Neoplasms of the oesophagus

Benign tumours

Benign tumours of the oesophagus are rare. The commonest is a leiomyoma (Fig. 50.45). Even a large leiomyoma may produce surprisingly mild symptoms. It is usually possible to enucleate the tumour at thoracotomy without breaching the mucosa. Small polyps of the oesophagus, such as granular cell tumours and fibrovascular polyps, may be found in the course of endoscopy for other conditions.

 Malignant tumours

Sarcoma

Sarcoma of the oesophagus is exceedingly rare, but leiomyosarcoma (Fig. 50.46) and rhabdomyosarcomas have been reported. Other types of sarcoma have been reported, but are confined to individual case reports.

Malignant melanoma

Malignant melanoma of the oesophagus is rare. It may represent secondary spread, but primary melanoma of the oesophagus does occur. It has a very poor prognosis.

Carcinoma of the oesophagus

Cancer of the oesophagus is the ninth most common cancer in the world. It is in general a disease of mid to late adult­hood with a poor survival rate. Only 5—10 per cent of those diagnosed will survive for 5 years. Despite this gloomy prognosis there are some encouraging signs of change.

Pathology and aetiology. Squamous cell cancer (Figs 50.47 and 50.48) and adenocarcinoma (Figs 50.49 and 50.50) are the commonest types. Squamous cell carcinoma usually affects the upper two-thirds of the oesophagus and adeno­carcinoma affects the lower one-third, but there are frequent exceptions to this rule. Oat cell cancer occurs occasionally. World-wide, it is squamous cell cancer that is commonest, but adenocarcinoma is the commonest type in most westernised countries and is increasing in incidence.

Geographical variation of oesophageal cancer

  The incidence of oesophageal cancer varies more than any other cancer and its epidemiology provides a fascinating story. Squamous cell cancer is endemic in the Transkei region of South Africa and in the Asian ‘cancer belt’ that extends across the middle of Asia from the shores of the Caspian Sea in northern Iran to China. The highest incidence in the world is in Linxian in Henan province in China where it is the commonest single cause of death with more than 100 cases per 100 000 population per annum. The cause of the disease in the endemic areas is not definitely known, but is probably due to a combination of fungal contamination of food with the production of a carcinogenic mycotoxin, together with nutritional deficiencies in the population. In Linxian, supplementation of the diet with beta-carotene, vitamin E and selenium has been shown to reduce the incidence of cancer.

Away from the endemic areas tobacco and alcohol are the major factors in the occurrence of squamous cancer. Incidence rates vary from less that five per 100 000 in whites in the USA to 26.5 per 100 000 in some regions of France. Contrary to popular opinion Japan does not have a particularly high incidence of oesophageal cancer.

The incidence of adenocarcinoma of the oesophagus is increasing

In many westernised countries the incidence of squamous cell cancer has fallen or remained relatively static, but the incidence of adenocarcinoma of the oesophagus has increased since the mid-1970s by 5—10 per cent per annum. The change is greater than that of any other neoplasm in this time. Adenocarcinoma now accounts for 60—75 per cent of all oesophageal cancers in several countries. The reason for the change is not understood, but is thought to be due to a changed incidence of Barrett’s oesophagus. There has been a similar increase in the incidence of carcinoma of the cardia of the stomach. This epidemiological change suggests that cancer of the cardia and adenocarcinoma of the oesophagus are, in fact, the same disease. The incidence of cancer of the gastric antrum has decreased markedly over the same period and now more than 60 per cent of all upper GI cancers involve the cardia or distal oesophagus. Survival rates remain poor although in several regions of the UK there has been a modest improvement in 5-year survival from 5 per cent to 10 per cent.

Involvement of coeliac axis nodes is a bad prognostic sign

The poor prognosis of oesophageal cancer is proof of its ability to spread. This may be locoregional or systemic. Locoregional spread occurs through the wall of the oesophagus into adjacent structures, along the length of the oesophagus in the submucosal lymphatics, and to regional lymph nodes. Locoregional spread is often discontinuous, i.e. distant regional lymph nodes may be invaded even when local nodes are free of tumour, and there may be satellite nodules in the oesophagus proximal to the main tumour. Spread to the coeliac axis nodes from a lesion in the intra­thoracic oesophagus is a bad prognostic sign and is regarded as metastatic (M) rather than nodal (N) disease in the TNM classification. Systemic spread is mainly to the liver and lungs, but practically any organ can be involved including brain, bone and skin.

