bronchial carcinoma - thorax.bmj.com · to record in all other aspects of the anamnesis, and ......

Thorax (1968), 23, 136.

Bronchial carcinomaB. T. LE ROUX'

From the Thoracic Unit, Edinburgh

From a retrospective review of 4,000 patients with bronchial carcinoma, managed in the ThoracicSurgical Unit in Edinburgh, the salient features of presentation, surgical management, and thenatural history in those unsuitable for surgical management are reviewed.

From a retrospective analysis of the case recordsof 4,000 patients managed over a period of 15years and established sooner or later to havebronchial carcinoma, it is possible to make certaingeneralizations.

Bronchial carcinoma is more common in men;only 10% of patients are women. The highestincidence of the disease was in the sixth andseventh decades; 10% of patients were 70 yearsor older when they were first seen; about one-fifthof women and fewer than one-sixth of men withbronchial carcinoma were younger than 50 yearsat the time of their first investigation. While theyearly incidence of the disease did not appear tobe changing significantly, the incidence amongthose over 70 years was rising and among thoseunder 50 years might have been falling. An in-creasing incidence of the disease in the aged wasprobably explicable on the grounds of a generallyaging population; a diminishing incidence in theyounger age group is recorded in many series, andan explanation is lacking.The survey, of which this is a summary, was

made in a region where coal mining is a commonoccupation, and of the patients in this series 15%were coal-miners, an occupational incidence whichis disproportionately high. It is generally main-tained that coal mining does not predispose speci-fically to bronchial carcinoma. That a largenumber of coal-miners were found to havebronchial carcinoma may be related to the factthat theirs is an occupational group regularlyexamined radiographically and probably moreoften than are most other occupational groups.Of every four patients with bronchial

carcinoma, three were known to have beencigarette smokers, and most of these were heavysmokers. Of the total number of patients, 3%smoked only a pipe and 4% claimed never to have

1P.resent address: Thoracic Unit, Wentworth Hospital. Durban,Natal, South Africa

smoked: among the women in the series 15 %claimed never to have smoked. Latterly thereemerged a noticeable reluctance among patientsto admit that they smoked, and it is now common-place for patients, when asked the question 'Doyou smoke?' to answer, quite honestly, that theydo not on the grounds that -they had stoppedsmoking with the onset of symptoms, or with thefinding of an abnormality on a chest radiograph,or even on the way to hospital on the morningof interrogation. In one-third of those who did notsmoke the tumour was squamous. The smokinghabits of 20% of the patients were not known, anincidence of omission in the records probablysignificant of the percentage of failure adequatelyto record in all other aspects of the anamnesis, andperhaps an argument in favour of demanding,from rotational junior staff who interrogatepatients, the completion of a detailed pro formarather than the recording of answers to thosevariable questions which they may remember toask. While it has been established that a largeproportion of those who developed bronchialcarcinoma had smoked cigarettes for some years,the precise relationship between cigarette smokingand bronchial carcinoma was not established. Itis clear that not smoking is not a guaranteeagainst the development even of squamous carci-noma, land that cigarette smoking is not the onlyfactor in the genesis of bronchial carcinoma.

Patients shown by investigation to havebronchial carcinoma presented in one of fiveways: (1) with one, some, or all of the cardinalsymptoms of respiratory disease-cough, sputum,haemoptysis, chest pain, dyspnoea, and wheeze:(2) without symptoms and because of anabnormality detected on a chest radiograph madefor routine purposes during mass miniatureradiographic surveys; during radiographicsurveys among coal-miners and other occupationalgroups; during the routine follow-up of patients

