0016-5107/92/3802-0158$03.00 GASTROINTESTINAL ENDOSCOPY Copyright © 1992 by the American Society for Gastrointestinal Endoscopy

Treatment of malignant strictures of the cervical esophagus by endoscopic intubation using modified endoprostheses L. A. Loizou, MD D. Rampton S. G. Bown, MD London, United Kingdom

Endoscopic intubation has traditionally been considered unsuitable as a means of palliating cervical esophageal carcinomas involving or within 2 cm of the cricopharyngeus sphincter muscle because of the potential problems of foreign body sensation and proximal prosthesis migration. We attempted to palliate eight such patients, three of whom had tracheo-esophageal fistulas by the endoscopic placement of modified Celestin endoprostheses; the floppy funnel of the prosthesis was positioned above the cricopharyngeus in the hypopharynx. Prosthesis placement and fistula occlusion was possible in all patients. Six patients had a significant long-term improvement in their dysphagia, managing a semi-solid (5 patients) or liquid diet (1 patient); two patients did not improve, despite accurate prosthesis placement, because of marked tracheal aspiration. Six patients reported no foreign body sensation; one patient had minor discomfort, and another moderate throat discomfort. Distal prosthesis migration occurred in two patients (replaced in 1 patient). Endoscopic intubation of high cervical esophageal carcinomas with specially modified endoprostheses is feasible and can provide worthwhile palliation of dysphagia and symptoms due to a tracheo-esophageal fistula. Foreign body sensation and proximal prosthesis migration did not prove troublesome. (Gastrointest Endosc 1992;38:158-164)

Statistics show that 7 to 10% of primary esophageal carcinomas are located in the cervical portion.1,2 Unfortunately, at presentation the majority of these tumors are incurable because of local invasion of vital structures or disseminated disease. 3 The primary aim of treatment, therefore, is the palliation of the distressing symptoms of dysphagia, tracheal aspiration, and those caused by the development of a tracheoesophageal (T-E) fistula. Palliative resection, if feasible, is a formidable procedure necessitating total esophagectomy and often laryngectomy with the formation of a permanent tracheostomy. Such surgery has a mortality risk of about 10% and a high rate of Received August 1, 1991. For revision October 2, 1991. Accepted December 26,1991. From the National Medical Laser Centre and Department of Gastroenterology, University College Hospital, London, United Kingdom. Reprint requests: Louis A. Loizou, MD, National Medical Laser Centre, The Rayne Institute, Room 103, 5 University Street, London WC1E 6JJ, United Kingdom.

158

local treatment failure; the median duration of palliation is approximately 9 months. 3- 5 High surgical risk patients and those with unresectable lesions have conventionally been treated with radiotherapy. The results of palliative radiotherapy, however, are very disappointing; local treatment failure occurs in 64 to 80% of patients and only 10% have satisfactory swallowing. 3,6,7 In patients with predominantly exophytic strictures, tumor photoablation using the Nd:YAG laser can provide reasonable palliation. 8 Where the tumor is predominantly extrinsic or fistulization into the trachea has occurred, effective palliation can only be achieved by the endoscopic placement of an endoprosthesis. This technique, although highly efficacious for tumors of the thoracic esophagus and gastric cardia, has been traditionally contraindicated for lesions of the cervical esophagus involving or within 2 cm of the cricopharyngeus sphincter muscle (CPSM) because of the potential problems of troublesome foreign body sensation, tracheal compression, and proximal GASTROINTESTINAL ENDOSCOPY

migration of the prosthesis into the hypopharynx with consequent laryngeal compression. 9 - 13 The limited published experience of intubation of high cervical lesions with standard or custom-made prostheses indicates that the concern about these potential complications is probably unjustified. 14• 15 Encouraged by these reports, we attempted to palliate patients with carcinomas of the high cervical esophagus (involving or within 2 cm of the CPSM) by endoscopic placement of modified Celestin endoprostheses. We report here our initial experience and discuss specific problems of the technique and its role in the management of cervical esophageal cancer.

and a non-traumatic, collapsible umbrella-shaped flange attached at the distal end of the prosthesis which is designed to prevent proximal migration (Fig. 1). The endoprostheses are available in four lengths (9.5, 12.5, 15, and 21 em) with