Early endoscopic diagnosis is the key to good results

Clinical features. Dysphagia is the usual presenting feature and is generally a sign of advanced disease. Weight loss is likewise a sign of advanced disease. An increasing number of cancers is diagnosed at a relatively early stage when the chances of cure are greater. Patients with early disease may present with rather nonspecific dyspeptic symptoms or a vague feeling of ‘something that is not quite right’ during swallowing. Some are diagnosed during screening of Barrett’s oesophagus, but it has to be confessed that the results of formal screening are still disappointing. The widespread use of endoscopy as a diagnostic tool is the major contributor to early diagnosis (Fig. 50.51). It should be emphasised that biopsies should be taken of all lesions (Fig. 50.52) no matter how trivial they appear. Small cancers are curable. Large ones usually are not. Some benign lesions can look surprisingly ‘malignant’ (Fig. 50.53).

Hoarseness due to recurrent laryngeal nerve palsy is a sign of advanced and incurable disease. Palpable lymphadenopathy in the neck is likewise a sign of advanced disease.

Staging and general assessment. Methodical assessment of the patient and the cancer is essential if the best results are to be achieved. Some patients are quite obviously not candidates for radical treatment even on the most superficial inspection and common humanity must always take precedence. Staging must not delay effective palliation.

The most important aspect of staging is a careful search for metastatic disease. Ultrasonography of the liver and comput­erised tomography (CT) scanning of the chest and abdomen is mandatory before resection. The purpose of CT scanning is to exclude metastatic disease from the lungs and liver. Despite considerable improvements in CT and magnetic resonance imaging these methods are still inaccurate for staging the primary lesion and for staging lymph nodes. Endoscopic ultrasonography, if available, is the best method for preoperative staging of oesophageal cancer. Bronchoscopy should be done in lesions of the upper or middle thirds where there is potential for tracheobronchial invasion (Fig. 50.54). Laparoscopy is useful for assessing adenocarcinoma of the distal oesophagus (Fig. 50.55), particularly if it is likely to extend below the phreno-oesophageal ligament. Laparoscopy is at its best for detecting transperitoneal spread and liver metastasis. It is an inexpensive technology that deserves to be widely used. At the cost of some time and effort it is also possible to sample lymph nodes.

Short-term nutritional support has doubtful benefits

  It is also important to assess the fitness of the patient to withstand major surgery. Respiratory and cardiovascular function are the most important aspects of the assessment. Nutritional assessment is also an important part of general care, but the value of preoperative nutritional support remains somewhat controversial. It is not possible to post­pone treatment for several weeks to allow full nutritional ‘resuscitation’ and short periods of nutritional support are of questionable benefit for reducing postoperative morbidity and mortality. If food intake is impaired the simplest means of providing adequate nutrition is by withdrawing all solid food and starting the patient on a high-protein liquid diet. It is surprising how much severe dysphagia improves when obstructing food material in the oesophagus has been cleared.

Treatment of malignant tumours

Principles

A gastrostomy should never be carried out as the ‘palliation’ for oesophageal cancer. Palliation in this disease demands relief of dysphagia. It is important that staging and general assessment are carried out speedily and humanely. Those who have incurable disease should not be submitted to needlessly aggressive treatment that simply prolongs the process of dying. Palliation of dysphagia can be achieved in a number of ways that do not unduly stress the patient. Pain is also a surprisingly common feature of advanced oesophageal cancer and requires careful attention. 

   Curative treatment involves radical surgery or radiotherapy. This should be carried out in a specialist centre with the necessary expertise. The role of palliative resection is debatable. It is wise not to embark on resection if it is clear beforehand that palliation is all that can be achieved. However, palliative resection may be appropriate if incurable disease is found when an operation is already well under way.

