AN CASES OF CARCINOMA OF THE …Thorax(1961), 16, 226. AN ANALYSIS OF 700 CASES OF CARCINOMA OF THE...

Thorax (1961), 16, 226.

AN ANALYSIS OF 700 CASES OF CARCINOMA OFTHE HYPOPHARYNX, THE OESOPHAGUS, AND

THE PROXIMAL STOMACHBY

B. T. LE ROUXFrom the Department of Clinical Surgery, University of Edinburgh

(RECEIVED FOR PUBLICATION JANUARY 1, 1961)

In the course of 12 years, 1,200 patients havebeen investigated for dysphagia in the ThoracicSurgical Unit in Edinburgh. Seven hundred ofthese patients were shown to have as the cause oftheir dysphagia a carcinoma of the hypopharynx,of the oesophagus, or of the proximal part of thestomach. All these patients were referred forinvestigation because of dysphagia, and dysphagiais the link between them. The review has beenundertaken at this time because 700 is a con-venient number and one sufficiently large to lendsome weight to the analysis.The Unit in Edinburgh serves a relatively closed

population of about one and a half million. Mostof the patients with dysphagia in the south-easternregion of Scotland are referred to this Unit forinvestigation. A sprinkling of patients withdysphagia comes from further north and furthersouth, but these probably compensate for the smallnumber with oesophageal carcinoma occurringwithin this region and not referred to this Unit.The number of patients investigated, therefore,probably represents fairly accurately the incidenceof proximal alimentary carcinoma in this popula-tion. A further advantage of the analysis is thatthe management of patients with carcinomatousdysphagia has been uniform throughout the periodunder review, so that the review is an accuratecriticism of a standard management and a

standard operation in the hands of a small numberof surgeons, all of them using the same technique.

In 20 patients, 3% of the total, the cause of thedysphagia was a squamous carcinoma of the hypo-pharynx (laryngopharynx); 19 of these patientswere women. In 382 patients, 54.5% of the total,the cause of the dysphagia was a squamouscarcinoma of the oesophagus; approximatelythree-fifths of these patients were men. In 26patients, just under 4% of the total, the carcinomawas undifferentiated and in the oesophagus; in

this group the sexes were equally represented. Inthe remaining 272 patients, just under 39% ofthe total, the cause of the dysphagia was an adeno-carcinoma, either visible oesophagoscopically, orin the proximal part of the stomach beyond therange of the oesophagoscope. An analysis ofthese 700 cases outlines the natural history ofcarcinoma of the proximal alimentary tract, andshows what influence resection in the hands of thisUnit may have in altering the natural course ofthe disease.

In this analysis adenocarcinoma and squamouscarcinoma are discussed separately. Adeno-carcinoma is probably primarily a gastric tumour,and all the squamous carcinomas in this serieswere in the oesophagus or hypopharynx. Theundifferentiated carcinomas were all in theoesophagus and did not behave in a wayimportantly different from the squamous tumours.Many of the squamous carcinomas involved thecardia, and many of the adenocarcinomasextended into the oesophagus, some as high asthe aortic arch. But, while many squamouscarcinomas in the middle third of the oesophagushad distal to them normal squamous epithelium,no adenocarcinoma was found in the middle thirdof the oesophagus with squamous epithelium distalto it, unless this was a submucous lymphaticmetastasis from a more distal tumour. There isalso a significant difference in the direction ofspread of squamous and adenocarcinoma, a con-sequence of site of occurrence, the cell typeprobably being unimportant.

In assessing operability and operative results,the oesophageal tumours are called high or low.A high tumour is one (a) which was approachedfrom the right, or (b) if approached from the leftnecessitated the making of an anastomosis abovethe level of the aortic arch, or (c) if a resectionwas not undertaken, one the proximal edge of

copyright. on June 11, 2020 by guest. P

rotected byhttp://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.16.3.226 on 1 S

eptember 1961. D

ownloaded from

http://thorax.bmj.com/

CARCINOMA OF THE OESOPHAGUS

which lay closer than 30 cm. to the upper alveoluswhen seen through the oesophagoscope. Whena low tumour was resected the anastomosis laybelow the level of the aortic arch, and theproximal limit of an irresectable low tumour was30 cm. or further from the upper alveolus.The frequency of the different tumours is shown

in Table I. In 271 patients a resection was not

TABLE I

Male Female Total

Squamous Carcinoma of theHypopharynx (20)

Resected .1 10 11Not submitted to operation - 9 9

Total 1 19 20Squamous Carcinoma of the

Oesophagus (382)Resected .127 109 236Irresectable .. 28 16 44Not submitted to operation 66 36 102

Total 221 161 382Undifferentiated Carcinoma of the

Oesophagus (26)Resected .6 6 12Not submitted to operation 7 7 14

Total 13 13 26Adenocarcinoma in the Oesophagus of

Proximal Stomach (272)(a) Visible Oesophagoscopically (180)

Resected.82 48 130Irresectable.20 4 24Not submitted to operation 19 7 26

Total 121 59 j 180(b) Not Visible Oesophagoscopically (92)

Resected.24 16 40Irresectable.19 3 22Not submitted to operation 18 12 30

Total .. 6 1 3 1 92

Total.700Males 418 (60%)Females.282Resections.429 (61%)

attempted or, if attempted, was not completed,an incidence of inoperability of 39%. Fourhundred and twenty-nine resections were under-taken (61 %). None of the 11 patients from whoma carcinoma of the hypopharynx was resected diedfrom operation. The operative mortality in theremaining 418 patients, all of whom had anoesophago-gastrectomy, was 30%. This mortality

Sn

z

04.UI.0 3

25

20 25

is analysed in a subsequent paper (le Roux, 1962a),and a further paper (le Roux, 1962b) is devoted tohaemorrhage complicating the late convalescenceof patients who had undergone oesophago-gastrectomy.The investigation of these 700 patients included

radiography with barium, and endoscopy. Theendoscopic instruments used were the laryngo-scope only in some with tumours in the hypo-pharynx, and the Negus oesophagoscope in allothers. A gastroscope was not used. When anadenocarcinoma of the proximal stomach isdescribed as beyond the range of endoscopicexamination, this means that it was not visiblethrough the long (45 cm.) adult Negus oesophago-scope.

SQUAMOUS CARCINOMA OF THE HYPOPHARYNXTwenty patients with dysphagia were shown to

have a squamous carcinoma of the hypopharynx.The age range is shown in Fig. 1. Nineteen ofthe 20 patients were women. This preponderanceof women is usual in carcinoma of the hypo-pharynx, and this is the only primary malignanttumour of the proximal alimentary tract with ahigher incidence amongst women than amongstmen.NOT RESECTED. Nine women were found to

have cervical glandular metastases when they werefirst seen. In three of these patients the tumourwas irradiated (Fig. 3). None survived longer thanseven months from the time of admission, andirradiation had no appreciable influence in theprolongation of life or in the restoration ofswallowing. Four of these nine patients had along history of intermittent dysphagia, one ofthem for 30 years and the others for more than20 years. They were anaemic and these areprobably examples of sideropenic dysphagia. Thecharacter of the dysphagia had changed in thesefour patients two to six months before they werereferred for advice. The average survival fromthe onset of dysphagia, or from the change incharacter of dysphagia, in these nine patients wasseven months, and the average duration of

19 FEMALES

-n 8530 35 40 45 50 55 60 65 70 75 80

AGE IN YEARSFIG. 1.-Age range in patients with squamous carcinoma of the hypopharynx.

* _

227

mm

I MALE

0 1 ms

if



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

1

-. -

7

>,

.4:

a-'_

__r...~J

FIG. 2 FIG. 3

FIGS. 2 and 3.-Each vertical line represents a patient, and these are arranged the youngest on the leftand the oldest on the right. (DXT is an abbreviation for deep x-ray therapy.)

dysphagia before admission to hospital was fourmonths. Death in eight was due to aspirationpneumonia and in one to haemorrhage.RESECTED.-Ten women and one man were

submitted to resection (Fig. 2): lateral pharyn-gectomy or cervical oesophagectomy in four andlaryngopharyngectomy in seven, leaving thesewith a permanent tracheotomy. Four patients,all women, survived long enough for the satis-factory fashioning of a cervical oesophagus fromskin, and these four patients are alive, 12 years,nine years, and five years (two patients) afterresection. They all swallow normally, one requir-ing dilatation of the skin tube at about yearlyintervals. The man and two women died fromrecurrence of tumour within a year of resection,one patient died from aspiration pneumonia twomonths after resection, and three patients diedfrom haemorrhage five weeks to five months afterresection. In two, haemorrhage was from erosionof the right common carotid artery by tumour

recurrence; in the third, haemorrhage was fromthe nasopharynx, where residual tumour was

found. Three of these patients, all women andall anaemic, had a long previous history of inter-mittent dysphagia (15 to 20 years), a change inthe character of the dysphagia precipitatingadmission. None of these survived operation bymore than eight months. In three of the fourpatients alive after resection, dysphagia had beenpresent for less than five months before resection,but in the fourth survivor dysphagia had beenpresent for 18 months. One patient dying fromtumour recurrence had had dysphagia for twoyears before resection.The average duration of dysphagia in these 11

patients was eight months, from onset in thosewithout a very long history, or from changingcharacter in those three with a very long history.Average survival from the onset of dysphagia inthose dying from recurrence, pneumonia, or

haemorrhage was 14 months.

.-A

IncS

Z 60

z

4. 12

lolS,oiIn

o,,p0-hi

0C0

228



j.com/


eptember 1961. D

ownloaded from


CARCINOMA OF THE OESOPHA G L S

In the case of one woman a carcinoma of thehypopharynx was irradiated 14 months before shereturned with a recurrence of dysphagia for whichshe was submitted to resection. She had haddysphagia for a month before she was irradiated;irradiation improved her swallowing but thetumour remained visible ; she had nearly totaldysphagia for a month before requesting furthertreatment ; she died five weeks after operationfrom haemorrhage from the nasopharynx.To sum up, of 20 patients with carcinoma of

the hypopharynx, 19 of them women, nine hadmetastases when first seen and these patients diedon an average seven months from the onset ofdysphagia. Eleven were submitted to operationand four of these (36°0) are alive more than fiveyears later. Seven of the 20 patients had sidero-penic dysphagia for many years, and all sevensuccumbed rapidly.

SQUAMOUS CARCINOMA OF THE OESOPHAGUSThree hundred and eighty-two patients with

dysphagia were shown to have a squamouscarcinoma of the oesophagus. The age and sexdistribution of these patients are shown in Fig. 4.The ratio of men to women is about 3: 2. Theoesophageal carcinoma was resected in 236patients (62c'). Forty-four patients (120°,') weresubmitted to thoraco-laparotomy with a view toresection but were found to have either anirresectable tumour or one from which dissemi-nation was so gross that resection was unjustifiedbecause of relatively mild dysphagia. In 102

141210

z

I-

4(I

lL0

acI

z

patients (2600 ) thoracotomy was not undertakenbecause they had metastases when first seen. orbecause they were too old or too ill or declinedoperation. Thoraco-laparotomy was undertakenin five patients in the knowledge that the tumourhad disseminated and resection was completed infour of these. In those patients in whom resectionwas not undertaken attempts to improve swallow-ing included the use of stomach or jejunum tobypass the tumour. intubation of the oesophagealstricture with a Simmonds' or Souttar's tube, andirradiation using either deep radiotherapy orradium bougies. The percentage of patients withsquamous carcinoma of the oesophagus found tohave a resectable tumour wvas the same as amongstpatients with adenocarcinoma.

