Esophageal Carcinoma Imaging

Updated: Jun 23, 2015
  • Author: from Memorial Sloan-Kettering – Raymond Thornton, MD; Chief Editor: Eugene C Lin, MD  more...
  • Print


Cancer of the esophagus remains a devastating disease because it is usually not detected until it has progressed to an advanced incurable stage. Modern imaging techniques, including barium esophagography, [1] contrast-enhanced computed tomography (CT), multidetector CT, [2] magnetic resonance imaging (MRI), endoscopic ultrasonography (EUS), [3] and positron-emission tomography (PET), are powerful tools in the detection, diagnosis, and staging of this malignancy. [4, 5, 6] See the following images:

Nonenhanced axial CT image at the level of the ori Nonenhanced axial CT image at the level of the origin of the great vessels demonstrates a nasogastric tube in place. The esophageal wall is thickened circumferentially and exerts mass effect on the posterolateral tracheal wall. Infiltration of the mediastinal fat adjacent to the great vessels is present. Findings are consistent with esophageal carcinoma with tracheal invasion.
Lateral barium esophagram demonstrates a large, in Lateral barium esophagram demonstrates a large, intraluminal, lobulated filling defect diagnosed as esophageal carcinoma during endoscopic biopsy.

The most common types of esophageal carcinoma are squamous cell carcinoma and adenocarcinoma. [7]

Early detection remains the elusive but essential goal of research. Only surgical resection at a very early stage has been shown to improve survival rates in patients with this disease. [8]

Preferred examination

Barium esophagography is a useful initial examination in the evaluation of a patient with esophageal complaints because it allows the assessment of esophageal morphology and motility. Esophageal endoscopy permits direct inspection and biopsy of the esophageal mucosa for histologic diagnosis. For the purpose of staging esophageal carcinoma, EUS, contrast-enhanced CT, and PET each offer unique information.

Limitations of techniques

Barium esophagography has optimal sensitivity for the detection of lesions when a double-contrast technique is used. Such technique requires that the patient be able to stand upright, which may not be possible with patients who are debilitated. For bulkier obstructive lesions, an air-contrast technique may not be possible, and a detailed mucosal examination may not be achieved distal to the obstruction.

A notable limitation of CT in diagnosis involves the characterization of lymph nodes. With CT scans, size criteria are used to determine possible metastatic involvement; however, lymph nodes may be enlarged because of infectious or inflammatory etiologies. Conversely, subcentimeter lymph nodes may harbor metastatic tumor.

Ultrasonographic examinations are highly operator dependent. Limitations of EUS with the standard Olympus diagnostic echo-endoscope (13 mm in diameter) include an inability to pass the malignant stricture with the transducer. This limitation results in an incomplete examination, which occurred in 40% of patients reported by Massari et al. [9] However, the use of a dedicated 8-mm-diameter esophagoprobe for EUS allows complete examination in most patients.

With PET, the resolution and cost remain the primary limitations. Subcentimeter foci of tumor metabolism may not be detected.



Barium esophagography is unique among esophageal studies for assessing both morphology and motility. Barium esophagography remains the study of choice for characterization of esophageal strictures. Esophageal carcinoma may demonstrate a variety of appearances on barium esophagrams. See the images below.

Anteroposterior barium esophagram demonstrates an Anteroposterior barium esophagram demonstrates an abrupt change in the caliber of the esophagus, with a long, irregular, annular stricture of the thoracic esophagus. Note the masslike shouldering at the proximal extent of the lesion at which filling defects are present within the dilated esophageal lumen. Findings are most consistent with esophageal carcinoma.
Lateral barium esophagram demonstrates an abrupt c Lateral barium esophagram demonstrates an abrupt change in the caliber and contour of the esophagus caused by an irregular circumferential stricture containing focal ulcerations. Findings are most consistent with esophageal carcinoma.
Lateral barium esophagram demonstrates marked irre Lateral barium esophagram demonstrates marked irregular narrowing of the esophageal lumen with focal ulcerations and almost complete obstruction; these findings are consistent with esophageal carcinoma.

Lesions may be annular and constricting; intraluminal, polypoid, or masslike; infiltrative; ulcerating; or varicoid. A mixed pattern is most common.

Early esophageal carcinoma may present as a small polypoid lesion or as coalescent plaques or nodules.

A double-contrast technique should be used for optimal sensitivity.

The length and location of the involved esophageal segment and the functional impairment resulting from the lesion should be reported.

