- Email: [email protected]
Resection for advanced cancer of the thoracic esophagus: Cervical or thoracic anastomosis?
Resection for advanced cancer of the thoracic esophagus: Cervical or thoracic anastomosis? Late results of a prospective randomized study During a 21/z-year period, 60 consecutive patients with cancer of the thoracic esophagus were randomized to undergo a cervical or thoracic anastomosis. The tumors were staged postoperatively (stage I, n = 2; stage II, n = 19; stage III, n = 9; and stage IV, n = 30) and were almost equally distributed between the two groups. The upper limit of three tumors was above the convexity of the aortic arch. The esophageal specimens were studied with regard to measurements of the tumor and of the resected esophagus. The microscopic aspects were evaluated by serial sections after vital staining. The prevalence of ignored plurifocal cancers, of submucosal infiltrations, and of distant areas of dysplasia in both groups was confirmed. Malignant invasions of esophageal sections were more frequent in patients undergoing thoracic anastomosis (10 versus 3), and diseased upper mediastinal lymph nodes were more frequent in those undergoing cervical anastomosis (17 versus 7). Mortality was equally divided between the two groups. Respiratory complications and recurrent laryngeal trauma were more frequent in patients having cervical anastomosis. Long-term survivors had stage NO disease, with a healthy esophageal section. Even though subtotal esophagectomy reduces the prevalence of microscopic esophageal wall invasion above the tumor and allows more complete unilateral exploration and resection of invaded lymph nodes, it offers no significant benefit concerning survival of patients with advanced cancer and malignant lymphadenopathy. (J 'fHORAC CARDIOVASC SURG 1992;103:784-9)
M. Ribet, MD,a
B. Debrueres, MD,a and M. Lecomte-Houcke, MD,b Lille, France
h e delay of symptoms, the multifocal sites of malignant disease, the long-running submucosal lymphatic drainage, and the disordered pattern of lymphatic spread account for the poor prognosis of thoracic esophageal cancer. Is it beneficial to lengthen the esophageal resection with a cervical anastomosis in all cases? To consider this question, we conducted a prospective study of partial esophagectomy with esophagogastric anastomosis in the thorax and of subtotal esophagectomy with anastomosis in the neck.
Patients and methods Between December 1983 and May 1986, we performed at random a cervical or a thoracic anastomosis on 60 patients. From Hopital Calmette, University Hospital, Lille," and the Department of Pathology, Lille Medical School," Lille, France. Received for publication Aug. 21, 1990. Accepted for publication Feb. 7, 1991. Addressfor reprints:M. Ribet, Hopital Calmette,F 59037 LilleCedex, France.
12/1/29024
784
There were 55 men and five women, aged from 34 to 70 years (mean 54.9), and the anastomosis was performed after an esophagectomy for cancer of the thoracic esophagus. The upper radiologic limit of the stricture was above the aortic arch in three patients, between the convexity of the arch and the inferior pulmonary vein in 40 patients, and under the hilum of the lung in 17 patients (Table I). In all cases there was a healthy-looking radiologic margin of esophagus below the clavicular level.* The preoperative work-up consisted of a clinical examination, a barium meal, and an examination of the pharynx, the larynx, the trachea, the bronchi, and the esophagus with both a flexible and a rigid endoscope, performed with the aid of general anesthesia, with biopsies of the tumor and of those areas of mucosa above the tumor marked by vital staining with toluidine blue. A chest film and computed tomographic scan, abdominal echography, respiratory and cardiac tests, and a biologic assessment completed the evaluation (Table II). Only patients having invasion of the respiratory mucosa or distant metastasis were eliminated from the trial. Endoscopic signs of tracheobronchial 'This study wasstarted beforethe creationof the LilleUniversity HospitalCommitteeof Ethics,whenpatients'informedconsentwasstill notnecessary. It wasjudgedacceptable, becausebeforethat date the techniquehad dependedon the preference ofthe particularsurgeon, with no objective evaluation.
Volume 103 Number 4 April 1992
Resection for advanced cancer of thoracic esophagus
Table I. Sites of upper limits of resected cancers Highest level of cancer
Above aortic arch Between aortic arch convexity and inferior pulmonaryvein Under hilum of lung
Anastomosis Cervical
Table II. Preoperative clinical evaluation of the patients Anastomosis
Thoracic
1 23
2
Patient evaluation
17
6
II
Mean age (yr) Mean lossof weight (%) Respiratory disease Cardiovascular disease FEV J <25% normal Pa02 at rest <70 mm Hg Liver steatosis Liver cirrhosis Renal insufficiency
compressions (n = 20), as well as radiologic evidence of tumoral contacts with adjacent structures (n = 39), were not considered to be contraindications. According to these conditions, 60 consecutive patients were included in the study. No preoperative adjuvant treatment was administered. Patients were selected at random for either cervical or thoracic anastomosis by the anesthetist, who drew a ticket specifying the type of anastomosis from a pool of 60 tickets before the procedure was begun. Two patients were excluded from the study because the intended procedure, a thoracic anastomosis, was impossible to perform. The two tickets labeled "thoracic anastomosis" were returned to the ticket pool. The surgical technique was identical for the two groups: I. Median celiotomy and preparation of a gastric tube vascularized by the right gastric and right gastroepiploic vessels after resection of the upper half of the lesser curvature and transparietal finger dilatation of the pyloric ring; resection of the lymphatic nodes situated on the left gastric, splenic, and hepatic chains and around the celiac trunk; creation of a feeding jejunostomy 2. Right posterolateral thoracotomy and esophageal dissection extending up to provide a 5 em margin of stretched esophagus above the cancer in the case of thoracic anastomosis and to the base of the neck in the case of cervical anastomosis; resection of the posterior mediastinal connective tissues; lymphadenectomy of the parahiatal, paraesophageal, subcarinal, and right paratracheal nodes; in case of incomplete tumoral resection, clipping of invaded zones; transhiatal pull-up of the gastric tube into the posterior mediastinum, to the level of the esophageal section in the case of thoracic anastomosis and to the base of the neck in the case of cervical anastomosis 3. For cervical anastomosis, right oblique cervical incision, esophageal freeing, and section at 2 to 3 cm under the pharynx; unilateral resection of lymph nodes of the recurrent laryngeal, subclavicular, jugular, and subdigastric chains 4. Esogastric terminolateral hand suture in two layers; introduction of a nasogastric tube Resection was considered to be complete in 36 patients; two wedge resections of the liver and two wedge resections of the lung were added to the esophagectomy for solitary metastases, which had not been detected before the operation; one lobectomy was done for a direct tumoral invasion of the lung and inferior pulmonary vein. Induction of anesthesia and performance of the operation took a mean of 6 hours 45 minutes for cervical anastomoses and 6 hours 15 minutes for thoracic anastomoses The esophageal resection specimens were immediately opened and spread out. The length of the shrunken esophagus,
785
Cervical
Thoracic
53.7 -7.7 22 10 4 14 3 2 0
56.1 -7.1 25 13 4 12 3 I I
FEV,. Forced expiratory volume in 1 second; Pa02. arterial oxygen tension.
