SURGICAL TREATMENT OF DYSPLASIA AND ADENOCARCINOMA

THE COLUMNAR-LINED ESOPHAGUS 0889-8553/97 $0.00 + .20

SURGICAL TREATMENT OF DYSPLASIA AND

ADENOCARCINOMA Tom R. DeMeester, MD

Patients with Barrett's esophagus are at an increased risk for the development of esophageal adenocarcinoma that is 30 to 125 times that of the normal population and, when expressed as adenocarcinomas per 100,000 Barrett's esophagus patients per year, is of the magnitude of 500 per 100,000. This risk is comparable to the risk of lung cancer in patients of a similar age and higher than the risk of 160 per 100,000, which has been reported for squamous cancer in China, where mass endoscopic surveillance programs have been instit~ted.~"

Consequently, it is currently recommended that patients with Barrett's esophagus who are medically fit be enrolled in a surveillance program and undergo 1 to 2 yearly endoscopy examinations with multiple biopsies. An acceptable protocol for these purposes requires obtaining four biopsy specimens, one from each quadrant of the esophagus, every 2 cm along the visible length of the Barrett's mucosa, with additional biopsy specimens from any abnormal- appearing area. Patients in surveillance programs are directed for further therapy if they develop low-grade or high-grade dysplasia or invasive adenocarcinoma.

BARRETT'S ESOPHAGUS WITH LOW-GRADE DYSPLASIA

Patients who are diagnosed as indefinite for dysplasia or as low-grade dysplasia should undergo a repeat endoscopic examination with meticulous examination of the Barrett's epithelium and documentation of the location of any areas of mucosal irregularity. Biopsy specimens are obtained along the length of the mucosa in the standard fashion, and the location of each biopsy specimen is recorded. After examination of the histologic specimens, the pres-

From the Department of Surgery, Division of Cardiovascular and Thoracic Surgery, Uni- versity of Southern California School of Medicine, Los Angeles, California

GASTROENTEROLOGY CLINICS OF NORTH AMERICA

VOLUME 26 - NUMBER 3 SEPTEMBER 1997 669

670 DeMEESTER

ence and extent of the dysplastic change can be documented. Because inflamma- tory atypia can be confused with low-grade dysplasia,32 patients who have not been receiving treatment for reflux disease should receive a 3-month course of intensive acid suppression, with omeprazole 20 to 40 mg daily, to reduce the active inflammation. After this course of treatment, the patient undergoes endoscopy again, and extensive biopsy specimens of the esophagus are obtained, paying particular attention to areas previously reported to show dysplastic change. If the low-grade dysplasia persists, the patient should undergo an antireflux procedure followed by endoscopic surveillance with multiple biopsy specimens every 6 months. In a few cases, the dysplastic changes regress, and surveillance can be performed less frequently.

BARRETT'S ESOPHAGUS WITH HIGH-GRADE DYSPLASIA

The optimal management of Barrett's esophagus patients found to have high-grade dysplasia is controversial. There are two schools of thought regarding this subject. The first proposes that high-grade dysplasia is a separate entity to invasive adenocarcinoma and that the condition does not invariably progress to adenocarcinoma.22, 33 It hypothesizes that high-grade dysplasia can be man- aged expectantly by close endoscopic surveillance, going to surgery only when adenocarcinoma is diagnosed. Surgical therapy is considered to have a high morbidity, and in a few isolated cases, mortality has occurred in patients who were free from invasive cancer in the resected ~pecimen.~, 28, 33 Fundamental to this hypothesis is the ability to distinguish high-grade dysplasia from early adenocarcinoma by endoscopic biopsy. Levine and colleagues22 reported good results with this policy. All patients underwent an aggressive biopsy protocol, however, and up to 12 biopsy specimens per centimeter of Barrett's esophagus were taken with large-size biopsy forceps rather than the traditionally accepted four quadrant biopsy specimens per 2 cm. Whether such a protocol is appropriate for clinical practice has been questioned,'", 41 and the authors have indicated that their protocol was not intended for routine Cameron and co- workersy undertook a prospective study of 23 patients with high-grade dysplasia who all underwent thorough preoperative biopsy evaluation using standard- sized biopsy forceps. Fifteen patients were resected, of whom only one had adenocarcinoma, and one patient with only dysplasia died in the postoperative period. Such observations favor a conservative approach to patients with high- grade dysplasia but may be suspect because of lack of data regarding the natural history of high-grade dysplasia.

Experienced surgeons have been struck by the difficulty differentiating high-grade dysplasia from adenocarcinoma on the basis of endoscopic biopsy examination. Studies by Pera and colleagues,Z8 Altorki and co-workers,5 Rice and co-~orkers,3~ and Streitz and associates4* indicated that patients who underwent esophagectomy for high-grade dysplasia had invasive adenocarcinoma up to 50% of the time. None of these reports documented the extent of the preoperative biopsy specimens and have come under criticism.

The author and colleagues reviewed their experience with 23 patients who were referred from endoscopic surveillance programs, of whom 15 after repeat endoscopy and multiple biopsy specimens of the Barrett's segment were operated with a diagnosis of high-grade dysplasia.*y Despite careful1 and systematic preoperative biopsies, done by the author and colleagues, 10 (66%) had unex- pected invasive adenocarcinoma in the resected specimen. Patients with mucosal

SURGICAL TREATMENT OF DYSPLASIA AND ADENOCARCINOMA 671

abnormalities were more likely to have adenocarcinoma, but a normal-appearing Barrett’s mucosa did not indicate the absence of underlying adenocarcinoma.