Once a malignancy is detected on barium examination, the radiologist must be careful to evaluate the remainder of the esophagus and stomach for synchronous lesions. Endoscopy should follow.

Degree of confidence

Levine et al reviewed 2484 barium esophagrams and found that endoscopy had been recommended in 26 patients (1%) because of findings suggestive of malignancy. Cancer was ultimately diagnosed in 11 patients, yielding a positive predictive value of 42%. The same investigators also retrospectively reviewed 50 cases of endoscopically proven esophageal carcinoma and found that lesions were present on barium esophagrams in 49 patients (98%). In addition, carcinoma was diagnosed or suggested in 48 patients (96%) on the basis of the barium study results. [10, 11]

False positives/negatives

The morphologic pattern of the lesion on barium esophagrams establishes the need to consider the differential diagnoses.


Computed Tomography

Contrast-enhanced CT plays an important role in the staging of esophageal carcinoma. Attention is directed to determining the extent of the local tumor; invasion of mediastinal structures; involvement of supraclavicular, mediastinal, or upper abdominal lymph nodes; and distant metastases. These observations are useful in distinguishing between T3 and T4 lesions and in determining the N and M status. See the images below.

Nonenhanced axial CT image at the level of the ori Nonenhanced axial CT image at the level of the origin of the great vessels demonstrates a nasogastric tube in place. The esophageal wall is thickened circumferentially and exerts mass effect on the posterolateral tracheal wall. Infiltration of the mediastinal fat adjacent to the great vessels is present. Findings are consistent with esophageal carcinoma with tracheal invasion.
Sagittal reformatted chest CT scan demonstrates ma Sagittal reformatted chest CT scan demonstrates marked, long-length, esophageal wall thickening abutting the posterior tracheal wall. Tracheal invasion by esophageal carcinoma was found at bronchoscopy.
Enhanced axial CT image demonstrates irregular wal Enhanced axial CT image demonstrates irregular wall thickening of the esophagus. A heterogeneously enhancing mass to the right of the esophagus represents a markedly enlarged metastatic lymph node. No significant loss of the fat plane is noted between the esophageal mass and the descending thoracic aorta, indicating the absence of aortic invasion. Small bilateral pleural effusions are present.
Nonenhanced axial CT image demonstrates a large pr Nonenhanced axial CT image demonstrates a large proximal esophageal mass with asymmetric impression on the posterior tracheal wall, suggestive of invasion. Recall that the posterior tracheal wall may bow inwardly (usually symmetrically) during expiration because of incomplete tracheal cartilage rings. Infiltration of the peri-esophageal fat is demonstrated, although no evidence of aortic invasion is present.
Enhanced axial CT image demonstrates a large esoph Enhanced axial CT image demonstrates a large esophagogastric junction carcinoma and several large liver lesions, with heterogeneous peripheral contrast enhancement and irregular margins consistent with metastases.

CT examination should extend from the thoracic inlet through the liver. Routine oral contrast material should be administered. This may be positive contrast agent, such as dilute barium, or a negative intraluminal contrast medium, such as water. A low-density 3% weight-for-weight esophageal barium paste may be administered immediately prior to scanning. Techniques for virtual esophageal endoscopy have also been described using effervescent granules and glucagon.

A sample helical CT (single detector) protocol includes the intravenous administration of 150 mL of 60% contrast material injected at a rate of 2-3 mL/s with a scanning delay of 50 seconds after the start of the injection; a pitch of 1.5 and a 5-mm section thickness may be used. For multidetector-row CT, 100 mL of 60% intravenous contrast medium may be injected at a rate of 3 mL/s with a scanning delay of 40 seconds after the start of the injection; a pitch of 6 and a 2.5-mm section thickness may be used. [12]

CT scans should be obtained during full inspiration when patients are evaluated for esophageal carcinoma. Inward bowing or the trachea and mainstem bronchi during expiration may produce the false impression of tracheobronchial invasion. [13, 14]

Key findings include the following:

