Файл: Advanced Imaging of the Abdomen - Jovitas Skucas.pdf

An esophageal carcinoma is generally first detected either with a barium study or endoscopy, while tumor invasion is evaluated with CT or endoscopic US. Because many of these tumors are hypervascular compared to normal esophageal tissues, postcontrast multidetector CT imaging during the arterial phase also appears useful for initial tumor detection, although currently little data are available in the literature to provide guidance.

Esophagography remains an excellent tool both for diagnosis and follow-up of an esophageal carcinoma. A double-contrast technique and meticulous attention to detail are necessary.

At times an early superficial carcinoma presents as an isolated focus of nodularity identified with a double-contrast esophagram. The differential diagnosis of such a finding includes Barrett’s esophagus, focal esophagitis, or even focal glycogenic acanthosis. A more extensive carcinoma is readily detected by a double-contrast barium study, but a biopsy is generally necessary to establish the diagnosis (Fig. 1.17).

Adenocarcinomas tend to have a nodular outline, and a technically high-quality study is necessary to detect smaller tumors (Fig. 1.18). Even when advanced, they tend to infiltrate intramurally, and a bulky intraluminal mass is found only in a minority (Fig. 1.19). Varicoid adenocarcinomas in the distal esophagus mimic varices, but unlike varices, these serpiginous cancers do not change shape and size with degree of esophageal lumen distention, and

B

Figure 1.17. Squamous cell carcinoma in mid-esophagus. A: CT identifies a poorly enhancing tumor infiltrating the esophagus (arrow). (Courtesy of Arunas Gasparaitis, M.D., University of Chicago.) B: Ulcerated carcinoma in mid-esophagus (arrow) in another patient.

most experienced radiologists readily differentiate these two conditions (Fig. 1.20). Adenocarcinomas developing in a setting of Barrett’s esophagus typically have an initial flat or plaque-like, irregular appearance and tend to be difficult to detect.

Staging

Squamous Cell Carcinoma: The tumor, node, metastasis (TNM) system is used for staging

Figure 1.18. Polypoid adenocarcinoma in mid-esophagus (arrows).

esophageal carcinomas (Table 1.4). Information obtained from imaging is combined with other data to achieve a preoperative stage useful for subsequent prognosis and therapy. From an immediate surgical viewpoint, however, a primary

Figure 1.19. Distal esophageal adenocarcinoma. CT reveals diffuse infiltration of surrounding tissues (arrow).

concern is not staging but whether a tumor is resectable or not.

Preoperative imaging is used to detect local and metastatic spread. Computed tomography complements rather than competes with endoscopy and barium studies, which are primarily diagnostic rather than staging modalities. Arterial phase multislice CT shows most tumors to be hypervascular to surrounding normal soft tissues; CT detects tumor intralu-

Table 1.4. Tumor, node, metastasis (TNM) staging of esophageal tumors

Primary tumor:

Lymph nodes:

Nx Regional lymph nodes cannot be assessed N0 No regional lymph node metastasis

N1 Regional lymph node metastasis

Distant metastasis:

Mx Distant metastasis cannot be assessed M0 No distant metastasis

M1 Distant metastasis

Source: From the AJCC Cancer Staging Manual, 6th edition (2002), published by Springer-Verlag, New York, NY, used with permission of the American Joint Committee on Cancer (AJCC), Chicago, IL.

minal extension, invasion, or displacement of tissue planes and blood vessels, obliteration of normal fat planes, and lymph node enlargement. The CT signs pointing toward irresectability are tumor-to-aorta contact angle >45 degrees; obliteration of the fat pad between the tumor, aorta, and spine; and tracheal invasion. Unfortunately, these signs have a high falsepositive rate and they should be interpreted judiciously. Some of these patients are cachectic, however, and have little fat that can serve as a landmark.

A CT-specific staging system is used by some to gauge intramural extension and adjacent structure involvement. Intravascular contrast defines major adjacent vessels. A lymph node transverse diameter of 10mm or more is one criterion used to gauge lymph node metastases, but the results are disappointing. Analysis of 7218 resected lymph nodes in patients with esophageal carcinoma revealed that only 28% of

ADVANCED IMAGING OF THE ABDOMEN

histopathologically proven nodes containing metastases were ≥10mm in size, whereas 35% of metastatic nodes were 5 to 10mm,and 36% were <5mm (61); looking at the reverse, of resected nonmetastatic lymph nodes 7% were >10mm and the authors concluded that CT cannot assess whether mediastinal lymph nodes are affected. Current data indicate that CT cannot reliably identify nodal involvement. It is useful, however, in detecting extensive spread.

