Esophageal carcinoma: pre-operative staging and evaluation of anastomotic recurrence

ooI6·5107/90/3602-0S11$02.oo GASTROINTESTINAL ENDOSCOPY Copyright © 1990 by the American Society for Gastrointestinal Endoscopy

Esophageal carcinoma: pre-operative staging and evaluation of anastomotic recurrence Charles J. Lightdale, MD, Jose F. Botet, MD New York, New York

Survival after surgical resection of esophageal carcinoma is highly related to stage. The latest staging classifications (UICC/AJCC, 1987/1988) use the TNM system. Accumulating data show endoscopic Ultrasonography (EUS) to be consistently more accurate than CT in pre-operative staging of depth of tumor invasion. Detailed images of the esophageal wall obtained by EUS allow accurate staging even in early cancer where CT is ineffective. EUS is also more accurate than CT in staging regional lymph nodes, but is less accurate than CT in staging distant metastases due to tumor stenosis in some patients and limited depth of field. EUS has also been shown to be accurate in diagnosing post-operative recurrence of cancer in the area of the surgical anastomosis. EUS represents a major advance in the clinical staging of esophageal cancer. (Gastrointest Endosc

199O;36:S11-S16)

Esophageal carcinoma is a highly lethal disease, with a dismal prognosis. l Although relatively uncommon in the United States, it is essentially endemic in several other areas of the world. In China, it is the second leading cause of death from cancer. 2 In 1990, it is expected that almost 10,000 Americans will be diagnosed with esophageal cancer and about 8500 will die of this disease. 3 Epidermoid carcinoma has been the predominant type, which in Western countries has been largely related to the use of tobacco and alcohol. Adenocarcinoma of the distal esophagus and esophagogastric junction, often linked to Barrett's epithelium, has recently accounted for an increasing percentage of cases. 4 In spite of technical advances in diagnosis, symptomatic esophageal cancer is almost always advanced at the time of presentation. 5 Although both medical and surgical treatments have also improved, there has been no change in the exceedingly poor outlook. 6,7 In the United States currently, 90% of patients diagnosed with esophageal cancer will die within 5 years, the great majority within the first 12 months.s Studies conducted in China, using mass screening with esophageal cytology, have indicated that earlier detection in Received December 1, 1989. Accepted January 23, 1990. From the Department of Medicine, Gastroenterology Service, and the Department of Medical Imaging, Memorial Sloan-Kettering Cancer Center and Cornell University Medical College, New York, New York. Reprint requests: Charles J. Lightdale, MD, Memorial Sloan· Kettering Cancer Center, 1275 York Avenue, Box 67, New York, New York 10021. VOLUME 36, NO.2, 1990

asymptomatic patients can produce better long-term results. 9 Endoscopic surveillance in patients with Barrett's esophagus, although not proven effective, has been used successfully to detect early malignant change. 4 Survival after surgical resection of esophageal carcinoma, both epidermoid and adenocarcinoma, is highly related to the stage or anatomical extent of the disease. 6 The cancer originates in the esophageal mucosa, progressively infiltrates the esophageal wall, and may penetrate adjacent structures or organs. The deeper the penetration into the wall, the more likely are concurrent metastases to regional lymph nodes or distant sites. Patients with more advanced local/regional disease are also more likely to have local recurrence postoperatively in the area of the surgical anastomosis. 6 • lo STAGING CLASSIFICATION

The International Union Against Cancer (UICC) and the American Joint Committee on Cancer (AJCC) published essentially identical recommendations for staging cancer of the esophagus in 1987 and 1988 (Table 1).11,12 The classifications are based on the TNM staging system: T-This indicates the depth of primary tumor invasion, recognizing that these carcinomas originate in the mucosa and invade progressively deeper layers of the gastrointestinal tract wall as they advance. N-This indicates the spread of cancer to specified Sll