  • Eccentric or circumferential wall thickening is greater than 5 mm.
  • Peri-esophageal soft tissue and fat stranding may be demonstrated.
  • A dilated fluid- and debris-filled esophageal lumen is proximal to an obstructing lesion.
  • Tracheobronchial invasion appears as displacement of the airway (usually the trachea or left mainstem bronchus) as a result of mass effect by the esophageal tumor. Absence of a fat plane between the airway and the esophageal mass cannot be used as an indication of invasion. Even in patients without esophageal carcinoma, a fat plane is usually not evident between the esophagus and left mainstem bronchus.
  • Note that the tracheal cartilage rings are incomplete posteriorly, which allows the posterior wall of the trachea to bow anteriorly during expiration; therefore, it is important that CT images be acquired during full inspiration.
  • Aortic invasion may be assessed in 2 ways.
    • The Picus method considers the arc of contact between the tumor and aorta. [15] Loss of the periaortic fat plane over less than 45° suggests no aortic invasion, whereas contact over 90° or more is predictive of invasion of the aortic wall. Contact between 45° and 90° is indeterminate. Accuracy with this method is 80%.
    • Obliteration of the triangular fat space between the aorta, esophagus, and spine is another predictor of aortic invasion.
  • A careful search for lymph nodes is an essential component of the interpretation. A short-axis diameter exceeding 1 cm is considered abnormal for lymph nodes in all mediastinal locations except those in the subcarinal region, in which 1.4 cm is the upper limit of normal. Because lymph nodes may harbor metastases without being enlarged, noting the location of any identified lymph nodes is important. In addition, remembering that lymph nodes may be enlarged because of inflammatory or infectious etiologies is important.
  • Esophageal carcinoma is often metastatic at presentation.
  • In a review of 838 patients with M1 disease, Quint et al found that metastases were diagnosed most commonly in the abdominal lymph nodes (45%); liver (35%); lung (20%); cervical and/or supraclavicular lymph nodes (18%); bone (9%); adrenal glands (5%); peritoneum (2%); brain (2%); or stomach, pancreas, pleura, skin or body wall, pericardium, or spleen (1% each). [16] The same group found that none of the cases with stage M0, as determined at chest and abdominal CT, were upstaged to M1 during either bone scanning or head CT. [16]
  • Reviewing a cohort of 116 cases of predominantly esophageal squamous cell carcinoma, Margolis et al determined that a solitary lung nodule found at the time of diagnosis is more likely to be a benign abnormality or a primary lung cancer than a metastatic lesion. [17] However, during terminal phases of the disease, lung metastases are increasingly common.

Degree of confidence

Multiple groups have reported on the sensitivity and specificity of CT in determining lymph node involvement. Its sensitivity for the detection of lymph node metastases is in the range of 60-80%. Its specificity is higher, approximately 90%.

False positives/negatives

An impression of tracheal invasion may be caused by expiratory bowing of the posterior trachea. Lymph node metastases may be present without enlarged lymph nodes, and lymph node enlargement may not be the result of malignancy. Lesions in solid organs identified at CT may represent primary benign or malignant processes or metastatic disease, and biopsy is frequently necessary to obtain histologic proof.


Magnetic Resonance Imaging

MRI presents the advantage of direct multiplanar imaging capabilities, which may be of particular use in assessing tracheobronchial, aortic, and pericardial invasion. Currently, MRI has not yielded other significant advantages compared with CT in the staging of esophageal carcinoma.

Recent research studies suggest that T2-weighted MRIs obtained with an endoluminal coil can reveal 7 layers of the esophageal wall: [18]

  • Epithelial layer (intermediate intensity)
  • Lamina propria and muscularis mucosa (low intensity)
  • Submucosa (high intensity)
  • Inner circular muscularis propria (low intensity)
  • Intermuscular structure (high intensity)
  • Outer longitudinal muscularis propria (low intensity)
  • Subadventitial structures (high intensity)

Degree of confidence

Preliminary studies have shown that the sensitivity and specificity of MRI for the determination of tumor invasion are equivalent to those of CT.

False positives/negatives

False-positive and false-negative findings are similar to those seen on CT scans.



Unlike CT, EUS allows visualization of the distinct layers within the esophageal wall. Alternating circumferential layers define the mucosal interface (hyperechoic), the mucosa (hypoechoic), the submucosa (hyperechoic), the muscularis propria (hypoechoic), and the adventitial interface (hyperechoic). Such resolution permits the distinction of T1, T2, T3, and T4 tumors. Esophageal carcinoma appears as a hypoechoic lesion disrupting the normal circumferential layers. However, with standard EUS equipment, often only 3 layers are seen in the esophageal wall: the mucosa and submucosa (hyperechoic), the muscularis propria (hypoechoic), and the adventitial interface (hyperechoic).