136

on 17 March 2019 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.23.2.136 on 1 M

arch 1968. Dow

nloaded from

http://thorax.bmj.com/

Bronchial carcinoma

known to have had pulmonary tuberculosis; or forinsurance or immigration purposes; (3) becauseof evidence of extension of or metastases frombronchial carcinoma, e.g., cerebral or osseousmetastases, hoarseness of voice, obstruction ofthe superior vena cava, or jaundice; (4) withnon-specific symptoms such as loss of weightand appetite, insomnia, and dyspepsia; (5) withwhat may be called the humoral or neural con-comitants of the disease-hypertrophic pulmonaryosteoarthropathy, myopathy, and neuropathy;and others such as Cushing's syndrome.Cough was a symptom in three of every four

patients with bronchial carcinoma, and in 40%cough was severe and one of the primary reasonsfor medical advice being sought. Cough was, innearly all instances, at some time productive. Inmany patients cough was not a new symptom andwas often attributed to smoking. Worsening ofan otherwise chronic cough, often in relation toone or several febrile episodes, was common.Haemoptysis was a symptom in more than half

(57%), and in 4% of patients haemoptysis was theonly symptom. Presentation for investigation in1% was prompted by a single haemoptysis. Inpatients with haemoptysis this symptom variedfrom occasional to daily. The complaint of dailyhaemoptysis-usually in the form of blood-streaking of the sputum in the early morning-was regarded as very suggestive of bronchialcarcinoma. There was commonly delay, evenamong doctors later shown to have bronchialcarcinoma, of months rather than weeks beforepatients with the complaint of haemoptysis pre-sented for investigation. Delay in investigation ofhaemoptysis by a thoracic surgeon was oftenoccasioned both by delay in referral for a surgicalopinion and delay on the part of the patient toreport the abnormality.

Pain related to bronchial carcinoma was notusually acute pleural pain and was commonly anintermittent ache, which lasted for hours at a time,was worse at night, and was aggravated byposture and relieved by activity. Severe pain wascommon in those patients shown by investigationto have invasion of the chest wall, but chest wallinvasion, vertebral invasion, and rib and vertebralmetastases were found in patients who deniedpain, and the complaint of pain was not un-common in patients found to have pulmonarytumours which were successfully managed surgi-cally and in whom pain could not be related topleural invasion or infection.

It was often not possible to equate the complaintof dyspnoea with diminution in measured

respiratory function; very nearly normal respira-tory function figures were recorded in patientsamong whose more distressing complaints wasdyspnoea; there was little relationship betweenthe size of a radiographic opacity (and presum-ably, therefore, the extent of deprivation offunctioning lung) and the severity of the dyspnoea.One in every four patients with bronchial

carcinoma gave a history either of an acutefebrile respiratory illness, often called influenza,from which recovery had been slow and in-complete, or of a series of acute febrile illnesseswith complete or incomplete recovery betweenepisodes. Cough with purulent sputum, chestpain, dyspnoea, and haemoptysis were oftenfeatures of these acute illnesses, but evanescentfeatures of which little cognizance had been takenduring the acute episode; in fact these featureswere taken for granted as part of the acuteepisode. This history-of failure to make ananticipated recovery from an acute respiratoryillness-is accepted as strongly suggestive of thediagnosis of bronchial carcinoma, as is dailyhaemoptysis.Wheeze was a presenting symptom in 2% of

patients, and in those who were not chronicasthmatics the complaint of wheeze could usuallybe related to a tumour that had extended in amain bronchus close to the main carina, and eveninto the trachea; and, as a prognostic feature,wheeze, like stridor, was often significant ofinoperability. Stridor, as distinct from wheeze, wasless common, and fewer than 1% of patients wereadmitted to hospital urgently stridulous and inneed of therapeutic bronchoscopy, at which thetumour had to be cored from the trachea or bothmain bronchi in order to establish a more adequateairway. While stridor with bronchial carcinoma isoften a herald of death, acute tracheo-bronchitis,even in an adult, may present with stridor asevidence of gross proximal airway obstruction andis always an indication to investigate stridor bybronchoscopy.There was a group of patients-15% of the total

-who presented not with respiratory symptomsbut with symptoms such as loss of weight andappetite, tiredness and lassitude, non-specificgeneral ill-health, dyspepsia, with recent difficultyin the control of previously easily manageddiabetes, and so on. In these, referral to a thoracicsurgical unit was usually the result of anabnormality having been detected on a chestradiograph made as one of many routine investi-gations undertaken in a search for a cause fornon-specific symptoms, or because an abnormality