Table 1. AJCC/UICC staging of esophageal

cancer",'2

Primary tumor (T) TX, Primary tumor cannot be assessed TO, No evidence of primary tumor Tis, Carcinoma in situ TI, Tumor invades lamina propria or submucosa T2, Tumor invades muscularis propria T3, Tumor invades adventitia T4, Tumor invades adjacent structures Regional lymph nodes (N)O NX, Regional lymph nodes cannot be assessed NO, No regional lymph node metastasis NI, Regional lymph node metastasis Distant Metastasis (M)b MX, Presence of distant metastasis cannot be assessed MO, No distant metastasis MI, Distant metastasis Stage grouping Stage 0 Tis NO MO Stage I TI NO MO Stage IIA T2 NO MO T3 NO MO Stage lIB TI NI MO T2 NI MO Stage III T3 NI MO T4 Any N MO Stage IV Any T Any N MI o Regional lymph nodes: cervical esophagus-the cervical nodes including the supraclavicular lymph nodes; all others distant; thoracic esophagus-mediastinal and perigastric lymph nodes; all oth· ers distant, including supraclavicular and celiac lymph nodes. • b Most common metastatic sites include liver, lungs, and pleura.

regional lymph nodes. Any regional lymph node metastasis is considered N1, and there is no longer a category N2 of more distant nodal spread. Lymph node metastases at the celiac axis are considered not regional, but distant metastases. M- This indicates distant metastases to lymph nodes outside specified regional nodes or to organs, such as liver, lung, and adrenal glands, not involved by direct extension from the primary cancer. It is important to note that the new UICCjAJCC classification does not recognize symptoms, circumferential extent, or length of the primary cancer to be stage determinants. Clinical, surgical, and pathological stages are based on the same TNM criteria of extent of disease. CLINICAL STAGING

The initial evaluation of patients with carcinoma of the esophagus includes a physical examination with particular attention to sites of possible nodal metastases in the lymph nodes of the supraclavicular areas and to the lungs and liver. Blood tests can provide an indication of possible distant metastases, particularly to the liver, and chest x-ray will allow assessment of possible pulmonary metastases and some idea of mediastinal structures, but far more information can be obtained by modern' imaging methods. Of potential 812

value for clinical staging are CT, magnetic resonance imaging (MRI), and endoscopic ultrasonography (EUS). CT scanning has been accurate for local extension to mediastinal structures such as the trachea, aorta, and pericardium and for distant metastases to liver, lung, and adrenals. However, CT has been disappointingly inaccurate in determining depth of wall penetration in earlier stages of disease and in assessing regional lymph node metastases. 13.14 MRI is in an early phase of evaluation, but seems to add little to CT for staging esophageal cancer. EUS, with its unique ability to define the gastrointestinal tract wall in detail, offers the most accurate method for evaluating the depth of cancer invasion (T). EUS also appears to be more effective in detecting abnormal regional lymph nodes (N) than CT or MRI. The limited penetration of high-frequency EUS lowers its utility for complete evaluation for distant metastases, where CT and MRI have been effective. ENDOSCOPIC ULTRASONOGRAPHY

Most experience with examination of the esophagus with EUS has been with the Olympus GF-UM2 (7.5 MHz) or GF-UM3 (switchable 7.5 to 12 MHz). The forward oblique optics of these instruments do not allow visual examination of the esophagus. EUS examination should be preceded by a standard upper gastrointestinal endoscopy with a diagnostic forwardviewing endoscope. Areas of interest are noted as distance from the incisors. The entire esophagus is best examined using the water-filled balloon technique!5,16 Areas of interest are located by advancing the instrument to distances from the incisors corresponding to those measured on standard endoscopy. Some searching within a localized area may be needed to image a small lesion. The endoscopist watches the screen and the length markings. For orientation, anatomical structures brought into view may be correlated with depth of insertion. Important anatomical structures for localization include the heart, aorta, aortic arch, and great vessels of the neck emerging from the aortic arch. The inferior vena cava, superior vena cava, azygous, and hemiazygous veins are useful landmarks as well, especially the azygous vein, as it progresses in a cephalad direction from posterior to anterior along the right side of the esophagus to enter the superior vena cava. The characteristic air-filled images of the trachea, carina, and mainstem bronchi and the double-layered pericardium are other key landmarks for esophageal cancer staging (Figs. 1 to 6)}7-19 Observing the image on the screen in real time, the endoscopist may help orientation by using torque on the shaft of the instrument. Thus, in the distal esophagus' the aorta may be placed in roughly the 5 o'clock or 6 o'clock position, which provides an image orienGASTROINTESTINAL ENDOSCOPY

1. The five-layer esophageal wall at the top of the photograph is disrupted by a near circumferential hypoechoic nodular cancer, which penetrates the entire thickness of the wall. There is a clear plane, however, between the cancer and the aorta. Stage T3.

