Tobacco Documents Online

T)iorax 11969),14) 446 J~~c{ 1~ ! Lung cancer in women DAVID J. B. ASHLEY AND H. DUNCAN DAVIES From Morrisron /lospiraf, Swansea 9 Eighty-threc instances of lung cancer in women are presented. The frequency of the condition is lower in women than in men but the tumours in women tend to be less well diffcrcntiatcd and are less likely to be resectable than in men. The frequency of extrathoraeic metastases at the time of diagnosis is greater in women than in men. It is suggested that the immunological defences of women are better than those of men bccausoof sex-linked genes, present in double dose in the female, which are concerned with the immunological mechanisms, and that the clinically apparent tumours are those which, by virtue of their greater intrinsic malignancy, can overcome the defenec mechanisms. Malignant disease of the lower respiratory tract is TA B LE I substantially less common in women than in men. RELATIVE FREQUENCYOF LUNG CANCER IN MENANn WOMEN les This difference in incidence in England and Wa increased between 1913 and 1943 but has remained more or less constant since then (Myddclton, 1965). The differing frequencies of the tumour in the two sexes are usually related to the lower frequency of squamous-celled carcinoma in women which in turn is attributed to the higher prevalence of cigarette smoking among men and the greater likclihood that men would work in an atmosphere of high air pollution. The present contribution is a study of the clinical, histological, and biological features of lung cancer in 83 instances of this tumour diag- noscd in women by clinical, ridiological, and histological examination at Morriston Hospital, Swansea, during the past 15 years and a compari- son of these features in 1,353 lung cancers in men (Ashleyand Davies, 1967). FREQUENCY The ratio between men and women in this series was 16: 1: if only histologically proven cases are included the ratio is 12: 1. The prnportion' of women is lower than in most of the other reported series (Table 1). and is lower than that for England and Wales as a whole. The total number of deaths in women recorded by the Registrar General as due to lung cancer in the six years 1958 to 1963 was 12,786 ; the total for the same six years in men was 78.235, a male to female ratio of 6-1 : 1. The age distribution of the female population differs markedly from that of the male and, if the age specific death rates for this tumour in men had Rdcrcnco Mck fFcmak Rmlo Ma-nn (1949) . ,. ., .. .. 9: 2 : 1 OntlandHilt(1952) ., ., ., ,. s u:t lli&na11 (1955) . .. .. m:1 Nknnlsnn, Fo., nnd aryce (1957) .. .. Il:l N'hiiw010961) .. .. B/:1 Edcsr and Menheimer (1962) .. .. 6:t z KreYbere (1962) Flnland .. Norwny .. .. .. .. .. z:l 8~6:1 Shfnton (1963) .. . 9^:1 vinccnt, S~tmrGcid, md Ackcrman (19651 .. 8~4 :{ TGls.cncs . . .. I .. 16:1 applicd to the female population. the number of deaths expected wou!d have bzcn 105,437, a ratio of 8-1 : 1. iThc relatively lower frequency of lung cancer in women in this series may be related to the cultural habits of the people of this part of Wales. possibly y to`.a lower prevalence of cigarette smoking among the women of Wales. This view is suppq,rtcd by the finding of a Standardized Mortality Ratio for lung cancer of 81 in men mtd 58 in women for the Welsh Hospital Region..which is coterminntlc with tht Principality of Wales (Registrar General, 1965). In 1963 there were 19,746 deaths from lung cancer among men in England and 3.558 such deaths among women, a male to female ratio of 5-55: 1, whcrcas in Wales there werc 1.011 deaths in men and 119 in womcn, a male to female ratio of 84 : 1(Rcpistrar General, 1965). This difference is h~ghby significant. ~ H1S7OLOGICAL TYPES OF LESION Material was available from either bronchial biopsy or open operation in 56 cases. The histo- 446 N: O W O W O O ~ ~ .