137



j.com/


arch 1968. Dow

nloaded from


B. T. le Roux

had been recognized during routine fluoroscopy ofthe chest as part of a barium study.Nearly 2% of patients with bronchial carcinoma

had at the time of their first presentation thecomplaint of dysphagia, and in half of thesedysphagia was the presenting complaint. Dys-phagia from malignant disease is nearly alwaysthe consequence of a primary alimentary tumour:but where it is the consequence of metastaticmalignant disease the primary tumour is nearlyalways bronchial.A small group of patients presented with the

complaint of recurrent peripheral venous throm-boses, often called thrombophlebitis migrans. Inthis group the incidence of operability was lowerthan average; the incidence of metastases atoperation was higher than average; the incidenceof death from pulmonary embolism, in con-valescence from an operation, high; and that oflong survival, low.

Bronchial carcinoma was found in patients inwhom presentation was a myopathy or neuro-pathy; because of Cushing's syndrome; becauseof exfoliative dermatitis, dermatomyositis, orpurpura; and with other evidences of generalsystemic disease. A patient who presents withspontaneous pneumothorax or an inguinal herniaand a chronic cough may well have a bronchialcarcinoma. Cough syncope and cough fracturesmay be the form of presentation in men.Apart from a small number of patients, too ill

at the time of admission to hospital for routineinvestigations or who died before investigationscould be completed, all patients were submittedto a routine series of investigations. The historyand clinical examinations were recorded. Routinechest radiographs were made and all patients wellenough to go to the radiology department wereexamined fluoroscopically, primarily for evidenceof phrenic paresis and displacement of thebarium-filled oesophagus. Tomograms and otherspecial films were made, in circumstances inwhich they were believed necessary, for diagnosticor record purposes. Bronchoscopy was a routineand it was the practice to make a biopsy from avisible tumour. In 58% of patients the diagnosisof bronchial carcinoma was confirmed atbronchoscopy. The itumours in these patients werecalled 'central' on the ground of visibility atbronchoscopy. In the remaining patients theappearances at bronchoscopy were normal, orbronchi were displaced but with the mucosanormal, and these patients were recorded ashaving tumours of 'peripheral' type. The purposeof bronchography is to demonstrate dilated

or obstructed bronchi, and bronchography wasnot often regarded as contributory in establishingthe diagnosis of bronchial carcinoma. There aretoo many other causes, of which mucus is the com-monest, of obstruction of bronchi beyond therange of bronchoscopic vision to justify accep-tance of 'bronchographic evidence of peripheralbronchial obstruction as evidence of carcinoma.Pleural effusion in relation to bronchial carcinomawas investigated by thoracentesis, examination ofthe pleural liquid both bacteriologically andhistologically, the replacement of the effusion withair, thoracoscopy, and pleural biopsy undervision. Biopsies were made of palpable cervicallymph nodes, and from other easily accessiblemetastases, in skin, subcutaneous and other softtissues, and in muscle. Sputum was routinelyexamined for tumour cells, but often not by anexpert specifically trained in the techniques in-volved. Symptoms, such as bone pain andhaematuria, were appropriately investigated; asearch was not routinely made in symptomlesspatients for cryptic metastases, by radiographicskeletal surveys or by routine pylography. Specialexaminations, such as pneumomediastinography,were occasionally undertaken in unusual circum-stances but were never a part of routine investiga-tion. Respiratory function, in particular the forcedexpiratory volume, was measured in patients inwhom surgical management was contemplated, butwas rarely found more informative than theclinical assessment of exercise tolerance by climb-ing stairs. An electrocardiogram was made pre-operatively in all patients over the age of 40 years.Routine pre-operative chest radiographs made onthe day before operation were usually antero-posterior films made with portable radiographicapparatus with the patient in bed, to provide amore precise base-line against which to judge thepost-operative films. A search for tubercle bacilliwas made in serial specimens of sputum from allpatients.With regard to the radiographic appearances of

bronchial carcinoma, in only three of 4,000patients were postero-anterior and lateral radio-graphs normal-repeatedly normal and normal toa number of independent experts. Presentation inthese three patients was with haemoptysis, in oneassociated with wheeze. The tumours were visibleat bronchoscopy, in a main bronchus in all, andalso in the trachea in one. Tubercle bacilli werefoundainAxtevutum of nearly 2% of patients withpulmonary carmnwoma, ani_Fcavitated pulmonaryttfi6rculosis was recognized radiographically as aconcomitant of bronchial carcinoma in 1 % of