Figure

tation similar to that for CT scans. Posterior appears at the bottom and anterior at the top of the picture, and left and right are reversed as if facing a supine patient. Using the hand controls to angle the transducer helps focus the ultrasound beam on areas of interest. The amount of water in the balloon can be varied by instillation and suctioning of water. Balloon decompression is often needed when pulling the transducer back through the cardia into the esophagus, and when passing the aortic arch. In the proximal esophagus, dilation of the balloon may cause discomfort due to pressure on the trachea, and only small amounts of water should be instilled in this area. Five layers of the esophageal wall are most easily seen in the distal esophagus and at the esophagogastric junction. In other areas, it is not always possible to image five esophageal wall layers, usually because the wall is too close to the transducer and not in best fOCUS. 15,16 Inflating the balloon with water will push the wall further from the transducer, but may at the same time compress the wall, causing poor definition of the first three layers and resulting in a three-layer wall image. Fortunately, the fourth hypoechoic layer is usually well defined. While the precise correlation between the five-layer EUS wall image and histological layers remains somewhat controversial, it has proven clinically useful to accept the fourth hypoechoic layer as the muscularis propria. 20 Early cancer has been differentiated from advanced on the basis of involvement of the fourth hypoechoic layer on EUS. 21 On EUS, esophageal cancer usually appears as a hypoechoic disruption of the normal esophageal wall image pattern. For clinical staging, we consider invasion of the first three wall layers to be Tl (cancer VOLUME 36, NO. 2, 1990

involving mucosa and submucosa), invasion of the fourth layer, T2 (cancer involving muscularis propria), invasion through the fifth layer, T3 (cancer penetrating the adventitia) (Fig. 1), and direct invasion of adjacent organs or structures, T4 (Figs. 2 and 3). Lymph nodes are seen as round or ellipsoid, discrete structures, which can usually be differentiated from blood vessels in real time (Figs. 4 to 6). Malignant infiltration makes lymph nodes irregularly hypoechoic, round, and sharply defined. Inflammatory nodes tend to be more elongated, echogenic, and homogeneous, with a less distinct circumference. The larger the node, the more likely it is to be malignant. Size of lymph nodes can be accurately measured on EUS. Lymph nodes may be imaged in the 2- to 3-mm range, but it can be difficult to interpret their significance. 22 • 23 Intraoperative and pathological studies have shown that lymph nodes less than 5 mm on EUS were not likely to be malignant, but micrometastases could be missed. 24 In addition, not all malignant lymph nodes are imaged. Aibe et a1. 23 reported the detection of 43% of periesophageal lymph nodes larger than 5 mm and 58% of nodes larger than 10 mm. Regional lymph nodes greater than 10 mm and round were malignant in 48%. Aibe et a1. 23 reported enhancement of lymph node detection by oral administration of an oil-water emulsion 3 hours before EUS. PRE-OPERATIVE STAGING WITH EUS

In 26 patients with esophageal cancer, Tio et a1. 21 were able to judge five of six to be locally resectable, defined by finding a clearly demarcated mass with or without positive lymph nodes. Palliative resectability was correctly assessed in 11 of 13 patients with the finding of a localized mass, but with distant abnormal lymph nodes. Local non-resectabilty defined by malig-

2. A large esophageal tumor mass is evident, extending through the adventitia, and eroding the pericardium, indicated by the shaggy, irregular interface. Stage T4. Figure

813

agreement with surgical pathology for T and N staging are shown in Figure 7. 25 -29 There seems little doubt from the accumulating data that the depth of tumor invasion of esophageal cancer can be determined with high accuracy with EUS, and that staging of regional lymph nodes is more accurate than with CT. Microscopic metastases may not be detected and in some cases it has been difficult to distinguish large inflammatory nodes from metastatic disease. EUS-guided needle biopsy may potentially improve accuracy in staging N.

4. A rounded, hypoechoic, sharply demarcated malignant lymph node in the sub-carinal area in a patient with a mid-esophageal epidermoid carcinoma. In the 9 to 11 o'clock position, the azygous vein is seen, moving from postenor to anterior as it approaches the superior vena cava at the level of the aortic arch.