Other concerns with continued surveillance for high-grade dysplasia are that the majority of patients progress to adenocarcinoma within a relatively short period of time, thereby negating the rationale for continuing surveillance. Hameeteman and associate^'^ in a prospective study showed five patients with Barrett’s adenocarcinoma who all progressed from low-grade to high-grade dysplasia before the development of cancer. The time interval from low-grade dysplasia to carcinoma was only 1.5 to 4 years. Robertson and co-w~rkers~~ showed that in two patients who developed adenocarcinoma in a setting of high-grade dysplasia the time interval from high-grade dysplasia to adenocarcinoma was only 11 to 14 months. The author considers high-grade dysplasia to be a marker for early esophageal adenocarcinoma and recommends that patients who are medically fit undergo esophagectomy with anticipation of permanent cure of the disease. Continued surveillance is reserved for older patients and those with coexisting medical conditions in whom resection would be under- taken only if adenocarcinoma develops.

In the author’s experienceF9 endoscopic surveillance for patients with Barrett’s esophagus followed by early surgical resection for high-grade dysplasia or intramucosal adenocarcinoma is an effective management policy because such patients have an improved survival compared to patients with newly recognized Barrett’s carcinomas who were not under surveillance. Figure 1 shows the significant survival advantage that was obtained in those patients in the surveyed group compared to nonsurveyed patients.

BARRETT’S ESOPHAGUS WITH ADENOCARCINOMA

Esophageal adenocarcinoma is a highly malignant tumor that carries a poor prognosis. An encouraging observation is that patients with Barrett’s adenocarci-

100

80

20

0 0 10 20 30 40 50 60

Months Surveyed 1 7 9 5 4 2 1

35 2 2 9 3 2 1 I Non Surveyed

Figure 1. Survival after esophageal resection in endoscopically surveyed and nonsurveyed patients. There was a significant difference between the groups (Log rank X2 = 5.8, R 0 . 0 5 ) . Bars indicate standard errors of the mean. Surveyed group includes 13 patients with adenocarcinoma and 4 with high-grade dysplasia. The number of patients followed each year are shown in the key box.

672 DeMEESTER

nomas, who are discovered on surveillance at an earlier stage, have an improved survival. Esophageal adenocarcinoma tends to spread early to the regional lymph nodes. Up to 30% of patients with intramucosal tumors have one to four lymph node metastases; when the tumor extends into the submucosa, the prevalence of node metastasis increases to 60%; and by the time the tumor has penetrated through the esophageal wall, almost 90% have metastases to the regional nodes. Surgical therapy of esophageal adenocarcinoma requires identi- fication of the purpose of the operation (i.e., cure of the disease or palliation of symptoms). Inclusion of patients who were operated for the purpose of palliation in the reporting of 5-year survival data encourages a pessimistic approach to the disease. All factors considered the anatomic extent of tumor at the time of diagnosis and the residual tumor status after resection (i.e., no residual tumor [&I, microscopic residual tumor [R,], and macroscopic residual tumor [RJ, are the strongest predictors of survival).l8

Selection of patients appropriate for curative therapy requires that they be physiologically fit, that is younger than 75 years of age with an ejection fraction of greater than 40% and a forced expiratory volume greater than 1.25 L. Patients can be staged preoperatively on the basis of endoscopy and endoscopic ultrasound to determine the depth of tumor invasion and to identify potentially involved regional nodes. Metastatic disease is excluded by abdominal and thoracic computed tomography scanning. In the future, biologic staging of the tumor may be helpful in identifying patients with a potential for long-term survival.

EXTENT OF RESECTION TO CURE DISEASE CONFINED TO THE MUCOSA

Controversy exists over how to manage tumors confined to the mucosa, that is, limited by the muscularis mucosa. Some authors have endoscopically resected squamous carcinomas after using endoscopic ultrasound to determine that the depth of the tumor was limited to the muc~sa . '~ Surprisingly, large areas of squamous mucosa can be resected without perforation or bleeding, leaving the smooth surface of the muscularis mucosa intact. Reepithelization of the large, artificially induced ulcer is usually complete in 3 weeks. So as not to miss a squamous cancer that has invaded deeper than expected, it is important to examine carefully the deep margins of the resected specimen and perform periodic endoscopic follow-up examinations with vital staining techniques. Mul- tiple and widespread or circumferential squamous lesions are not applicable to this technique for fear of developing a stricture on healing. In this situation, those acquainted with endoscopic resection advocate an esophagectomy.

Although most experience with endoscopic mucosal resections has been with high-grade squamous cell dysplasia, a similar approach (i.e., endoscopic ablation) is being advocated to remove metaplastic areas of Barrett's mucosa.8 In this situation, not only must the abnormal mucosa be removed, but also reflux must be prevented to allow rehealing of the esophagus with squamous epithelium. Investigators have also applied mucosal ablation technology to areas of high-grade dysplasia and intramucosal carcinoma in Barrett's mucosa.15, 27! 37

There are several factors that must be considered in determining whether this method is appropriate for high-grade dysplasia or intramucosal cancer.

Several studies have shown that high-grade dysplasia and intramucosal tumors (i.e., invasive cancer limited by the muscular mucosa) are quite different in their biologic behavior from more advanced submucosal Vessel


invasion and lymph node metastasis do not occur in severe dysplasia, are uncommon in the intramucosal tumors, but are the rule in submucosal tumors. As a consequence, the 5-year survival for severe dysplasia and intramucosal tumors differs significantly from that for submucosal tumor~.2~, 29 A critical issue to be resolved is whether an intramucosal tumor can be correctly discriminated from a submucosal tumor before surgery. Accuracy of determining the depth of tumors confined to the esophageal wall using endoscopic ultrasound is question- able. In the author's experience, the differentiation of intramucosal tumors from submucosal tumors by endoscopic ultrasound is poor. Only one in five intramucosal tumors can be correctly classified, although all can be correctly classified as being limited to the esophageal wall. The resolution of the present- day endoscopic ultrasonographic systems may not be sufficient to differentiate predictably the fine detail of tumor infiltration when it is limited to the esophageal wall.30 Some have suggested that the endoscopic appearance of a mucosa lesion can be helpful in determining its depth. Obvious elevated or depressed lesions or ones that appear nodular usually have invaded down into the submucosa or deeper. Although these efforts are noble, it must be concluded that at present there is no dependable way of determining whether a tumor extends beyond the muscularis mucosa before its removal. This difficulty becomes even more understandable if it is realized that the distance from the epithelium to the muscularis is only 2 to 3 mm.