Table III. Measurements and aspects of resected esophagus and tumors Anastomosis Macroscopy
Cervical
Thoracic
Mean resectedesophagus (em) Extremes (em) Mean supratumoral margin (em) (on shrunken esophagus) Mean tumor length (em) Obstructive and ulcerated Ulcerated (perforated: 2) Infiltrating Obstructive Stemmed
14.6 (10.5-18) 5.01
11.9 (8-15) 2.83
5.08
17 7 (I) 5 1 0
5.3 12 3 (I) 10 3 2
the size of the tumor, and the distances between the tumor, the cardia, and the upper section were measured; the whole mucosa was stained with toluidine blue dye to prepare 15 to 60 sections (mean 45); all resected lymphatic nodes were prepared for sections. A map of the esophagus and nodes was made. This pathologic work-up took 10 hours. The patients were kept in the intensive care unit for 24 to 72 hours and were supported with a ventilator. A feeding jejunostomy was used on the second or third day. The anastomosis was opacified on the sixth day, and the nasogastric tube was then removed. In case of a leak, the tube was left in place until opacification showed that the anastomosis had sealed off. The mean hospital stay, including the preoperative period, was 24.2 days for patients with cervical and 16.6 days for those with thoracic anastomoses. Radiotherapy was started between 5 and 6 weeks after the operation (cervical, n = 23; thoracic, n = 27). Esophageal dilations were necessary for four patients having a cervical anastomosis (one after a clinical leak, one after a subclinical leak, and two with no history of leak) and for one patient having a thoracic anastomosis (after a subclinical leak). The results of the two types of anastomosis were studied by the X2 test, modified by the Yates coefficient for small numbers; to study survival, we used the log-rank test and the KaplanMeier actuarial method (R. Beuscart, MD, Department of Statistics, Lille University Hospital).
786
The Journal 'of Thoracic and Cardiovascular Surgery
Ribet, Debrueres, Lecomte-Houcke
Table VI. Postoperative staging of the tumors (Unione Internazionale Contro if Cancro 1987)
Table IV. Microscopic study Anastomosis Microscopy Poorly differentiated tumor Moderately differentiated tumor Well-differentiated tumor Juxtatumoral dysplasia Distant dysplasia Submucosal infiltration Plurifocal cancers Invasion of superior section (Mucosal infiltration) (Submucosal infiltration) (In situ cancer) (Lymphatic embolus)
Cervical
25
26
I 4
4
o
26
24 4 6 4 10 (4) (0) (6) (0)
5
7 2 3 (I) (I) (0) (I)
Table V. Lymph node invasion Anastomosis Invaded lymph nodes Abdominal (celiac) Mediastinal
Cervical Jumping metastasis Mediastinal or abdominal
Cervical
Cervical
Anastomosis
Thoracic
Thoracic
13
15
(8)
(II)
17 7
7 0
4
5
2
0
Difference in mediastinal N + concerning the upper area of the mediastinum.
Results The data on the specimens are summarized in Table
III. Two cancers had perforated the mediastinum, with an abscess in one patient. Tumoral invasion of adjacent structures, mentioned in the operative record for 24 patients, was confirmed in 23 patients. The tumor had invaded several organs in one patient and affected the membranous surface of the trachea or bronchi, or both (9), the aorta (4), the thoracic duct (9), the azygos vein (2), the auricular opening of a pulmonary vein (2), the left pulmonary artery (1), the left recurrent laryngeal nerve (1), the pericardium or the posterior parietal pleura (8), and the median surface of the lung (3). The microscopic structure of the tumor and the appearance of the neighboring or distal mucosa and of the upper esophageal section are detailed in Table IV. The results of abdominal, mediastinal, and cervicallymphadenectomies are given in Table V. The resected cancers are staged in Table VI. Nine patients died, four of an anastomotic leak (two cervical, two thoracic), one of a perforated ulcer in the gastric transplant (thoracic anastomosis), two of respira-
Staging
I IIa lIb III IV
Cervical
Thoracic
I 10
4
1 8 I 5
15
15
0
There were 23 T4 (13 CA, IOTA), 19 celiac N+, and 7 cervical N+ cancers, and 4 solitary metastases were classified as M l.