Figure

Figure 3. A, An esophageal tumor mass is seen compressing the posterior wall of the trachea, marked by a bright (hyperechoic) line in the 9 to 12 o'clock position. B, An image from a slightly higher position shows erosion of the cancer into the wall of the trachea. Stage T4.

nant infiltration into adjacent structures (pericardium, bronchus, major vessels) was correctly judged in six of seven patients. In an updated study involving 74 patients, Tio et a1. 25 compared EUS staging with CT, using the standard of surgical pathology and the 1987 UICC classification. In staging T, EUS accuracy was 89% compared with CT at 59%. The improved accuracy of EUS was most notable in the early stages. The overall accuracy in staging regional lymph nodes for EUS was 80% compared with CT at 51 %. At Memorial Sloan-Kettering Cancer Center, we are carrying out a prospective study of pre-operative staging of esophageal carcinoma using the 1988 AJCC classification. Preliminary results from our first 50 patients are similar to those of the Amsterdam group.26 In staging T, EUS accuracy was 92% compared with CT at 60%, and for N, EUS accuracy was 88% compared with 74% for CT. Combined results from five reports indicating 814

The great vessels emerging from the aortic arch are evident in the 12 to 3 o'clock position, surrounded by mottled, hypoechoic lymph nodes containing metastases from a distal esophageal epidermoid carcinoma. The vertebral column is at 6 o'clock and the superior vena cava at 10 o'clock.

Figure 5.

GASTROINTESTINAL ENDOSCOPY

Figure 6. Recurrent adenocarcinoma of the esophagus (associated with Barrett's epithelium) is seen as a hypoechoic interruption of the esophagogastric anastomosis from 6 to 7 o'clock. The tumor is confined to the wall (T2), but a 6-mm lymph node metastasis is evident at 1 o'clock to the left of the trachea.

Agreement With Surgical Pathology

100 87%

84 %

T

N

_EUS

~CT

Figure 7. Combined results in a total of 297 patients from five reported studies 25 - 29 where EUS and CT were used preoperatively to stage esophageal carcinoma.

Inability to pass a stenotic esophageal cancer with the EUS instruments currently in use (Olympus GFUM3, diameter 13 mm; optics 65 degrees forward oblique) is a limitation. Even with a 10-mm prototype blind probe, Tio et a1. 25 were unable to pass a cancer in 26% of cases. Our experience has been similar, with other reports indicating even more difficulty with inability to pass a malignant stenosis in 48 to 62% of cases, but this is clearly related to patient selection and stage. 80,81 Because of inability to pass the tumor, EUS staging of celiac axis lymph node metastases may be limited. Tioet a1. 25 found an EUS accuracy of 68% compared with 82% for CT in staging lymph nodes at the celiac axis. Thinner mechanical sector scan probes VOLUME 36, NO.2, 1990

and forward-viewing ultrasound endoscopes are in a developmental stage and may largely resolve this problem. Our preliminary results also showed that EUS was less accurate in staging distant metastases at 78% than CT at 90%.26 The greatest overall staging accuracy in our experience was achieved using both modalities: CT to stage for M and EUS for T and N. EUS represents a major advance in the clinical staging of esophageal cancer. Further studies are needed to determine whether the increased staging accuracy afforded by EUS will have an impact on patient outcome. More accurate clinical staging might allow better patient selection for operation or nonoperative management of esophageal cancer, such as endoscopic laser therapy. Accurate staging could also facilitate the application of combined modality approaches using radiation and chemotherapy. EVALUATION OF ANASTOMOTIC RECURRENCE

In spite of aggressive surgery, locally recurrent upper gastrointestinal cancer is a common problem. Gignoux et a1.82 recently reported 67% of recurrent esophageal cancer to have a local component and 32% to be local alone. In the series of Gunderson and Sosin,33 87% of patients with recurrent gastric cancer had a local component, and in 53% this was the sole site of recurrence. Symptoms due to locally recurrent cancer, however, are essentially indistinguishable from those due to post-operative dysmotility, inflammation, or fibrous stricture. 84 Exploratory surgery is diagnostic in most cases of recurrent malignancy, but has high morbidity, and is reported to be infrequently curative. 88 ,34 Endoscopy with biopsy is diagnostic only if the recurrent cancer involves the mucosa or superficial submucosa, but in many patients recurrence primarily occurs on the serosal adventitia side of the gastrointestinal tract wall. 10,33 Efficacy of extracorporeal imaging tests, most notably CT, is limited after esophageal or gastric resection by varying anastomotic morphology, and artifacts created by movement, retained metallic clips, and staples. 35 An accurate diagnosis of recurrent anastomotic cancer provides critical prognostic guidance for appropriate patient management, and earlier diagnosis has the potential to improve surgical and oncological results. At Memorial Sloan-Kettering Cancer Center, we used EUS with the Olympus GF-UM2 and GF-UM3 (7.5 MHz/12 MHz) to examine the upper gastrointestinal tract in 40 patients who had resection of esophageal or gastric cancer and symptoms suggesting recurrence.:16 There were 24 patients with recurrent cancer in the area of the surgical anastomosis (based on endoscopic biopsy in 16, repeat endoscopy in 2, and surgery after negative endoscopy in 6) and 16 patients without anastomotic recurrence. With EUS, locally recurrent cancer was correctly identified by nodular S15