SQUAMOUS CARCINOMA OF THE OESOPHAGUS NOTTREATED BY OPERATION. One hundred and tw opatients wvith squamous carcinoma of theoesophagus were not submitted to operation.Sixty-six of these wvere men and 36 were women.The youngest patient was 47 and the oldest 86.The age incidence in the two sexes is shown inFig. 5. All these patients are dead. In all ofthem the tumour was visible endoscopically. andhistological confirmation of the diagnosis wasobtained by biopsy of tissue obtained atoesophagoscopy. In 60',, the proximal edge ofthe tumour in the lumen of the oesophagus wasless than 30 cm. from the upper alveolus. Inthose tumours that were resected the proximaledge of 36°, lay less than 30 cm. from the upperalveolus. This rather arbitrary division into high

MALE (i

AGE IN YEARSFIG. 4.-Age and sex distribution of patients X ith squamous carcinoma of the oesophagus.

-))



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

MALE (

I m|fffI|LA)

zuJ 4

q Wa.L-0

co4

z2

FEMALE ( )

20 25 30 35 40 45 50 55 60 65 70 75 80 85AGE IN YEARS

FIG. 5.-Age and sex distribution of patients with squamous carcinoma of the oesophagus not submitted to operation.

and low oesophageal tumours serves its purpose

in demonstrating the lower incidence of resect-ability in the more proximal tumours. Theanatomical explanation for this may be adequate;the proximal half of the oesophagus is in suchclose contact with the trachea and bronchi thatearly invasion of these is likely.

All these patients gave a history of dysphagia.In two women this history of dysphagia was

unusual. One had all the clinical, radiographic,and endoscopic features of cardiospasm for nineyears when she was originally investigated. Asingle dilatation restored swallowing so nearly to

normal that she did not attend again. Three years

later she was again referred for investigation, thistime because of hoarseness. She was shown to

have a right recurrent laryngeal nerve palsy anda squamous carcinoma in the middle third of theoesophagus. Dysphagia had not recurred afterthe single dilatation of the cardia three years

before. The second woman, with an unusualstory, presented at the age of 74 with increas'ngdysphagia for three months and respiratorysymptoms. She was shown to have a tracheo-oesophageal fistula through squamous carcinoma,and she died from aspiration pneumonia; she hadbeen irradiated elsewhere 14 years before for a

I)T O 26i24

22

X:Z 3 18r1

LU

<I< 14

_ I 14

> 6

vI0 4

U) L _

NUMBER OF PATIENTSFIo. 6.

histologically proven squamous carcinoma of theoesophagus at the level of the aortic arch;irradiation had partially relieved her dysphagia atthat time and she had lived on liquids and pureesfor 14 years.

Since no operation was undertaken, the dura-tion of dysphagia in this group is synonymouswith survival from the onset of dysphagia (Fig. 6).Nine patients died within three months of theonset of dysphagia. Omitting the two patientswith the unusual histories, the three longestsurvivors lived 22 months from the onset ofdysphagia, one without any treatment and twowith radiotherapy. Again omitting the twowomen mentioned above, the average survivalfrom the onset of dysphagia in 100 patients withsquamous carcinoma of the oesophagus notsubmitted to operation was eight months.Survival is a little longer than in those with an

adenocarcinoma not submitted to operation. Theaverage duration of dysphagia before admissionto hospital was just over four months, and thisis not significantly different from that in theother groups discussed. Dysphagia was the firstsymptom recognized in all these patients; in 20hiccup was another early symptom, and in 30pain was a major complaint. Those with trachealor bronchial invasion and those with totaldysphagia, altogether 56 patients, had importantrespiratory symptoms.The reasons for not undertaking resections in

these patients were as follows:Thirty-four patients were shown histologically

to have invasion of the trachea or one or othermain bronchus. The trachea was invaded in 15,the left stem bronchus in 12, the right in six, andboth stem bronchi in one. One patient withtracheal invasion had, in addition, cervicalglandular metastases.

I

230



j.com/


eptember 1961. D

ownloaded from



Twenty-six patients had histologically provencervical glandular metastases, 15 on the left side,seven on the right, and in four there were glandson both sides of the neck. Two of these patientshad an additional contraindication, namely,tracheal invasion in one and interruption of theright recurrent laryngeal nerve in the other.

Seven patients had a recurrent laryngeal nervepalsy (five left and two right); only one of thepatients with a right recurrent palsy had rightcervical glandular metastases.

Six patients had clinically palpable hepaticmetastases, subsequently confirmed histologicallyafter necropsy.

Three patients had histologically provencutaneous metastases.

Twenty-six patients were not submitted tooperation because of non-metastatic contra-indications. In six, resection was not undertakenbecause of their poor general condition; therespiratory reserve in a further six was so limitedthat they were thought unable to tolerate thoraco-laparotomy; and in 14 resection was not under-taken because of their advanced years.Two patients declined treatment.This makes a total of 102 patients (two having

two distinct contraindications) who were notsubmitted to operation.

Seventeen patients were treated by radiotherapy.From Fig. 7 it will be seen that nearly all thelonger survivors from the time of admission tohospital were irradiated. In nine of these 17dysphagia was unaffected by irradiation, and thesenine patients deteriorated rapidly. In the other

a

lI,,,_j

I.nZ.-iz

z

JO 6

O-

z

.n>all0ztiJ

02 1Zv

eight dysphagia was relieved, three survivingirradiation by more than a year, but nonesurviving two years. In four a radium bougiewas used; all these were treated more than 10years ago; none of them survived treatment morethan five months, and two died from massivehaemorrhage.

In 35 patients dysphagia was partially relievedby intubation with a Simmonds' or Souttar's tube.Either of these tubes allows of the passage of onlyliquids, and there is consequently no indication fortheir insertion through the irresectable carcinomaof a patient able to swallow liquids withoutdiscomfort. Death in three of these patients wasprecipitated by perforation of the oesophagus atsubsequent attempts at intubation, necessitated byblockage of a previous tube. Those intubated didnot survive a significantly longer, or shorter,period than those whose dysphagia was not totaland who were therefore not intubated. Threepatients intubated were, in addition, irradiated.

Information regarding the mode of death inmany who died at home is scanty, but, relyinglargely on the memory of the general practitionersconcerned, a terminal respiratory illness was verycommon. Those patients with a fistula into therespiratory tract all died as a consequence ofaspiration pneumonia, and nearly all these patientsdied in hospital. Seven patients, including twotreated with radium bougies, died from massivehaemorrhage.Summary.-One hundred and two patients with

squamous carcinoma of the oesophagus were notsubmitted to operation; 60% had high tumours,

SOUAMAOUS CARCINOMA i) DXT =DEEP X RtAY THERAPY | FNOT SUBMITTED TO OPERATION 1P E

YEAR - -

FIn. 7.-Each vertical line represents one patient, and these are arranged the youngest on the left and the oldest on the right.

231



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

mmAIiLL0

% 53

z I

20 25 30 35 40 45 55 60

AGE IN YEARSFIG. 8.

and invasion of the trachea or main bronchi wasthe commonest reason for not attempting resec-tion. Radiotherapy temporarily relieved dysphagiain about half of those in whom this form of treat-ment was used. Average survival from the onsetof dysphagia was eight months. A terminalrespiratory illness was very common.

SQUAMOUS CARCINOMA FOUND IRRESECTABLE ATTHORACO-LAPAROTOMY. There were 44 patientssubmitted to thoraco-laparotomy with a view toresection (16 women and 28 men) in whom thetumour was found unsuitable for resection. Theage and sex distribution of these patients areshown in Fig. 8. Two men in this group wereunusually young, 23 and 25 years of age. Resec-tion was not undertaken in 12 because of invasionof the aorta; in seven because of invasion of thelungs or bronchi; in six because of hepaticmetastases; in four because of diffuse pleuralmetastases; in three because of cardiac invasion;and in a further three because of invasion of theposterior abdominal wall. In nine resection was

abandoned because of what is loosely referred toin the operative notes as " mediastinal invasion."

In seven patients an attempt was made tore-establish swallowing by fashioning a bypass,using jejunum in three, stomach in three, andcolon in one.At operation three patients were found to have,

in addition to the oesophageal tumour, a carci-noma of the stomach at the pyloric end: in allthree the gastric tumour was an adenocarcinoma.Operation was precipitated in three patients byperforation of the oesophagus at oesophagoscopy.In these three an attempt was being made tointubate the tumour to promote hydration as apreliminary to resection. In one thoraco-laparo-tomy was undertaken in the presence of known

cervical glandular metastases in an attempt torestore swallowing. In 10 the tumour was at thelevel of the aortic arch or higher and the approachwas therefore from the right. In the remaining 34a left thoraco-laparotomy was made. In 17 theproximal edge of the tumour, as seen endoscopic-ally, was less than 30 cm. from the upper alveolus,an incidence of " high " oesophageal tumours of40%, which is a little higher than that amongstthe squamous tumours for which resection wascompleted.

Fourteen patients died as a direct consequenceof operation, an operative mortality of 32% (Fig.9). Two died from pulmonary embolism andseven from respiratory infection. Three patientsfor whom a bypass had been made died fromdisruption of the anastomosis. The remaining twodied as a consequence of rupture of the oeso-phagus which became clinically obvious late in thefirst post-operative week but which had probablybeen initiated by the attempts at mobilizing theoesophagus.

Thirty patients survived operation. Swallowinghad been restored in four of these by a bypassprocedure. A further 10 were intubated with aSimmonds' or Souttar's tube and left hospital ableto swallow liquids. Three patients treated about10 years ago were treated with a radium bougie,and one, in addition to being intubated, wasirradiated.

All these patients are dead. The duration ofpre-operative dysphagia and the length of post-operative survival are shown in Fig. 9. The longestsurvivors include the four patients with the bypassprocedure. Only one patient survived operationby more than a year. Survival from the onset ofdysphagia is shown in Fig. 10. Two patientssurvived only one month from the onset of

UQI~-z SJ 3_L,

MALE (

FEMALE (

65 70 75 so 85

232



j.com/


eptember 1961. D

ownloaded from



>0

I>0g

O¢XI

0n

z -0002 CK tZ LA

r fi ff I

D U. >.a O a3

FIG. 9.-Squamous carcinoma found to be irresectable. Patients dying from operation are indicated by the dot in the transverse gap labelledoperation.