tory insufficiency after a previous pulmonary lobectomy (one cervical, one thoracic), and two of inhalation pneumonitis (two cervical). Morbidity consisted of seven anastomotic leaks in the neck, six of which were subclinical and sealed up rapidly, and one subclinical leak in the thorax. There were 21 episodes of bronchopulmonary infection after cervical anastomosis and II after thoracic anastomosis. There were three temporary and three permanent cases of vocal paralysis after cervical anastomosis and one after thoracic anastomosis. Survival data, hospital mortality included, were as follows: 27 patients were alive at I year (45%),19 at 2 years (31.6%), and 18 at 3 years (30%). The median survival was 9 months after cervical anastomosis and 12 months after thoracic anastomosis. With a minimum follow-upof 4lf2 years and a maximum of 7 years, eight patients are still alive (13.3%), equally divided between the cervical and the thoracic anastomosis groups; to the time of writing, fivehave survived more than fiveyears and three more than 6 years (Table VII). Table VIII summarizes the patterns of recurrence that are known for 29 patients; two patients died of another cancer, one died of a noncancerous disease, and II of an unknown cause considered as cancerous. Discussion Attempts to devise more radical operations for esophageal cancer have followed three directions: subtotal esophagectomy in all patients,' en bloc resection.? and more complete lymphadenectomy.' This is a prospective study of the first procedure. In the following paragraphs we compare the pathologic and clinical aspects of the two groups for which there was no significant preoperative difference. I. Considering the spread of malignant disease through the mucosa, preoperative stains and biopsy specimens of the supratumoral mucosa's 5 were deceiving:four zones that appeared to be malignant corresponded to one plurifocal cancer, one mucosal malignant infiltration, one
Volume 103 Number 4 April 1992
submucosal infiltration, and one area of severe dysplasia. However, serial sections of the specimens confirmed a greater frequency of plurifocal cancers (6/60), of submucosal infiltrations (13/60), and of distant areas of dysplasia (9/60). The discovery of this lesion is correlated with the meticulousness of the pathologist," It confirms the shortcomings of the preoperative endoscopic findings and the theoretic oncologic importance of the supratumoral esophagus concerning the extent of resection. For that matter, cervical anastomosis has allowed an average gain of 2.18 em in length on the shrunken esophagus,' with 27 sections devoid of any microscopic malignancy; there were only 20 disease-free segments in the thoracic anastomosis group, despite the fact that there were fewer lower tumor sites in the group having cervical anastomosis. The frequency of this invasion is noticeable, because in three cancers only the upper limit was above the aortic arch (one cervical and two thoracic anastomoses), with I and 2 em margins of resection on the shrunken specimens'' for the two thoracic anastomoses; eight invasions involved sections for tumors whose upper limits were under the convex portion of the aortic arch. 2. The length and width of the tumors did not differ between the two groups of patients. We went as far as possible in considering indications for operation: Examination with a rigid tracheobronchoscope, which remains for us one of the most valuable preoperative investigations, showed bulging of the posterior tracheobronchial wall in 20 patients. (There was actual invasion in 9.) Although the surgeon was satisfied with 36 resections, only 12 tumors were stage Tl or T2. 9 3. To determine nodal status (N) we studied 877 nodes, an average of 14.6 nodes for each patient; 179 nodes were invaded, and in 38 patients they were classified N +. Cervical anastomosis showed the prevalence of diseased cervical lymph nodes (7/30); it also improved the results of the superior mediastinal lymphadenectomy (17/30 N + versus 7/30 in the thoracic anastomosis group), but on both sites the operation was unilateral. In the abdomen the figures were logically similar in the two groups.l'v!' 4. Hospital mortality also was similar in the two groups. It was nil for the first 26 patients and very heavy for the rest of the series. There is no excuse for this, but the question arises regarding the negative aspect of operating under the constraint of randomization: One of the deaths in the group undergoing thoracic anastomosis was related to our attempting a difficult, highly situated anastomosis. The mortality in this series (15%) is roughly twice the standard mortality observed in our unit (8%), and we admit to having the poorest sampling of the northern France population, marked by unemployment
Resection for advanced cancer of thoracic esophagus
787
Table VII. Late survival Anastomosis Stages
Cervical
Thoracic
TI NO MO (I) T3 NO MO (lIa) T4 NO MO (III)
2 (66 rna; 52 rna) 2 (53 rna; 72 rna)
I (70 rna) 2 (50 rna; 78 rna) I (81 rna)
Table VIII. Patterns of recurrence Anastomosis Causes of the death
Cervical
Thoracic
Metastasis Metastasis + mediastinal recurrence Mediastinal recurrence Cervical recurrence
6 I
0
9
5
5
3
0
and alcohol and tobacco abuse (Table 11). Such a mortality makes for consideration of some palliative treatment other than esophagectomy, and it alters the oncologic aspects of long-term results.V 5. The clinical severity of anastomotic leaks was the same in cervical and thoracic anastomoses (two each), but subclinical leaks were more frequent in the neck (six versus one). This observation confirms the lower frequency of leakage in thoracic anastomosis, but it shows that cervicalleaks can lead to fatal mediastinal infections. Respiratory complications were significantly more common in patients having a cervical anastomosis (x 2 = 6.69; P = 0.01) and probably were due to a more extended dissection of the trachea and to recurrent laryngeal nerve trauma with swallowing difficulties. 6. Survival was not significantly influenced by age (p = 0.96), but in our experience age became a pejorative factor at 75 years. Survival was decreased, but not significantly, by a loss of weight of more than 10% (p = 0.04). It was insignificantly improved in patients in whom the tumor was at a low level (p = 0.15) and a healthy esophageal section was obtained (p = 0.13). It was not modified by the volume of the resected tumor in this series, in which 48 of 60 cancers were stage T3 or T4. Survival was ameliorated in NO cases, but insignificantly, for the same reason-38 cancers were N+. However, all long-term survivors had NO MO disease, and this confirms the observations made for esophageal cancer.P: 14 as well as for bronchial cancer, that N is more important to prognosis than T (hospital mortality having not modified the proportion of NO MO cases, i.e., 20 of 60 before and 17 of 51 after). Actuarial survival rate for patients with stage I and II
788
Ribet, Debrueres, Lecomte-Houcke
disease was impossible to consider in this series. For stage III it was 47.9% ± 10.1% at 1 year and 21.5% ± 9.51% at 3 years. For stage IV it was 35% ± 9% and 18% ± 8.5%. There was no significant difference between the two stages, despite the advantage noticed when there was no cervical lymph node involvement, a point that could be checked only in patients having a cervical anastomosis. Survival was not influenced by the type of operation. IS The patterns of recurrence were not modified by the condition of the esophageal section. However, two observations were made: There were cervical recurrences only in cases ofcervical anastomosis (due to surgical trauma ?), and the long-term survivors had a healthy esophageal section. 16, 17 It must be noted that seven of the eight long-term survivors had moderately differentiated or well-differentiated cancers even though there were only nine such tumors versus 51 poorly differentiated cancers.l'' Quality oflife was comfortable, with good swallowing, gain of weight, and return to normal activity (albeit temporary) for nine patients after cervical anastomosis and 13 patients after thoracic anastomosis. Last, this randomized trial illustrates that hospital mortality and morbidity should not be excluded when one is judging results of different types of treatments 19,20 but must be kept to a minimum to evaluate the operation.