hypoechoic thickening at the anastomosis (Fig. 6) in 23 of 24 patients with one false negative; absence of anastomotic recurrence was correctly diagnosed in 13 of 16 with three false positives (sensitivity, 95%; specificity, 80%; positive predictive accuracy, 88%; and negative predictive accuracy, 92%).36 In 6 of our 23 patients with recurrent upper gastrointestinal cancer diagnosed by EUS, endoscopy with biopsy and brush cytology was negative. 36 Dancygier and Classen:17 have noted a false positive anastomotic thickening on EUS in one patient probably due to inflammation. The development of EUS-guided needle biopsy should improve the specificity of EUS in the diagnosis of post-operative locally recurrent upper gastrointestinal cancer. CT scans were not useful for diagnosis of anastomotic recurrence. EUS also diagnosed recurrent cancer in local lymph nodes not seen on CT (Fig. 6). On the other hand, CT was more effective in diagnosing distant metastases to the lung, liver, and peritoneum.:l6 High-frequency EUS has a limited depth of penetration and is not adequate for evaluation of sites distant from the gastrointestinal tract, but is highly sensitive for diagnosis of recurrent cancer at the surgical anastomosis.

16. 17. 18. 19. 20. 21.

22.

23. 24.

25. 26.

REFERENCES 1. Earlam R, Cunha-Menlo J. Oesophageal squamous cell carcinoma: a critical review of surgery. Br J Surg 1980;67:381-90. 2. Yang GR, Qui SL. Endoscopic surveys in high-risk and lowrisk population for esophageal cancer in China with special references to precursors of esophageal cancer. Endoscopy 1987;19:91-5. 3. Silverberg E, Lubera J. Cancer statistics 1989. CA 1989;39:321. 4. Hameeteman W, Tytgat GNJ, HouthoffHJ, van den Tweel JG. Barrett's esophagus: development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:1249-57. 5. Lightdale CJ, Winawer SJ. Screening diagnosis and staging of esophageal cancer. Semin Oncol 1984;11:101-12. 6. Skinner DB, Ferguson MK, Soriano A, Little AG, Staszak VM. Selection of operation for esophageal cancer based on staging. Ann Surg 1986;204:391-401 7. Leichman L, Herskovic A, Leichman C, et al. Nonoperative therapy for squamous cell cancer of the esophagus. J Clin Oncol 1987;5:365-70. 8. Kelsen D, Bains M, Hilaris B, Martini N. Combined modality therapy of esophageal cancer. Semin OncoI1984;11:169-77. 9. Huang GJ. Early detection and surgical treatment of esophageal cancer. Jpn J Surg 1981;11:399-405. 10. Kasai M, Mori S, Watanabe T. Follow-up results after resection of thoracic esophageal carcinoma. World J Surg 1978;2:543-50. 11. TNM classification of malignant tumors. International Union Against Cancer. 4th ed. Berlin: Springer-Verlag, 1987:40-2. 12. Manual for staging of cancer. American Joint Committee on Cancer. 3rd ed. Philadelphia: JB Lipppincott, 1988:63-7. 13. Moss AA, Schnyder P, Thoeni RF, Margulis AR. Esophageal carcinoma: pretherapy staging by computed tomography. AJR 1981;136:1051-6. 14. Halvorsen RA, Thompson WM. Computed tomographic staging of gastrointestinal tract malignancies: Part I. Esophagus and stomach. Invest RadioI1987;22:2-16. 15. Caletti G, Bolondi L, Barbara L. Instrumentation and scanning

816

27. 28. 29.

30. 31.

32. 33. 34. 35. 36.