MATERIALS AND METHODS

Patients

Placement of a modified Celestin endoprosthesis was performed in eight patients with malignant strictures located in the high cervical esophagus, involving or within 2 em of the CPSM. Six patients had histologically proven primary squamous cell carcinomas and two had anastomotic recurrence following transhiatal esophagectomy for squamous cell carcinoma of the thoracic esophagus. All of the patients were considered to be inoperable because of advanced disease (metastases, local invasion), high anesthetic risk (cachexia, age, severe concomitant disease), or recurrence following surgery. Six patients had received prior external beam radiotherapy and four had received Nd:YAG laser therapy; laser therapy was initially successful in two, both patients (patients 1 and 6) managing a solid diet for 15 and 23 weeks, respectively. The anatomy of the tumor was documented in all patients by a contrast swallow examination and in some by CT scanning. At presentation, half the patients could only swallow liquids and the others experienced difficulty swallowing their saliva. Patient and tumor details are shown in Table 1. Illustrative case histories are presented below. Figure 1. Left, Commercially available Celestin (Medoc) en-

Endoprostheses

Celestin endoprostheses (Medoc, Ltd.) are made of latex rubber with a nylon spiral incorporated in the wall to prevent kinking when bent; they have a proximal inkwell funnel to collect food and liquid boluses and prevent distal migration

doprosthesis. Right, Celestin endoprosthesis modified by amputation of its funnel 3 mm proximal to the junction with the shaft and replacement with a distal flange removed from another tube; the latter functions as a soft, collapsible funnel for placement in the hypopharynx.

Table 1. Patient and tumor details·

o

b

Patient

Age

Sex

Tumor

1 2 3 4 5 6 7 8

73

M M M F F F M M

l"b

75 70 75 67 68

70 65

I" I" Anastom Anastom I" I" I"

Tumor position (em)

CPSM (em)

T-E fistula

Previous radiotherapy

Previous laser

19-24 12-19 16-21 16-19 15-18 16-22 14-17 14-20

17 12 15 14 13 15 14 14

Yes No No No No Yes No Yes

Yes Yes Yes No Yes No Yes Yes

Yes No No Yes No Yes No Yes

All tumors were squamous cell carcinomas. 1", primary; Anastom, anastomotic.

VOLUME 38, NO.2, 1992

159

internal and external diameters of 11 and 15 mm, respectively. Commercially available Celestin prostheses were modified for placement in the cervical esophagus as follows. The funnel was amputated 3 mm proximal to its junction with the shaft using a scalpel blade and replaced with a distal flange removed from another tube in order to provide a funnel which is soft and collapsible; as the amputated end of the tube fits tightly inside the new funnel, it was not necessary to fix the two parts together using glue (Fig. 1). The 3-mm rim of reinforced funnel left behind by the amputation fits snugly inside the new soft funnel and serves as a retainer ring to prevent distal migration of the modified prosthesis. The height of the new funnel measured from the proximal end of the shaft of the prosthesis is 2 em. Endoscopic technique

Endoscopic examinations were performed with a pediatric upper gastrointestinal endoscope (Olympus GIF-PI0). Careful measurements of the position of the CPSM, the upper and lower limits of the tumor, and the site of any T -E fistula were made in order to enable accurate positioning of the prosthesis. Patients were not routinely bronchoscoped to assess for possible airway invasion prior to intubation. Strictures impassable even with the pediatric endoscope were dilated to 10 to 12 mm using Celestin dilators under fluoroscopic control to enable complete passage of the endoscope through the tumor. At a second procedure after an interval of 2 to 3 days, the stricture was dilated further with Celestin bougies to 16 to 18 mm and the modified prosthesis placed using the Nottingham introducer under fluoroscopic control. 9 The junction of the shaft and funnel of the prosthesis was placed just proximal to the CPSM with the funnel left in the hypopharynx posterior to the larynx, its upper lip lying at the level of the arytenoid cartilages. As the funnel of the modified Celestin tube is soft and collapsible, it is partially compressed in an antero-posterior direction as it lies in the hypopharynx (Fig. 2). During the act of swallowing, the anterior and upward movement of the larynx enables the funnel to expand to virtually its natural resting size in anticipation of the oncoming food bolus. Careful positioning of the cervical prosthesis is crucial; as we found it impossible to accurately place a prosthesis under general anesthesia with an endotracheal tube in situ, all intubations were performed under conscious sedation. Post-intubation, chest and neck x-rays, and a contrast swallow were performed within 24 hours in order to detect and localize possible perforations, assess tube function, and ascertain T -E fistula occlusion. Illustrative case histories