0

Z 21

O VA <

zO <

CD0 f(A LL

NUMBER OF PATIENTSFIG. 10.

dysphagia and both died as a consequence ofoperation. The patient who survived more thana year following thoraco-laparotomy was also thelongest survivor from the onset of dysphagia, 20months. The average survival from the onset ofdysphagia was just under seven months and theaverage duration of dysphagia before operationwas four months.

Dysphagia was the first symptom in all thesepatients. The level at which the patient appre-ciated the obstruction had little bearing on thesite of the tumour, the patients appreciatingobstruction at one of three levels-in the supra-

sternal notch, at midsternal level, or in theepigastrium. Obstruction was usually appreciatedat a level proximal to the tumour, and occasionallyat a level distal to the tumour. In addition todysphagia, pain, usually related to swallowing butsometimes not, and experienced in the epigastrium,in the suprasternal notch, at midsternal level or

between the scapulae, was an obtrusive symptomin nine. In two patients hiccup was an annoyingsymptom which developed at the onset ofdysphagia but which had disappeared by the timethe patients were admitted for investigation.Summary.-Forty-four patients with squamous

carcinoma of the oesophagus were found atthoraco-laparotomy to have tumours unsuitablefor resection. The operative mortality was 32%and average survival from the onset of dysphagiain these patients was a little less than in the groupnot submitted to operation.SQUAMOUS CARCINOMA RESECTED. - Two

hundred and thirty-six patients with squamouscarcinoma of the oesophagus were submitted toresection, 109 women and 127 men. The ageincidence in the two sexes is shown in Fig. 11.The youngest was 25 years of age and the oldest82. The average duration of pre-operativedysphagia was four to five months but nearer fourmonths in the men and five months in the women.There is no significant difference in the averageduration of dysphagia before resection in compari-son with the duration of dysphagia in thosepatients who presented with metastases or in thosewith tumours found irresectable at thoracotomy,or in patients with adenocarcinoma. Symptomsin this group were no different from those withunresected tumours.

THE OPERATION.-After resection of the tumouran oesophago-gastric anastomosis was made in 219

233



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

M.J.

AGE IN YEARSFIG. 11.

patients and an oesophago-jejunal anastomosis in16. In one patient continuity was re-establishedwith a plastic tube in the mediastinum. She diedfrom metastases eight months after resection. In44 patients the tumour was approached througha right tnoraco-laparotomy, or through a rightthoracotomy after preliminary mobilization of thestomach through a left paramedian laparotomy.In 192 the approach was from the left. In 53in whom a left thoraco-laparotomy was used theanastomosis was made above the level of theaortic arch, the stomach lying lateral to the arch.The patients in whom a right approach was usedand those in whom a supra-aortic anastomosiswas made on the left are regarded as having ahigh oesophageal tumour, and these constitute360% of the squamous tumours resected. If thesomewhat arbitrary level of the proximal limit ofthe tumour as seen endoscopically is used todifferentiate high from low tumours, and this levelis fixed at 30 cm., then again in this series 36%of the tumours are high. While these percentagesare identical, there is a little difference in therepresentation of patients in the two groups. Thisdifference is less than 5% and is explicable. Thelevel of 30 cm. is not correlated with the heightof the patient or the position of the aortic arch;there is some variation in factors influencing thedecision to make an anastomosis below the arch,both from surgeon to surgeon and at differenttimes, and where the tumours are mobile theproximal edge can be made to move 3 or 4 cm.so that the level can be varied by pressure fromthe endoscope. The endoscopic evidence of theproximal limit of the tumour has nevertheless beenfound more dependable than the radiographicevidence. Only those tumours obviously at orabove the level of the aortic arch were approachedfrom the right.

OPERATIVE MORTALITY.-Eighty patients diedas a direct consequence of operation; 40 weremen and 40 were women, and the operativemortality was 340%. The operative mortalityamongst men (32%) was lower than amongstwomen (37 %). Most of the operative deathsamongst the women were in those over 65 years ofage (Fig. 12), of whom 50% died from operation.Amongst the men there is a fairly even scatter ofoperative deaths under 70 years of age, but over

70 the operative mortality was 50% (Fig. 13). Ofthe 44 patients in whom the approach was fromthe right, 15 (340%) died from operation, anoperative mortality for this approach correspond-ing with that of all squamous carcinomas whichwere resected. Of the 53 patients in whom a

supra-aortic anastomosis was made, 23 (430%)died from operation. The operative mortality forpatients with " high " tumours-a combination ofthose approached from the right and those inwhom the anastomosis was above the aortic arch-is therefore 39% against an operative mortalityof 30'() amongst the resections for " lowsquamous tumours, since of 139 patients in whomthe anastomosis was made at or below the levelof the lower border of the aortic arch, 42 died.Amongst the resections for adenocarcinoma noneof the few patients whose tumour was approachedfrom the right or in whom the anastomosis wasmade above the level of the aortic arch died fromoperation.

Seventeen operative deaths, nine of whichoccurred in women, were a consequence ofleakage from the anastomosis, and a further fiveoperative deaths were from haemorrhage, a con-sequence of erosion of the aorta, a late sequel, itis believed, to leakage from the anastomosis.Three patients approached from the right and sixpatients in whom a supra-aortic anastomosis were

V%,

z

I-

Q.

0au

z

234



j.com/


eptember 1961. D

ownloaded from


3t

0

2

ZlIQ_> t

* 5

7 26- -ff

O

a

FIG. 12.-Each vertical line represents a patient, and these are arranged youngest on the left, oldest on the right. Patients dying fromoperation are indicated by the dot in the transverse gap labelled operation. The red lines represent patients who are alive at the presenttime or who have died more than five years after operation but not from metastases. The vertical black lines above the operativegap represent patients who have died from metastases or who have died less than four years after operation from causes other than metastases.

30

2

0

SQUAMOUS CARCINOMA RESECTED

MtALE(i OPERATIVE

FIG. 13.-Each vertical line represents a patient, and these are arranged youngest on the left, oldest on the right. Patients dying fromoperation are indicated by the dot in the transverse gap labelled operation. The red lines represent patients who are alive at the presenttime or who have died more than five years after operation but not from metastases. The vertical black lines above the operative gaprepresent patients who have died from metastases or who have died less than four years after operation from causes other than metastases.

,4~~~~~~~~>g*AS-UWUS Ujt a§., 0I- 0 U X

!~~~~~ S1II1l1, 111*o1l; 1l TSWIOS iSiodSi510 IRESECTE

F E 1 L E>O a>O> >̂ T

- _- _- A A- 0J

12~~~~111

SQUAMOUS CARCINOMA RESECTED

FEMALE 109 OPERArIvE4 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~MORTAILITY4037/.

3



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

made died from leakage from the anastomosis.Fifty-three per cent. of inadequate anastomoseswere made, therefore, in patients with hightumours, whereas high tumours comprised only36% of the squamous carcinomata resected.

Thirteen patients, eight of them women, diedfrom pulmonary embolism, and 19, 13 of themmen, died most directly from pulmonary infection.Three patients, two of them approached from theright, died from empyema thoracis with an intactanastomosis and one patient died from necrotizingenteritis on the fourth post-operative day. Threewomen died with an uncontrolled leak of chyleinto the right pleural space; pulmonary infectioncontributed to their death. In all three deathswith chylothorax a supra-aortic anastomosis hadbeen made, but this relationship may be fortuitoussince, of five other chylothoraces which were notassociated with death, only one was in a patientwith a supra-aortic anastomosis.

Five patients, three of them with high tumours,remained hypotensive after operation and failedto respond to resuscitative measures, dying moreor less directly from surgical " shock " during thefirst three post-operative days. Two women diedfrom reactionary haemorrhage during the firsteight hours after operation from sources notdetermined at necropsy.Three men and one woman died from coronary

artery disease during the first four post-operativedays; three of them had post-mortem evidenceof recent coronary artery occlusion and the fourthhad grossly diseased coronary arteries and died incardiac failure. Two men died with renalinsufficiency, one being found at necropsy to havea chronic destructive inflammatory lesion in bothkidneys and the other acute tubular necrosis,possibly related to a period of hypotension in theearly post-operative phase. One woman and oneman died during the first two post-operative daysfrom primary respiratory insufficiency, withrespiratory infection contributing to their death.They had gross emphysema and, in retrospect,should probably not have been subjected toresection. One woman died from hepatic failureconsequent upon ligation of the hepatic arterywhich was involved in a mass of tumour. Oneman died four hours after operation with anunrecognized right tension pneumothorax; in himoesophagectomy through a left thoraco-laparo-tomy had been associated with transmediastinalresection of part of the right lower lobe which wasdirectly invaded by the tumour. One woman diedwith a left hemiplegia from cerebral thrombosisand another died with severe parotitis as the only

lesion found to account for her death. It is clearthat a large proportion of the operative deaths inthis series are the result of faulty technique. Someof the others are a consequence of faulty selectionof patients, but selection is always biased infavour of resection because of the need to relievedysphagia.

METASTASES AT OPERATION.-In 145 patients(61 %) histological confirmation of glandularmetastases in the abdomen or in the mediastinumwas obtained. In 53 evidence of spread otherthan glandular invasion was found, but only fourof these 53 did not have, in addition, glandularmetastases. In all, 63% of patients in whom asquamous carcinoma was resected had spreadbeyond the oesophagus in comparison with morethan 70% of patients in whom an adenocarcinomawas resected. The commonest sites of directinvasion were the posterior wall of either stembronchus, the posterior wall of the trachea, andthe lower lobe of either lung, more commonly theright, in relation to the pulmonary ligament. In29 patients tracheal, bronchial, or pulmonaryinvasion was found. Resection was undertakenin nine found at operation to have hepatic meta-stases, and two of these did not have glandularmetastases. These nine with hepatic metastaseswere relatively young and had severe dysphagia.Tumour was left adherent to the aorta in sevenpatients, and elsewhere in the mediastinum in six.Diffuse peritoneal spread was found in one, anda renal metastasis in another.

