Chevalier Jackson (1865 - 1958)
A good and easily accessible start for studying oesophageal cancer is to search Wikipedia. The present subheadings track the author’s search for its incidence, diagnosis, staging, treatment outcome and survival. Only by collecting all the patients in a community can the treatment’s effect be correctly assessed. If one treatment is selectively chosen its outcome can be compared with other treatments or none at all. If two treatments are combined either before, during or after surgery the analysis becomes complex. In scientific experiments it is essential to change only one factor between tests. Most of the following work between 1980 and 1991 concerns surgery and radiotherapy for squamous cell oesophageal cancer together with the management of oesophageal strictures. The author is not an expert on chemotherapy.
In England and Wales in 1978 (72- ) there were just under 4,000 deaths from cancer of the oesophagus giving a figure for epidemiology of about 9 per 10,000 for males and 6 for females. The sources of information were International Classification of Disease (ICD) 8th and 9th Editions, Mortality Statistics, Cancer Registration for England and Wales, Survival Data from the Office of Population Censuses and Surveys (OPCS), Hospital Activity Analysis (HAA) and the 10 % sample called Hospital Inpatient Enquiry (HIPE) as well as World Health Statistics. It was clear that the term oesophageal cancer included many cases of stomach or gastric cancer which is predominately adeno-carcinoma compared with the oesophageal squamous cell cancer (74). The term also included, at the top end, cases of pharyngeal and upper or cervical oesophageal disease with different treatments from the others surgically treated by oesophagogastrectomy. Gastric cancer was decreasing but oesophageal cancer was increasing, possibly due to increased incidence of adenocarcinoma at the lower end.
A further analysis (80) in 1984 studied oesophageal cancer in the North East Thames Region including East London with a population of 3.7 million, 3,000 surgical beds and an estimated 286 new cases of oesophageal cancer each year. About 50 % of 444 patients admitted died in hospital, but in spite of this there was little pathology or histological data. 73 (25 %) had an oesophagogastrectomy of whom 17 died in hospital at a mean time of 27 days post-operatively, with an overall operative mortality of 35 %. If death in hospital after 30 days (7) was excluded, the post-operative mortality reduced to 21 % (an underestimate of 42 %). 55 (20 %) had radiotherapy and 80 (27 %) had intubation. This study provided the incentive to improve data collection and analyse what the best treatment was for each particular patient based on diagnosis and staging.
TOPOGRAPHY, MORPHOLOGY TNM STAGING
ICD 9 (1979) was a curious mixture of coding but after admitting defeat actually had prepared the way forward to SNOP and SNOMED. Not only did it encourage total lack of discipline by including code 150.9 oesophagus “unspecified” to which about 50 % of clinicians subscribed but it ran simultaneously cervical, thoracic and abdominal part for the 10 inch (25 cm) long oesophagus and added upper third, middle third and lower third as extra alternatives, which confused everyone.
Luckily most clinicians would now agree with the oesophagus being divided into four mutually exclusive parts – cervical and upper, middle or lower thirds. Since every oesophageal cancer patient can easily have a barium swallow (simple) or endoscopy (more complex), this should obtain details for topography or anatomy, leaving only the arguments about the gastro-oesophageal junction as to whether it should be called stomach, oesophagus, cardia or oesophageal junction.
The main information from histopathology should indicate whether it is squamous carcinoma or adenocarcinoma, but again there is a junctional area of doubt. TNM staging can also be done by imaging (CT, MRI or PET scans) which is essential for pre-treatment staging. Post surgical treatment staging depends on whether all the tumour has been removed (curative resection) or some could not be removed (palliative resection) either because of longitudinal spread of the primary tumour (a surgical problem) or lateral spread into the local tissues (usually not possible to excise surgically).
With this background, treatment decisions should be made rational, based on pre-treatment staging as well as the age and general condition of the individual patient remembering that 35 % are over 75 years old at presentation. Thoracotomy without resection merely for inspection as to operability should not occur with the present imaging techniques available. Bypass procedures leaving the tumour in situ should be operations of the past. Most resection operations should aim to be curative rather than palliative and every operation should be staged by histology. Only when post treatment staging is stated can results, expressed as one year survival rates (1YSR), 2 year and 5 years (5YSR) be compared. When preoperative chemotherapy has been given then it becomes impossible to be accurate comparing the results of surgeons and their operations, because the individual’s response to this adjuvant chemotherapy may be totally variable. The neologism neoadjuvant applies to that chemotherapy performed before a surgical operation. If given alone or after the operative therapy the new word is not used. The same problem arises when radiation therapy is used prior to surgery. It will be seen later that the one chance to compare surgery against radiotherapy for the treatment of “operable” squamous cell oesophageal cancer in a prospective trial was not supported. Future studies using preoperative neoadjuvant therapy will be impossible to analyse scientifically, because too many variables have been changed simultaneously.