138



j.com/


arch 1968. Dow

nloaded from


Bronchial carcinoma

patients, bi1ateraly,--and_--a-- b-ar -to surgicalmanageme_f_lf ccinomna _-some. The radio-graphic changes of pneumoconiosis were common-place, and the radiographic features of progressivemassive fibrosis were recognized in 1 % ofpatients. Pleural effusion was a concomitant ofbronchial carcinoma in 10%. Obstructive emphy-sema, the consequence of bronchial carcinoma,was uncommon.

In 7% of patients with bronchial carcinomathere were related cardiovascular abnormalities,sufficiently gross in one-tenth of these to precludesurgical management. Operative mortality wasquadrupled in those with systemic hypertensionwith electrocardiographic changes, or with ahistory of angina on effort, or in those who hadpreviously had a myocardial infarct. Of thosepatients in whom cardiovascular disease andbronchial carcinoma coexisted, who weremanaged surgically and who died in convalescence,half died from an acute cardiovascular episode-myocardial infarction-or in acute congestivecardiac failure.Of all patients with bronchial carcinoma, 55%

were unsuitable for surgical management whenfirst they were seen, most because of clinical,radiographic, fluoroscopic, or bronchoscopic evi-dence of metastases and the remainder because ofevidence other than metastases of unsuitability forsurgical management or because of death duringinvestigation. The remaining patients (45%) wereregarded after investigation as suitable forexploratory thoracotomy with a view to resectionof pulmonary carcinoma. Nearly one-fifth of thesepatients, 8% of the total, were found to have atumour, the resection of which was technicallyimpossible or therapeutically pointless; and pul-monary resection was completed in approximately37% of the total. Pneumonectomy was undertakentwice as often as lobectomy.The pulmonary tumour in about 55% of

patients with bronchial carcinoma was squamousin type. Undifferentiated tumours, including 'oat-cell' carcinoma, accounted for approximately 40%of the series. Adenocarcinoma was relatively un-common and in most series accounts for less than10% and in this series for no more than 5% ofthe total. Alveolar-cell or bronchiolar carcinomais rare and accounted for less than 1 %. In anylarge series mixed tumours are common-that is,tumours in which a biopsy from the primary maybe called squamous, sections of the whole tumourmay show different histological varieties, media-stinal or cervical glandular metastases may beundifferentiated, and an hepatic or cerebral

metastasis found at necropsy may show the histo-logical morphology of, say, adenocarcinoma. Allpossible permutations of the histological varietiesmay be found in the same primary tumour and inmetastases. Oat-cell carcinoma is said to be thetumour the histological appearance of which ismost consistent throughout the primary and inmetastases.Among those tumours found suitable for