Of further concern is that up to 30% of patients with intramucosal tumors have lymph node metastases, although the number of involved nodes per patient is usually five or fewer." Akiyama and have reported that even though the number of involved nodes may be small, they can spread to distant nodal regions including cervical and abdominal nodes. Of concern is that this jump metastasis can occur in up to 10% of patients (i.e., nodes adjacent to the tumor can be free of disease, while one or two distant nodes are involved).

Even though the number of involved nodes in patients with intramural tumors may be small and the pattern of nodal metastasis is unpredictable, there appears to be some relationship between the location of the primary tumor in the esophagus and the anatomic site of lymph node metastasis. Upper esophageal cancers commonly involve cervical and mediastinal nodes, lower esophageal cancers, mediastinal and abdominal nodes, and middle esophageal cancers commonly involve cervical, mediastinal, and abdominal nodes. Consequently, if an extensive lymphadenectomy is done for intramucosal tumors, it should include the neck and mediastinum for upper esophageal cancers and the mediastinum and abdomen for lower esophageal cancers. Cancers of the middle esophagus need the most extensive lymphadenectomy (i.e., of the cervical, mediastinum, and abdominal lymph nodes).

It appears that intramucosal cancer (i.e., tumors limited by the muscularis mucosa) may not be completely resected or ablated by endoscopic surgery, particularly if there is unrecognized penetration of the tumor into the submucosa. Further, there is a small but consistent prevalence of lymph node metastasis in these patients. At present, it is undecided whether a transhiatal esophagectomy without lymphadenectomy or an en bloc esophagectomy with systematic lymphadenectomy should be done for these patients. Most surgeons agree that when a tumor penetrates the submucosa, tumor cells can spread intramurally several centimeters from the primary site and metastasize to the regional lymph nodes. In view of the difficulty of differentiating between intramucosal and submucosal tumors and that these are the most curable tumors, it is the author's policy that patients suspected of having an intramucosal tumor should have a total removal of the thoracic esophagus with a systematic lymph node dissec-

674 DeMEESTER

tion.3, 14, 16, 39 Gastrointestinal continuity is reestablished by pulling the stomach up into the neck and performing an esophagogastrostomy. This opinion is not universally accepted, and many surgeons perform a transhiatal esophagectomy without a systematic lymphadenectomy.

EXTENTOFRESECTIONTOCUREDISEASEBEYONDTHE MUCOSA

Some have advocated that the current treatment strategies for carcinoma of the esophagus that extends beyond the mucosa (i.e., the muscularis mucosa) limit the role of surgery to removing the primary tumor, with the belief that this is sufficient for early tumors and the hope that adjuvant therapy will increase cure rates of more advanced tumors by destroying systemic and unre- sected regional disease. This approach emphasizes the concept of biologic deter- minism, that is, that the outcome of treatment in esophageal cancer is deter- mined at the time of diagnosis and that surgical therapy aimed at removing more than the primary tumor is not helpful. This approach considers lymph node metastases, regardless of the number involved, as simply markers of systemic disease and that the systematic removal of involved regional lymph nodes is not beneficial. Based on this concept is the belief that the removal of the primary tumor by transhiatal esophagectomy results in the same survival as a more extensive en bloc resection.

The concept of an extended en bloc resection for carcinoma of the esophagus began with a report by Logan in 1963.= He was the first surgeon to apply the traditional concepts of surgical oncology to esophageal cancer. Logan described an en bloc resection for carcinoma arising in the distal one third of the esophagus and the gastric cardia. The 5-year survival rate of 16% was better than any of the other reports up to that time, but the 21% mortality rate associated with the procedure hindered widespread acceptance of the operation. Skinner3y and Akiyama and colleagues3 subsequently modified the procedure, reduced the mortality to 11% and 5.2%, and increased the 5-year survival rates to 18% and 42%. Subsequent modifications regarding the selection of patients and the extent of gastric and esophageal resection have led to further reductions in mortality rates, while preserving the improved s~rvival.~,

Performing an adequate en bloc resection requires that three incisions be made in the following order: (1) right posterolateral thoracotomy for en bloc dissection of the esophagus and mediastinum, closure of the thoracotomy, repo- sitioning of the patient in the recumbent position; (2) upper abdominal midline incision for en bloc dissection of the stomach; and (3) left neck incision for mobilization and division of the cervical esophagus. Some surgeons add an inverted-T sternotomy above the articulation of the third rib to dissect out the superior mediastinal nodes along the left recurrent laryngeal nerve. The dis- sected specimen is removed transhiatally from the thorax. For tumors of the mid-third and upper third of the thoracic esophagus, gastrointestinal continuity is reestablished usmg a gastric pull-up with a cervical esophagogastric anastomosis. For visible tumors of the lower third of the esophagus and cardia, the author adds a proximal gastric resection to obtain adequate distance below the tumor to accommodate submucosal spread of the tumor within the margins of the resection. Gastrointestinal continuity is reestablished using an interposed colon based on the left colic artery and inferior mesenteric vein and placed isoperistaltically between the cervical esophagus and distal stomach.I6

The concept of a transhiatal esophagectomy was introduced in 1933 when


Turne? first successfully used this approach in a patient with esophageal cancer. This technique was reintroduced by Akiyama' in 1981 and was popular- ized in the United States by Orringer26 after his report of a large series of patients in 1984. As it is applied today, this procedure involves separation of the esophagus from the adjacent mediastinal structures through an abdominal and left neck incision, largely by blunt dissection. Gastrointestinal continuity is reestablished by bringing the stomach up through the posterior mediastinal tunnel with an anastomosis to the cervical esophagus in the neck.