Warned by these studies concerning the requisite data needed a review of the literature on oesophageal squamous cell cancer surgery (oesophagectomy) was done analysing over 80,000 patients in 122 papers, excluding those with no histological data and adenocarcinoma originating from stomach (62). This was an audit trying to discover the truth in the form of a meta-analysis in 1980 without using statistical analysis. The neologism had been invented in 1976. It combined the results of several studios which by themselves had no statistical power due to their small sample size. Of any 100 patients, 58 were explored and 39 of these had the tumour removed of whom 13 (30 %) died. Of the 26 leaving hospital 18 survived one year and 4 were alive at 5 years (6.5 % of those resected and 4 % of the original 100). Operative mortality ranged from 1 to 83 %, but presumably because editors did not accept papers with 100 % mortality. The only reliable survival data based on good post-treatment staging came from Japan. The patients in this series were selected and it was not possible to assess the true incidence of operable oesophageal squamous cell cancer in the community. Later studies have suggested with more complete staging, that Stage I – tumour, no glands = 5 %, Stage II – tumour 5 cm or less, loco-regional glands = 20 %, Stage III and IV – large tumours, distant lymph nodes, metastases = 75 %. Theoretically only Stage I and II would have a chance for curative resection leaving no tumour nor involved glands. But it was clear from this large literature study that many Stage III or IV cases had been included.
Treatment by radiation therapy or radiotherapy, both radical and palliative, was studied in a similar analysis of the literature, 8,500 patients in 49 articles (63). The 5YSR was 6 % compared with 4 %, and the usual reservations due to patient selection are the same as with surgery. The study confirmed that a result similar to surgery but without an operation and the risk of death could be obtained. This fact that good long term survival could be obtained by radiotherapy in older patients over 70 with Stage I or II disease who were unfit for surgery had persuaded the author to use radiotherapy for all operable squamous cell cancers of the oesophagus regardless of age (98). In operable Stage II (no Stage I found) patients there was a 46 % 1YSR and 14 % 5YSR. In Stage III and IV which included palliative radiotherapy there was a 16 % 1YSR and 4 % 5YSR.
This was achieved with neither operation nor mortality.
If radiotherapy is used the clinician must be capable of performing oesophageal dilatation for strictures with or without a later intubation. Debulking of large intraoesophageal tumours can be done by radiotherapy, which leaves the majority of the narrowing as virtually a benign stricture.
Treatment by diaphermy or lasers is only a temporary method of reducing intraluminal obstruction.
The literature was examined for designs used to dilate benign strictures (68). There is an overwhelming number of tubes for permanent intubation of either recurrent strictures or tracheo-oesophageal fistulae which need blocking (see x-rays to right). These were drawn and described in another review article (71). But the simplest, best and cheapest tube that could be made in your workshop out of winding 1 mm stainless steel wire on a simple rod with a lip was the Souttar tube (61). He was a famous thoracic surgeon at The London Hospital who like almost no other surgeon also had a degree in engineering from Oxford. Dilatation and intubation are absolutely essential skills for the oesophageal surgeon, and should not be handed over to any other person.
RIGID OESOPHAGOSCOPE SOUTTAR TUBE
In the last thirty years several surgical procedures such as hiatal hernia repair and vagotomy/pyloroplasty have almost disappeared, with numbers decreasing well below those done in the 1950 and 60s. This has been due to the introduction of H2 receptor antagonists and proton pump inhibitors. Operations for benign oesophageal stricture have now also almost completely disappeared. Norman Tanner had noted that a partial gastrectomy done below the diaphragm could cure almost all. But conservative management with oesophageal dilatation has achieved satisfaction for most. To this purpose the author designed and developed a stainless steel rigid oesophagoscope (57), a set of Teflon (Polytetrafluoroethylene) bougies and a new version of the Souttar tube (61).
The previous oesophagoscope universally found in UK hospitals was the Negus scope, but the new development was improved, reverting to the spherical lumen and distal illumination of Chevalier Jackson’s scope, machined from tubing on a metal lathe. This was longer, had a spoon shaped end for easier entry, the light channel was on the outside leaving the lumen easy to clean and a maximum area for bougies and tubes, the proximal end had a safe handle and electrical light fittings similar to Seward’s Lloyd-Davies sigmoidoscope.