Patient 1. A 73-year-old man had radical external beam radiotherapy for squamous cell carcinoma of the cervical esophagus 18 months previously, after which he was able to eat a virtually normal diet until he developed progressive dysphagia 6 weeks prior to referral to University College Hospital, by which time he was only managing a semi-solid diet. At endoscopy an exophytic esophageal carcinoma extending from 19 to 24 em from the incisors was found; the CPSM was located at 17 em. It was decided to manage the patient by endoscopic Nd:YAG laser tumor ablation. Following the initial course of laser therapy (three treatments), he 160

Figure 2. A, Endoscopic view of the soft funnel of the modified prosthesis in the hypopharynx posterior to the larynx. B, Note that the funnel is partially compressed in the antero-posterior direction and that its upper lip lies at the level of the arytenoid cartilages.

was able to eat a virtually normal diet. This symptomatic improvement was maintained over the next 12 weeks by repeat laser treatments at monthly intervals. By 15 weeks he was only managing to take liquids and had developed a hoarse voice and a cough, especially on attempting to swallow liquids. Indirect laryngoscopy confirmed a right recurrent laryngeal nerve palsy. At endoscopy, there was muralf extrinsic stricturing of the tumor segment with very little luminal tumor and a I-em T-E fistula at 22 em, findings confirmed by contrast swallow study (Fig. 3, A and B). The stricture was dilated and a modified Celestin prosthesis placed as described above with the whole of the funnel above the CPSM, its upper lip lying at the level of the arytenoid cartilages. A post-intubation contrast swallow study demonstrated free flow through the prosthesis with no tracheal aspiration and complete occlusion of the fistulous tract (Fig. 3, C and D). The patient complained of a minimal foreign body sensation in the throat which wore off completely over the next few weeks. He managed a semi-solid (pureed) diet until his death 19 weeks after placement of the modified prosthesis. GASTROINTESTINAL ENDOSCOPY

Figure 3. Gastrografin swallow (patient 1). A, Lateral view showing an irregular, almost completely occluding, stricture of the cervical esophagus starting at the CPSM (C6). A nasogastric feeding tube has been passed through the stricture. B, Lateral film (taken following virtually complete clearance of contrast from the esophagus) showing an oval-shaped T-E fistula at the level of TY2 (arrow). At endoscopy, the fistula measured 1 cm in diameter. C and D, Lateral views of repeat Gastrografin swallow taken following placement of a modified Celestin endoprosthesis showing complete occlusion of the fistula and free flow of contrast through the prosthesis into the distal esophagus. The top lip of the funnel is situated above the CPSM in the hypopharynx at the level of C5.

Patient 4. A 75-year-old woman, an insulin dependent diabetic, had a transhiatal esophagectomy for squamous cell carcinoma of the thoracic esophagus 6 months previously. Three months after the operation, she developed recurrent dysphagia and anastomotic tumor recurrence was documented at endoscopy. She was managed by repeated dilations with short-lasting symptomatic benefit after each procedure. Her dysphagia progressed and she was unable to swallow her own saliva; she was referred to University VOLUME 38, NO.2, 1992

College Hospital for further management. At endoscopy, a 3-cm long circumferential anastomotic recurrence was seen 2 cm below the CPSM (14 cm). The stricture was dilated and all exophytic tumor ablated using Nd:YAG laser therapy at three separate sessions. As the symptomatic result of laser therapy was poor, a modified Celestin endoprosthesis was placed with its funnel above the CPSM in the hypopharynx. Following intubation, she developed stridor which resolved spontaneously after 48 hours. A contrast swallow study showed excellent tube function with no tracheal aspiration. She reported no foreign body sensation and managed a pureed tube diet until her death 7 weeks later. Patient 8. A 65-year-old man with a locally invasive (on CT scanning) squamous cell carcinoma of the mid-esophagus had palliative radiotherapy 3 months previously with only slight transient symptomatic benefit. His dysphagia deteriorated rapidly and by the time of his referral to University College Hospital he was unable to swallow his own saliva. Endoscopy showed a tight malignant exophytic stricture extending from 25 to 29 cm .from the incisors and a skip nodule of tumor at 20 cm. The stricture was dilated and both tumor areas treated with Nd:YAG laser in two sessions. Following this he was able to swallow liquids without difficulty. He presented again 3 weeks later with a hoarse voice and complete dysphagia. Bilateral vocal cord paralysis was documented at indirect laryngoscopy. Repeat endoscopy showed extrinsic compression of the cervical esophagus starting at the CPSM (14 cm) and extending down to 20 cm (presumably by involved lymph nodes); at this level, there was an obvious T-E fistula (1 cm in diameter) at the site of the tumor nodule treated previously with laser. There was moderate luminal narrowing through the distal tumor segment. The proximal esophagus was dilated with difficulty to 18 mm and a long modified endoprosthesis placed with its funnel above the CPSM in the hypopharynx. A contrast swallow study showed free flow into the tube and through to the stomach with complete occlusion of the fistula; there was, however, considerable aspiration into the trachea. He complained of no foreign body sensation and could swallow his saliva; as he was troubled by tracheal aspiration on attempting to swallow a liquid diet, his oral nutritional intake was minimal. He was, therefore, given enteral feeding through a percutaneous endoscopic gastrostomy until his death 6 weeks later.