Local lymph nodes are also demonstrated by using EUS. Nodes are considered malignant if they are round, if they are hypoechoic, and if they have well-defined borders. Usually, benign nodes are hyperechoic and less well defined.

Degree of confidence

T-stage accuracy with EUS is in the range of 79-94%. Accuracy is better for T3 or T4 lesions than for T1 or T2 lesions. Massari et al showed that a 12-MHz transducer outperforms a 7.5-MHz transducer, with accuracy for T staging of 94% and 82%, respectively. N-stage accuracy is in the range of 69-90%. Distant metastases cannot be reliably identified with EUS because of the limited field of view. [9] N-stage accuracy is in the range of 69-90%; according to Rasanen et al, the technique probably outperforms CT and PET in the detection of locoregional lymph node metastasis. [19]

At this time, only EUS is useful in distinguishing T1 and T2 lesions.


Nuclear Imaging

PET is quickly becoming a standard oncologic imaging modality. The technique is useful not only for the primary detection of tumor and metastases but also for the further characterization of abnormalities discovered by using other imaging modalities. [20, 21, 22]

2-[Fluorine 18]-fluoro-2-deoxy-D-glucose (FDG) is the most commonly used radiopharmaceutical. FDG is taken up by cells and trapped within them by means of phosphorylation. FDG decays inside the cell, with a half-life of approximately 110 minutes, while emitting positively charged electrons called positrons. A positron travels a short distance in tissue before colliding with an electron. Collision results in annihilation of the 2 particles and the emission of photons. The detection of the photons by using ring detectors is the basis for image formation. [23]

In a systematic review and meta-analysis, 18F-FDG and PET/CT were found to be a reliable imaging modality for detecting recurrent esophageal cancer after treatment. Pooled estimates of sensitivity and specificity for 18F-FDG PET or PET/CT in diagnosing recurrent esophageal cancer were 96% (95% confidence interval [CI]: 93%-97%) and 78% (95% CI: 66%-86%), respectively. According to the study, however, histopathologic confirmation of 18F-FDG or PET/CT suspected lesions was required because of a considerable false-positive rate. [24]

Radiopharmaceuticals other than FDG can be used in PET imaging. Carbon-11 choline has received particular attention. Choline, a component of the cell membrane, is taken up by actively dividing cells.

In theory, this agent provides some advantages over FDG. Uptake of FDG requires that the tumor derive a portion its energy supply from glycolysis, an anaerobic condition that exists when the oxygen supply to some part of the tumor is low. For example, this circumstance may occur when a tumor outgrows its arterial supply. On the other hand, the uptake of11 C-choline requires only the presence of cell division. Cardiac uptake of FDG may be significant, even in the fasting state, and this characteristic complicates the interpretation of findings in the mediastinum. Conversely, cardiac uptake is not significant with11 C-choline.

11 C-choline PET scanning has been shown to outperform FDG PET scanning in the detection of malignant mediastinal lymph nodes. With this agent, tumor-containing mediastinal lymph nodes as small as 4 mm have been identified. The short half-life of11 C-choline (approximately 20 min) will likely limit its use to major academic centers. The agent is not useful in the abdomen because of intense liver uptake. FDG PET scanning remains the technique of choice for the detection of metastatic abdominal lymph nodes.

Degree of confidence

Flamen et al performed a prospective study to compare the staging of esophageal carcinoma with FDG PET versus the combination of CT and EUS in 74 patients. [25] For T stages, FDG PET scanning demonstrated increased activity in the primary tumor in 70 patients (95% sensitivity). Four proven T1 lesions had negative findings on FDG PET scans.

The sensitivity of FDG PET in assessing nodal metastasis is reportedly 33-83%, but studies have shown the superiority of FDG PET to CT and EUS for determining the N status. FDG PET is more sensitive than CT for the detection of distant metastases. [26]

In a study by Flamen et al, FDG PET had higher accuracy in diagnosing stage IV disease compared with the accuracy of CT and EUS combined; the results led to upstaging of the disease in 11 of 74 patients and downstaging of the disease in 5 of the patients. [25] Kobori et al have shown that the combination of FDG and11 C-choline PET permitted the identification of 85% of the cases with metastatic lymph nodes. [27]

False positives/negatives

FDG PET scans may fail to demonstrate T1 primary lesions. Inflammatory processes cause increased uptake of FDG. In addition, patients with no esophageal pathology may have minimal esophageal FDG uptake, particularly in distal areas.