Bougies were made of 7 cm long solid Teflon (polytetrafluorethylene) which has the advantage of a smooth slippery surface; the shaft was thin stainless steel and the handle small with the size in French Gauge (FG) clearly seen. At smaller diameters Teflon can be bent and holds its shape for entry into a small hole. The Teflon is easily cleaned and withstands autoclaving. Most strictures can be dilated at least to 39 French Guage (FG), equivalent to the half-inch that Chevalier Jackson said was ideal (12.5 mm diameter equals about 39 mm circumference). French Gauge is the accepted measurement and is measured in mm of circumference, 2∏r, where ∏ equals 3.142. However, the majority of strictures will eventually accept the full size of the small oesophagoscope to pass through at 60 FG (20 mm diameter), after repeated dilatations.
The large scope was only used for intubation after the smaller scope had been safely passed first and had an internal diameter for the largest Souttar tube measuring externally 16 mm. However, the usual size of tube barrel measured 12 mm external diameter, 10 mm internal diameter and 70 mm in length with a collar 2 mm wider and 10 mm in length.
MEDICAL RESEARCH COUNCIL UK TRIAL
The Medical Research Council UK trial Cancer Therapy Committee set up a Working Party on oesophageal cancer with the author as Chairman in 1982 (109). The first four years (1982-6) were spent in collecting data from the surgeons performing 2,000 oesophagogastrectomies each year (both general and thoracic surgeons). 1,000 of these were done for the 4,000 cases of oesophageal cancer (25 %) and 1,000 for the 12,000 gastric adenocarcinoma cancer (7 %). It was decided to concentrate on the definitely proven squamous cell cancer of the oesophagus in which it was known that radiotherapy was potentially useful (98). Equally important was that pre-treatment staging should be limited to Stage I and II (25 % of the total) as operable and resectable tumours. The decision had been made in 1986 to start a Phase III prospective randomised trial to compare the effects of surgery versus radiotherapy with the results measured by 1YSR, 2YSR and 5 YSR. Protocols were written and the trial started in January 1987. 30 centres had promised 100 cases/year for four years. But only 31 patients were entered by 16 centres in the first 18 months and the trial was stopped in June 1988. Did it fail? Possibly yes, because it stopped and we shall never know which is the best treatment.
No it did not fail. Disappointed at having to stop the trial upon which so many people had spent so much time and energy, I met a wise expert on clinical trials who asked me what the PURPOSE of any trial was, forgetting the statistical results, for which he was a real expert. His answer was that the main purpose of a trial was to alter the climate of opinion about treatment, get people to question their present actions and change them. The failure, or the stopping of the trial, therefore did not indicate that everybody had changed to surgery, but precisely the opposite. The threshold for advising surgery increased because more pre-treatment staging was done and alternative therapies, including radiotherapy and chemotherapywere used. It was also realised that in a system like the NHS, where surgeons are not paid for doing operations, it was sensible to refer oesophageal cancer to expert specialists. So the trial was actually a success because all the preliminary data collection discussions and dissemination of knowledge had changed the climate of opinion without the planned trial ever being completed.
PRESENT STATE OF PLAY
In 1993 I was asked to write a chapter entitled “Surgical treatment of carcinoma of the Esophagus” (124). This contains not only a description of the standard optimum oesophago-gastrectomy operative procedure but a summary of staging, classification, TNM staging, preoperative assessment and results. In retrospect it summaries the progress made by oesophageal surgeons in the previous 25 years, 1968-93, including works by the author and many other thoracic and general surgeons. The operation description is unique in that it states the main complications and what can be done surgically to minimise them, but never completely avoid them. The operative mortality had decreased from 35 % in a local study (80), 30 % in a literature study (62) to a recommended best 10 % in 1980 which in turn has lead to the aim of between 1 and 5 % nowadays. It can never be zero in a large series because the operation is a dangerous procedure.
The author has always as a general surgeon (74) tried to work closely with thoracic surgeons and is of the opinion that general surgeons must always have a thoracic surgical training to be called oesophageal surgeons. Ideally they should also be physicians, epidemiologists, accountants, pathologists and engineers like Souttar and Chevalier Jackson. There is now a danger in the 21st Century that gastroenterologists, radiotherapists and chemotherapists are taking over oesophageal cancer care. They are not uniquely trained in wide enough expertise for this role, so this wave of control will and must change to a more general doctor who can look after the whole patient with all his complex problems associated with a very serious pathological condition.
74. General surgical workload in England and Wales.(PDF)
Allen-Mersh T G, Earlam R J
Brit Med J 287:1115-8 (1983) PMID 6414594