx indicates any T stage; Nx, any N stage.|
Single-contrast barium studies have an overall sensitivity of 75% in the detection of gastric cancer, but double-contrast barium examinations have a sensitivity of 90-95% in gastric cancer detection, comparable to that of endoscopy. Computed tomography (CT) scanning, magnetic resonance imaging (MRI), and endoscopic ultrasonography (EUS) are used in the staging, but not usually in the primary detection, of gastric cancers. [2, 3, 4, 5, 6, 7, 15, 16, 17, 18]
This section examines radiologic findings in early and advanced gastric carcinoma.
Early gastric cancer
Double-contrast barium upper GI examination is widely recognized as the radiologic technique of choice for diagnosing early gastric cancers. These lesions are confined to the mucosa or submucosa and are classified into 3 types:
Type I - Elevated lesions that protrude more than 5 mm into the lumen
Type II - Superficial lesions that are elevated (IIa), flat (IIb), or depressed (IIc)
Type III - Shallow, irregular ulcers surrounded by nodular, clubbed mucosal folds
Gastric carcinomas are occasionally seen on plain abdominal radiographs as abnormalities in the gastric contour or as soft-tissue masses indenting the gastric contour. Rarely, mucin-producing carcinomas may show areas of punctate calcification.
On barium studies, gastric carcinomas may be polypoidal, ulcerative, or infiltrating lesions.
Polypoid carcinomas (an example of which appears below) are lobulated masses that protrude into the lumen. They may contain 1 or more areas of ulceration.
With ulcerated carcinomas, an irregular crater is located in a rind of malignant tissue. Seen en face, tumor nodules may be in the adjacent mucosal folds; the mucosal folds that converge to the edge of the ulcer may be blunted, nodular, or clubbed from tumor infiltration. Seen in profile, these lesions are intraluminal, whereas benign ulcers project beyond the contour of the stomach. The radiating folds associated with a benign ulcer are regular and extend close to the ulcer margins (see the image below). Endoscopy and biopsy are generally required to confirm or rule out malignancy in most cases of gastric ulcers.
Infiltrating carcinomas result in irregular narrowing of the stomach, with nodularity or spiculation of the mucosa. (See the image below.)
Scirrhous carcinomas typically cause irregular narrowing and rigidity of the stomach, giving rise to the typical linitis plastica, or "leather bottle," appearance (as seen in the image below). Although some are lobulated lesions in the fundus or body, others consist of thickened, irregular mucosal folds and nodularity without significant narrowing.
Carcinomas of the cardia are often missed during single-contrast examinations. In double-contrast studies, normal anatomic landmarks are obliterated and replaced by a plaquelike lesion with nodularity or ulceration. The distal esophagus is often involved. (See the images below).
Submucosal spread of tumors may result in pseudoachalasia or secondary achalasia with tapered, beaklike narrowing of the distal esophagus and infiltration of the gastric cardia (demonstrated in the images below).
Degree of confidence
As previously stated, single-contrast barium studies have an overall 75% sensitivity in the detection of gastric cancer, while double-contrast barium examinations have a 90-95% sensitivity in gastric cancer detection. Thus, the double-contrast studies have a sensitivity comparable to that of endoscopy.
The appearance of gastric carcinomas on barium studies must be distinguished from the appearances of benign gastric ulcers and polyps, gastric lymphomas, and focal gastritis. Malignant stromal tumors may also cause confusion. Although linitis plastica is usually caused by gastric carcinoma, it can also be caused by metastatic breast cancer.
In rare cases, radiation therapy, Crohn disease, tuberculosis, sarcoidosis, and syphilis may simulate gastric carcinoma, and primary esophageal adenocarcinoma may invade the stomach. Gastric varices and inadequate distension may mimic tumors of the gastric fundus.
CT is used preoperatively primarily to determine the stage and extragastric spread of a gastric carcinoma. This information is vital in deciding between palliative surgery and curative radical surgery (ie, identifying patients who would not benefit from radical surgery). Additionally, CT is used to monitor a patient's response to treatment. [8, 9, 10, 17, 18]
Detection of gastric carcinoma is improved by using thin-section sequences and helical or multidetector-row CT.  When thin collimation is used, near-isotropic imaging of the stomach is possible, allowing high-quality multiplanar reformation and 3-dimensional reconstruction of gastric images. An intravenous contrast medium is used, along with water or gas as a negative intraluminal agent (the results of which are seen in the image below). Prone views improve visualization of tumors of the cardia and distal stomach.
Helical scanning allows for a biphasic technique.  The early arterial phase is used to assess enhancement of the gastric wall; the later portal venous phase is used to assess the liver parenchyma for metastases.
CT scans may show the following:
Polypoidal mass with or without ulceration
Focal wall thickening with mucosal irregularity or ulceration
Wall thickening with the absence of normal mucosal folds (infiltrative lesions)
Focal infiltration of the gastric wall (see the image below)Carcinoma of the lesser curve. Note the focal mural thickening due to a tumor plaque.
Variable thickening of the wall and marked contrast enhancement (typical of scirrhous lesions)
Mucinous carcinomas, which have low attenuation due to their high mucin content and which may contain calcification
The depth of tumor invasion is not accurately assessed with CT.
Tumor invasion of the perigastric fat is seen as soft-tissue stranding. Tiny 4- to 8-mm nodules may be observed. These may coalesce into sheets of tumor in advanced cases.
Direct extension of the tumor is relatively common. The pancreas is invaded via the lesser sac; the transverse colon, via the gastrocolic ligament; and the liver, via the gastrohepatic ligament.
Longitudinal spread to the distal esophagus occurs in as many as 60% of patients with carcinoma of the cardia (see the images below). However, the duodenum is involved in only 5-20% of antral carcinomas.