management by pulmonary resection, most weresquamous ; among those tumours found unsuitablefor management by exploratory thoracotomy, thenumber of undifferentiated tumours was in excessof the number of squamous tumours. The opera-tive mortality of exploratory thoracotomy withoutpulmonary resection was 6%. Nearly threequarters of patients found by exploration to havean irresectable pulmonary carcinoma had diedwithin a year of operation. Nearly half thepatients, in whom management of pulmonarycarcinoma was by pneumonectomy, were found tohave hilar glandular metastases. The operativemortality of pneumonectomy was 12%, and theoperative mortality rate increased with age andwas 20% in those 65 years or older. The operativemortality for right pneumonectomy was nearlytwice that for left pneumonectomy. The threecommon causes of death in the post-operativeperiod after pneumonectomy were pulmonaryinfection, coronary thrombosis, and pulmonaryembolism, together accounting for three-quartersof the deaths in this group. The survival rate inthe first three years after pneumonectomy for pul-monary carcinoma was significantly higher inthose for whom resection was undertaken forsquamous carcinoma than in those for whomresection was undertaken for undifferentiatedcarcinoma. Thereafter survival rates for the twotypes of tumour ran a roughly parallel course. Onein every three patients who died from metastasesin the first two years after pneumonectomy diedwith cerebral metastases. Of all patients in whombronchial carcinoma was managed by pneumon-ectomy 20% were alive more than five years afteroperation. While most of those who survived along time were not shown to have mediastinalglandular metastases at operation, the finding ofmetastases in mediastinal glands is not a bar tolong survival. Following pulmonary resection forbronchial carcinoma, survival for more than fiveyears is not an insurance against the developmentof metastases at a later date, or against the growthof a new primary tumour.

In nearly 70% of those patients in whombronchial carcinoma was managed by lobectomy,

139



j.com/


arch 1968. Dow

nloaded from


B. T. le Roux

the tumour was squamous in type. The overalloperative mortality for lobectomy was approxi-mately half that of pneumonectomy and was 14%in those 65 years and older, and more than 20%in those 70 years and older. Right upper lobectomycarried the highest lobectomy mortality; 'sleeve'resection-usually of the right upper lobe withpart of the stem bronchus-was a variety oflobectomy relatively free from special complica-tions. The long survival rate after lobectomy forbronchial carcinoma was twice that for pneumon-ectomy. The fact that lobectomy for carcinomais less commonly complicated by death in earlyconvalescence and more commonly followed bylong survival than is pneumonectomy does notestablish lobectomy for carcinoma as a 'better'operation than pneumonectomy. When a pul-monary carcinoma is manageable by lobectomythe prognosis is better, because of the small size ofthe tumour, which in turn relates to the biologyof the tumour and the rate of growth or early, andtherefore often chance, recognition, and to theposition of the tumour in relation to the pul-monary hilum.

Bronchial carcinoma in a small number ofpatients was managed 'by segmental resection.Segmental resection is not regarded as an accept-able alternative to lobectomy in the managementof established bronchial carcinoma, and thecircumstances in which segmental resection wasundertaken were unusuaL Pulmonary carcinomamanaged by segmental resection was always peri-pheral in type and the lesion had often beeninterpreted, by palpation at thoracotomy, astuberculous on the grounds of hardness, relatedscarring, or multiplicity of nodules. Occasionallythe segment in which the tumour lay was separatedfrom neighbouring segments by an anatomicallycomplete fissure, usually the apical segment of thelower lobe or the lingular segment of the leftupper lobe, and in these circumstances it may bejudged as effective to manage the tumour bysegmental resection as by lobectomy. In rarercircumstances limitation of respiratory reserve, inpatients in whom thoracotomy has been under-taken with a view to making the most limitedresection possible to establish the diagnosis ofcarcinoma, may be an indication for managementof a peripheral pulmonary lesion by segmentalresection. Among patients in whom bronchialcarcinoma was managed by segmental resection-and in all these the tumour was small and peri-pheral-long survival was as common as afterpneumonectomy, and late death was more com-monly the consequence of coronary artery

insufficiency or respiratory failure than ofmetastases.Those patients who presented without symptoms,

and with an abnormality detected on a chestradiograph made for an unrelated purpose, weremore often suitable for surgical management ofbronchial carcinoma than were those who pre-sented because of symptoms, but not all weresuitable for surgical management. In this groupthe number with peripheral tumours was higherand the number with undifferentiated tumourslower than among those who presented withsymptoms. Approximately one-quarter of thosewho presented without symptoms and with anabnormality demonstrated on a routine chestradiograph made for an unrelated purpose were un-suitable for or declined surgical management.Operative mortality and the rate of irresectabilityat exploratory thoracotomy-both about 5%-were lower than among those in whom bronchialcarcinoma had declared itself before radiologicalinvestigations were undertaken.Those patients who declined surgical manage-