Surgical resection probably gives the best results for all forms of oesophageal cancer. Radical radiotherapy can cure both types of cancer, but poses technical problems at the lower end of the oesophagus. Tumours that involve the stomach are not generally accepted for radiotherapy. The results of radiotherapy have been improved by concurrent chemotherapy, so-called chemoradiotherapy. Some have suggested that this development may challenge the dominant place of surgery for cure.

Surgery

Beware of satellite nodules proximal to the primary lesion

Curative surgery involves resection of an appropriate length of the oesophagus together with any involved stomach and the locoregional lymphatics. There is controversy about the length of oesophagus that should be resected. Some surgeons feel that operative trauma should be reduced to the minimum by resecting only enough oesophagus to clear the tumour. Others advocate subtotal oesophagectomy pioneered in the UK by McKeown on the grounds that generous proximal clearance gives the best chance of clearing satellite nodules in the submucosal lymphatics and gives the best postoperative function with the least tendency to gastro-oesophageal reflux. Restoration of continuity is almost always achieved by transposition of the stomach and oesophagogastric anastomosis. Colon or, less commonly, small intestine may be inter­posed between the oesophageal remnant and the stomach, but is a more major undertaking with a higher postoperative mortality. There is also controversy about the extent of lymph node dissection that should be done. Akiyama has pioneered the concept of three field lymph node dissection, an ultraradical operation involving extensive removal of the regional lymph nodes in the abdomen, chest and neck.

Lesions of the cardia that do not involve the oesophagus to any significant extent may be dealt with by extended total gastrectomy to include the distal oesophagus or by proximal gastrectomy and distal oesophagectomy.

Subtotal oesophagectomy. A variety of approaches is pos­sible. The most commonly used is the Ivor Lewis (or. Lewis—Tanner) approach. The stomach is first mobilised through a midline incision and then the oesophagus is approached from a right thoracotomy (Fig. 50.56). This approach is preferred by most surgeons because the access for resection and anastomosis is not hampered by the aortic arch as it is through a left thoracotomy. In the classic McKeown operation a third incision is made in the neck. However, in practice it is possible to remove just as much oesophagus by dividing the oesophagus in the chest at the thoracic inlet. A neck incision is required if a lymph node dissection is to be done or if there are technical difficulties with an anastomosis at the thoracic inlet. It is often stated that an oesophagogastric anastomosis is safest when done in the neck. In fact this only applies to an anastomosis that lies in the anterior part of the neck as for reconstruction with stomach or colon that has been brought up retrosternally or subcutaneously. The usual oesophagogastric anastomosis in the posterior neck drops back into its normal anatomical position and usually lies at the thoracic inlet. Any leak therefore has consequences that are just as serious as any other intrathoracic anastomosis.

 Carefully preserve the blood supply of the stomach — venous and arterial

Mobilisation of the stomach must be done with care as it is essential to have a tension-free, well-vascularised stomach for transposition. The left gastric artery, the short gastric vessels and the left gastroepiploic are all divided. The viability of the transposed stomach mainly depends on the right gastroepiploic artery and vein with a small contribution from the right gastric. It should be noted that venous drainage is as important as arterial supply and it is essential to perform an accurate anatomical dissection that preserves the right gastroepiploic vein as well as the artery. The stomach is divided to remove the cardia and the upper part of the lesser curve including the whole of the left gastric artery and its associated lymph nodes.

Right thoracic approach gives easy access to the oesophagus

The approach to the oesophagus through the right chest is relatively straightforward. The azygos vein is divided and this allows easy access to the whole of the intrathoracic oesophagus. A thoracotomy with entry above the fifth rib gives best access to the mid-mediastinum and the thoracic inlet. The oesophagus is divided just below the thoracic inlet. Since most lesions are in the lower third or middle third this usually gives adequate proximal clearance of at least 5 cm. If there is any doubt about clearance frozen sections should be taken from the resection margins and it may be helpful to open the oesophagus and assess whether the lesion is well circumscribed or whether it is diffusely infiltrating. Carci­nomas of the upper thoracic oesophagus are almost always incurable at the time of diagnosis and invasion of the trachea is common. If one of these lesions is resectable it is essential to use an incision in the neck and to resect more of the oesophagus than is customary in the operation of subtotal oesophagectomy.