Part of the pericardium was resected because ofinvasion by tumour in 17 patients. Resection wasundertaken in three with histologically provencervical glandular metastases and in a fourth witha left recurrent laryngeal nerve palsy presumed tobe a consequence of mediastinal invasion. Thesefour were young, had severe dysphagia, and werein good general condition. Resection in thepresence of known cervical glandular metastasesenabled a 39-year-old woman to swallow normallyfor three years after resection while she cared fora young family. Of the other three in whomresection was undertaken despite known meta-stases, one with cervical glandular metastases diedfrom operation and the other two each surviveda year.Whereas 63% of all patients in whom a

squamous tumour was resected had operativeevidence of metastases or invasion, 73% of thosewith " high " tumours had evidence of spread atoperation, and only 55% of those with "low "tumours had spread at operation. Resection was

236



j.com/


eptember 1961. D

ownloaded from



undertaken in 87 patients (37%) in the absence ofhistological evidence of dissemination at the timeof operation, and, of these, 25 died as a conse-quence of operation. One hundred and fifty-sixpatients survived operation, 62 of them withouthistologically confirmed spread. Twenty-three ofthese 62 later died from metastases, 10 died fromnon-metastatic causes, and 29 are alive.OPERATION IN UNUSUAL CIRCUMSTANCES.-Four

patients in whom resection was undertaken wereknown to have two foci of squamous carcinomain the oesophagus with normal oesophagealmucosa between them. Examination of theresected specimen showed submucous lymphaticextension of the primary tumour and the secondmalignant mucosal ulcer was therefore ametastasis. Two of these patients died fromoperation and two died from metastases duringthe second post-operative year. Resection wasundertaken in two patients in whom dysphagiahad been relieved for a short period by radio-therapy, only to recur. Three men gave a longhistory of intermittent dysphagia with a changein symptoms precipitating their admission tohospital. These three probably had benign,peptic strictures of the oesophagus. All survivedoperation, two died from metastases within a yearof operation, one survived 10 years and then diedafter resection of an abdominal aortic aneurysm.A woman gave a history of 17 years of inter-mittent dysphagia; she had a web at thecrico-pharyngeus and a squamous carcinoma inthe distal oesophagus, and she was anaemic; sheis probably an example of sideropenic dysphagiacomplicated by carcinoma other than in thehypopharynx. In three in whom resection wasundertaken an additional malignant tumour wasfound: an adenocarcinoma of the distal part ofthe stomach in two and in the third themicroscopic glandular and splenic changes ofchronic lymphatic leukaemia. In two patients theleft recurrent laryngeal nerve was sacrificed atoperation to clear glandular metastases.

NON-FATAL COMPLICATIONS.-Five of the eightpatients whose convalescence was complicated bychylothorax survived. The chylous effusion wason the right side in all these, and they were treatedby intercostal intubation with drainage into awater-seal. Two patients survived leakage fromthe alimentary tract into the left pleura. One,in whom a leak was demonstrated on the ninthpost-operative day, was re-explored and found tohave a perforation of the intrathoracic gastrictube 5 cm. distal to the oesophago-gastrostomy.This was regarded as a perforated peptic ulcer,

but may have been the result of trauma to orischaemia of the gastric wall. The perforationwas closed and the patient's recovery thereafterwas uneventful. The other was found atre-exploration on the seventh day to havea disrupted oesophago-gastric anastomosis; theproximal oesophageal stump was delivered on tothe surface of the skin in the left side of the neck,just above the clavicle, and the gastric stump wasexteriorized at the left costal margin and usedas a gastrostomy for purposes of feeding. Shewas left with these fistulae while the left empyemawas managed and she improved over severalmonths. Alimentary continuity was thenre-established through the anterior mediastinumusing jejunum, and she is well two years aftercompletion of these surgical procedures.THE FATE OF THE SURVIVORS AFTER OPERATION

-One hundred and fifty-six patients survivedoperation, 62 of these not having been found atoperation to have histologically demonstrableevidence of dissemination. One hundred and eightpatients, 67 men and 41 women, died subsequentlyfrom metastases. Ten have died from non-metastatic causes, six of them more than fiveyears after resection. Thirty-eight patients, 24women and 14 men, are alive.Of the 108 patients who died from metastases

after resection, 59 died during the first year, 33men and 26 women (Figs. 12 and 13). During thesecond year after resection a further 34 (24 menand 10 women) died from metastases, and duringthe third year after resection 11 patients (sevenmen and four women) died from metastases.During the fourth year after resection one womandied from metastases and three men died five,six, and seven years after resection respectively,also with metastases. Of those who died frommetastases, eight who died within a year ofoperation, seven who died in the second year afteroperation, seven who died in the third year afteroperation, and one dying seven years afteroperation had not been shown to have metastasesat the time of operation. Therefore, 23 of the 62survivors of operation who had not been foundto have evidence of spread at operation died frommetastases.A further 48 patients need to be accounted for.

Five are alive less than a year after resection, andtwo of these had glandular metastases at operationwhile three had no evidence of spread. Five havesurvived more than a year after resection, but notmore than two years; two of these had metastasesat operation and three did not. Five havesurvived more than two years after operation but

237



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

have not yet survived three years; only one ofthese five had glandular invasion at operation, theother four not being shown at operation to haveevidence of spread. Four are alive more thanthree and less than four years after resection, andnone of these had glandular metastases at opera-tion. Nineteen patients are alive more than fouryears after resection. Of these 19, two are alivein the fourth post-operative year, one in the fifth,three in the sixth, four in the seventh, two in theeighth, two in the ninth, two in the tenth, andthree in the eleventh post-operative years. Fourpatients surviving more than four years afterresection had evidence of glandular metastases atoperation, and one of these is alive seven andanother nine years after resection.

Thirty-eight of the 48 patients previouslyunaccounted for are alive; the remaining 10have died. Three women have died from non-metastatic causes more than five years afteroperation: one from bronchopneumonia at theage of 86, having survived resection by five years,one from old age at 86, having survived resectionnearly seven years, and one from a cardiac lesionat 50, having survived resection by five years.Four men have died from non-metastatic causesmore than four years after resection. One diedseven years after resection at the age of 58 fromcerebral thrombosis; one died six years afterresection at 67 from coronary thrombosis; onedied nearly five years after resection from massivebleeding from an adenocarcinoma of the stomach,resection having been for a squamous carcinomaof the oesophagus; and one died 10 years afterresection as a consequence of an operation for ananeurysm of the abdominal aorta.Three patients have died from non-metastatic

causes less than four years after resection. Awoman died from ruptured diverticulitis of thecolon in the second year after resection; a mandied from an adenocarcinoma of the rectum inthe third year after resection; and a man diedfrom bronchial carcinoma in the fourth year afterresection. This last-mentioned man presented twoyears after resection with a ruptured aneurysm ofthe abdominal aorta distal to the renal arteries;this aneurysm was resected and replaced with aDacron graft; he recovered rapidly from thisprocedure; no intra-abdominal metastases hadbeen found at laparotomy. Two years later hepresented with haemoptysis and was shown tohave a squamous carcinoma in the right upperbronchus. His original oesophageal tumour hadbeen a squamous carcinoma, but withoutglandular metastases. Because he was then 75

and with a poor respiratory reserve, he was notsubmitted to resection of the bronchial tumour;he died seven months after a course of radio-therapy. At necropsy the bronchial tumour wasbarely recognizable, having been satisfactorilyshrunk by radiotherapy, and there were noevidences of spread in the mediastinum; theoesophago-gastrostomy looked normal; he diedwith severe coronary arteriosclerosis. He is thesecond man in this series to develop an abdominalaneurysm after oesophago-gastrectomy. Theother died 10 years after oesophago-gastrectomyand 10 days after resection of an abdominalaneurysm, continuity again being established withDacron. He died from respiratory infection, andat necropsy there was no evidence of metastasesfrom the oesophageal carcinoma. The Dacrongraft was patent and cleanly lined by a smooth,polished membrane, presumably composed ofplatelets and fibrin.Of 236 patients submitted to resection and of

156 who survived operation, 17 are alive at thepresent time and more than five years afteroperation. Six other patients have died morethan five years after resection from non-metastaticcauses. Of the 236 resections, 128 were under-taken more than five years ago, and of these 128patients, 37 died from operation. Only 91patients have had the opportunity of survivingfive years. If the five-year survival rate iscalculated to include only those at present alive,then 13% of those submitted to operation, or19% of those who survived operation, aresurvivors for five years. But survival fromoesophago-gastrectomy does not confer immor-tality, and it is justifiable to include those nowdead, but alive at five years, provided they didnot die from metastases. The five-year survivalrate is therefore 18%, calculated against thenumber submitted to operation, and 25 %,calculated against those surviving operation. Inall but one of the 23 patients alive at the end offive years the distal part of the stomach was usedto re-establish continuity after resection of thetumour; in one jejunum was used. Approximatelyone anastomosis in every 13 was made usingjejunum.RECURRENCE OF DYSPHAGIA.-Dysphagia re-

curred before death in 33 of the 108 patientsdying with metastases, and in 32 of these evidenceof glandular invasion had been found at opera-tion. In 14 of those dying with recurrence ofdysphagia the original tumour had been high, andin 19 low. Fifteen died with recurrence ofdysphagia during the first year after resection:

238



j.com/


eptember 1961. D

ownloaded from



eight of these had had high tumours, seven low,one of these seven without glandular metastases.Fourteen died with recurrence of dysphagiaduring the second year after resection, four withhigh and 10 with low tumours originally; andfour, two of each, died during the third year afterresection. In three dysphagia recurred nearly ayear before death, and was treated by intubationof the recurrence at the anastomosis; in most ofthe remainder dysphagia recurred within twomonths of death and was only one of severalmanifestations of recurrence or spread in thesepatients. Malignant dysphagia recurred, therefore,in 21% of those surviving operation.

In three patients still alive and believed not tohave metastases the oesophago-gastric stoma hasrequired dilatation because of a fibrous stricture,probably of peptic origin. A fourth patientrecorded earlier as dying six years after resectionwith metastases developed a peptic stricture at theoesophago-gastrostomy within a year of resection.He learned to use a gum-elastic bougie andmaintained adequate swallowing with this for fiveyears. A retired railwayman, he was granted freepassage on the railways, and made the journeyregularly from his distant home to Edinburgh,entertaining his fellow passengers by his demon-stration of " sword-swallowing." He then beganto feel less well, and asked that his stricture bedealt with surgically. The stricture was resected,and at this operation glandular metastases werefound; he died some weeks later from urinaryinfection, but with fairly extensive metastases.

FATE OF THOSE WITH HIGH TuMOURS.-Ninety-seven patients were found to have "high"tumours, and 38 (39%) of these died from opera-tion. During the first year after operation 26from whom high tumours had been resected diedfrom metastases, eight of these with recurrenceof dysphagia. During the second year afterresection 11 died with metastases, four of themwith recurrence of dysphagia; and during thethird year after resection four died frommetastases, two with recurrence of dysphagia.One died from metastases six years and one sevenyears after resection. Sixteen are alive, eight ofthem more than five years after resection, andthree of these eight had glandular metastases atoperation. Two of these are worthy of moredetailed record.

One, a man who is alive six years after resec-tion, was submitted at the age of 64 to resectionof a squamous carcinoma, 23 cm. from the upperalveolus, through a right approach; glands wereinvaded. A year later he was readmitted

T

seriously ill with a mass of glands in the left sideof the neck, shown histologically to containsquamous carcinoma, and with stridor, the conse-quence of squamous carcinoma invading theposterior wall of the trachea and left mainbronchus at the main carina. To improve hisairway a considerable quantity of tumour wasresected with biopsy forceps through the broncho-scope, and he was irradiated. He is well, at theage of 70, six years after resection and five yearsafter irradiation, without clinical evidence ofmetastases, swallowing normally, and with normalbronchoscopic appearances. The second patientof particular interest in this group of long-surviving high tumours was the youngest sub-mitted to resection, a woman, then unmarried, 25years of age, with a squamous carcinoma at 25cm. from the upper alveolus. A supra-aorticanastomosis was made from the left; glands werenot invaded; and she is alive, married now andwith a family, six years after resection. Of theeight patients alive less than five years afterresection of a high tumour, five had glandularmetastases at operation and three did not. Oneof these last-mentioned was the woman mentionedabove who survived leakage from the anastomosis,exteriorization of oesophagus and stomach andultimate restoration of continuity.