ment of a peripheral pulmonary opacity found atroutine radiography offered an opportunity, notjustifiable in other circumstances, of observingradiographically the varied rate of growth of pul-monary carcinoma and the time interval betweenthe recognition of a peripheral shadow and thedevelopment of symptoms and metastases. Survivaltime from operation in a small group, whodeclined the advice offered when first they pre-sented with an abnormal routine film but who lateraccepted surgical management when they haddeveloped symptoms, was compared with thesurvival time in those without symptoms whoaccepted surgical management, and in those whopresented for the first time with symptoms andwere managed surgically. The number of patientsin the first group was too small to say anythingmore than that the survival time varied over theentire range from long survival with ultimatedeath without metastases to inoperability withrapid decline despite irradiation, and death withmetastases.Some hazards were attached to routine radio-

graphy for whatever purpose. Observer error inthe interpretation of routine miniature films hasbeen assessed, and the incidence of error is nothigh, especially if experts collaborate in the inter-pretation of routine radiographs. Notwithstanding,when a patient previously radiographed for routinepurposes presents with new respiratory symptomsor with a second, now abnormal, film made forsimilar routine purposes, and the original film,

140



j.com/


arch 1968. Dow

nloaded from


Bronchial carcinoma

passed as normal, is scrutinized with the know-ledge gained from recognition of an abnormalityin the later film, a similar abnormality in the earlierfilm is often recognized. This can be of value inreaching a decision on management-for examplein making more likely the diagnosis of a chronicinflammatory lesion such as tuberculosis; inassessing the rate of natural growth and, thereby,the prognosis of malignant tumours; and inestablishing that other lesions such as chondro-adenomata do in fact grow. Furthermore, a patientroutinely radiographed at arbitrary intervals, sayyearly, who develops respiratory symptoms suchas haemoptysis in the interval between radio-graphs, is inclined to ignore these symptoms inthe knowledge that a recent previous film has beenpassed as normal and another film will be maderelatively shortly. But while these are hazards thatattach to routine radiography, the overall advan-tage of this technique has been clearly established.The number of patients found, by chest radio-graphy prompted by the development of a pul-monary complication after an elective surgicalprocedure other than thoracic, to have a pul-monary lesion other than the acute post-operativeone, and probably to which the acute post-opera-tive lesion could be related, is large, and there cannever be any justification for not routinely makinga chest radiograph before performing any electivesurgical procedure in an adult. Where a thoracicsurgical unit exists in close relation to generalsurgical charges and the opportunity for educationis taken, a large number of patients will be foundin any one year in general surgical wards to havesurgical pulmonary abnormalities an argumentin favour of collaboration between the branches ofsurgery and against the isolation of special unitsin peripheral hospitals.

Cervical lymph glandular metastasis was themost common single clinical evidence of dissemi-nation of bronchial carcinoma. In approximately10% of all patients with pulmonary carcinoma,cervical nodes only on the same side as the primarytumour were the site of metastases. In approxi-mately 2% cervical glandular metastases werebilateral, and in another 2% cervical glandularmetastases were contralateral. In nearly half thosewith ipsilateral cervical glandular metastasis thiswas the only evidence of metastasis on standardclinical and special investigation. Where cervicalglands were palpable in patients with bronchialcarcinoma, these were found by biopsy to be thesite of metastasis in three of every four patients.Scalene node biopsy in patients in whom a glandcould not be felt was an unrewarding exercise.

Cervical glandular metastasis from a left pulmon-ary primary-not only in the lower lobe-to theright side of the neck was not more common thanwas a metastasis from a right pulmonary primarycarcinoma to the left side of the neck. Bilateralcervical glandular metastases were only occasion-ally the only evidences of spread of bronchialcarcinoma, and most patients with bilateral cervi-cal glandular metastases died within a few weeksof investigation.