Conclusions Cervical anastomosis after esophagectomy for cancer in the thorax allows resection of undetected supratumor mucosal lesions. Fewer microscopically invaded esophageal sections are obtained through this route and more, albeit not all, lymph nodes can be resected from the neck and upper part of the mediastinum. Mortality is the same as in thoracic anastomosis, but subclinical leaks, respiratory morbidity, and recurrent laryngeal nerve trauma are more prevalent. In a series composed of 2 stage I, 19 stage II, 9 stage III, and 30 stage IV tumors, survival was not influenced by the anastomotic site. Ivor Lewis partial esophagectomy combined with postoperative radiotherapy remains a good palliative treatment and can result in long remissions in patients with advanced tumors, situated below the convexity of the aortic arch, with diseased lymph nodes. 13, 21 Subtotal esophagectomy with cervical anastomosis is indicated for technical reasons for tumors growing above the dome of the aortic arch or for tumors at any level without invasion beyond the muscle coat, when abdominal and mediastinal lymph nodes are not invaded macroscopically or on frozen sections. This pro-
The Journal of Thoracic and Cardiovascular Surgery
cedure reduces the prevalence and consequences of a microscopic invasion of the supratumoral mucosa or submucosa and of the esophageal section. REFERENCES I. Reboud E, Pradoura JP, Giudicelli R, Fuentes P. Multicentricite du cancer de I'oesophage. Chirurgie 1983; 109: 41-6. 2. Skinner DB. En bloc resectionfor neoplasms of the esophagus and cardia. J THORAC CARDIOVASC SURG 1983; 85:59-71. 3. Akiyama H, Tsurumaru M, Kawamura T, Ono Y. Principlesof surgical treatment for carcinoma of the esophagus: analysis of lymph node involvement. Ann Surg 1981; 194:438-45. 4. Couraud L, Hefez A, Velly JF, Levy F, Pierchon MS. Contribution a l'etude des cancers oesophagiens multifocaux par impregnation au bleu de toluidine. Implications therapeutiques. A propos de 50 cas. Ann Chir 1985;39: 211-3. 5. Seitz JF, Monges G, Navarro P, Giovannimi M, Gauthier A. Depistage endoscopique des dysplasies et des cancers infra-cliniques de l'oesophage. Resultats d'une etude prospectiveavec colorationvitale par Ie bleu de toluidine chez 100 patients alcoolo-tabagiques. Gastroenterol Clin Bioi 1990;14:15-21. 6. Manolard AM, Tourneux J, Lechaux M, Marnay J, Mandard Jc. A study method of esophago-gastrectomy specimens in esophagealcancer. Acta Endoscop 1981;II :293-8. 7. Richelme H, Bourgeon A, Ferrari C, Lebreton E, Limouse B, Birtwhistle Y. Etude radiologique de I'oesophage cervico-thoracique. Ann Chir 1983;37:597-8. 8. Sin KF, Cheung HC, Wong J. Shrinkage of the esophagus after resection for carcinoma. Ann Surg 1986;203: 173-6. 9. Domergue J, Kasse A, Daures JP, et al. Survie along terme descancersde l'oesophageopere.Definition des parametres qui influencentla survie. Ann Chir 1988;42:402-8. 10. Perrachia A, Bardini R, Castoro C, et al. La lyrnphadenectomie dans le traitement du cancer de l'oesophage intrathoracique. Ann Chir 1990;44:9-17. 11. Lasser P, Elias D. Technique des curages ganglionnaires etendus pour cancers de l'oesophage. J Chir (Paris) 1989; 126:40-4. 12. Eeftinck Schattenbeck M, Obertop H, Mud HJ, EijkenboomWMH, Van AudelJG, Van Houten H. Survivalafter resection for carcinoma of the esophagus. Br J Surg 1987; 74:165-8. 13. King RM, Pairolero PC, Trastek VF, Payne WS, Bernatz PE. Ivor Lewisesophago-gastrectomy for carcinomaof the esophagus: early and late functional results. Ann Thorac Surg 1987;44: 119-22. 14. Ribet M, Sobecki L, Pruvot FR, Giard-Lefevre S. Etude anatomo-clinique et therapeutique des resections oesophagiennes pour cancer de l'etage thoracique. J Chir (Paris) 1988;125:401-7.
Volume 103 Number 4 April 1992
15. Chasseray YM, KiroffGK, Buard JL, Launois B. Cervical or thoracic anastomosis for esophagectomy for carcinoma. Surg Gynecol Obstet 1989;169:55-62. 16. Tam PC, Chung HC, Ma L, Siu KF, Wong J. Local recurrences after subtotal esophagectomy for squamous cell carcinoma. Ann Surg 1987;205:189-94. 17. Barbier PA, Luder PJ, Schiipfer G, Becker CD, Wagner HE. Quality of life and patterns of recurrence following trans-hiatal esophagectomy for cancer: results of a prospective follow-up in 50 patients. World J Surg 1988;12: 270-6. 18. Sugimachi K, Inokuchi K, Kuwano H, Kai H, Okamura T.