i Lrulg cancer in women logical classification used was that of Ashley and Davics (1967) in which five types of lung cancer were recognired-well and poorly diffcrcntiated squamous carcinoma, well and poorly d10'cren- tiated adcnocarcinoma, and undifferentiated car- cinonm. Examples of the carcinoid (argcntaffin) tumour of the bronchus were excluded as it was considered that these form an intrinsically different tumour type. The data are presented in Table II, in which they arc subdivided by the type of differentiation, TABLE II /IISTOLOGICAL TYPES OP LUNG CANCER TYM or Dincrcmiotian Fem.ka No. % Males /. S.tuan.nut .. ., IS 23 60 Glnndular . .. .. 16 28 11 UnJiffcremiated .. .. 27 48 29 DrF.re nfdinerrntiatian woll Jiil:aentlumd ~ .. 7 125 18 5 Ponrlv Jln:rcn+iu+ed ,. 34 15 UnJinercmi~+eJ .. .. Jo 53.5 35 whether glandular, squamous, or undifrerentiated, and secondly by the degree of differentiation of the lcast well differentiated part of the tumour. The proportions of the different types in women :vc compared with those derived from a pre- viously reported survey of 666 cases of lung cancer in men (Ashley and Davies, 1967). The rare lesion, adcnosquamous carcinoma, in which both gland- ular and squamous differentiation are seen in the same tumour, was not represented in this series. There was a significantly lowcr, proportion of tumours of squamous type in the women nnd a significant excess of glandular and of undifferen- tiated carcinomata. The proportion of cases in which the least well differentiated part of the tumour was classified as undifferentiated was also significantly higher in the women. Similar observations to these have been madc in ntany reported series (Mason, 1949; Shinton, 1963: Hanbury, 1964: Vincent, Satterfield, and Ackerman, 1965). The deficit of cases - of squamous-celled carcinoma may be attributed to the lower frequency of chronic irritation in the female bronchi, whether the result of cigarette smoking or of chronic bronchial inflammation. The lower degree of differentiation in tumours in women was a feature which was noted by Nicholson, Fox, and Bryce (1957), Whitwell (1961), Rrcyberg (1962), and Shinton (1963) al- though Vincent et al. (1965) found the proportion of undifferentiated tumours to be equal in the two sCXCS. I + 447 AGE INCIDENCE The age incidcnce in this series is given in Table 111. Th: mean age at diagnosis was 56'1 yean, whereas in men the mcan ngc at diagnosis in those presenting with the symptom cough was 58-3 years, and in men in whom the tumour was diag- nosed as a result of routine radiography the mean TABLE Il] AGEINCIDENCE A¢c Grnup <40 yean 47-50 51-50 61-70 >70 No. 4 16 34 25 4 age was 60-7 years. The age at diagnosis has been noted to be lower in women than in men in a i number of other series (Mason, 1949 ; Nicholson ei al., 1957 ; Whitwcll, 1961). This is a particularly surprising finding because the female population is relatively great_r in the older age groups than the malc and, in It lesion which shows an increas- ing incidence with increasing age, as does lung cancer, a relatively higher average age would be expected in women than in men. The histological features of lung cancer in women include a greater proportion of ill- diffcrcntiated eareinomata than in men, and as these lesions tend to occur in men at anearlier age than those of more differentiated type (Ashley and Davies, 1967) it is possible tl•it the age dif- fcrencc may bs due to some factor such as this. Another possibility is that the habit of cigarette smoking is eommoner in the younger women than I in the older and therefore that the frequency of 11 lung cancer may be relatively higher in younger women than in older women. . SYMPTOMS The four major symptoms attribdtable to lung' cancer are cough, dyspnoea, chest pain, and hnemoptysis. The distribution of symptoms in the patients in this series is set out in Table IV and TABLE IV SYMFTOMS AT THE TIME OF DIAGNO515 Fcm.ks No. % Males N 0 L CnuFh . „ .. 56 67-5 61 00 D1'sPnoen .. .. 26 ]1 47 ( Cbnt pain -, -. 37 45 46 Hm~n,nplritx .. .. 25 30 39 2 tJ Rou4ne raJieeraGh Sccond:+riu .. .. ., 6 25 30 SI tP4 1 00 M ' ' 0