In the transhiatal procedure, there is no specific attempt made to remove lymph node-bearing tissue in the posterior mediastinum. In contrast, the en bloc esophagectomy removes the tumor covered on all surfaces with a layer of normal tissue, and long lengths of foregut above and below the lesion are resected to incorporate submucosal spread of the tumor. To be consistent, this means that for patients with a tumor in the lower third of the thoracic esophagus or cardia, the proximal two thirds of the stomach should also be resected. Appropriate cervical, mediastinal, and abdominal lymph node dissections are included using an en bloc technique to remove potentially involved regional lymph nodes. Arguments to support the more extensive esophagectomy, gastrectomy, and lymph node dissection are listed in Table 1.

The author advocates a selective approach to patients with cancer of the esophagus in regard to the extent of the resection. Deciding between a curative en bloc esophagectomy or a palliative, transhiatal esophagectomy is based on the location of the tumor, age of the patient, the patient's physiologic fitness, the extent of disease on endoscopic ultrasound, and intraoperative staging (Fig. 2).13 En bloc esophagectomy is an option for patients who are physically fit and whose tumor is located below the carina and clinically assumed to be limited to the esophageal wall or to have fewer than five nodes involved. If patients are physically unfit, a palliative transhiatal esophagectomy is done. Intraoperative staging is done because inaccuracies with existing clinical staging methodology still exist. It is based on the observation that patients with a tumor that penetrates through the esophageal wall and metastasizes to five or more regional lymph nodes have advanced disease with poor prognosis, and an extended resection is not appropriate. It requires a surgical approach that allows intraoperative switching from a curative en bloc to a palliative resection. A curative en bloc dissection is abandoned and a palliative transhiatal esophagogastrectomy is done if intraoperative findings reveal cavitary spread of the tumor, extension of the tumor through the mediastinal pleura, multiple gross lymph node metastases, or microscopic evidence of lymph node involvement at the margins of an en bloc resection (i.e., low paratracheal, porta hepatis, subpancreatic, or periaor- tic lymph nodes).

To investigate the relationship between the type of resection performed and survival in patients with esophageal cancer, the author and colleagues reviewed their experience with en bloc and transhiatal esophagectomy for carcinoma arising in the esophagus below the level of the carina, including the gastric cardia in the patients whose tumor is clinically assumed to be limited to the wall or to have fewer than five nodes involved.'6 Overall, survival was significantly better in patients having en bloc resection (41%) than in patients having transhiatal resections (14%) ( W.001). Using the WNM system of postoperative histologic staging, patients were grouped into those with limited disease, defined as intramural tumors, that is, confined to the wall of the esophagus and associated with fewer than five involved nodes. Survival was significantly better for patients with limited disease after en bloc resection, compared to transhiatal resection (75% versus 20%, W.01). When all patients with intramural tumors,

676 DeMEESTER

Table 1. RATIONALE FOR EN BLOC ESOPHAGECTOMY

Reasons for a More Extensive Esophagectomy Injection of submucosal contrast medium shows that the length of longitudinal lymph

At least 10 cm of grossly normal esophagus proximal to the tumor must be resected to

Spatial relation indicates that for an adequate proximal margin a cervical anastomosis

flow is 6 x the transverse flow

prevent local recurrence

is almost always needed Reasons for a More Extensive Gastrectomy for Tumors of the Lower Third of the Esophagus or Cardia

No barrier to submucosal lymphatics between esophagus and stomach at the cardia Tumor cells in submucosal lymphatics can result in intragastric recurrence if too little of

the stomach is resected Spatial relationships of the stomach do not allow for both adequate distal tumor

margins and sufficient residual stomach to perform a cervical anastomosis

Survival of lung cancer patients with metastases to the hilar lymph nodes, i.e., a cancer that also metastasizes to mediastinal lymph nodes, is dependent on removal of involved nodes

Patients with esophageal carcinoma and lymph node metastases are cured by resection. It is extremely rare for patients with lymph node metastases to be cured without surgical removal

Patients with esophageal and cardia cancer, similar to those with head and neck cancer, can die from lymph node metastasis alone

Asian surgeons, who are incessant data keepers, accept unconditionally the benefit of lymph node dissection on survival in patients with carcinoma of the esophagus or stomach

43% of patients with esophageal carcinoma who have histologically node-negative disease have histochemical node-positive disease. Further, after a median observation time of 12 months, patients with histochemical node-positive disease had a significantly shorter disease-free survival. On the basis of this finding, it is believed that when nodes are reported to be histologically free of tumor, more disease than what is currently appreciated is removed or left behind, depending on the extent of resection

Reasons for a Lymph Node Dissection

regardless of lymph node status, were evaluated, survival was significantly better after en bloc resection (Pc.05). This was so even though the incidence of fewer than five lymph nodes involved was the same in both groups. Conversely, when all patients with involvement of fewer than five lymph nodes, regardless of the depth of tumor penetration, were evaluated, survival was significantly better after en bloc resection (R.005). This was so even though the incidence of transmural tumors penetration was the same in both groups. It appears on the basis of these studies that en bloc esophagectomy is most beneficial to patients with tumors limited to the esophageal wall or with fewer than five lymph nodes involved.