Axillary lymph glandular metastasis frombronchial carcinoma without chest wall invasionwas very rarely the only clinical evidence ofmetastases and in two patients the primary tumourand the invaded axillary glands were on oppositesides. A radiographic opacity in the vicinity ofthe azygos vein and inseparable from the media-stinum, in patients in whom the pulmonaryprimary was right- or left-sided, was good evidenceof lymph glandular metastasis and of unsuitabilityfor surgical management. Where the radiographicappearances were those of lymphangitis carcino-matosa, dyspnoea was usually the predominantsymptom and death was rapid. Unilateral radio-graphic features of lymphangitis carcinomatosawere more commonly right- than left-sided; wherethe changes were bilateral there was usually aright-sided predominance; unilateral radiographicfeatures of lymphangitis carcinomatosa werealmost always evidence of spread from a primarypulmonary tumour rather than from an extra-thoracic primary tumour.Three per cent of patients found on investiga-

tion to have pulmonary carcinoma presentedbecause of cerebral metastases. Hemiplegia,epilepsy, and personality changes were the threemost common forms of presentation; speechdefects, cerebellar symptoms, and headache wereother common presenting symptoms. Only oncewas it found rewarding to excise a cerebral meta-stasis and thereafter to manage the pulmonaryprimary surgically. Where suspicion, howeverslight, was raised regarding the existence of cere-bral metastasis, the presence of which was uncon-firmed by investigation, the passage of time, oftenbrief, usually confirmed the suspicion.Approximately 2% of patients were found to

have osseous metastases as the only evidence ofdissemination of bronchial carcinoma at the timeof their first investigation; three in every four ofthese presented with the complaint of bone painand without new respiratory symptoms. The com-monest sites of osseous metastases were the ribsand the vertebrae. In approximately the same per-centage of patients, hepatic metastases were found

141



j.com/


arch 1968. Dow

nloaded from


B. T. le Roux

as the only evidence of spread of tumour at thetime of first investigation.Nearly 5% of patients presented with the

clinical features of obstruction of the superiorvena cava. More of these were women, of a

younger than average age, and the incidence ofundifferentiated tumours was higher, than among

those who presented in other ways. Obstructionof the superior vena cava and cervical lymphglandular metastases, bilateral disproportionatelyoften, were commonly associated. A right vocalcord palsy in patients with bronchial carcinomawas rare, and when it was found it was usuallyin association with obstruction of the superiorvena cava, a right upper pulmonary carcinoma,and right cervical lymph glandular metastases. Itwas usually possible quickly to relieve the symp-toms and signs of obstruction of the superior venacava by irradiation; the recurrence of cavalobstruction was common and usually a herald ofdeath. Obstruction of the superior vena cava was

nearly always associated with a right-sided pul-monary carcinoma, usually in the right upperlobe.

Fluoroscopy demonstrated, in approximately5% of patients, that bronchial carcinoma wasunsuitable for surgical management-more oftenbecause of displacement of the barium-filledoesophagus than because of interruption of a

phrenic nerve. Interruption of the left recurrentlaryngeal nerve as evidence of mediastinalinvasion was found in approximately 4% ofpatients. The pulmonary opacity in patients witha left recurrent laryngeal nerve palsy was left-sided in nearly all, and in the left upper lobe or

at the left pulmonary hilum in most. Histologicalconfirmation of the diagnosis of bronchial carci-noma was achieved, at the time of their firstinvestigation, in only about one-third of patientswho presented with a left cord palsy. Where thehistology of the tumour was established, squamoustumours predominated. Survival time from theonset of hoarseness of voice was often long.Irradiation of a pulmonary carcinoma in patientswith a hoarse voice did not restore the voice.Proximal extension of tumour, seen at broncho-

scopy and usually established by biopsy, pre-cluded surgical management of pulmonarycarcinoma in nearly 5% of patients. Bronchialcarcinoma was judged unsuitable for surgicalmanagement when the main carina was seen tobe invaded or where the main carina, while notulcerated, was broad and the main bronchisplayed; where the trachea was involved (unlessthis involvement was limited to the right lateral