RESULTS

Successful placement of a modified Celestin endoprosthesis with the funnel proximal to the CPSM in the hypopharynx was possible in all eight patients with high cervical esophageal carcinomas. Treatment results and survival are shown in Table 2. Swallowing ability improved in six patients, from a mean dysphagia grade of3.5 pre-intubation to a mean dysphagia grade of 2.2 post-intubation. This symptomatic improvement persisted until death in five patients (mean survival, 13 weeks; range, 4 to 30 weeks) and for 16 weeks in a patient (patient 5) not referred back, to University College Hospital after distal migration of the prosthesis; for the remaining 10 weeks of her life 161

Table 2. Treatment results and survival following intubation with modified endoprostheses· Patient

DG preintubation

DG postintubation

1 2b 3b 4 5 6 7 8

3 3 3-4 4 4 3 4 3-4

2 2 3-4 2 2 2 3 3-4

Time laser --> intubation (wk)

Occlusion of T-E fistula

15

Yes

2 23

Yes

2

Yes

Survival postintubation (wk) 19 30 2 7 26' 4 6 6

Dysphagia grade (DG) was measured on a 5-point scale; 0 = no dysphagia, 1 = able to eat most solids,2 = able to eat semi-solids, 3 = able to swallow liquids, 4 = complete dysphagia. b Dilation-induced perforation. , Palliation with endoprosthesis for 16 weeks; tube displaced distally in an attempt to clear a food bolus at an outside hospital. Patient managed until death by repeat dilations. a

her dysphagia was adequately palliated by repeat dilations. The treatment failures (patients 3 and 8) were functional as swallowing ability did not improve despite successful and correct prosthesis placement; both had bilateral vocal cord paralysis and marked pharyngo-tracheal aspiration on a contrast swallow study. The latter patients died as a result of progressive malnutrition and aspiration pneumonia; the remaining patients died of cachexia. None of the deaths were attributable to a prosthesis complication (perforation, hemorrhage). The high-placed modified prosthesis was well tolerated. Four patients complained of no foreign body sensation in the throat, three patients of minor discomfort which subsequently disappeared completely in two, and only one (patient 3) of continuous moderate discomfort in the throat. Stridor following intubation developed in two patients; in both of these cases it resolved spontaneously within 72 hours. A tracheostomy did not prove necessary in any of the eight patients. Distal migration of the modified endoprosthesis into the stomach occurred in two patients, one spontaneously (patient 3) and the other following an attempt at an outside hospital to clear the prosthesis blocked by food (patient 5). The prosthesis was successfully replaced in the first patient; the second was not referred back to University College Hospital for further management. Prosthesis food bolus obstruction occurred in only one patient (patient 5). Subcutaneous neck emphysema following dilation to 18 mm with Celestin bougies developed in two patients (patients 2 and 3) with very tight, fibrous strictures who had been treated in the past with radiotherapy; following placement of the prosthesis, no leak could be demonstrated on contrast swallow study, and they made an uncomplicated recovery on conservative management. 162