Oesophagogastric anastomosis may be performed equally well by hand or stapler. Both methods require attention to detail. In experienced hands serious anastomotic leakage should be uncommon (significantly less than 5 per cent). Minor leakage detectable by contrast radiology is more common, but should not disturb the patient. The significance of these minor leaks is debated. Most surgeons still prefer to keep the patient nil by mouth for 5—7 days and then perform a contrast swallow If small leaks are to be detected it is essential to use barium for the examination. Water-soluble contrast media miss 50 per cent of anastomotic leaks. If leakage is detected the patient is kept nil by mouth until it has sealed.

Postoperative nutritional support remains controversial. There is general agreement that parenteral feeding is associated with more nosocomial infection, including pneumonia, than enteral feeding. It is also expensive. If nutritional support is given a feeding jejunostomy is probably the best method.

Transhiatal oesophagectomy (without thoracotomy). This approach has been popularised by Orringer in the USA and Pinotti in Brazil. The stomach is mobilised through a midline abdominal incision and the cervical oesophagus is mobilised through an incision in the neck. The diaphragm is then opened from the abdomen and the posterior mediastinum is entered. The lower oesophagus and the tumour are mobilised under direct vision and the upper oesophagus is mobilised by blunt dissection. This approach can provide an adequate removal of the tumour and lymph nodes in the lower mediastinum, but it is not possible to remove the nodes in the mid or upper mediastinum. It may be a useful procedure for lesions of the lower oesophagus, but may be hazardous for a middle third lesion that may be adherent to the bronchus or to the azygos vein.

Left thoracoabdominal approach. A long skin and muscle incision is made on the left side with entry into the chest above the seventh rib and removal of a short segment of costal cartilage (Fig. 50.56). The diaphragm is incised and the oesophagus and stomach are removed. Some surgeons advocate a left thoracoabdominal approach together with an incision in the neck for subtotal oesophagectomy, the ‘Birmingham’ approach popularised by Matthews.

Thoracoscopic oesophagectomy. Oesophagectomy may be done by thoracoscopy or by the hybrid technique of video-assisted thoracic surgery (VATS) in which a combination of endoscopic and conventional instruments is used through small thoracic incisions. Thoracoscopic oesophagectomy in the prone position to minimise injury to the collapsed lung has been pioneered by Cuschieri. At present this is still an evolving technique and its place is not yet established. The procedure takes longer than open surgery and postoperative morbidity is still a problem.

Gastro-oesophageal reflux following oesophagogastric resection. Gastro-oesophageal reflux may be a major problem following any operation that involves resecting the cardia. Postoperative reflux may present with the typical symptoms of GORD or with a peptic stricture at the site of the ana­stomosis. However, the presentation may be different with a miserable patient who fails to thrive following the operation and who is then suspected of having recurrent cancer. This atypical presentation is particularly common following total gastrectomy with an inadequate reconstruction.

Reflux may be a problem following resection

Symptoms may be atypical

Reflux may be limited or avoided by:

1. subtotal oesophagectomy and gastric transposition high in the chest. The vertical stomach empties rapidly and functions as a barrier to reflux;

2. resection of a generous portion of proximal stomach if an anastomosis is made to the lower oesophagus. This reduces gastric secretion;

3.   Roux-en-Y reconstruction with a long ascending jejunal limb (5 0—60 cm);

4. interposition of jejunum or colon (Fig. 50.57)..

The old Braun or omega reconstruction with an entero­enteroanastomosis below an oesophagojejunostomy is men­tioned only to be condemned. The side-to-side anastomosis does not divert bile and pancreatic secretion satisfactorily from the oesophagus and several studies have shown that these patients fare badly.

Pyloroplasty. Pyloroplasty or pyloromyotomy is an option following oesophageal resection. It may avoid early problems with gastric emptying, but the stomach always seems to recover its function even without pyloroplasty. If delayed gastric emptying is a problem in the early postoperative period erythromycin, which is a motilin agonist, seems to be the best therapy.