• • 448 David J. B. Ashley and !f. Dunean Davies V is compared with the data from 1,353 men sutTer- T A B L L• V ing from this disease. The proportions of women complaining of cough and chest pain did not diffcr significantly from the proportions of men pom- plaining of these symptoms. Dyspnoca and hacmo- ptysix were less common in women than in men. Shortness of brcath in patients with lung cancer is often the result of a relatively small decrease in respiratory capacity in an individual who already suffers front respiratory impairment, often as a consequence of chronic bronchitis and emphysema. Chronic lung infection of this type is less common in women than in men and it is probable that the lower prevalence of dyspnoca in these patients is related to their general respira- tory health rather than to specific biological differences in the tumours. Hacmoptysis, on the other hand, is more often associated with tumours showing squamous differentiation than those of glandular or undifrerentiated pattern (Ashley and Davies, 1967), and the low proportion of women complaining of this symptom may be related to the lower frequency with which lung cancer is sqlclmous in women. The proportion of patients in whom the tumour came to light as the result of routine radiography was small in both mcn and women and did not differ significantly in the two sexes. There was a significant excess of women in whom extrathoracic metastases were detectable at the time of diagnosis. Six of these had bony secondary deposits at the time of diagnosis. Among the 56 patients in whom histological con- firmation of the diagnosis was available, 18 had cxtrathoracic secondary deposits at the time of diagnosis. The numbers of men with secondary deposits in each of the histological types was recorded in the previous paper (Ashley and Davies, 1967): when these rates were applied to the women in each of the histological types a total of seven with metastases would have been expected, BRONCHOSCOPIC APPEARANCES Bronchoscopy was carried out in 78 of the paticnts in this series (Table V) and tumour was found in 51 (65 ;). In a comparable series of men broncho- scopic evidence of pulmonary neoplasm was ob- tained in 55%. This difference was not statistically significant. OPERABILITY Surgical excision of the tumour offers the only present hope of cure in cases of lung cancer. Eleven of the patients in this series had lesions RRONCI/OSCOPIC APPEARANCLG ANI> OfCRAilIL17Y Is rnnahmonrf Pashlw - Ncymi.e Nat dun< ^ Tulal Not rc.ntnMc .. 18 /9 S 72 Re.ccmhle .. I 8 - u Tuul .. 51 21 7 83 which were resected (134%). This is a lower pro- portion than in the scrics of men (21%). The difference approaches statistical significance. Two patients refused lhoracotomy, and in a further seven the lesion was found at thorncotnmy to be inoperable. The proportion of resectable eases in both the men and women is lower than reported by others. It is suggested that this is because almost all patients suspected of having bronchial lumours are sent for investigation to the joint medical and surgical chest units at Morriston Hospilal, and selection of polcntially operable cases at other hospitals is nt a minimum. A lower resecutbility rate in women was also noted by Nicholson and his colleagues (Nicholson et al., 1957) and they attributcd this to the higher frequency of ill differentiated tumours in their female patients. In the present series data on the operability of the different histological types of tumour were avail- able from the men (Ashley and Davics, 1967). When the male percentages were applied to the numbers of the different types of tumour in the female serics a total of 18 cases might have been expected to have resectable lesions, whereas only lI did so. This difference in the proportion of patients with resectable lesions in the groups of men and women with histoloBically classifiable tumours was statistically significant. In the men it was noted that the site of the lesion, w•hether accessible to the bronchoscope or r not, had an effect on the.Taasibility`of surgical excision (Ash)ey and Davies, 1967): The expected ' frequency of operable fumours if the resection rates for men in the two groups, bronchoscopy positive and bronchoscopy negative, had applied were calculated. Eleven cases in this scrics were opcrtblc ; 16•4 would have been expected. The lower operability •rate is therefore not the con- sequence of a higher frequency of more proximal tumours. The presence of extrathoracic metastases is a bar to thoracic surgical treatment. The higher pro- portion of women who had such metastascs at the time of diagnosis also influences the proportion of women with operable lesions. '