DISCUSSION

The palliation of patients with high cervical esophageal carcinomas is a formidable management problem. Nd:YAG laser treatment of predominantly exophytic tumors in this location can provide reasonable palliation of dysphagia, although the functional result is worse than for tumors of the thoracic esophagus and gastric cardia. 1B,17 It appears that laser therapy may predispose to or accelerate the formation of an esophago-tracheal fistula, especially if previous radiotherapy has been given and complete ablation of the tumor is attempted. 18 For tumors that have fistulized and those that are predominantly extrinsic, the only possible means of providing effective palliation is the endoscopic placement of an endoprosthesis. The conventional teaching, however, is that intubation is contraindicated for cervical esophageal tumors because of the potential problems of foreign body sensation, tracheal compression, or proximal migration of the prosthesis. 9- 13 Using our specially modified Celestin endoprosthesis with the soft collapsible funnel positioned above the CPSM, foreign body sensation did not prove troublesome. Of the eight patients intubated, only two reported throat discomfort (one moderate, one mild) which was tolerable and did not necessitate removal of the prosthesis. This lack of significant foreign body sensation in patients with advanced tumors is probably due to local infiltration of the nerves innervating the hypopharynx, CPSM, and upper esophagus resulting in hypo/anesthesia; this propensity to infiltrate local neural structures is reflected by the frequent occurrence of unilateral or bilateral vocal cord paralysis. In addition, previous radiotherapy or surgery also impair normal sensation. The soft, collapsible funnel of the modified Celestin prosthesis also helps to reduce hypopharyngeal irritation and the GASTROINTESTINAL ENDOSCOPY

severity of the foreign body sensation compared with unmodified prostheses. Goldschmid et a1. 15 reported that of eight patients with high cervical carcinomas intubated with custom-made or modified Atkinson (Key Med) endoprostheses, only two complained of mild but tolerable foreign body sensation; in both these patients the funnel was placed in the hypopharynx. 15 In the report by Maercke et aI.l 4 of the six patients with high cervical tumors intubated with unmodified Atkinson (Key Med) prostheses, two experienced foreign body sensation severe enough to necessitate removal of the tube; the one patient in whom the funnel of the prosthesis was placed in the hypopharynx did not complain of throat discomfort. 14 The earlier limited published experience with intubation for cervical carcinomas using surgically placed unmodified Celestin prostheses 12 and endoscopically placed Procter-Livingstone prostheses 13 was unfavorable. In the present series, prosthesis placement provided good palliation of dysphagia, most patients managing a soft pureed diet; the experience of Goldschmid et aI.l5 and Maercke et aI.l 4 was very similar. In the presence of recurrent laryngeal nerve palsy, the functional result is poor despite accurate positioning of the endoprosthesis because of significant aspiration into the trachea and associated dysmotility of the hypopharynx. Proximal migration of the prosthesis did not occur in our patients. The design of the proximal end of the prosthesis with a 3-mm rim of reinforced funnel acting as a retainer ring proved effective in preventing distal migration. Stridor development, another of the traditional concerns about intubation of cervical esophageal tumors, was seen in only two patients; this was transient and resolved within 72 hours. Routine bronchoscopic evaluation of the trachea for tumor invasion and external compression prior to intubation was not performed. We adopted a policy of performing bronchoscopy only if a patient developed significant persistent stridor post-intubation. However, as tracheal obstruction following prosthesis placement may cause sudden death, such a policy of close monitoring and timely intervention will not eliminate fatalities from this complication. A safer approach would be to assess all patients for tracheal narrowing prior to considering intubation. As CT scanning is a sensitive means of detecting tracheal tumor involvement,19,2o we would recommend this as the initial investigation followed by bronchoscopy in patients with scans showing tracheal displacement or compression. If significant narrowing of the tracheal lumen is demonstrated bronchoscopically, esophageal intubation should only be performed after the formation of a tracheostomy or stenting of the trachea with modified Montgomery tubes or expandable metal stents;21,22 tracheostomy is indicated only for compression of the upper trachea. The attending nursing and medical staff should be VOLUME 38, NO. 2, 1992

instructed to report the development of respiratory distress following intubation to the gastroenterology team early, so that endoscopic removal of the prosthesis if deemed necessary will not be delayed; in cases of imminent respiratory arrest the prosthesis can be removed rapidly by grasping it in the hypopharynx with Magill forceps under direct laryngoscopic vision. Both dilation induced perforations occurred in patients with very fibrotic strictures who had previously received radiotherapy, in one despite gradual dilation in two sessions. The perforation site was adequately occluded in both cases by correct placement of the prosthesis, as no leak could be demonstrated on a contrast swallow study, and the patients made an uncomplicated recovery. In order to minimize the risk of full-thickness splitting of tight fibrotic strictures, dilation to the diameter necessary for intubation should be performed gradually at multiple sessions. It can be concluded that endoscopic intubation should not be absolutely contraindicated for high cervical esophageal carcinomas. Our initial experience with a specially modified Celestin endoprosthesis positioned with its funnel proximal to the CPSM in the hypopharynx has been encouraging; there was no significant procedure-related morbidity or mortality, the prosthesis was well tolerated without troublesome foreign body sensation, and it provided worthwhile palliation of dysphagia and symptoms due to a T -E fistula in a group of patients who would otherwise have suffered terribly until their death.