Radiotherapy for cure

Radiotherapy may be a useful alternative to surgery, especially in unfit patients

Radical radiotherapy can produce long-term survival in oesophageal cancer (Fig. 50.58). Although traditionally used for squamous cell cancer radiotherapy may also be effective for adenocarcinomas. There has been no formal comparison of the results of radiotherapy and surgical resection and it is therefore impossible to make dogmatic statements about the relative merits of each form of treatment. At present surgical resection is generally regarded as producing the best survival rates and quality of life. Many surgical series have reported 5-year survival rates between 20 per cent and 35 per cent, with an average figure of 25 per cent. Survival following radio­therapy with a typical UK case mix is between 9 per cent and, exceptionally, 19 per cent. The average appears to be about 10 per cent. However, the technology of radiotherapy continues to improve and the advent of chemoradiotherapy is a significant advance, if at the cost of significant morbidity. The survival figures make it abundantly clear that neither surgery nor radiotherapy is a particularly effective treatment. There is therefore a real need for improvements in early diagnosis, at present the best hope of improving survival, and improve­ments in treatment with the use of multimodality treatment.

Chemotherapy

With the advent of regimens containing cis-platinum, chemo­therapy for oesophageal cancer has improved considerably. Chemotherapy never cures the disease, but can produce worthwhile shrinkage of disease in up to 60 per cent of cases. The best responses are seen in squamous cell cancers. Survival is extended modestly.

Multimodality treatment

Randomised prospective studies of preoperative and post­operative radiotherapy have not shown any improvement in survival. Thus far there is no evidence that perioperative chemotherapy improves survival, but the results of studies using modern combination chemotherapy are awaited. Significant improvement in survival will only be achieved by a treatment that has a powerful effect on systemic disease.

Palliative treatment

Surgical resection and external beam radiotherapy may be used for palliation, but are not suitable when the expected survival is short, as most of the remainder of life will be spent recovering from the ‘treatment’. Surgical bypass is likewise too major a procedure for use in a patient with a limited life expectancy. A wide variety of relatively simple methods of palliation is now available that will produce worthwhile relief of dysphagia with minimal disturbance to the patient.

Palliation should be simple and effective

Intubation has been used for many years following the invention of the Souttar tube made of coiled silver wire (Fig. 50.59). This was superseded by the Celestin tube whose design gave a better quality of swallowing and was safer to insert. The Celestin tube was originally designed to be inserted by oesophagoscopy, to place a plastic rod in the stomach, followed by laparotomy and gastrostomy, to retrieve the rod and pull the tube down into place. The development of methods of intubation that could be used with a flexible endoscope was a major advance pioneered by Atkinson of Nottingham. The Atkinson tube is still in use. It is made of silastic with a nylon spiral reinforcement and has a distal retaining flange to prevent proximal displacement. It is inserted over a guidewire with a specially designed introducer. There are now many other designs of semi rigid tubes for palliation including the Procter—Livingstone tube which is popular in South Africa where there is a very high incidence of oesophageal cancer.

The technology of intubation has now moved on with the invention of various types of expanding metal stent (Fig. 50.60). These are inserted under X-ray or endoscopic control. The stent is restrained in the collapsed state during insertion and then released when it is in the correct position. Expanding stents produce a better lumen for swallowing than rigid tubes, but are relatively expensive.

Endoscopic laser treatment may be used to core a channel through the tumour. It produces worthwhile improvement in swallowing, but has the disadvantage that it has to be repeated every few weeks. Lasers may also be used to un­block a stent that has been blocked by tumour overgrowth.

Brachytherapy is a method of delivering intraluminal radiation with a short penetration distance (hence the term brachy) to a tumour. An introduction system is inserted through the tumour and the treatment is then delivered in a single session lasting for 20 minutes or so. The equipment is expensive to purchase, but running costs are low

Other methods of palliation that can be given endoscop­ically include bipolar diathermy (the BICAP probe), argon beam plasma coagulation and alcohol injection.

Malignant tracheo-oesophageal fistula

Malignant tracheo-oesophageal fistula is a sign of incurable disease and life expectancy is short. Some have advocated surgical bypass and oesophageal exclusion, but this is a major procedure. An expanding metal stent is probably the best treatment, but semi rigid prosthetic tubes may also be used (Fig. 50.61).

Postcricoid carcinoma

Postcricoid carcinoma is considered in the section on neoplasms of the pharynx (Chapter 43).