Resection for advanced cancer of thoracic esophagus
789
Patterns of recurrence after curative resection for carcinoma of the thoracic part of the esophagus. Surg Gynecol Obstet 1983;157:537-40. 19. Mitchell RL. Abdominal and right thoracotomy approach as standard procedure for esophagogastrectomy with low morbidity. J THORAC CARDIOVASC SURG 1987;93:205-11. 20. Mansour KA, Downey RS. Esophageal carcinoma: surgery without preoperative adjuvant chemotherapy. Ann Thorac Surg 1989;48:201-5. 21. Couraud L, Yelly JF, Clerc P, Martigne C. Experience of partial esophagectomy for cancer. From a follow-up of 366 cases. Eur J Cardiothorac Surg 1989;3:99-104.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OFTHORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1992 issues from the Publisher, at a cost of $71.00 for domestic, $93.97 for Canadian, and $89.00 for international subscribers for Vol. 103 (JanuaryJune) and Vol. 104 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact MosbyYear Book, Inc., Subscription Services, 11830 Westline Industrial Drive, S1. Louis, Missouri 63146-3318, USA; phone (800) 325-4177, ext. 4351, or (314) 453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.
M. Ribet, MD,a
B. Debrueres, MD,a and M. Lecomte-Houcke, MD,b Lille, France
h e delay of symptoms, the multifocal sites of malignant disease, the long-running submucosal lymphatic drainage, and the disordered pattern of lymphatic spread account for the poor prognosis of thoracic esophageal cancer. Is it beneficial to lengthen the esophageal resection with a cervical anastomosis in all cases? To consider this question, we conducted a prospective study of partial esophagectomy with esophagogastric anastomosis in the thorax and of subtotal esophagectomy with anastomosis in the neck.
Patients and methods Between December 1983 and May 1986, we performed at random a cervical or a thoracic anastomosis on 60 patients. From Hopital Calmette, University Hospital, Lille," and the Department of Pathology, Lille Medical School," Lille, France. Received for publication Aug. 21, 1990. Accepted for publication Feb. 7, 1991. Addressfor reprints:M. Ribet, Hopital Calmette,F 59037 LilleCedex, France.
12/1/29024
784
There were 55 men and five women, aged from 34 to 70 years (mean 54.9), and the anastomosis was performed after an esophagectomy for cancer of the thoracic esophagus. The upper radiologic limit of the stricture was above the aortic arch in three patients, between the convexity of the arch and the inferior pulmonary vein in 40 patients, and under the hilum of the lung in 17 patients (Table I). In all cases there was a healthy-looking radiologic margin of esophagus below the clavicular level.* The preoperative work-up consisted of a clinical examination, a barium meal, and an examination of the pharynx, the larynx, the trachea, the bronchi, and the esophagus with both a flexible and a rigid endoscope, performed with the aid of general anesthesia, with biopsies of the tumor and of those areas of mucosa above the tumor marked by vital staining with toluidine blue. A chest film and computed tomographic scan, abdominal echography, respiratory and cardiac tests, and a biologic assessment completed the evaluation (Table II). Only patients having invasion of the respiratory mucosa or distant metastasis were eliminated from the trial. Endoscopic signs of tracheobronchial 'This study wasstarted beforethe creationof the LilleUniversity HospitalCommitteeof Ethics,whenpatients'informedconsentwasstill notnecessary. It wasjudgedacceptable, becausebeforethat date the techniquehad dependedon the preference ofthe particularsurgeon, with no objective evaluation.
Volume 103 Number 4 April 1992
Resection for advanced cancer of thoracic esophagus
Table I. Sites of upper limits of resected cancers Highest level of cancer
Above aortic arch Between aortic arch convexity and inferior pulmonaryvein Under hilum of lung
Anastomosis Cervical
Table II. Preoperative clinical evaluation of the patients Anastomosis
Thoracic
1 23
2
Patient evaluation
17
6
II
Mean age (yr) Mean lossof weight (%) Respiratory disease Cardiovascular disease FEV J <25% normal Pa02 at rest <70 mm Hg Liver steatosis Liver cirrhosis Renal insufficiency
compressions (n = 20), as well as radiologic evidence of tumoral contacts with adjacent structures (n = 39), were not considered to be contraindications. According to these conditions, 60 consecutive patients were included in the study. No preoperative adjuvant treatment was administered. Patients were selected at random for either cervical or thoracic anastomosis by the anesthetist, who drew a ticket specifying the type of anastomosis from a pool of 60 tickets before the procedure was begun. Two patients were excluded from the study because the intended procedure, a thoracic anastomosis, was impossible to perform. The two tickets labeled "thoracic anastomosis" were returned to the ticket pool. The surgical technique was identical for the two groups: I. Median celiotomy and preparation of a gastric tube vascularized by the right gastric and right gastroepiploic vessels after resection of the upper half of the lesser curvature and transparietal finger dilatation of the pyloric ring; resection of the lymphatic nodes situated on the left gastric, splenic, and hepatic chains and around the celiac trunk; creation of a feeding jejunostomy 2. Right posterolateral thoracotomy and esophageal dissection extending up to provide a 5 em margin of stretched esophagus above the cancer in the case of thoracic anastomosis and to the base of the neck in the case of cervical anastomosis; resection of the posterior mediastinal connective tissues; lymphadenectomy of the parahiatal, paraesophageal, subcarinal, and right paratracheal nodes; in case of incomplete tumoral resection, clipping of invaded zones; transhiatal pull-up of the gastric tube into the posterior mediastinum, to the level of the esophageal section in the case of thoracic anastomosis and to the base of the neck in the case of cervical anastomosis 3. For cervical anastomosis, right oblique cervical incision, esophageal freeing, and section at 2 to 3 cm under the pharynx; unilateral resection of lymph nodes of the recurrent laryngeal, subclavicular, jugular, and subdigastric chains 4. Esogastric terminolateral hand suture in two layers; introduction of a nasogastric tube Resection was considered to be complete in 36 patients; two wedge resections of the liver and two wedge resections of the lung were added to the esophagectomy for solitary metastases, which had not been detected before the operation; one lobectomy was done for a direct tumoral invasion of the lung and inferior pulmonary vein. Induction of anesthesia and performance of the operation took a mean of 6 hours 45 minutes for cervical anastomoses and 6 hours 15 minutes for thoracic anastomoses The esophageal resection specimens were immediately opened and spread out. The length of the shrunken esophagus,
785
Cervical
Thoracic
53.7 -7.7 22 10 4 14 3 2 0
56.1 -7.1 25 13 4 12 3 I I
FEV,. Forced expiratory volume in 1 second; Pa02. arterial oxygen tension.