There are four possible explanations why transhiatal esophagectomy or a standard transthoracic esophagectomy fails to achieve the results obtained with en bloc esophagectomy in patients with tumors in the distal esophagus or cardia and disease that extends beyond the mucosa: (1) the potential dissemination of tumor cells during the dissection of the thoracic esophagus, (2) an inadequate distal tumor margin in an effort to preserve a lengthy stomach with sufficient blood supply to perform a cervical anastomosis; (3) the transfer of unrecognized perigastric metastatic nodes into the thorax with the gastric pull-up, (4) the probability that the limited or blunt thoracic dissection leaves residual nodal


- Four or Less Lymph Nodes Involved

Age 4 5 , FEV rl .25

Intraoperative Palliative Transhiatal Esophagogastrectomy

Assumed, Limited Disease

Curative En Bloc

Figure 2. Selection criteria for en bloc or transhiatal esophagectomy.

disease in the mediastinum. This may particularly be so in patients who are histologically free of lymphatic metastasis but on histochemical studies have evidence of microlymphatic metastasis. An en bloc esophagectomy effectively eliminates all of these potential causes of recurrence.

As the understanding of the pathology of esophageal cancer improves and experience with resection increases, evidence is accumulating to indicate that for patients with an intramural tumor in the distal third of the esophagus or cardia, the best chance for cure is an en bloc esophagectomy and proximal gastrectomy with gastrointestinal continuity reestablished with a colon interposition. For patients with a similar tumor in the middle or upper third of the esophagus, the best chance for cure is an en bloc esophagectomy and a cervical lymph node dissection with gastrointestinal cohtinuity reestablished with a gastric pull-up. Table 2 summarizes the extent of resection for tumors extending various depths into the esophageal wall.

ROLE OF ADJUVANT THERAPY

The proposal to use adjuvant chemotherapy in the treatment of esophageal cancer evolved when it became evident that most patients develop postoperative systemic metastasis without local recurrence. This observation lead to the hypothesis that undetected systemic micrometastases were present at the time of diagnosis and if effective systemic therapy were added to local regional therapy, survival should improve.

678 DeMEESTER

Table 2. RECOMMENDED SURGICAL THERAPY FOR ESOPHAGEAL CARCINOMA

Lesion Resection

Small confined areas of high-grade dysplasia (intraepidermal lesions)

Widespread or circumferential area high-grade dysplasia (intraepidermal lesions)

T'umor invading through the basement membrane but not through the muscularis mucosa (intramucosal cancers)

Tumor deeper than muscularis mucosa but not through the esophageal wall (intramural cancers)

Tumor extending through the muscularis propria (transmural cancers)

Endoscopic mucosal resection or

Esophagectomy without thoracotomy ablation

Esophagectomy without thoracotomy or en bloc resection with node dissection. Reconstruction with gastric pull-up

En bloc esophagectomy with systematic lymphadenectomy of the cervical, mediastinal, and abdominal nodes. Superior mediastinal dissection must include the node along the left recurrent nerve. (For lower third esophageal and cardia cancers, omit cervical superior mediastinal node dissection but include proximal stomach and spleen in the resection. For upper and middle third esophageal cancers, omit splenic artery dissection and preserve the spleen.) Reconstruction with gastric pull-up for upper and middle third tension and colon interposition for lower third and cardia lesions

Same as for intramural tumors unless 2 5 lymph nodes are assumed to be involved, in which case a palliative transhiatal esophagectomy is done

This hypothesis has been supported by the observation of epithelial tumor cells in the bone marrow in 37% of patients with esophageal cancer who were resected for cure. These patients had a greater prevalence of relapse at 9 months after surgery compared to those patients without such ~ells.4~ Such studies emphasize that hematogenous dissemination of viable malignant cells occurs early in the disease and that systemic chemotherapy may be helpful if the cells are sensitive to the agent. If the cells are resistant, systemic Chemotherapy may be a hindrance because of its immunosuppressive properties. Unfortunately, current technology is not able to test tumor cell sensitivity to chemotherapeutic drugs. Consequently, the choice of drugs rest solely on their clinical effectiveness against grossly similar tumors.

The decision to encourage preoperative rather than postoperative chemotherapy was based on the perceived ineffectiveness of chemotherapeutic agents when used after surgery and animal studies suggesting that agents given before surgery were more effective. The claim that patients who received chemotherapy before resection were less likely to develop resistance to the drugs was only opinion and unsupported by data. Similarly the claim that drug delivery is enhanced because blood flow is more robust before the patients undergo surgical dissection is also flawed in that if enough blood reaches the operative site to heal the wound or anastomosis, the flow should be sufficient to deliver


chemotherapeutic drugs. An agreed-on benefit that can be attributed to preoperative chemotherapy in esophageal carcinoma is its ability, if effective, to facilitate surgical resection, particularly squamous cell tumors above the level of the carina. Reducing the size of the tumor may provide safer margin between the tumor and the trachea or larynx and allow an anastomosis to a tumor-free cervical esophagus just below the cricopharyngeus. An involved margin at this level usually requires a laryngectomy to prevent subsequent local recurrence.

Preoperative chemoradiotherapy using the drug combinations of platinum with 5-fluorouracil has been reported by several investigators to be beneficial in both adenocarcinoma and squamous cell esophageal carcinoma but, particularly for squamous cell carcinoma, by increasing the number of complete pathologic responses of the primary tumor before resection. There have been only five randomized prospective studies using this strategy, three with squamous cell carcinoma, one with both squamous cell carcinoma and adenocarcinoma, and one with only adenocarcinoma (Table 3).6* 20, 25' 45. 46 All but one have shown no survival benefit with preoperative chemoradiotherapy over surgery alone. Most authors report substantial morbidity and mortality to the treatment. Despite this report, many have been encouraged by the observation that some patients who have had a complete response have remained free of recurrence at 3 years.

