Nearly all localized primary esophageal carcinomas are detected first by clinical symptoms of dysphagia or in the context of endoscopy performed as part of an evaluation of the upper gastrointestinal (GI) tract for GE reflux—or in sequential management of patients with Barrett’s esophagus. PET has not been used as a primary screening method for esophageal carcinoma. The precursor lesion to many cases of adenocarcinoma of the esophagus is Barrett’s esophagus. It usually has a low total volume that would not be expected to be detected using whole-body PET imaging devices, although on rare occasions it can be (11). Similarly, some inflammations of the distal esophagus can have mildly increased FDG uptake (11). Several studies have shown that the vast majority of primary esophageal cancers that are first diagnosed by other methods are detectable by [¹⁸F]-FDG PET, with sensitivities in the 90% to 100% range for T2 to T4 tumors (12,13,14,15,16). When PET results have been reported to be falsely negative in esophageal cancers, this was usually due to small tumor volume, such as in stage T1 primary lesions. Furthermore, some adenocarcinomas of the gastric cardia demonstrate only low FDG uptake and can be falsely negative on FDG PET even at advanced tumor stages. This is likely related to their growth pattern and mucin production (17).

Diagnostic challenges in esophageal cancer include determining whether there is abnormal or physiologic uptake at the GE junction. There may be some uptake in this location normally, so detecting small esophageal cancers can be problematic as they can be lost in the normal spectrum of mild FDG uptake in the distal esophagus. For these reasons, it is probable that early low-volume esophageal cancer can be much more easily detected by direct visualization using an endoscope or by careful barium studies than by PET.

The detection of nodal metastases with PET is dependent on the volume of tumor in the metastasis, the intensity of tracer uptake in the lesion, the background tracer activity (as well as the injected dose of radiotracer and the performance and resolution of the scanner), and the interpretation criteria. CT detects nodal metastases based on their size, and what constitutes a “positive or negative” node by CT differs with different observers. The larger the node, the more likely it is to represent a node involved by cancer, although the optimal cutoff size for “positive or negative” nodes is not clear in esophageal cancer.

Both PET and CT can fail to detect small metastases of esophageal cancer to locoregional lymph nodes. Lesions smaller than 5 mm are not usually detected on PET with FDG, which is consistent with other detection challenges encountered with current FDG PET technology. To achieve high sensitivity with CT, small lymph nodes must be called positive. For example, 5-mm and larger nodes may be called abnormal in some CT studies, whereas others may choose a 10-mm cutoff. The smaller the cutoff for node size, the more likely cancer will be detected, but at the price of a lower specificity. This results in a considerable range in the sensitivity and specificity of PET for assessing tumor involvement in regional lymph nodes and an even greater range in accuracy of CT (Table 8.13.3).

In general PET is not exceptionally sensitive, but it has a high specificity for the detection of locoregional lymph node metastases (typical sensitivities of 30% to 80%, but with specificities of 80% to 90%). A meta-analysis published in 2004 evaluating data from 421 patients reported a pooled sensitivity and specificity of FDG PET of 51% and 84%, respectively. Table 8.13.3 provides an overview of the largest studies used in the meta-analysis as well as of more recently published studies (18). By contrast, reported CT sensitivity/specificity pairs range from (31% to 86% to 87% to 14%), meaning CT can be either very insensitive or reasonably specific (although not as specific as PET) or very sensitive and extremely nonspecific. When directly compared, in several studies PET was an equivalent or more accurate method for nodal staging than CT (Table 8.13.3).

EUS has been reported to be more sensitive (70% to 80%) than either PET or CT for staging regional nodes and it is reasonably specific (70% to 80%) (15,19,20). EUS is also valuable for assessing tumor size and depth of invasion. According to a systematic review including a total of 2,508 patients, the diagnostic accuracy of EUS for differentiation of stages T1 and T2 from stages T3 and T4 was 91% (21). There are no data showing that PET or CT is accurate in evaluating the primary tumor stage of esophageal cancers. Neither can resolve the individual layers of the esophageal wall that form the basis of the T-staging system. A disadvantage of EUS is its operator dependency. Furthermore, EUS may be technically impossible if the tumor causes a stenosis that cannot be passed by the endoscope (21). There is some interest in using minimally invasive surgical techniques for staging esophageal cancer. Specifically, sentinel node dissection is currently being evaluated as a minimally invasive technique to improve lymph node staging (22).