Table III. Measurements and aspects of resected esophagus and tumors Anastomosis Macroscopy
Cervical
Thoracic
Mean resectedesophagus (em) Extremes (em) Mean supratumoral margin (em) (on shrunken esophagus) Mean tumor length (em) Obstructive and ulcerated Ulcerated (perforated: 2) Infiltrating Obstructive Stemmed
14.6 (10.5-18) 5.01
11.9 (8-15) 2.83
5.08
17 7 (I) 5 1 0
5.3 12 3 (I) 10 3 2
the size of the tumor, and the distances between the tumor, the cardia, and the upper section were measured; the whole mucosa was stained with toluidine blue dye to prepare 15 to 60 sections (mean 45); all resected lymphatic nodes were prepared for sections. A map of the esophagus and nodes was made. This pathologic work-up took 10 hours. The patients were kept in the intensive care unit for 24 to 72 hours and were supported with a ventilator. A feeding jejunostomy was used on the second or third day. The anastomosis was opacified on the sixth day, and the nasogastric tube was then removed. In case of a leak, the tube was left in place until opacification showed that the anastomosis had sealed off. The mean hospital stay, including the preoperative period, was 24.2 days for patients with cervical and 16.6 days for those with thoracic anastomoses. Radiotherapy was started between 5 and 6 weeks after the operation (cervical, n = 23; thoracic, n = 27). Esophageal dilations were necessary for four patients having a cervical anastomosis (one after a clinical leak, one after a subclinical leak, and two with no history of leak) and for one patient having a thoracic anastomosis (after a subclinical leak). The results of the two types of anastomosis were studied by the X2 test, modified by the Yates coefficient for small numbers; to study survival, we used the log-rank test and the KaplanMeier actuarial method (R. Beuscart, MD, Department of Statistics, Lille University Hospital).
786
The Journal 'of Thoracic and Cardiovascular Surgery
Ribet, Debrueres, Lecomte-Houcke
Table VI. Postoperative staging of the tumors (Unione Internazionale Contro if Cancro 1987)
Table IV. Microscopic study Anastomosis Microscopy Poorly differentiated tumor Moderately differentiated tumor Well-differentiated tumor Juxtatumoral dysplasia Distant dysplasia Submucosal infiltration Plurifocal cancers Invasion of superior section (Mucosal infiltration) (Submucosal infiltration) (In situ cancer) (Lymphatic embolus)
Cervical
25
26
I 4
4
o
26
24 4 6 4 10 (4) (0) (6) (0)
5
7 2 3 (I) (I) (0) (I)
Table V. Lymph node invasion Anastomosis Invaded lymph nodes Abdominal (celiac) Mediastinal
Cervical Jumping metastasis Mediastinal or abdominal
Cervical
Cervical
Anastomosis
Thoracic
Thoracic
13
15
(8)
(II)
17 7
7 0
4
5
2
0
Difference in mediastinal N + concerning the upper area of the mediastinum.
Results The data on the specimens are summarized in Table
III. Two cancers had perforated the mediastinum, with an abscess in one patient. Tumoral invasion of adjacent structures, mentioned in the operative record for 24 patients, was confirmed in 23 patients. The tumor had invaded several organs in one patient and affected the membranous surface of the trachea or bronchi, or both (9), the aorta (4), the thoracic duct (9), the azygos vein (2), the auricular opening of a pulmonary vein (2), the left pulmonary artery (1), the left recurrent laryngeal nerve (1), the pericardium or the posterior parietal pleura (8), and the median surface of the lung (3). The microscopic structure of the tumor and the appearance of the neighboring or distal mucosa and of the upper esophageal section are detailed in Table IV. The results of abdominal, mediastinal, and cervicallymphadenectomies are given in Table V. The resected cancers are staged in Table VI. Nine patients died, four of an anastomotic leak (two cervical, two thoracic), one of a perforated ulcer in the gastric transplant (thoracic anastomosis), two of respira-
Staging
I IIa lIb III IV
Cervical
Thoracic
I 10
4
1 8 I 5
15
15
0
There were 23 T4 (13 CA, IOTA), 19 celiac N+, and 7 cervical N+ cancers, and 4 solitary metastases were classified as M l.