represents the one positive outcome and deserves special comment. Several concerns have arisen about this trial. First, before the results are accepted, it must be kept in mind that three other studies have not shown any benefit to this approach, and the report of this study is only an interim analysis at 3 years. Things could change with further follow-up. If one more death occurs in the multimodal group, the P value would go from .01 to .03 and with two deaths to greater than .05.7 Second, the number of early-stage tumors was low in the surgery group and hence a worse survival than what is generally reported after resection. In contrast, nothing is known about the initial stage of the multimodal therapy group. The number in each arm is small, and there could be a stage bias. Third, there is no clear account of the kind of surgical resection performed (i.e., &, R,, or RZ). Fourth, withdrawals from the protocol may have resulted in a selection that favored the outcome of multimodal therapy in that 10 patients were withdrawn from that arm compared to only 1 from the surgical arm. Five of the 10 withdrawals from the multimodal therapy arm completed a full course of chemotherapy. Fifth, 51 patients (45%) of the original 113 evaluated were excluded from randomization, which questions if the studied population is an accurate representation of the disease as seen in the clinics.

The study reported by Walsh and

Table 3. ESOPHAGEAL CARCINOMA: RANDOMIZED PREOPERATIVE CHEMO- RADIOTHERAPY VERSUS SURGERY ALONE

Survival Author Year n=CIS Cell Type Regimen' cvs

NygaardZ5 1992 47/41 Squamous P, 8, 35 Gy NS Le PriseZo 1994 41/45 Squamous P, 5-FU, 20 Gy NS Apinop6 1994 35/34 Squamous P, 5-FU, 40 Gy NS Urba45 1995 50/50 Squamous + adenocarcinoma P, 5-FU. V, 45 Gy N S W a l ~ h ~ ~ 1996 48/54 Adenocarcinoma P, 5FU, 40 Gy P= .01

*Regimen: P = cisplatin; 5-FU = 5-fluorouracil; B = bleomycin; V = vinblastine. C = Preoperative chemotherapy; S = Surgery only; NS = not significant.

680 DeMEESTER

Table 4. RESULTS OF NEOADJUVANT THERAPY IN ADENOCARCINOMA OF THE ESOPHAGUS

No. of CR Institution Year Patients Regimen' (W Survival

M. D. Anderson 1990 35 P,E,5-FU 3 42% at 3 y SLMC 1992 18 P,5-FU,RT 17 40% at 3 y Vanderbilt 1993 39 P,E,S-FU,RT 19 47%at 4 y Michigan 1993 21 P,VBL,5-FU,RT 24 34% at 5 y MGH 1994 16 P,S-FU 0 42% at 4 y MGH 1994 22 EAP 5 58% at 2 y

"Regimen: P = cisplatin; E = etoposide; 5-FU = 5-fluorouracil; RT = radiation therapy; VBL = vinblastine; A = Adriamycin (doxorubicin).

CR = Pathologic complete response; SLMC = St. Louis University Medical Center; MGH = Massachusetts General Hospital.

From Wright CD, Mathisen DJ, Wain JC, et al: Evolution of treatment strategies for adenocarcinoma of the esophagus and gastroesophageal junction. Ann Thorac Surg 58:i 574-1 579, 1994; with permis- sion.

Caution must be exercised in emphasizing the effects of chemotherapy with a multimodal approach in that the addition of preoperative radiation therapy to chemotherapy elevates the complete response rate and inflates the benefit attributed to chemotherapy. With chemoradiation, the complete response rates for adenocarcinoma range from 17% to 24%. When radiation was removed, the complete response fell to 0% to 5%, which suggests that the effects of chemotherapy are negligible (Table 4).47 If radiotherapy is the factor responsible for improved response rate, surgery alone could do the job as well because numerous studies in the past have shown that the combination of surgery and radiation does not provide any beneficial survival ad~antage. '~

Most medical and radiation oncologists, dismayed by the high local recurrence and distant failure rates after transhiatal or standard transthoracic esophagectomy, have called into question the relevance of surgery in the treatment of esophageal cancer. In their minds, it has become a medical disease despite sound evidence to the contrary. Surgeons who perform a more extended en bloc resection and lymphadenectomy for cure of appropriately staged disease have shown gratifyingly low local recurrence rates in long-term survivors."

The ultimate question is should a patient with carcinoma of the esophagus go through two to three cycles of chemotherapy on the 5% chance that they may get a complete response in the primary tumor and little evidence that a lesser response will control systemic disease. Prudence encourages going directly to surgery and avoiding the morbidity associated with preoperative chemotherapy, particularly chemoradiation therapy. Some studies have shown that the rates of infection, anastomotic breakdown, incidence of adult respiratory distress syndrome, and long-term use of a respirator were greater in patients receiving adjuvant therapy as compared with surgery alone.38

Most treatment failures in patients with adjuvant therapy are due to distant disease. This fact underscores the negative answer to the original hypothesis supporting its use and emphasizes that systemic therapy requires further im- provement before it can be unconditionally recommended to patients. Rather than calling into question the relevance of surgery, oncologists need to answer the question, "Is chemotherapy with the newer agents, given at the time of systemic recurrence after resection, effective in prolonging survival?"31 Perhaps

Barium S

wal

low

, End

osco

py

Clin

ical

Eva

luat

ion

Figu

re 3

. Su

gges

ted

glob

al a

lgo-

rit

hm fo

r th

e m

anag

emen

t of

car-

ci

nom

a of

the

eso

phag

us. (

Mod

i- fie

d fro

m D

eMee

ster

TR

, Kim

mey

M

B,

Koz

arek

R

A,

Levi

n B,

Sp

echl

er S

, Ty

tgat

GN

J: G

uide

- lin

es

in

the

Nin

ety-

fifth

Ann

ual

Am

eric

an G

astr

oent

erol

ogic

al A

s-

soci

atio

n’s

Clin

ical

Pra

ctic

e Se

c-

tion

sym

posi

um

on e

soph

agea

l ca

ncer

, N

ew O

rlean

s, M

ay 15-

18, 1

994;

cour

tesy

of T

om R

. De-

Mee

ster

, MD

.) B

Tum

or S

tagi

ng

(CT

ches

t and

abd

omen

, en

dosc

opic

ultr

ason

ogra

phy)