37.

techniques. In: Kawai K, ed. Endoscopic ultrasonography in gastroenterology. Tokyo: Igaku-Shoin 1988:1-17. Tio TL. Endosonography in gastroenterology. Berlin: SpringerVerlag, 1988:3-8. Strohm WD, Classen M. Anatomical aspects in ultrasonic endoscopy. Scand J GastroenteroI1984;19(suppI94):21-33. Sivak MV Jr, George C. Endoscopic ultrasonography: preliminary experience. Scand J Gastroenterol 1984;19(suppl 94):519. Tio TL, Tytgat GNJ. Atlas of transintestinal ultrasonography. Aalsmeer, The Netherlands: Drukkenj Mur-Kostverloren BV, 1986. Kimmey MB, Martin RW, Haggitt RC, Wang KY, Franklin DW, Silverstein FE. Histologic correlates of gastrointestinal ultrasound images. Gastroenterology 1989;96:433-41. Tio TL, den Hartog Jager FCA, Tytgat GNJ. The role of endoscopic ultrasonography in assessing local resectability of oesophagogastric malignancies. Accuracy, pitfalls, and predictability. Scand J GastroenteroI1986;21(suppI123):78-86. Tio TL, Tytgat GNJ. Endoscopic ultrasonography in analyzing peri-intestinal lymph node abnormality. Preliminary results of studies in vitro and in vivo. Scand J Gastroenterol 1986;21(suppI123):158-63. Aibe T, Ito T, Yoshida T, et al. Endoscopic ultrasonography of lymph nodes surrounding the upper GI tract. Scand J Gastroenterol 1986;21(suppl 123):164-9. Yasuda K, Kiyota K, Mukai H, Cho E. Anatomical aspects of endoscopic ultrasonography. In: Kawai K, ed. Endoscopic ultrasonography in gastroenterology. Tokyo: Igaku-Shoin 1988:14058. Tio TL, Cohen P, Coene P, Udding J, den Hartog Jager FCA, Tytgat GNJ. Endosonography and computed tomography of esophageal carcinoma. Gastroenterology 1989;96:1478-86. Lightdale CJ, Botet JF, Zauber AG, Brennan MF. Endoscopic ultrasonography in the staging of esophageal cancer: comparison with dynamic CT and surgical pathology [Abstract]. Gastrointest Endosc 1990;36:191. Murata Y, Muroi M, Yoshida M, et al. Endoscopic ultrasonography in the diagnosis of esophageal carcinoma. Surg Endosc 1987;1:11-6. Souquet JC, Valette PJ, Berger F. Endosonography (ENS) and the preoperative staging of esophageal squamous cell cancer [Abstract]. Gastrointest Endosc 1988;34:208. Grimm H, Sollano J, Hamper K, Noar M, Soehendra N. Endoscopic ultrasound (EUS) of esophago-gastric cancer: a new requirement for preoperative staging [Abstract]. Gastrointest Endosc 1988;34:176-7. Ziegler K, Sanft C, Semsch B, et al. Endosonography is superior to computed tomography in staging tumors of the esophagus and the cardia. Gastroenterology 1988;94:A517. Palazzo L, Ponsot P, Gayet B, et al. Endoscopic ultrasonography and computed tomography in esophageal and cardial malignancies: Prospective study of 29 surgical cases. Gastroenterology 1988;94:A339. Gignoux M, Roussel A, Paillot B, Gillet M, et al. The value of preoperative radiotherapy in esophageal cancer: Results of a study of the E.O.R.T.C. World J Surg 1987;11:426-32. Gunderson LL, Sosin H. Adenocarcinoma of the stomach: areas of failure in reoperation series (second or symptomatic look). Int J Radiat Oncol Bioi Phys 1982;8:1-11. Allum WH, Hockey MS, Fielding JWL. Gastric remnant recurrence-detection and implications for the management of gastric cancer. Clin Oncol 1984;10:333-9. Mauro MA, Lee JKT, Heiken JP, Balfe DM. Radiologic staging of gastrointestinal neoplasms. Surg Clin North Am 1984;64:6788. Lightdale CJ, Botet JF, Kelsen DP, Turnbull AD, Brennan MF. Diagnosis of recurrent gastric cancer at the surgical anastomosis by endoscopic ultrasound. Gastrointest Endosc 1989;35:407-12. Dancygier H, Classen M. Endoscopic ultrasonography in esophageal diseases. Gastrointest Endosc 1989;35:220-5.

GASTROINTESTINAL ENDOSCOPY