REFERENCES 1. Goodner JT. Treatment and survival in cancer of the cervical esophagus. Am J Surg 1979;118:673-5. 2. Parker EF, Moertel GC. Carcinoma of the oesophagus: is there a role for surgery? Dig Dis 1978;23:730-6. 3. Collin CF, Spiro RH. Carcinoma of the cervical esophagus: changing therapeutic trends. Am J Surg 1984;148:460-6. 4. Kagewawa T, Yamana H, Ando N. Analysis of surgical treatment for carcinoma situated in the cervical oesophagus. Surgery 1985;97:150-6. 5. Kron IL, Joob AW, Levine PA, et aJ. Blunt esophagectomy and gastric interposition for tumours of the cervical oesophagus and hypopharynx. Am 8urg 1986;52:140-1. 6. Burdette WJ, Jesse R. Carcinoma of the cervical oesophagus. J Thorac Cardiovasc 8urg 1972;63:41-53. 7. Pearson JG. Esophageal cancer: value of radiation therapy. JAMA 1974;227:181-3. 8. Loizou LA, Bown 8G. Clinical laser applications in the gastrointestinal tract. In: Pounder RE, ed. Recent advances in gastroenterology. Chap 7. London: Churchill Livingstone, 1988:4787. 9. Ogilvie AL, Dronfield MW, Ferguson R, Atkinson M. Palliative intubation of oesophagogastric neoplasms at fibreoptic endoscopy. Gut 1982;23:1060-7. 10. Tytgat GNJ, Bartelsman JFWM, den Hartog Jager FCA, Huibregtse K, Mathus-Vliegen EMH. Upper intestinal and biliary tract endoprosthesis. Dig Dis Sci 1986;31:578-768. 11. Boyce HW Jr. Medical management of esophageal obstruction and esophageal-pulmonary fistula. Cancer 1982;50:2597. 12. Kairaluoma MI, Jokinen K, Karkola P, Larmi TKI. Celestin tube palliation of unresectable esophageal carcinoma. J Thorac Cardiovasc Surg 1977;73:783. 13. Hegarty MM, Angorn IB, Bryer JV, Henderson BJ, Le Roux

163

14. 15. 16. 17. 18.

164

BT, Logan A. Pulsation intubation for palliation of carcinoma of the oesophagus. Br J Surg 1977;64:160. Maercke PV, Kometz B, Bueb G. Pharyngeal placement of an oesophageal prosthesis in advanced squamous cell carcinoma. Endoscopy 1985;17:127-8. Goldschmid S, Boyce HW Jr, Nord HJ, Brady PG. Treatment of pharnygoesophageal stenosis by polyvinyl prosthesis. Am J GastroenteroI1988;83:513-8. Fleischer DE, Sivak MV Jr. Endoscopic Nd:YAG laser therapy as palliation for esophagogastric cancer: parameters affecting initial outcome. Gastroenterology 1985;89:827-31. Buset M, des Marez B, Batze M, et al. Palliative endoscopic management of obstructive oesophagogastric cancer: laser or prosthesis? Gastrointest Endosc 1987;33:357-61. Ell C, Demling L. Laser therapy oftumour stenoses in the upper

19. 20.

21. 22.

gastrointestinal tract: An international inquiry. Lasers Surg Med 1987;7:491-4. Picus D, Bafle DM, Koehler RE, Roper CL, Owen JW. Computed tomography in the staging of oesophageal carcinoma. Radiology 1983;146:433-8. Thompson WM, Halvorsen RA, Foster WL Jr, Williford ME, Postlethwait RW, Korobin M. Computed tomography for staging esophageal and gastroesophageal cancer: re-evaluation. AJR 1984;141:951-8. Westaby S, Jackson JW, Pearson FG. A bifurcated silicon rubber stent for the relief of tracheo-bronchial obstruction. J Thorac Cardiovasc Surg 1982;83:414-7. George PJM, Erving JD, Mantell BS, Rudd RM. Covered expandable metal stent for recurrent tracheal obstruction. Lancet 1990;335:582-4.

GASTROINTESTINAL ENDOSCOPY