One limitation of endoscopic US in staging esophageal cancers is that often the probe cannot be passed through a carcinoma and the full extent of a tumor cannot be assessed. Also, more distant involvement cannot be evaluated. Endoscopic US often does suggest whether a carcinoma is limited to the mucosa or submucosa, or extends into adjacent tissues. Using endoscopic US criteria, a blinded evaluation of patients with esophageal malignancy attempted to differentiate between tumors limited to the esophageal wall (T1-2) and extraesophageal (T3-4) involvement (62); the accuracy of muscularis disruption and tumor border irregularity was low in differentiating between intraand extraesophageal disease. On the other hand, tumor maximal thickness was accurate in predicting extraesophageal extension; tumors limited to the esophagus measured 8 ± 1mm, but extraesophageal tumors had a thickness of 16 ± 2mm.

In a multicenter retrospective study of 79 patients with invasive (T4) esophageal carcinoma, endoscopic US accuracy was 88% in detecting tumor invasion, whereas CT accuracy was 44% (p = .0002) (63); of interest is that in this group of invasive cancers (consisting of both adenocarcinomas and squamous cell carcinomas), overall mortality was not significantly different whether surgery was performed or not.

Gauging lymph node involvement by endoscopic US is problematic. Although some studies have reached a nodal staging sensitivity of 92% (64), others have achieved staging sensitivities of only 60% to 80%. Nevertheless, endoscopic US is more accurate than CT in staging node invasion.

More proximal esophageal squamous cell carcinomas metastasize to cervical lymph nodes. Neck US using a 7.5- or 10-MHz transducer appears useful, keeping in mind that US simply detects lymph nodes, and nodes above a certain

ESOPHAGUS

size, such as 5mm or so in their long dimension, are assumed to be metastatic. Some studies use a short-to-long node dimension ratio >0.5 as an indicator of metastasis.

In patients with an intrathoracic esophageal squamous cell carcinoma, 1.5 T, T1-weighted sagittal MRI could not detect lesions subsequently found to be less than stage T2 (65); detected T3 lesions had a mean diameter of 25mm, whereas T4 lesions were 34mm. Nevertheless, future MRI staging of esophageal cancers appears promising. Thus in evaluating depth of mural invasion, T2-weighted MRI (at 1.5 T) agreed in 95% of patients with suspected superficial esophageal carcinoma with final histopathologic evaluation (66); in 5% of specimens MRI staging was higher.

Positron emission tomography has a role in staging both esophageal squamous cell carcinoma and adenocarcinoma. Its spatial resolution is insufficient to evaluate local extension, but it is often more specific than CT or US in detecting nodal involvement, keeping in mind that activity in a primary tumor can mask adjacent lymph node uptake. In 36 patients with a newly diagnosed esophageal cancer, 18F-FDG PET detected abnormal uptake in all (67); in those undergoing a curative esophagectomy, PET and CT correctly identified extent of nodal involvement in 76% and 45% of patients,respectively. The current primary role for PET is in excluding stage IV disease and in following patients postresection. One should keep in mind that esophagitis and adjacent inflammation result in a false-positive PET scan,and this study thus should be obtained only several months or longer after radiation therapy or resection.

Although thoracoscopic staging has been proposed, it is not widely employed.

Adenocarcinoma: Compared to postoperative staging, endoscopic US correctly staged tumor depth invasion in 87% of patients compared to 40% for CT (68) (Fig. 1.21). It was also significantly more accurate (73%) than CT (33%) in staging node invasion. Even if the endoscope probe could not pass through a tumor stricture, US is able to stage. Falsepositive findings are due to edema or fibrosis, whereas scattered infiltration accounts for falsenegative ones.

Lymph node metastasis is uncommon if an adenocarcinoma is limited to the mucosa, is present in <20% of patients when a cancer

extends to the submucosa, and increases further with deeper tumor extension (Fig. 1.22).

Complications

Esophageal neoplasms necrose and fistulize to adjacent structures, including trachea and major bronchi (Fig. 1.23). Imaging detection of tumor invasion into the tracheoesophageal fat plane increases the likelihood of future tracheoesophageal fistula formation. Unusual ones include an esophagosubarachnoid fistula (69) and even lower extremity subcutaneous emphysema, with psoas muscles presumably serving as a pathway (70).

Some neoplastic fistulas are iatrogenic. These tissues are friable, and even passage of a feeding tube can result in a perforation and fistula.

Therapy

Therapeutic options available for cure consist of surgical resection, at times in combination with radiotherapy or chemotherapy. Combined chemoand radiation therapy before esophagectomy for invasive esophageal or gastroesophageal junction cancer did not influence distant metastases but did control local recurrence, achieving a 5-year local tumor control rate of 90%, compared to 64% after esophagectomy only (71); on the other hand, postoperative mortality was 16% in the chemoand radiation therapy group versus 6% after esophagectomy only.

Some surgeons recommend esophagectomy to all patients able to withstand resection, keeping in mind that such an approach provides no survival benefit; only those patients who are too ill for surgery are referred to palliation, an approach of questionable value.

For cure: Flat esophageal carcinomas and areas of dysplasia are amenable to endoscopic mucosal strip resection. This procedure holds promise as long as a lesion is limited in depth. In general, survival does not depend on tumor location within the esophagus.