wall and this was usually found in relation to atumour of the right upper bronchus); where thetrachea was compressed or scabbard-shaped;where the tumour would be demonstrated histo-logically to involve bronchi, usually the mainbronchi, of both sides; where tumour obstructeda main bronchus, within 05 cm. of the maincarina on the right and within 1P5 cm. of the maincarina on the left, provided that involvement ofthis degree of proximity was medial and not onlylateral; and where, by the independent assess-ment of two experts, a main bronchus was foundto be abnormally rigid, an assessment rarelyaccepted as the only basis for not undertakingsurgical management in a patient suitable in allother respects for this form of treatment. Adeno-carcinoma rarely extended proximally in the waysoutlined. A disproportionately large number oftumours which extended proximally were right-sided.

Invasion of the chest wall, not uncommonlyunsuspected clinically, was recognized radio-graphically in some 3% of patients. In one-thirdof these resection was undertaken because chestwall invasion was lateral; in another third thetumour was apical and associated with thefeatures of Pancoast's syndrome; in the lastthird invasion was medial, either anteriorly orposteriorly.

Where pleural effusion was associated withpulmonary carcinoma, this was routinely inves-tigated by thoracoscopy. This demonstrated thepresence of pleural metastases as the only evidenceof extra-pulmonary extension of bronchial car-cinoma in approximately 2% of patients. Halfthe patients with pleural metastases had peripheraltumours, and a disproportionately large numberwere women. Adenocarcinoma metastasized tothe pleura disproportionately often. Most patientswith pleural metastases died within six monthsof the diagnosis having been made, in contra-distinction to patients with primary pleuraltumours, in whom long survival was not uncom-mon. Pleural effusion, even if sanguineous, inassociation with bronchial carcinoma was notsynonymous with pleural metastases and was anindication for investigation, not a reason fordespair.

Hypertrophic pulmonary osteoarthropathy wasa concomitant of bronchial carcinoma in approxi-mately 1% of patients. In this group of patientsthe tumour was found disproportionately oftento be an adenocarcinoma. The incidence of peri-pheral tumours and the rate of resectability among

142



j.com/


arch 1968. Dow

nloaded from


Bronchial carcinoma

patients who presented with hypertrophic pul-monary osteoarthropathy was disproportionatelyhigh, and the incidence of long survival dispro-portionately low. Resection of the pulmonarylesion, or irradiation, relieved joint pain. Theappearance of intra-thoracic recurrence or meta-stases in the late post-operative period was, in a

small proportion of the patients who presentedoriginally with hypertrophic pulmonary osteo-arthropathy, associated with a return of thesymptoms and signs of this concomitant ofbronchial carcinoma, but this phenomenon was

never observed in patients who had not initiallypresented with hypertrophic osteoarthropathy andwho later developed intra-thoracic metastases.

Diminished respiratory reserve constituted theonly bar to the surgical management of bronchialcarcinoma in approximately 4% of patients. Morethan three-quarters of the patients in this groupwere older than 59 years when first they were seen,whereas in the series as a whole fewer than half

the patients were older than 59 years. In elderlypatients with poor respiratory reserve and smallperipheral tumours, death, more often than not,was with rather than from bronchial carcinoma.In approximately another 4% of patients there wasfound a non-metastatic contraindication to surgi-cal management other than diminution in respira-tory reserve-for example, extensive pulmonarytuberculosis, chronic asthma, gross pneumoconi-osis, bullous emphysema, myocardial disease,advanced age alone, or frailty.Approximately 2% of patients who were being

investigated for a lesion later established to be a

bronchial carcinoma died before investigationscould be completed. In many of these the terminalevent was not directly attributable to bronchialcarcinoma; at necropsy most had evidence ofwidespread metastases. Techniques used in theinvestigation of 4,000 patients with bronchialcarcinoma resulted in the death of two patients,a mortality from investigation of 005%.

143



j.com/


arch 1968. Dow

nloaded from


bronchial carcinoma - thorax.bmj.com · to record in all other aspects of the anamnesis, and ......

Documents