tory insufficiency after a previous pulmonary lobectomy (one cervical, one thoracic), and two of inhalation pneumonitis (two cervical). Morbidity consisted of seven anastomotic leaks in the neck, six of which were subclinical and sealed up rapidly, and one subclinical leak in the thorax. There were 21 episodes of bronchopulmonary infection after cervical anastomosis and II after thoracic anastomosis. There were three temporary and three permanent cases of vocal paralysis after cervical anastomosis and one after thoracic anastomosis. Survival data, hospital mortality included, were as follows: 27 patients were alive at I year (45%),19 at 2 years (31.6%), and 18 at 3 years (30%). The median survival was 9 months after cervical anastomosis and 12 months after thoracic anastomosis. With a minimum follow-upof 4lf2 years and a maximum of 7 years, eight patients are still alive (13.3%), equally divided between the cervical and the thoracic anastomosis groups; to the time of writing, fivehave survived more than fiveyears and three more than 6 years (Table VII). Table VIII summarizes the patterns of recurrence that are known for 29 patients; two patients died of another cancer, one died of a noncancerous disease, and II of an unknown cause considered as cancerous. Discussion Attempts to devise more radical operations for esophageal cancer have followed three directions: subtotal esophagectomy in all patients,' en bloc resection.? and more complete lymphadenectomy.' This is a prospective study of the first procedure. In the following paragraphs we compare the pathologic and clinical aspects of the two groups for which there was no significant preoperative difference. I. Considering the spread of malignant disease through the mucosa, preoperative stains and biopsy specimens of the supratumoral mucosa's 5 were deceiving:four zones that appeared to be malignant corresponded to one plurifocal cancer, one mucosal malignant infiltration, one
Volume 103 Number 4 April 1992
submucosal infiltration, and one area of severe dysplasia. However, serial sections of the specimens confirmed a greater frequency of plurifocal cancers (6/60), of submucosal infiltrations (13/60), and of distant areas of dysplasia (9/60). The discovery of this lesion is correlated with the meticulousness of the pathologist," It confirms the shortcomings of the preoperative endoscopic findings and the theoretic oncologic importance of the supratumoral esophagus concerning the extent of resection. For that matter, cervical anastomosis has allowed an average gain of 2.18 em in length on the shrunken esophagus,' with 27 sections devoid of any microscopic malignancy; there were only 20 disease-free segments in the thoracic anastomosis group, despite the fact that there were fewer lower tumor sites in the group having cervical anastomosis. The frequency of this invasion is noticeable, because in three cancers only the upper limit was above the aortic arch (one cervical and two thoracic anastomoses), with I and 2 em margins of resection on the shrunken specimens'' for the two thoracic anastomoses; eight invasions involved sections for tumors whose upper limits were under the convex portion of the aortic arch. 2. The length and width of the tumors did not differ between the two groups of patients. We went as far as possible in considering indications for operation: Examination with a rigid tracheobronchoscope, which remains for us one of the most valuable preoperative investigations, showed bulging of the posterior tracheobronchial wall in 20 patients. (There was actual invasion in 9.) Although the surgeon was satisfied with 36 resections, only 12 tumors were stage Tl or T2. 9 3. To determine nodal status (N) we studied 877 nodes, an average of 14.6 nodes for each patient; 179 nodes were invaded, and in 38 patients they were classified N +. Cervical anastomosis showed the prevalence of diseased cervical lymph nodes (7/30); it also improved the results of the superior mediastinal lymphadenectomy (17/30 N + versus 7/30 in the thoracic anastomosis group), but on both sites the operation was unilateral. In the abdomen the figures were logically similar in the two groups.l'v!' 4. Hospital mortality also was similar in the two groups. It was nil for the first 26 patients and very heavy for the rest of the series. There is no excuse for this, but the question arises regarding the negative aspect of operating under the constraint of randomization: One of the deaths in the group undergoing thoracic anastomosis was related to our attempting a difficult, highly situated anastomosis. The mortality in this series (15%) is roughly twice the standard mortality observed in our unit (8%), and we admit to having the poorest sampling of the northern France population, marked by unemployment
Resection for advanced cancer of thoracic esophagus
787
Table VII. Late survival Anastomosis Stages
Cervical
Thoracic
TI NO MO (I) T3 NO MO (lIa) T4 NO MO (III)
2 (66 rna; 52 rna) 2 (53 rna; 72 rna)
I (70 rna) 2 (50 rna; 78 rna) I (81 rna)
Table VIII. Patterns of recurrence Anastomosis Causes of the death
Cervical
Thoracic
Metastasis Metastasis + mediastinal recurrence Mediastinal recurrence Cervical recurrence
6 I
0
9
5
5
3
0
and alcohol and tobacco abuse (Table 11). Such a mortality makes for consideration of some palliative treatment other than esophagectomy, and it alters the oncologic aspects of long-term results.V 5. The clinical severity of anastomotic leaks was the same in cervical and thoracic anastomoses (two each), but subclinical leaks were more frequent in the neck (six versus one). This observation confirms the lower frequency of leakage in thoracic anastomosis, but it shows that cervicalleaks can lead to fatal mediastinal infections. Respiratory complications were significantly more common in patients having a cervical anastomosis (x 2 = 6.69; P = 0.01) and probably were due to a more extended dissection of the trachea and to recurrent laryngeal nerve trauma with swallowing difficulties. 6. Survival was not significantly influenced by age (p = 0.96), but in our experience age became a pejorative factor at 75 years. Survival was decreased, but not significantly, by a loss of weight of more than 10% (p = 0.04). It was insignificantly improved in patients in whom the tumor was at a low level (p = 0.15) and a healthy esophageal section was obtained (p = 0.13). It was not modified by the volume of the resected tumor in this series, in which 48 of 60 cancers were stage T3 or T4. Survival was ameliorated in NO cases, but insignificantly, for the same reason-38 cancers were N+. However, all long-term survivors had NO MO disease, and this confirms the observations made for esophageal cancer.P: 14 as well as for bronchial cancer, that N is more important to prognosis than T (hospital mortality having not modified the proportion of NO MO cases, i.e., 20 of 60 before and 17 of 51 after). Actuarial survival rate for patients with stage I and II
788
Ribet, Debrueres, Lecomte-Houcke
disease was impossible to consider in this series. For stage III it was 47.9% ± 10.1% at 1 year and 21.5% ± 9.51% at 3 years. For stage IV it was 35% ± 9% and 18% ± 8.5%. There was no significant difference between the two stages, despite the advantage noticed when there was no cervical lymph node involvement, a point that could be checked only in patients having a cervical anastomosis. Survival was not influenced by the type of operation. IS The patterns of recurrence were not modified by the condition of the esophageal section. However, two observations were made: There were cervical recurrences only in cases ofcervical anastomosis (due to surgical trauma ?), and the long-term survivors had a healthy esophageal section. 16, 17 It must be noted that seven of the eight long-term survivors had moderately differentiated or well-differentiated cancers even though there were only nine such tumors versus 51 poorly differentiated cancers.l'' Quality oflife was comfortable, with good swallowing, gain of weight, and return to normal activity (albeit temporary) for nine patients after cervical anastomosis and 13 patients after thoracic anastomosis. Last, this randomized trial illustrates that hospital mortality and morbidity should not be excluded when one is judging results of different types of treatments 19,20 but must be kept to a minimum to evaluate the operation.