Relative to their representation in the series ofsouamous carcinomas as a whole, more patientswith high tumours died from operation andduring the first year after operation frommetastases than did patients with low tumours.Thereafter the survival rate of high and lowtumours is the same. The higher death rate frommetastases in the first year after operation in thosewith high tumours is probably a corollary of thehigher incidence of metastases at operation inthose with high tumours.

INCIDENCE OF EFFECTIVE PALLIATION.-Thepalliative value of resection in those patientsdying from metastases must be assessed againstthe average survival time of those patients withsquamous carcinoma of the oesophagus notsubmitted to operation. In the group of 100typical examples of squamous carcinoma notsubmitted to operation it has been shown earlierthat the average duration of dysphagia beforeadmission to hospital was four months, and thataverage survival from the onset of dysphagia waseight months. Eighty patients in whom asquamous carcinoma was resected died fromoperation; in these the average duration ofdysphagia before operation was four months(Fig. 14a), and, since they died from operation,

239



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

40L363636

34

o 3230

z 280 26I 24

222 200cc

02

44

a>

0~; 12

oI lL' z a7. __

IV av pu Q 50 60 70g)

(a)

average survival from the onset of dysphagia was

also four months. No palliative benefit accruedto them. Fifty-nine died from metastases duringthe first year after resection, and, of these 59, 15died with recurrence of dysphagia. The averagesurvival from the onset of dysphagia in this groupof 59 was just over 12 months, the longest survivalbeing 34 months and the shortest five months(Fig. 14b). The restoration of swallowing foreven a limited period was probably regarded bythese patients as valuable, but as regards durationof survival little palliative benefit can be claimed,especially since 25'/ ) died with recurrence ofdysphagia. Resection in 139 patients (80 + 59) can

therefore be regarded as little more than an

episode in the natural course of the disease, i.e.,in 60), of resections undertaken the naturalhistory of the disease was unaltered. Thirty-fourpatients died from metastases during the secondyear after resection, 14 of these 34 dying withrecurrence of dysphagia. The average survivalfrom the onset of dysphagia in this group of 34was 23 months, the longest survival being 40months and the shortest 15 months (Fig. 14c).Significant palliative benefit, i.e., increase insurvival time from the onset of dysphagia, can

be claimed for these patients, and for all thosedying subsequently from metastases.

Slummary. - Two hundred and thirty-sixsquamous tumours of the oesophagus were

resected; 360% were high tumours. Thirty-fourper cent. of patients died as a consequence ofoperation; 630,b had evidence of spread atoperation. Forty-six per cent. of those submittedto operation (69% of those who survived opera-tion) died from metastases, more than half ofthem during the first 12 months after resection.

a

4

(c)

Ten per cent. of those submitted to operation(15')Y, of those who survived operation) were alivefive years after resection, although some of thesehave since died from non-metastatic causes.

Three patients died more than five years afterresection from metastases. Ten per cent. of thosesubmitted to operation are alive less than fiveyears after resection, or have died from non-

metastatic causes during the first five post-operative years.

UNDIFFERENTIATED CARCINOMA OF THE

OESOPHAGUSTwenty-six patients investigated because of

dysphagia were shown to have an oesophagealcarcinoma histologically called undifferentiated.The sexes were equally represented. Thirteen ofthese patients had metastases when first seen, andone had coronary artery insufficiency of suchseverity that thoraco-laparotomy was precluded.The age incidence of the 14 patients not submittedto operation is shown in Fig. 15. The sexes in thisgroup are again equally represented.

NOT RESECTED.-Of the 13 patients withmetastases, four had histologically proven spreadto left cervical glands, two had tracheal and threeleft main bronchus invasion confirmed histologic-ally, two had hepatic metastases palpable clinicallyand later demonstrated at necropsy, and two haddiffuse pulmonary metastases also confirmed atnecropsy. In nine of these 13 patients theproximal edge of the tumour was seen oesophago-scopically less than 30 cm. from the upper

alveolus.Four of these 14 patients were treated by

irradiation (Fig. 16). One, the man in whom

In

0

z

0

ui

I>

g

(b)NUMBER OF PATIENTS

FIG. 14.

n 7n 9_

240

40



j.com/


eptember 1961. D

ownloaded from



thoraco-laparotomy was precluded because ofcoronary artery disease, is alive seven months afterthe onset of dysphagia and four months aftercompleting radiotherapy. The longest survivorfrom the onset of dysphagia-28 months-was

irradiated. The third patient irradiated had anunusually long history of dysphagia-three years-before admission to hospital, but survived irradia-tion by only two months. The fourth patientirradiated died three months after admission to

Undifferentiated Carcinomas ~ ~ *m a s

0 j

U wW J(A) 41w 5cc E _- U

6

6

7

7

*UmsE *E E20 25 30 35 40 45 50 55

FIG. 15.

RESECTED a19

eOQ|9 A

I YEAR

I !iI11WI1

OPERATIVE

3ORTALITY4

60 65 70 75 80 85

FIG. 16.-Each vertical line represents a patient. The patient represented by the shorter red line althoughalive has metastases at the present time. The longer red line represents a patient dying from a cause

other than metastases six years after operation. DXT is an abbreviation for deep x-ray therapy.The patients are arranged the youngest on the left, oldest on the right.

3

3

33 Ow IzIj) rLLaJw

* -

30InII..z0

z-i

lie

n

I-

40I

ZzQ3

wz

CA

a.oi

WL -mmdii..

241

I



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

hospital. Three patients with total dysphagia wereenabled to swallow liquids by the introduction ofa Souttar's tube.The average duration of dysphagia before

admission to hospital in the 14 patients not sub-mitted to resection was seven months; if thepatient with the unusually long history ofdysphagia is excluded, this figure is reduced tosix months. The average survival from the onsetof dysphagia in this group is 11 to 12 months.RESECTED.-Twelve patients (six men and six

women) with an undifferentiated oesophagealcarcinoma were submitted to resection (see Fig.16). The age incidence is shown in Fig. 15. Inonly three of these was the proximal edge of thetumour less than 30 cm. from the upper alveolus.In 11 an oesophago-gastric anastomosis was madeand in one an oesophago-jejunal anastomosis.Two of the oesophago-gastrostomies were supra-aortic, and a third was made through a rightthoraco-laparotomy. Four patients died fromoperation, an operative mortality of 33%. Onedied from leakage from the anastomosis, one fromhaemorrhage, a consequence of erosion of theaorta, one from pulmonary infection, and one, inwhom an extensive gastrectomy had been made,jejunum being used to restore continuity, died onthe tenth day from peritonitis, a consequence ofrupture of the duodenal stump. One patient diedsuddenly a month after leaving hospital, twomonths after resection. The cause of death is notknown, but was not, so far as can be ascertained,from haemorrhage.

In 10 of the 12 patients glandular metastaseswere found at operation. The four operativedeaths and the sudden death two months afterresection were all in patients who had glandularmetastases. Four patients died from metastases:one two months after resection, one nine monthsafter resection, and two at 20 and 21 monthsrespectively. These four patients had glandularinvasion demonstrated at operation. A fifthpatient died from metastases 29 months afterresection, and in this patient no evidence of spreadhad been found at operation. One man withglandular invasion at operation is alive 14 monthsafter resection but is ill with clinical evidence ofhepatic metastases. One patient died fromcoronary thrombosis six years after resection. Heis the second patient who did not have evidenceof spread at operation. Dysphagia did not recurin any patient submitted to resection.

Excluding the patient dying from coronarythrombosis, and limiting the prognosis of thepatient alive with hepatic metastases to three

months, the average survival from the onset ofdysphagia in those undergoing resection is 12months, including those dying from operation, andthis does not differ from the duration of survivalin those not submitted to operation; excludingthose dying from operation the average survivalfrom the onset of dysphagia is 20 months. Theaverage duration of dysphagia before resectionwas four months.

Thus, to summarize, 26 patients were found tohave an undifferentiated carcinoma of the oeso-phagus. The sexes in this group were equallyrepresented. Thirteen were not submitted tooperation because of metastases and in theseaverage survival from the onset of dysphagia was11 to 12 months. Resection was undertaken in 12,in 10 with glandular metastases. One patientsurvived more than five years, four (33%) diedfrom operation, one is alive with metastases, andthe remainder have died with metastases. Theaverage estimated survival time from the onset ofdysphagia in those dying from operation andmetastases is no different from those not resected,but this average is heavily biased by the operativemortality. Resection is justified by three patientssurviving more than 18 months and a fourth sixyears from a series of 12.

ADENOCARCINOMA IN THE OESOPHAGUS ANDPROXIMAL STOMACH

Of the 700 patients found to have a primarycarcinoma of the proximal alimentary tract duringinvestigation for dysphagia, 272 (39%) were shownto have an adenocarcinoma. The age and sexdistribution of these patients are shown in Fig. 17.The incidence amongst men is a little more thantwice that amongst women. The tumour wasvisible oesophagoscopically in 180 and was notvisible through the oesophagoscope in 92.Resection was undertaken in 170 patients (62%).an incidence of resectability the same as that forsquamous carcinoma of the oesophagus. Forty-six patients were submitted to thoraco-laparotomywith a view to resection but were found to haveeither an irresectable tumour or one from whichdissemination was so gross that resection was un-justified because of the relative mildness ofdysphagia. Operation was therefore frustrated in17% of patients with adenocarcinoma in com-parison with 12% of patients with squamouscarcinoma. In 56 patients (21%), thoracotomywas not undertaken because they had metastaseswhen first seen, or because they were too old ortoo ill or declined treatment, or because they diedduring the course of investigation.

242



j.com/


eptember 1961. D

ownloaded from



MALE (

FEMALE (8)

SO 85 8920 25' 30 35 40 45 50 55


EI .a .ML=I..

MALE(@

IFEMALE(3

m22 30 35 40 45 50 55 bO 65 70 75 80 85


ADENOCARCINOMA NOT TREATED BY OPERATION.-Fifty-six patients with an adenocarcinoma inthe oesophagus or proximal stomach were notsubmitted to operation. The youngest was 45years of age and the oldest was 89. The agerange in the two sexes is shown in Fig. 18. Allthese patients are dead. The tumour was visibleoesophagoscopically in 26 and was beyond therange of the oesophagoscope in 30.

All these patients had dysphagia. Since opera-tion was not undertaken the duration of dysphagiain this group is synonymous with survival timefrom the onset of dysphagia (Fig. 19). Twopatients died within two months of the onset ofdysphagia. One survived for 26 months and theaverage duration of dysphagia was seven months.The average duration of dysphagia before

0

:

0-

I.

0

I

z0

I


admission to hospital (Fig. 20) was between fourand five months. There was no significantdifference in the duration of dysphagia in thosewho had visible tumours and in those who did not.