i Lung cancer in woinen . tiMt1KIN0 III}TORY . A history relevant to cigarettc smoking was avail- ahle in 56 cases ; 23 of these women (41'g,) werc ~~ non-sntokers. The mcan ugc of the snlokcrs (55 ycmx) was slightly lower than that of the non- smokers (57'2 years). The proportion of non- smokers in this scrics agrees well with the 40% recorded by Doll and Hill (1952) and the 36 % recorded by Hanbury (1964) and is greatly in excess of the proportion of non-smokers in men (2-S",,;). No significant diRerences were observed in thc distribution of histological types of tumour among the smokers and non-smokers (Table VI). This obscrvation agrecs with the finding in the TA13LE VI . SMOKING HISTORY v (n) ryre lrumm.r Squamoar f Olandul.r ndincrc Un- Toml iiamd Non-smokers 2 6 6 2J cfparenc ~ moken 7 6 12 33 ^-`_- (b) A n~n~nrn/rnanrrnCon.um ed Ggarnto <5/duy I 7 V 6-IS It I rb-l5 9 >25 • ,'Hcavy' ~ 2 men of this series but disagrees with the observa- tions of Kreyberg (1962) in men and of Vincent el af. (19651 in women (Ashley and Davies, 1967). It is our view that cigarette smoking is an activ: contributory factor in the oncogenesis ofpul- monary neoplasm but that the type of differentia- tion is dependent on the presence or absence of squamous mctaplasia in the bronchial trcc and that this in turn is related to smoking and also to chronic infection and to atmospheric pollution. . 449 TABI.E VII 1'old nw-IIt• DATA Ali. At 6 ma ntb. At 12 mn mha A t 2 4 mo.nh. Toul 26(41 %) 1 4 (22 %) 4 (6J'/.) 61 10 7 2 `- /0 36 21 6 73 35% 13-5% 5.1% rescctablc lesions is too small for assessment. Twenty pcr cent of these women were alive two years ;dter lung resection ; in the mcn 40% were still alive at two years: this difference does not approach statistical significance. In the inoperable cases there was a slightly higher proportion of women alive at each of the three survival periods studied, but by two years thc difference was small. A poorer general prognosis in women than in men :was noted in the M:mchcster series (Nicholson cr af., 1957) but a better survival at five years was noted in operablc lung cancers in women; by Bigmtll and Moon (1955) and by Edcrer and Mcrshcimcr (1962). Nicholson and his colleagues (1957) attributed the poorer prognosis to the higher proportion of poorly differentiated tumours among women, as these tumours have often been shown to be intrinsically more malignant (Whitwell. 1961 : Shinton, 1963 ; Ashley and Davies, 1967). The data presented above, how- ever, suggest that lung cancer in women is less likely to be operable than in men when allowance is made for the histological type of neoplasm. I DISCUSSION_ Consideration of the data presented above from our series of cases and from the reported series of lung cancer in women allows a concept of the character of this disease in women to be formed. Lung canccr is less common.. in women than in men, it is lcss likcly to be of the squlmous-cell FOLLO\4-UP DATA type, and more likely to be of'the glandular type , or to be undifferentiated. The dcgree of diHercn- The conventional assessment of treatment of ' tiation is less than in men. The tumour is likely to malignant disease by study of the five-year sur- vival rate is not applicable to a tumour of such poor prognosis as lung cancer. In this series follow- up from the date of diagnosis was available in 73 instances. Those patients who died within one month of thoracotomy, whether the tumour was resectable or not; have been excluded from the analysis because of the possibility that the stress of operation may have hastened death in these cases (Table VII). The number of patients with occur at a somewhat earlier age and is more likely to present symptoms at a time when extrathoracic metastases are already present. The tumour is as likely to be accessible to the endoscopist as in men and is less likely tb be accessible to surn cal exci- sion. The prognosis of lung cancer in women is worse than in men. This is a curious situation. In women lung cancer is less common, but when it occurs is less likely to be tesectable. It has been shown that the IJ O O W O W 00 rn ~ -I