Conclusions Cervical anastomosis after esophagectomy for cancer in the thorax allows resection of undetected supratumor mucosal lesions. Fewer microscopically invaded esophageal sections are obtained through this route and more, albeit not all, lymph nodes can be resected from the neck and upper part of the mediastinum. Mortality is the same as in thoracic anastomosis, but subclinical leaks, respiratory morbidity, and recurrent laryngeal nerve trauma are more prevalent. In a series composed of 2 stage I, 19 stage II, 9 stage III, and 30 stage IV tumors, survival was not influenced by the anastomotic site. Ivor Lewis partial esophagectomy combined with postoperative radiotherapy remains a good palliative treatment and can result in long remissions in patients with advanced tumors, situated below the convexity of the aortic arch, with diseased lymph nodes. 13, 21 Subtotal esophagectomy with cervical anastomosis is indicated for technical reasons for tumors growing above the dome of the aortic arch or for tumors at any level without invasion beyond the muscle coat, when abdominal and mediastinal lymph nodes are not invaded macroscopically or on frozen sections. This pro-
The Journal of Thoracic and Cardiovascular Surgery
cedure reduces the prevalence and consequences of a microscopic invasion of the supratumoral mucosa or submucosa and of the esophageal section. REFERENCES I. Reboud E, Pradoura JP, Giudicelli R, Fuentes P. Multicentricite du cancer de I'oesophage. Chirurgie 1983; 109: 41-6. 2. Skinner DB. En bloc resectionfor neoplasms of the esophagus and cardia. J THORAC CARDIOVASC SURG 1983; 85:59-71. 3. Akiyama H, Tsurumaru M, Kawamura T, Ono Y. Principlesof surgical treatment for carcinoma of the esophagus: analysis of lymph node involvement. Ann Surg 1981; 194:438-45. 4. Couraud L, Hefez A, Velly JF, Levy F, Pierchon MS. Contribution a l'etude des cancers oesophagiens multifocaux par impregnation au bleu de toluidine. Implications therapeutiques. A propos de 50 cas. Ann Chir 1985;39: 211-3. 5. Seitz JF, Monges G, Navarro P, Giovannimi M, Gauthier A. Depistage endoscopique des dysplasies et des cancers infra-cliniques de l'oesophage. Resultats d'une etude prospectiveavec colorationvitale par Ie bleu de toluidine chez 100 patients alcoolo-tabagiques. Gastroenterol Clin Bioi 1990;14:15-21. 6. Manolard AM, Tourneux J, Lechaux M, Marnay J, Mandard Jc. A study method of esophago-gastrectomy specimens in esophagealcancer. Acta Endoscop 1981;II :293-8. 7. Richelme H, Bourgeon A, Ferrari C, Lebreton E, Limouse B, Birtwhistle Y. Etude radiologique de I'oesophage cervico-thoracique. Ann Chir 1983;37:597-8. 8. Sin KF, Cheung HC, Wong J. Shrinkage of the esophagus after resection for carcinoma. Ann Surg 1986;203: 173-6. 9. Domergue J, Kasse A, Daures JP, et al. Survie along terme descancersde l'oesophageopere.Definition des parametres qui influencentla survie. Ann Chir 1988;42:402-8. 10. Perrachia A, Bardini R, Castoro C, et al. La lyrnphadenectomie dans le traitement du cancer de l'oesophage intrathoracique. Ann Chir 1990;44:9-17. 11. Lasser P, Elias D. Technique des curages ganglionnaires etendus pour cancers de l'oesophage. J Chir (Paris) 1989; 126:40-4. 12. Eeftinck Schattenbeck M, Obertop H, Mud HJ, EijkenboomWMH, Van AudelJG, Van Houten H. Survivalafter resection for carcinoma of the esophagus. Br J Surg 1987; 74:165-8. 13. King RM, Pairolero PC, Trastek VF, Payne WS, Bernatz PE. Ivor Lewisesophago-gastrectomy for carcinomaof the esophagus: early and late functional results. Ann Thorac Surg 1987;44: 119-22. 14. Ribet M, Sobecki L, Pruvot FR, Giard-Lefevre S. Etude anatomo-clinique et therapeutique des resections oesophagiennes pour cancer de l'etage thoracique. J Chir (Paris) 1988;125:401-7.
Volume 103 Number 4 April 1992
15. Chasseray YM, KiroffGK, Buard JL, Launois B. Cervical or thoracic anastomosis for esophagectomy for carcinoma. Surg Gynecol Obstet 1989;169:55-62. 16. Tam PC, Chung HC, Ma L, Siu KF, Wong J. Local recurrences after subtotal esophagectomy for squamous cell carcinoma. Ann Surg 1987;205:189-94. 17. Barbier PA, Luder PJ, Schiipfer G, Becker CD, Wagner HE. Quality of life and patterns of recurrence following trans-hiatal esophagectomy for cancer: results of a prospective follow-up in 50 patients. World J Surg 1988;12: 270-6. 18. Sugimachi K, Inokuchi K, Kuwano H, Kai H, Okamura T.
Resection for advanced cancer of thoracic esophagus
789
Patterns of recurrence after curative resection for carcinoma of the thoracic part of the esophagus. Surg Gynecol Obstet 1983;157:537-40. 19. Mitchell RL. Abdominal and right thoracotomy approach as standard procedure for esophagogastrectomy with low morbidity. J THORAC CARDIOVASC SURG 1987;93:205-11. 20. Mansour KA, Downey RS. Esophageal carcinoma: surgery without preoperative adjuvant chemotherapy. Ann Thorac Surg 1989;48:201-5. 21. Couraud L, Yelly JF, Clerc P, Martigne C. Experience of partial esophagectomy for cancer. From a follow-up of 366 cases. Eur J Cardiothorac Surg 1989;3:99-104.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OFTHORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1992 issues from the Publisher, at a cost of $71.00 for domestic, $93.97 for Canadian, and $89.00 for international subscribers for Vol. 103 (JanuaryJune) and Vol. 104 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact MosbyYear Book, Inc., Subscription Services, 11830 Westline Industrial Drive, S1. Louis, Missouri 63146-3318, USA; phone (800) 325-4177, ext. 4351, or (314) 453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.