(A

zLU

IL

0z

II

9753

CA

l-z

<8

u. 60

z2

5

3j(A

zLU

FLL

0

z64

2

. ..-= m

I~~~~~~~~~~~I

-t

243-

I

0 s 0m



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

IN HOSPITALz0 LiI-

.-J

towLI M ADENO-CARCINOMAw c

z _ NOT SUBMITTEDO'n TO OPERATION

n tI

YOUNGEST+- PATIENTS ARRANCACCORDING TO AGE

FIG. 20.-Each vertical

Dysphagia was the first symptom in 60% of thisgroup, and was as often the first symptom in thosewho did not have visible tumours as in those whohad. Forty per cent. had symptoms before theonset of dysphagia-loss of appetite and weight,vague abdominal discomfort, and in one patienthaematemesis-and these symptoms preceded theonset of dysphagia by three to six months. Theduration of dysphagia before admission to hospitaland survival following investigation is shown inFig. 20.The reasons for not undertaking resection in

these patients were as follows:Fourteen patients were found clinically to have

hepatic metastases and in all these the presenceof hepatic metastases was confirmed at necropsy.

Fourteen patients had cervical glandular meta-stases, 10 on the left and four on the right;histological confirmation of spread of the tumourto cervical glands was obtained in all these cases.

Seven patients had evidence of transcoelomicspread, five presenting with ascites and two witha mass in the pelvis felt on rectal examination.Two patients had histologically proven

cutaneous metastases.One patient had metastases at the umbilicus.

GE[E

hIr

IIII|I | 1I| IN HOSPITAL

D -OLDEST

ie represents a patient.

One patient presented with hoarseness of voicein addition to dysphagia and in him the leftrecurrent laryngeal nerve was interrupted.Three patients had radiological evidence of

diffuse pulmonary metastases and in two of thesethis diagnosis was later confirmed at necropsy.Three patients died during investigation from

massive bleeding from a fundal carcinoma; atnecropsy all had evidence of widespread dis-semination of the tumour.One patient died from pulmonary embolism

during the course of investigation; in her thetumour was seen to have invaded the posteriorabdominal wall at necropsy.

Eight patients were not submitted to operationbecause of their advanced years, their poorgeneral condition, or because of so limited arespiratory reserve that they were thought unableto tolerate thoraco-laparotomy.Two patients declined treatment.Of the 26 endoscopically visible tumours, 22

were visible at the distal end of the oesophagus,between 35 and 41 cm. from the upper alveolus.In one patient, in addition to the distal tumour.adenocarcinoma was obtained from a biopsy takenfrom a nodule at the level of the aortic arch;

244



j.com/


eptember 1961. D

ownloaded from



MALE 0

m m .Lm mm -M

55 60 65 70 75

FEMALE (0

mI

s0 85

AGE IN YEARSFIG. 21.-Age and sex distribution of patients with tumours visible and not visible endoscopically.

this was regarded as evidence of submucouslymphatic extension of the tumour since it was

separated by normal squamous epithelium fromthe more distal tumour. In four patients thetumour was seen endoscopically 24 to 27 cm. fromthe upper alveolus. In these patients there was

radiographic evidence of a hiatus hernia, and atnecropsy the tumour was growing in a viscuswhich had the appearances of stomach while lyingin the mediastinum.Only one patient was irradiated in this group;

resection was not undertaken because his respira-tory reserve was so small, and he died a year afterthe onset of dysphagia, five months after com-

pleting a course of radiotherapy. The tumourwas intubated with a Souttar's or Simmonds' tubein 12 of those with endoscopically visible tumoursto enable liquids to be swallowed. Tumours inthe gastric fundus are rarely suitable forintubation, but these patients seldom have totaldysphagia.

Information regarding the mode of death inthose who died at home is not accurate, but manyappear to have died from the general effects ofwidely disseminated malignant disease rather thanfrom some specific manifestation of the tumour,such as aspiration pneumonia because of totalobstruction of the oesophagus or invasion of therespiratory tract. No patient was discarded fromoperation because of evidence of invasion of thetrachea or main bronchi, while this was thecommonest single reason for not undertaking an

operation in patients with squamous carcinoma ofthe oesophagus.

Fifty-six patients with adenocarcinoma of theoesophagus were not submitted to operation. Theaverage survival from the onset of dysphagia was

again seven months; hepatic and glandular meta-stases were the commonest reasons for not under-taking resection.

ADENOCARCINOMA FOUND IRRESECTABLE AT

THORACO-LAPAROTOMY.-Forty-six patients (sevenwomen and 39 men) who were submitted tothoraco-laparotomy were found to have an

irresectable tumour. In 24 the tumour had beenvisible endoscopically; in 22 the oesophagus was

normal and the operation was undertaken becauseof the symptom of dysphagia in conjunction withbarium evidence of a tumour in the gastric fundus.The age and sex distribution of the patients in

whom the tumour was endoscopically visible andof those in whom the tumour was not visible are

shown in Fig. 21.

Visible Endoscopically.-In those patientswith endoscopically visible tumours resection was

not undertaken because of extensive hepatic meta-stases in six, because of diffuse peritoneal spreadin six, and because of both diffuse peritonealspread and hepatic metastases in four. Thetumour was fixed to the posterior abdominal wallin two and in six invasion of the aorta and thelung precluded resection. In one, a man of 50,thoraco-laparotomy was undertaken in thepresence of known cervical glandular metastasesin an attempt to restore swallowing.To restore swallowing, the tumour was bypassed

in six patients, by the use of jejunum in five andof stomach in one. Five of the six in whom a

bypass was made died from operation (Fig. 22a),four of these from leakage from the anastomosisand one from respiratory infection. Two othersdied as a direct consequence of operation, bothfrom respiratory infection. The operativemortality in this group of 24 patients is 29%.Swallowing was restored in the patient whosurvived a bypass procedure. This was the man

with cervical glandular metastases and he diedtwo months after operation from progress ofdisease. In the youngest patient in this group, a

z 3I

U101.60 4

2

. m I

I

- |~~~~~~~~~~~~~13r% li C., I^ 12C A^ A C C^

245

.du dp iri ib 4U 45 50



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

-4

D

An

4

wLL>

0.p

tol_

F- WzL

0

ZI

zo

cc

I~0.

Lootn <I 0

Z IO X.M wn>.z 0

Z >O p

cc w

0 O

(a) (b)FIG. 22.-Each vertical line represents a patient, and these are arranged youngest on the left, oldest on the right. The dots in

the transverse gap labelled operation represent patients who have died as a consequence of operation.

man of 35, swallowing was restored by irradiationafter thoraco-laparotomy and this patient is stillalive seven months after thoraco-laparotomy. Ofthe patients who survived operation, later to diefrom progress of disease, none survived longerthan seven months. The duration of pre-operativedysphagia and post-operative survival in thisgroup of patients is shown in Fig. 22a. In the 23patients who have died dysphagia had been presentin two for only two months; both of these were

operative deaths. One patient survived 22 monthsfrom the onset of dysphagia (Fig. 23a). Theaverage survival from the onset of dysphagia wasnine months. The average duration of dysphagiabefore operation was seven months. The patientstill alive has survived 31 months from the onsetof dysphagia.

Not Visible Endoscopically.-In those patientsin whom the tumour was not visible endoscopicallyresection was not undertaken because of extensivehepatic metastases in six, because of diffuse peri-

toneal spread in nine, and because of both diffuseperitoneal spread and hepatic metastases in five.In two patients fixation of the tumour to theposterior abdominal wall precluded resection.There was no intrathoracic contraindication toresection. In three swallowing was restored bybypassing the tumour with jejunum (Fig. 22b).Two of these survived for two months and one forfive months after operation, dying then fromprogress of disease but still able to swallow. Threedied as a consequence of operation, two fromrespiratory infection and one from an undeter-mined cause, an operative mortality in this groupof 22 patients of 14%. All these 22 patients aredead. One survived only two months from theonset of dysphagia, dying as a consequence ofoperation. The longest survivor lived for 14months from the onset of dysphagia (Fig. 23b).The average survival from the onset of dysphagiawas seven and a half months. The averageduration of dysphagia before operation was fourmonths.

OPEPRATIVE. MORtTALITY3

. <

0.0.~

JAII0oI~.c~ I t

I 1 111111 1I

ADENO-CARCINOMANOT VISIBLE OESOPHAGOSCOPICA-LLY FOUND TO BE IRRESECTABLE

L1JL50 - 9 0 - 7a1170 2MjLj36 49 -So +

246



j.com/


eptember 1961. D

ownloaded from


in('a

U)I- C

Z It

Z in4

U,

PATIEN(a)


U)

z00

CIV

In_

Q

is 20 2S 5 10ITS

FIG. 23.

Three of the patients with oesophagoscopicallyvisible tumours found irresectable at thoraco-laparotomy had large hiatus herniae and in thesethe tumour was visible endoscopically at the lowerborder of the aortic arch. In one patient in whomthe tumour was visible at 22 cm. from the upperalveolus there was no evidence of hiatus herniaat thoracotomy, but there was a very large tumourextending from the gastric fundus to above theaortic arch and invading the aorta.To sum up, 46 patients with an adenocarcinoma

in the oesophagus or proximal stomach werefound at thoraco-laparotomy to have tumoursunsuitable for resection. The operative mortality

In

wzS- 3

U-0

cox 4

D 2z

was 22%. The average survival from the onsetof dysphagia was a little longer than in those notsubmitted to operation.

ADENOCARCINOMA RESECTEDNot Visible Oesophagoscopically.-In 40 of the

170 patients in whom an adenocarcinoma wasresected the tumour was not visible oesophago-scopically. Sixteen of these were women and 24were men. The age incidence in the two sexes isshown in Fig. 24. The average duration ofdysphagia in these 40 patients was five months,not significantly different from the duration inthose presenting initially with a contraindication

MALE 24

* U

FEMALE 16

0 25 30 35 40 45 50 55 60 65 70 75 80 85


247

25ATIENTS

(b)

Jm- m _

I

.r



j.com/


eptember 1961. D

ownloaded from


248

p .;

-4

PF-;e

B. T. LE ROUX

cx

>0> 004Dul la° @

LU noU X La. .<0..

l Ir T: I--r-IrI-I iII,.t-.t&[.. XtTr ;, I 'I

iI

l

I- -I1I

ill ! l 11 I|i| 1uu1I

i,- '. i . N-NOIMA RESECTED

t.. ESESQPH AGO S COP IC ALMORTALITY

128°loI0;40

;

FIG. 25.-Each vertical line represents a patient, and these are arranged the youngest on the left andthe oldest on the right. The dots in the transverse gap labelled operation represent patients dyingfrom operation. Vertical red lines represent patients at present alive or patients who have diedfrom causes other than metastases more than three years after operation. The vertical black linesabove the operative gap represent patients who have died from metastases.

to resection or those found unsuitable for resectionat thoraco-laparotomy. An oesophago-gastro-stomy was made in 19 of these and an oesophago-jejunostomy in 21.