Photodynamic therapy cures most early squamous cell carcinomas (Tis and T1a) of the esophagus (72).

Traditional therapy for squamous cell esophageal carcinomas involves resection (where applicable), radiation therapy, and chemotherapy. Patients with stage I and II

28

A

C

E

ADVANCED IMAGING OF THE ABDOMEN

B

D

Figure 1.21. Esophageal endosonography. A: Image of superficial cancer of gastric cardia (arrows). The outer hyperechoic layer, representing muscularis propria, is intact. B: T1 tumor with an intact outer hypoechoic layer. C: Cancer invading muscularis propria. The outer hyperechoic layer is discontinuous and this tumor is thus stage T2. D: Tumor invading beyond the serosa; it is thus stage T3. Enlarged nodes were also present. E: Gastric cardia tumor extends through serosa and invades diaphragm (arrows).The tumor is thus stage T4. (Courtesy of Eric François, M.D., Centre AntoineLacassagne, Nice, France.)

Figure 1.22. Endosonography reveals numerous hypoechoic nodes surrounding the esophagus, (Courtesy of Eric François, M.D., Centre Antoine-Lacassagne, Nice, France.)

cancer achieve greater 5-year survival either with surgery only or surgery plus adjuvant chemoradiotherapy than with primary radiation therapy only.

Either a transhiatal esophagectomy or a transthoracic approach is used to resect a distal esophageal or gastric cardia adenocarcinoma. Both techniques appear to be equally safe and have comparable complications and survival rates.

The 5-year survival for patients undergoing resection of an esophageal adenocarcinoma, without postoperative residual tumor and without lymph node involvement, is about 60–70%.

Computed tomography using a negative oral contrast agent and endoscopy appear comparable in evaluating the response to chemotherapy

ESOPHAGUS

Figure 1.23. Cancer ulcerating into carinal lymph nodes (arrow).

of gastroesophageal junction adenocarcinomas

(3).

Although of limited application, transjugular intrahepatic portosystemic shunting (TIPS) can relieve portal hypertension in a cirrhotic patient with esophageal varices and an esophageal adenocarcinoma, allowing subsequent cancer therapy.

Palliation: Therapy of patients with cancer invading adjacent structures is controversial. Palliation of dysphagia in most patients with either squamous cell carcinoma or adenocarcinoma consists of stenting, radiotherapy, laser therapy, or serial dilation of the malignant stricture. Some centers report prolonged survival after chemotherapy, radiotherapy, and esophagectomy combined with regional node dissection.

Indications for stenting an inoperable esophageal carcinoma are obstruction or fistula. The procedure of choice for an inoperable carcinoma, especially in underdeveloped countries, is often a plastic stent inserted under endoscopic guidance without general anesthesia and requiring only a short hospital stay (73).

Numerous commercial stents are available and new ones continue to be introduced. These

stents are inserted using either endoscopic or fluoroscopic guidance. Self-expanding metal stents provide dysphagia palliation for a majority of patients. Covered stents provide better long-term palliation than uncovered stents. Comparing Wallstents and knitted nitinol stents, a multicenter study found that proce- dure-related mortality, early complication rate, and severe persistent pain after stent placement were higher with Wallstents (74); on the other hand, stent dysfunction, reintervention rate, and cost were lower in the Wallstent patients. Mean survival time of patients with malignant dysphagia is longer with covered than with uncovered stents.

In addition to treating strictures, covered metallic stents also are effective in treating inoperable malignant perforations and tracheoesophageal fistulas. Both covered Wallstents and Gianturco stents are popular. A contrast study shortly after the procedure evaluates stent position with respect to the fistula. Therapy with Wallstent endoprostheses or Gianturco stents of esophagorespiratory fistulas or perforations in patients with unresectable esophageal cancer led to closure of 90% of fistulas and all perforations (75).

What is the effect of a magnetic field on various metallic stents? Magnetic resonance imaging quality was not degraded by the titanium alloy Ultraflex stent and nitinol covered Wallstent (76); stainless steel Gianturco stent and modified Gianturco stent have an appreciable attraction force and torque. In particular, the Gianturco stent is pulled toward the head with a force 7¥ gravity, although whether this is enough of a risk for dislodgment is unknown. In addition, the stainless steel Gianturco stents produce gross artifacts around the stent.

Stent-related delayed complications tend to be more common than immediate complications and consist of recurrent fistula, tumor ingrowth, and tracheal compression by tumor (Fig. 1.24). Complications appear to be similar regardless of whether a stent is used primarily to treat a stricture or to bypass a fistula (Table 1.5). Gastroesophageal reflux is common if the stent straddles the gastroesophageal junction; at times resultant esophagitis is sufficiently severe to mimic an obstruction. Tumor ingrowth and overgrowth after stenting result in recurrent dysphagia and can be treated by balloon dilation or insertion of a second stent. Stents