Overall, the accuracy of determining the T-stage with CT is approximately 66%.
Lymph node metastases occur in approximately 80% of patients with gastric cancer (see the images below). The frequency is related to the size and depth of the tumor; local perigastric nodes are involved first, followed by the regional (celiac, hepatic, left gastric, splenic) and distant (left supraclavicular and axillary) nodes.
CT depicts 75% of nodes larger than 5 mm in diameter, but it does not depict tumor in normal-sized nodes. CT is not useful in distinguishing between enlarged nodes due to reactive changes and those due to tumor.
The local nodes (N1) are located in the prepyloric region and in the gastrocolic and gastrohepatic ligaments. These nodes are removed by performing the standard gastrectomy procedure. The regional nodes (N2), located in the porta hepatis, hepatoduodenal ligament, and peripancreatic region, are not removed by the gastrectomy procedure; thus, their detection is more important.
Because the portal vein drains the stomach, the liver is the most common site for hematogenous metastases (see the first image below). Less common sites are the lungs (see the second image below), adrenal glands, and kidneys. Bony and cerebral metastases are uncommon.
Intraperitoneal and omental metastases are common in advanced gastric cancer. They consist of nodules, localized fluid collections, and irregular thickening and stranding of the mesentery and omentum. Ascites and small-bowel obstruction may occur.
Gastric carcinoma is the most common primary tumor to metastasize to the ovaries. These ovarian metastases are usually bilateral and are known as Krukenberg tumors.
Degree of confidence
For accuracy rates in CT scanning (vs those in EUS), see Ultrasound. 
CT scanning has several pitfalls. A pseudomass as a result of a normal gastroesophageal junction may be seen, underdistension of the stomach may simulate wall thickening, and T2 and T3 lesions may be difficult to distinguish.
Moreover, loss of a fat plane between the gastric wall and the left lobe of the liver may be seen. Loss of a fat plane between the tumor and pancreas may be due to an inflammatory reaction. In cachectic patients, a loss of fat planes may simulate direct organ invasion.
Small nodes may contain tumor, and large nodes may result from inflammatory causes. Perigastric nodes may not be observed if the stomach is not well distended.
CT may fail to depict tiny omental and peritoneal deposits; small pelvic deposits may be overlooked as well.
Magnetic Resonance Imaging
MRI studies in which a breath-hold, fast-imaging technique and water were used showed accuracy rates comparable to those of helical, biphasic CT scanning. The fast-imaging technique was superior to CT in detecting serosal invasion. 
Degree of confidence
In T staging, the accuracy of MRI is 73%, compared with 67% for CT. In N staging, the accuracy of MRI is 55%, compared with 59% for CT.
MRI is limited by the presence of respiratory and peristaltic artifacts, the lack of suitable oral contrast media, and a higher cost than that of CT scanning.
The primary role of transabdominal ultrasonography (US) is to detect liver metastases. These metastases are usually hyperechoic, but they may be hypoechoic. CT scanning and endoscopic ultrasonography (EUS) are complementary. CT scanning is used first to stage the gastric carcinoma; if no metastases and no invasion of local organs are found, EUS is used to refine the local stage. The depth of tumor invasion is not accurately assessed with CT, and the investigation of choice for this indication is EUS. 
Gastric carcinomas are occasionally identified during US of the upper abdomen.
EUS has improved the accuracy of local staging of gastric carcinomas. Its role is to assess the depth of local invasion and the presence or absence of perigastric nodes. Unlike CT and MRI, EUS can depict individual layers of the gastric wall, with a rotating high-frequency probe inserted via an endoscope. EUS is limited to an area 5 cm from the probe. It cannot be used to assess distant metastases or nodes more than 5 cm away from the probe.
The gastric wall is visualized as 5 concentric bands:
Mucosa - Echogenic
Muscularis mucosa - Hypoechoic
Submucosa - Echogenic
Muscularis propria - Hypoechoic
Serosa - Echogenic
A gastric tumor is demonstrated as a hypoechoic mass with varying mural invasion. Its depth may be overestimated because of the inflammatory response around the tumor, or it may be underestimated because of microscopic spread.
With T1 tumors, wall thickening is limited to the mucosa and submucosa. Regarding N staging, involved nodes are rounder and more hypoechoic than normal nodes.
Degree of confidence
In the detection of liver metastases, sensitivities as high as 85% have been reported.
Overstaging is due to the peritumoral inflammatory response.
In T staging, EUS is 89-92% accurate, and CT is 43-65% accurate; however, the accuracy of CT increases with the use of the helical biphasic technique. In N staging, EUS is 60-85% accurate, and CT is 48-70% accurate. Inflammation may cause enlarged nodes. EUS has a high specificity (90%) but low sensitivity (53-80%) because it has a range of 5 cm from the gastric wall for nodes of normal size; thus, it does not permit assessment of the full extent of lymphadenopathy. Involved small nodes are not detected.
Intraoperative US and laparoscopy have an accuracy of 81% in T staging and an accuracy of 93% in N staging; however, the necessary equipment and expertise are not widely available.
Fluorodeoxyglucose (FDG) positron-emission tomography (PET) scanning may be useful in the staging and postoperative assessment of gastric carcinomas. FDG-PET depicts the primary tumor, but involved perigastric lymph nodes are not identified separately from the primary tumor. Thus, the role of PET scanning is limited in staging. The use of combined PET-CT scanning may improve diagnostic accuracy. [14, 16]
Indium-111 (111 In) – labeled monoclonal antibody has been used for intraoperative imaging to detect nodes, with an accuracy of 72%.
FDG-PET may be useful in evaluating patients with recurrent gastric cancer; findings can help to localize the disease when CT findings are not diagnostic. Imaging evaluation with PET may also impact the clinical management of patients with recurrent gastric cancer.