Eleven patients died as a consequence ofoperation, an operative mortality of 28% (Fig. 25).Four of these patients died because of disruptionof the anastomosis, in all four instances an

oesophago-jejunostomy. Three patients died fromrespiratory infection. Two died within 24 hours

of operation and their behaviour during thisperiod suggests that they died from surgical" shock." They remained hypotensive throughoutthe short period of post-operative survival andfailed to respond to resuscitative measures. Onepatient died 10 days after operation with a rightchylothorax which probably contributed signifi-cantly to her death. One died from intestinalobstruction; the obstruction was to the ante-colicloop of jejunum used to re-establish continuity.

cxC_1

-1

-- -I I I t

I

-L copyright. on June 11, 2020 by guest. P


j.com/


eptember 1961. D

ownloaded from



Resection was undertaken in six patients whowere found at operation to have hepatic meta-stases, in three with peritoneal metastases and ina further three with metastases in the omentum.In these 12 histological evidence of spread wasobtained. It is anomalous to have abandonedresection in many described in the previous sectionbecause of hepatic or peritoneal metastases, andyet to undertake resection in the presence of thesemanifestations of spread in the present series of40 patients. The anomaly is explained by theseverity of the dysphagia in a few patients inwhom resection was undertaken-some of themwere unable to swallow even liquids withoutdifficulty-and by the extent of neoplastic spreadin others, resection having been undertaken inthe presence of a single small hepatic metastasisand abandoned because of multiple large hepaticmetastases.

In 20 patients in whom resection was under-taken glandular metastases were found either inthe abdomen or in the mediastinum. Since twohad both hepatic and glandular metastasesresection was undertaken in only 10 patients whowere not shown at the time of operation to havespread of the tumour beyond the stomach, i.e.,75% of patients had evidence of dissemination.

Twenty-nine patients survived operation (Fig.25). Eighteen of these died subsequently frommetastases, 12 in the first year after operation,four in the second year, and two in the thirdyear. Fourteen of these 18 patients were shownat operation to have metastases and four werebelieved not to have metastases at the time ofoperation. These four died respectively withcerebral metastases, with a recurrence in themediastinum producing dysphagia, and two withhepatic metastases. In four of the 18 patientsdying from metastases dysphagia recurred beforedeath six, 21, 24, and 31 months after operation.Nine patients are alive at the present time (Fig.

25). Three are alive more than 10 years afterresection. Two are alive, one six and one fiveyears after resection, and four are alive less thantwo years after resection. One of the patientswho is alive more than 10 years after resection hadglandular metastases at operation. Three of thefour alive less than two years after resection hadglandular metastases at the time of operation, asdid the patient alive five years after resection.Four, at present alive, did not have evidence ofspread at the time of operation.Two patients have died from causes other than

operation or metastases. A woman in whomresection was undertaken at the age of 78 died

eight years later in cardiac failure. She did nothave metastases at the time of operation. A manin whom resection was undertaken at the age of70 died four years later from a combination ofalcoholism and bronchopneumonia. He had notbeen shown to have metastases at the time ofoperation.

All these 40 patients had had dysphagia beforeoperation. Other symptoms such as loss of weightand appetite and abdominal pain had precededdysphagia in 30 of them, in one by a year and inthe others from three to six months. Hiccuphad been a symptom in seven before theiradmission to hospital. One woman had perniciousanaemia.The average survival from the onset of

dysphagia in the group of 29 patients dying eitheras a consequence of operation or later from meta-stases was 12 months, the longest survivor dying36 months from the onset of dysphagia (Fig. 26a).The average survival from the onset of dysphagiaof the 18 patients who died after operation frommetastases was 16 months (Fig. 26b). Whetheror not operative deaths are included when calculat-ing survival from the onset of dysphagia in thosein whom a resection was undertaken and who arenot at present alive or who did not die from someother cause, the survival in this group from theonset of dysphagia is longer than the seven monthsin those who were not submitted to operation orthe nine months in those in whom the tumour wasfound irresectable, and is significantly longer inthose dying from metastases. Of the 40 resections,25 were undertaken more than five years ago, andof these 25 patients seven died from operation.The five-year survival rate amongst the patientswho may so far have had the opportunity ofsurviving five years is therefore 24% of thosesubmitted to operation and 33% of those survivingoperation.

In summary, 40 adenocarcinomas not visiblethrough the oesophagoscope were resected.

Eleven patients (28%) died as a direct conse-quence of operation.

Eighteen patients (45%) died from metastasesthree to 36 months after operation, an averagepost-operative survival of 16 months. In fourdysphagia recurred.

Six patients (15%) were alive five years afteroperation, and one of these died eight years afteroperation from an unrelated cause.One patient died four years after operation from

an unrelated cause.Four patients are still alive, but less than two

years after resection.

249



j.com/


eptember 1961. D

ownloaded from


B. T. LE ROUX

0It

5 [0 [5

(a)

LL0

w1U)z0

wI

0

() IZ 0U)'O0.

>

0 0

z

20 25 30NUMBER OF PATIENTS

FIG. 26.

5 10 15 20

(b)

Visible Oesoplago.s opi all/ - One hundredand thirty of the 170 adenocarcinomas resectedwere visible oesophagoscopically, 48 in women

and 82 in men. The average duration of pre-operative dysphagia in these 130 patients was a

little less than five months: nearly six months inthe women and just over four months amongst themen. Here again there is no significant differencein the duration of pre-operative dysphagia fromits duration in those in whom a contraindication to

thoracotomy was found when first they presentedor in those found to be unsuitable for resection at

U)

7z

43I

LI)im

z

thoraco-laparotomy. The age incidence amongstthe two sexes is shown in Fig. 27.

THE OPERATION. An oesophago-gastrostomiiywas made in 108 patients and an oesophago-jejunostomy in 22. In all but three the approachwas through a left thoraco-laparotomy. In all butfour of those approached froni the left theanastomosis was made at a level which variedfrom the lower pulmonary vein to the lowerborder of the aortic arch. In four patients in

whom a left thoraco-laparotomy was made the

MALE (i)

FEMALE (i)6

4

2

20 25 30 35 40 45 50 55 60 65 70 75 80 85

AGE IN YEARS

FIG. 27.

LA-0

(/)z0

wI

H

0LL

U)

Iz0

z

IJ

4x

0.T-

0.

.. NI

I

2)50

LL



j.com/


eptember 1961. D

ownloaded from



stomach was brought up lateral to the aortic archand a supra-aortic oesophago-gastrostomy wasmade. A right thoraco-laparotomy was made inthree patients and in these the oesophago-gastro-stomy lay above the level of the azygos vein.Seven of the 130 resections may therefore be saidto have been for " high " tumours, in comparisonwith 10 tumours lying closer than 30 cm. tothe upper alveolus amongst 50 patients withendoscopically visible adenocarcinomata either notsubmitted to resection because of a contraindica-tion or found unsuitable for resection at thoraco-laparotomy. In both squamous-cell carcinomaand adenocarcinoma there is a higher rate ofirresectability amongst the high tumours.

_ ~~~~~~~~~~~~~~~~~~I

,, YEARS--C.,

c: ,

I-0

44

7-

wI YEAR

I< <

0.

>0 r<a.-

Z) g ADENO-

z >.LD X ,_VISIBLE1w

OPERATIVE MORTALITY.-Thirty-two patientsdied as a consequence of operation (10 womenand 22 men), an operative mortality of 25%(Figs. 28 and 29). The operative mortalityamongst men (27%) was higher than amongstwomen (21 %). The overall operative mortalitywas lower than in resection for squamouscarcinoma, and the higher mortality amongstwomen with squamous carcinoma was herereversed. Amongst women most of the operativedeaths were in the age group 60-70. Amongstmen most of the operative deaths were in thoseover 65 and, as with squamous carcinoma, 50%of men over 70 submitted to resection died directlyas a consequence of operation. None of the

FIG. 28.-Each vertical line represents a patient, and these are arranged the youngest on the left andthe oldest on the right. The dots in the transverse gap labelled operation represent patients dyingfrom operation. Vertical red lines represent patients at present alive. The vertical black linesabove the operative gap represent patients who have died from metastases.

251



j.com/


eptember 1961. D

ownloaded from


B T. LE ROUX

_7.IeT;

soi

a . lPTAia0o a

aa3YE A ,'

I tj

. ,EA.r_.4

'E A i l'~~-;txt--1~7|T-r~T t-.or-;-4X- ;- ..

£ 6'

2ADENO-'APCINOMA tESE'VISIBLE ,)ESOPH-AGOSCOP('ASi.i.

-8gi-....;-- M.

FIG. 29.-Each vertical line represents a patient, and these are arranged the youngest on the left and the oldest on the right. Thedots in the transverse gap labelled operation represent patients dying from operation. Vertical red lines represent patientsat present alive. The vertical black lines above the operative gap represent patients who have died from metastases.

patients with "high" adenocarcinomas of theoesophagus died from operation.

Nine operative deaths, one in a woman, werea consequence of leakage at the anastomosis; ineight from an oesophago-gastrostomy and in onefrom an oesophago-jejunostomy. Four patients,all men, died from haemorrhage, the result oferosion of the aorta. This complication ofoesophagectomy is discussed in detail elsewhere,but it is closely related to leakage from theanastomosis. Four patients, three of them women,died from pulmonary embolism, and four others,all men, died from pulmonary infection. Threepatients died as a consequence of a cardiac lesion,two from electrocardiographically and histologi-cally established myocardial infarction, and onein cardiac failure from an undetermined cause.Five patients died during the first 24 hours afteroperation, three from a cause which includesoperative shock (they remained hypotensive andfailed to respond to resuscitative measures) andtwo from reactionary haemorrhage, in one ofthese from an inadequately ligated left gastric

artery and in one from an undetermined source.Three patients died from infection: one from apyocyaneus peritonitis during a period when thisorganism was uncontrolled in the unit, laternecessitating closure of the operating theatre fora month; one from an empyema with an intactanastomosis, and one from necrotizing enteritis.METASTASES AT OPERATION.-In 83 patients

(64%) histological confirmation of glandularmetastases in the abdomen or in the mediastinumwas obtained. Twenty patients in whom hepaticmetastases were found at operation were sub-mitted to resection to restore swallowing, andseven of these 20 did not have glandular meta-stases. Peritoneal spread was found in three,extensive invasion of the right lung in two,invasion of the transverse colon necessitatingpartial colectomy in one, and in two macroscopictumour was left in the mediastinum. All the last-mentioned eight patients had, in addition, glandularmetastases. Seventy per cent. of patients withendoscopically visible adenocarcinomata under-going resection had dissemination at operation,

252

I

a

.i

9



j.com/


eptember 1961. D

ownloaded from



resection being undertaken in only 40 out of thetotal of 130 in the absence of histologically provendissemination of tumour, and eight of these 40died as a consequence of operation. Ninety-eightsurvived operation, 32 of these without evidenceof spread of the tumour at operation. Of these32, 23 died subsequently from metastases and nineare alive at the present time.SYMPTOMS: OPERATION UNDER UNUSUAL CIR-

CUMSTANCES.-All of the 130 patients withendoscopically visible adenocarcinomas haddysphagia before operation. Other symptoms,such as loss of weight and appetite, had precededdysphagia in 45, a number significantly smallerthan amongst the adenocarcinomata not visibleendoscopically. In two patients symptoms otherthan dysphagia had preceded dysphagia by a year;in the remainder other symptoms precededdysphagia by three to six months. Hiccup hadbeen a symptom in eight before admission. Infour men there was a long history of intermittentdysphagia-15 to 25 years-with a recent changein the character of the dysphagia. In one of thesethe radiographic appearances were those of long-standing cardiospasm, and in the other three therewere the radiographic and operative features of anhiatus hernia; in these three patients thecarcinoma may have developed in relation to abenign peptic stricture at the oesophago-gastricjunction. In six others there was a large hiatushernia visible radiologically and demonstratedwith barium.