[Adv

ance

d D

iseas

e I

Thro

ugh

the

wall

and

mul

tiple

ly

mph

node

met

asta

sis

/I

\

I La

te D

isea

se o

r p

zz

q

Si

ifica

nt C

omor

bidi

t Tu

mor

susp

ecte

d no

t to

be th

roug

h the

wal

l an

d/or

less

than

5 ly

mph

nod

es in

volv

ed

- n

Distant

orga

n met

asta

sis

I Tr

ansh

iata

l eso

phag

ecto

my w

ith

the

poss

ibili

ty o

f pre

oper

ativ

e ch

emor

adia

tion

if un

der t

he a

ge of

75

- Im

min

ent c

ardi

ac p

ulm

onar

y or

he

patic

failu

re

Cur

ativ

e Sur

gery

En

blo

c Resection

I C

hem

orad

iatio

n

Trea

tmen

t Failure o

r

Seve

re de

bilit

y Lo

cal r

ecur

renc

e Iln

rese

ctab

le pr

oxim

al

Airw

ay fi

stula

or

1)is

tanl

met

asta

sis

Adv

ance

d di

seas

e Nu m

etasta

ses

or b

lccd

mg

tum

or

unre

scct

ablc

prim

ary

No

loca

l rec

urre

nce

I C

ompl

ete e

xcisi

on po

ssib

le

I tu

mor

or l

ocal

recurrence

1 1

1 1

Supp

ortiv

e Car

e Pa

lliat

ive S

urge

ry

Lase

r Abl

ativ

e The

rapy

st

at

Che

mot

hera

py

682 DeMEESTER

the newer agents are more effective at this stage if patients have not received chemotherapy before. Current data support using chemoradiotherapy preoperatively to reduce tumor size in a young person with squamous cell carcinoma above the carina and chemotherapy as salvage therapy for patients who have not had previous chemotherapy and develop recurrent systemic disease after surgical resection. A global management algorithm that incorporates this use of chemoradiotherapy is shown in Figure 3.

References

1. Akiyama H: Esophagectomy without thoracotomy. Ann Surg 13:109-121, 1981 2. Akiyama H, Tsurumaru M, Udagawa H, et al: Principles of surgical treatment for

carcinoma of the esophagus: Analysis of lymph node involvement: Ann Surg 194:438- 446, 1981

3. Akiyama H, Tsurumaru M, Udagawa H, et al: Radical lymph node dissection for cancer of the thoracic esophagus. Ann Surg 220364-373, 1994

4. Akiyama H, Tsurumaru M, Udagawa H, et a1 Strategy for resectable esophageal carcinoma. J Thorac Cardiovasc Surg (in press)

5. Altorki NK, Sanagawa M, Little AG, et al: High grade dysplasia in the columnar lined esophagus. Am J Surg 161:97-99, 1991

6. Apinop C, Puttisak P, Preecha N A prospective study of combined therapy in esophageal cancer. Hepatogastroenterology 41:391-393,1994

7. Badwe RA, Vaidya JS, Bhansali M: Multimodal therapy for esophageal adenocarcinoma. N Engl J Med 336:374-376, 1996

8. Berenson M, Johnson T, Markowitz NR, et al: Restoration of squamous mucosa after ablation of Barrett’s esophageal epithelium. Gastroenterology 1041686-1691, 1993

9. Cameron AJ, Carpenter HC, Trastek V F Barrett’s esophagus with high grade dysplasia: Is resection required? Gastroenterology 106:A375, 1994

10. Clark GWB, Ireland AP, DeMeester TR Dysplastic Barrett’s: Is continued surveillance appropriate? [letter]. Gastroenterology 106:1128-1129, 1994

11. Clark GWB, Peters JH, Ireland AP, et al: Nodal metastasis and sites of recurrence after en bloc esophagectomy for adenocarcinoma. Ann Thorac Surg 586464354, 1994

12. DeMeester TR, Barlow AP: Surgery and current management for cancer of the esophagus and cardia: Part 11. Curr Probl Surg 25:535-605, 1988

13. DeMeester TR, Zaninotto G, Johansson KE: Selective therapeutic approach to cancer of the lower esophagus and cardia. J Thorac Cardiovasc Surg 95:42-54, 1988

14. Fujita H, Kakegawa T, Yamana H, et a1 Mortality and morbidity rates, postoperative course, quality of life, and prognosis after extended radical lymphadenectomy for esophageal cancer: Comparison of three field lymphadenectomy with two field lymphadenectomy. Ann Surg 222654462, 1995 .

15. Gossner L, Sroka R, Stolte M, et al: Barrett’s esophagus with severe dysphagia and/ or superficial cancer: Internal results of photodynamic therapy using B-aminolevulinic acid. Gastroenterology 110A521, 1996

16. Hagen JA, Peters JH, DeMeester TR Superiority of extended en bloc esophagogastrectomy for carcinoma of the lower esophagus and cardia. J Thorac Cardiovasc Surg 106:850-859, 1993

17. Hameeteman W, Tytgat GNJ, Houthoff HJ, et a1 Barrett‘s esophagus: Development of dysplasia and adenocarcinoma. Gastroenterology 96:1249-1256,1989

18. Hermanek P: pTNM and residual tumor classification: Problems of assessment and prognostic significance. World J Surg 19:184-190, 1995

19. Inoue H, Endo M, Takeshita K, et al: Endoscopic resection of early-stage esophageal cancer. Surg Endosc 5:59-62, 1991