In a man of 55 resection was undertaken torestore swallowing despite the presence of histo-logically established cervical glandular metastases.This man died 22 months after resection, anddysphagia recurred three months before death;he had swallowed normally and lived his normallife for 18 months after resection. One womanwas shown pre-operatively to have an unusuallylarge left pericardial coelomic cyst.RADIOTHERAPY.-Two men were referred for

resection after irradiation. One had been exploredthrough the abdomen elsewhere in 1955, and hadbeen regarded as having an irresectable tumourwhich was irradiated with some relief ofdysphagia. Dysphagia became complete in 1957,and the tumour was resected in the presence ofmediastinal glandular metastases. In January,1960, he developed intestinal obstruction, the con-sequence of herniation of small gut through thegastric hiatus in the diaphragm; the obstructionwas relieved and the hiatus repaired, and he iswell and swallows normally. No evidence ofspread was found at this last operation. Experi-

ence with this man and others, explored throughthe abdomen and regarded as having irresectabletumours, most of them not irradiated, and laterreferred for the relief of total dysphagia, hasconfirmed the contention that proximal gastrictumours cannot be properly assessed through apurely abdominal approach. The other patientwhose tumour was irradiated before resection wasyoung (36 years) and was regarded elsewhere ashaving an irresectable tumour without exploration.

In him irradiation improved swallowing for ayear. When dysphagia became intolerable thetumour was resected but he died from metastasesfive months later. One patient was irradiated forrecurrence of dysphagia 30 months after resectionand survived irradiation by only three monthswithout relief of dysphagia.NON-FATAL COMPLICATIONS.-Major post-opera-

tive complications with survival were few. Twosubphrenic abscesses and one empyema weredrained. One patient developed clinical andradiographic evidence of leakage from theanastomosis into the right pleural space; thepleural space was intubated, and methylene bluetaken by mouth appeared rapidly in the water-seal. This patient was managed conservatively;hydration was maintained by venoclysis for somedays and then a jejunostomy was made; hesurvived and is still alive 10 years after resection,being one of three survivors from clinicallyrecognized leakage from the anastomosis. Oneman, aged 45, an unusually muscular individualand an active mountaineer, six months afteroesophago-gastrectomy with gastric replacement,developed intermittent severe substernal painrelieved by vomiting; these symptoms wererelated to exercise. He was shown to have amobile intrathoracic gastric tube and his symptomswere attributed to recurring torsion of this tube.The stomach was fixed at a second operation, atwhich no metastases were found, and he is wellsix years after his original resection.THE FATE OF THOSE WHO SURVIVED OPERATION.

-Ninety-eight patients survived operation, 32 ofthese without dissemination of tumour at the timeof operation. Eighty-two patients, 47 men and35 women, died subsequently from metastases.Two men died during the fourth year after opera-tion from non-metastatic causes; 11 men andthree women are at present alive.Of the 82 patients (29 men and 12 women) who

died from metastases after resection, 41 diedduring the first post-operative year (Figs. 28 and29). During the second year after resection 29patients (13 men and 16 women) died from meta-

253



j.com/


eptember 1961. D

ownloaded from


-b415i.11~~~~~~~~~~~~3

i 2

(a)

i

0

zI

,,,00I- z

ALJ

a

26

24 -22fl201

4',,

4I 4

031

Z 3

,, 3

3AI 31(I' 21

0 2i242i

zo 1 e

I;a

-jLca

35 40 45


stases, and during the third year after resectionII patients (four men and seven women) diedfrom metastases. One man died from metastasesduring the fourth year after resection, and twoother men died during the fourth year afterresection, one from cerebral thrombosis and onefrom coronary thrombosis. In both these therehad been spread at the time of resection.

Eleven men and three women are alive at thepresent time. Of these 14 patients, four are alivea year or less following resection, and two othersare alive less than two years after resection. Fiveof the six alive less thani two years after resectionhad glandular metastases. One man is alive fourvears after resection and seven patients (five menanid two women) are alive five years or more afterresection. Two of these have very nearly com-

pleted their fifth year after resection, one is aliveseven years and one eight years after resection,two are alive 10 years, and one is alive I I yearsafter resection. None of the patients survivingfive years or longer after resection had dissemina-tion of tumour at the time of resection. Sevenpatients, of 61 who underwent resection and of48 who survived resection more than five yearsago, may be regarded as five-year survivors, a five-year survival rate of 12', calculated against thenumber submitted to operation and of 15",calculated against the number surviving operation.Continuity was re-established in five. who havesurvived five years or more, using stomach, andin two using jejunum. Continuity had been re-

(c)

established using jejununm in onie out of everx~~~~~~i Jfive resections.

RECURRENC-E OF DYSPHAGIA. Dysphagiarecurred in 20 of the 98 patients who survivedoperation, and all 20 had spread at the time ofresection. Eight died with recurrence of dysphagiaduring the first year after resection, seven diedwith recurrence of dysphagia during the secondyear after resection, and one is alive during thesecond year after resection, but has a recurrenceof tumour, and four died during the third yearafter resection with recurrence of dysphagia. Innearly all these patients dysphagia recurred lessthan two months fronm the time of death, andrecurrence at the anastomosis was only one ofseveral evidences of recurrence or spread in thesepatients. One patient alive five years afterresection has a peptic stricture at the oesophago-gastrostomy requiring dilatation at about yearlvintervals.

INCIDENCE OF EFFECTIVE PALLIAJ ION. Thepalliative value of resection in patients dying fromzmetastases must be assessed against the average

survival of those patients with an adenocarcinonmaproducing dysphagia who were not submitted tooperation. In this group of 56 patients the averageduration of dysphagia before investigation was

shown earlier to be four to five months, and

average survival from the onset of dysphagia wasseven months. Thirty-two patients died fromoperat;on: in these the average durationi of

254 B. T. LE ROUX



j.com/


eptember 1961. D

ownloaded from



dysphagia before operation was four months (Fig.30a), and since they died from operation, averagesurvival from the onset of dysphagia is also fourmonths, and no increase of survival time accruedto them. Forty-one died from metastases duringthe first post-operative year, and of these 41, eightdied with recurrence of dysphagia. Averagesurvival from the onset of dysphagia in this groupof 41 was 1 1 months, the longest survival being23 months from the onset of dysphagia and theshortest four months (Fig. 30b). As withsquamous carcinoma, there is no increase insurvival time and resection in 73 patients (32+41)did not prolong life and can be regarded only asan incident in the natural course of the disease,i.e., in .56% of resections undertaken for adeno-carcinoma visible through the oesophagoscopethe natural history of the disease was not sub-stantially altered. Twenty-nine patients died frommetastases during the second year after resection,seven dying with recurrence of dysphagia.Average survival from the onset of dysphagia inthis group was 22 months, the longest survivalbeing 44 months and the shortest 17 months (Fig.30c). Significant increase in the duration ofsurvival can be claimed for these patients and forall those dying subsequently from metastases.

Summary.-One hundred and thirty oesophago-scopically visible adenocarcinomas of theoesophagus were resected.

Twenty-five per cent. of the patients died as aconsequence of operation; 70% had evidence ofspread at operation.

Sixty-three per cent. of those submitted tooperation (84% of those who survived operation)died from metastases, exactly half of these duringthe first 12 months after resection.

Five per cent. of those submitted to operation(7% of those who survived operation) are alivefive years or more after resection.

Seven per cent. of those submitted to operationare alive less than five years after operation orhave died from non-metastatic causes.

SUMMARYSeven hundred patients with dysphagia were

found to have as the cause for this symptom a

carcinoma in the proximal alimentary tract.Twenty squamous carcinomas involved the

hypopharynx; 19 of these were in women, and of11 patients submitted to resection four are alivefive years later.Three hundred and eighty-two patients with

squamous carcinoma of the oesophagus wereinvestigated and 62% were submitted to resection.

u

Invasion of the trachea and bronchi was thecommonest reason for not undertaking resectionin the remainder. Of those resected, 34% diedfrom operation and 10% were alive five yearslater. In 60% of those submitted to operation,the natural course of the disease, as judged bythe behaviour of those patients considered unsuit-able for resection, was unaltered. The prognosiswas worse in those with high oesophagealtumours.

Twenty-six patients with undifferentiatedcarcinomas were investigated, 12 were submittedto resection and one of these survived more thanfive years.Two hundred and seventy-two patients with

adenocarcinomas in the oesophagus or proximalstomach were investigated and 62% of these weresubmitted to resection. Hepatic and glandularmetastases were the commonest evidences ofspread which precluded resection in the remainder.Of 40 patients with tumours in the gastric funduswho were submitted to resection, six were alive fiveyears after operation, 11 having died from opera-tion. Resection was undertaken in 130 patientswith an adenocarcinoma visible in the oesophagus;25% died from operation, and 5% were alive fiveyears later. In 560% of visible adenocarcinomasresected the natural course of the disease wasunaffected. Of all patients with adenocarcinomasubmitted to resection, 8% were alive five yearsafter resection.Four hundred and eighteen oesophago-

gastrectomies were undertaken. The operativemortality was 30% and anastomotic failureaccounted for nearly one-third of these deaths.Of the 418 resections, 220 were undertaken more

than five years ago, and, of these 220 patients, 60died from operation. The five-year survival rateamongst those who may have had the opportunityof surviving five years is 17% of those submittedto operation and 23 % of those who survivedoperation. Twenty-nine per cent. of the 418patients submitted to resection or 37% of thosewho survived resection died from metastasesduring the first post-operative year. Effectivepalliation was achieved in 33% of those submittedto operation or 43 % of those who survivedoperation. Forty per cent. of those submitted toresection derived significant or lasting benefit fromoperationi, and 60% did not. Dysphagia due torecurrence developed in 14% of those submittedto operation, or 19% of those who survivedoperation.

REFERENCESle Roux, B. T. (1962a). J. roy. Coll. Surg. Edinb. To be published.--(1962b). Brit. J. Surg. To be published.

255



j.com/


eptember 1961. D

ownloaded from


AN CASES OF CARCINOMA OF THE …Thorax(1961), 16, 226. AN ANALYSIS OF 700 CASES OF CARCINOMA OF THE...

Documents

Transcript of AN CASES OF CARCINOMA OF THE …Thorax(1961), 16, 226. AN ANALYSIS OF 700 CASES OF CARCINOMA OF THE...