20. Le Prise E, Etienne P, Meunier 8, et al: A randomized study of chemotherapy radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus. Cancer 73:1779-1784, 1994


21. Lerut T, DeLeyn P, Coosemans W, et a1 Surgical strategies in esophageal carcinoma with emphasis on radical lymphadenectomy. Ann Surg 216583-590,1992

22. Levine DS, Haggitt RC, Blount PL, et a1 An endoscopic biopsy protocol can differentiate high grade dysplasia from early adenocarcinoma in Barrett’s esophagus. Gastroen- terology 105:40-50, 1993

23. Logan A: The surgical treatment of carcinoma of the esophagus and cardia. J Thorac Cardiovasc Surg 46:150-161, 1963

24. Nishimaki T, Tanaka 0, Suzuki T, et al: Tumor spread in superficial esophageal cancer: Histopathologic bases for rational surgical treatment. World J Surg 17766772, 1993

25. Nygaard K, Hagen S, Hansen HS, et a1 Pre-operative radiotherapy prolongs survival in operable esophageal carcinoma: A randomized multicenter study of pre-operative radiotherapy and chemotherapy. World J Surg 16:1104-1110, 1992

26. Orringer MB: Transhiatal esophagectomy without thoracotomy for carcinoma of the thoracic esophagus. Ann Surg 200:282-288,1984

27. Overholt B, Panjehpour M, Rose TE: Photodynamic therapy for treatment of early adenocarcinoma in Barrett’s esophagus. Gastrointest Endosc 39:73-76, 1993

28. Pera M, Trastek VF, Carpenter HA, et al: Barrett’s esophagus with high grade dysplasia an indication for esophagectomy? Ann Thorac Surg 54:199-204, 1992

29. Peters JH, Clark GWB, Ireland AP, et al: Outcome of adenocarcinoma arising in Barrett’s esophagus in endoscopically surveyed and nonsurveyed patients. J Thorac Cardiovasc Surg 108:813-822, 1994

30. Peters JH, Hoeft SF, Heimbucher J, et al: Selection of patients for curative or palliative resection of esophageal cancer based on preoperative endoscopic ultrasonography. Arch Surg 129:534-539, 1994

31. Raoul JL, LePrise E, Meunier B, et al: Combined radiochemotherapy for postoperative recurrence of esophageal cancer. Gut 37174-176,1995

32. Reid BJ, Haggitt RC, Rubin CE, et a1 Observer variation in the diagnosis of dysplasia in Barrett’s esophagus. Hum Pathol 19166178,1988

33. Reid BJ, Weinstein WM, Lewin KJ, et a1 Endoscopic biopsy can detect high grade dysplasia or early adenocarcinoma in Barrett’s esophagus without grossly recognizable neoplastic lesions. Gastroenterology 9481-90, 1988

34. Rice TW, Falk GW, Achkar E, et al: Surgical management of high grade dysplasia in Barrett’s esophagus. Am J Gastroenterol 88:1832-1836, 1993

35. Robertson CS, Mayberry JF, Nicholson DA, et al: Value of endoscopic surveillance in the detection of neoplastic change in Barrett’s oesophagus. Br J Surg 75:760-763, 1988

36. Sabile J, Rice TW, Goldblum JR, et al: Superficial esophageal carcinoma. Ann Thorac Surg 602396-902, 1995

37. Sampliner RE, Jaffe P: Malignant degeneration of Barrett’s esophagus: The role of laser ablation and photodynamic therapy. Dis Esoph 8104-108, 1995

38. Schlag PM, for the Chirurgische Arbeitsgemeinschaft fuer Onkologie der Deutschen Gesellschaft fuer Chirurgie Study Group: Randomized trial of preoperative chemotherapy for squamous cell cancer of the esophagus. Arch Surg 1271446-1450,1992

39. Skinner DB En bloc resection for neoplasms of the esophagus and cardia. J Thorac Cardiovasc Surg 85:59-71, 1983

40. Spechler SJ: Endoscopic surveillance for patients with Barrett’s esophagus: Does the cancer risk justify the practice? Ann Intern Med 106902-904, 1987

41. Spechler SJ, Goyal R K Cancer surveillance in Barrett’s esophagus: What is the end point? [letter]. Gastroenterology 106275-277, 1994

42. Streitz JM, Andrews CW, Ellis F H Endoscopic surveillance of Barrett’s esophagus: Does it help? J Thorac Cardiovasc Surg 105:383-388, 1993

43. Thorban S, Rodeu JO, Nekarda H, et a1 Disseminated epithelial tumor cells in bone marrow of patients with esophageal cancer: Detection and prognostic significance. World J Surg 20567-573,1996

44. Turner G G Excision of the thoracic oesophagus for carcinoma, with construction of an extra-thoracic gullet. Lancet 2:1315, 1933

45. Urba S, Orringer M, Turrisi A, et al: A randomized trial comparing transhiatal esophagectomy to preoperative concurrent chemoradiation followed by esophagectomy in

684 DeMEESTER

locoregional esophageal carcinoma. Proceedings of American Society of Clinical Oncol- ogy Annual Meeting, Los Angeles, 1995

46. Walsh TN, Noonan N, HolIywood D, et aI: A comparison of muhimodal therapy and surgery for esophageal adenocarcinoma. N Engl J Med 335:462467,1996

47. Wright CD, Mathisen DJ, Wain JC, et al: Evolution of treatment strategies for adenocarcinoma of the esophagus and gastroesophageal junction. Ann Thorac Surg 58:1574- 1579, 1994

Address reprint requests to Tom R. DeMeester, MD

University of Southern California School of Medicine Department of Surgery, Division of Cardiovascular and Thoracic Surgery

1510 San Pablo Street, Suite 514 Los Angeles, CA 90033-4612

Documents

SURGICAL TREATMENT OF DYSPLASIA AND ADENOCARCINOMA