0 450 David J. D. Ashley and II. Duncan Davies difference in resectabi(ity cannot be explained on the bnsu of the histological type of tumour or on the anatomicnl localifation in the bronchial tree. The sil;nificantly higher proportion of palicnls in whom cxlr:uhoracic mctasL•Iscs were present :u the time of diagnosis :1llcels Ibe overall prognosis but cannot be explained on the basis of the histo- logical type of lesion. An alternative hypothesis is put forward.- . One at Ieast of the genes concerned in the immunological defence mechanism of the body is carried in the X chromosome (McKusick, 1962) ; mutation at this locus leads to the clinical condi- tion of agammaglobulinacmia in which there is a deficiency of a circulating immunoglobulin and also a dcfieicney of plasmaccll reaction to foreign antigens. By virtue of their double complement of X chromosome women have a lower frequency of asammaglobulinacmia and, by inference, a bctlcr immunological capacity. It is suggested that those tumours of the lung which become clinically apparcnt arc the ones in which the multiplying neoplastic cells arc able to overcome the immuno- logical defences of the body which would other- wise treat them as foreign cells and destroy them by a similar reaction to that seen after tissue . transplantation. In women, with a better defence mechanism, fewer tumours would overcome the defences but those which did would be intrin- sically more malignant and less amenable to therapy. This hypothesis would account for the lower proportion of lung cancers in women, which can- not be completely explained by the lower, propor- tion of women who are cigarette smokers. It would account for the reduced frequency with which the tumours when they occur are resectable, for the higher proportion of neoplasms of histo- logically undifferentiated type, and for the g_nerally poorer prognosis in lung cancer in women. The reported better survival after lung resection in women (Bignall and Moon. 1955 ; Edcrer and Mcrsheimer, 1962) is explicable if in these cases a balance between the immune de-. c fences and lho advancing tumour is held rather more sccurcly in the case of the less 'malignant' lumoqrs in women than in men. A similar explanation has been invoked (Avhlcy, 1967) to explain the dilierences in the natural hiti- tory of lung cancer in mincrs and in non-mincrs. It was then suggested that the immunological defences of lhe lung which had been invaded by dust might be -in a state of enhancement and that when neoplastic changes occurred the abnormill cells might quickly be rceognized as 'not selt' and destroyed before clinically apparent neoplasia could develop. This work was supported by a research grant from the Welsh Hospital Board. We are indebted to our colleagues Dr. E. A. Danino, Dr. T. W. Davies, and Mr. C. J. Evans for accexs to their case notes. I REFERENCES Ashlcy, D. J. B. (1967). Lung crnccr in mircn. TAcrox, 23.87. and Davies, H. D. (196)). Czncer of the lung: histnlo8y cnd biological bchuviur. Conrrr fPAilnrL). 20.165. Bignnll. l. R. (1955). Bronchial eurcinoma: survey of 317 Dsti<nu. ionrrl, 1. 786. - ned Moon, A. J. (1955). Surrivnl after lung rneclion for bron- cMzt carciooma. TLn.nx, 10. 181. Doll, R., and Hill. A. B. (1952). A smdy ef the aeliology ofea.cfnnma of the Iong, aril. mM. J., 2,127 1. Edercr, F., and Mershdmer, W. L. (1962). Sea di6erenecs in the surviwl of lung cancer patients. Cmmrr U'Nilad.), 15, 423. Hnnhcry. W. J. (196e). Bronehoecnk cardnema in v.onxn. TGeres. ' 19, 333. KrcYherp. L. (1962). Hiamlogiral tunp CnnrerTjDes-Ana push. smad. SunnL 157. McKusick. V. A. (1962). On the %chromosome nrman. Quare. Her. aiol., 37. 69. Mason. G. A. (1949). Cancer of the lung: review ofa thousand eases. Lnncrq 2. 587. Myddclton, G. (1965). Clrcinoma of bronchus. )bfd., 2,796. Nichnlson, W. F., Foa, M.. and Bryc<, A. G: (1959). Fevie. o(9I0 scs of bronchial carcinoma wish results of treasmcet. Ibid., 1, 296. RcgistmrGeneral(1965). SmtinsiealRevicxnfEnglandand Walcafer the ycar 1963. Part I. Tubles, Medlcal. H.M.5.0., Loodnn. Shimnn, N. K. (1963) Dilferences in biological characteristics of various Aistolopieal typcs oflawcr resDiratory tract tumours. Pril- J. Canrrr, 17. 222. 1 Vincent, T. K. S.uerfield, J. V„ and A~ke+m.n: L'\~'f1963). Carei- nnma of ahc lun8 in womcneCmrcr. ('idilad.), 18 55,9. Whirwell. F. (I961) Rhe kfasopntholnpy of,lunk a~ncer in Lircrnool: the specifialty of "e histological cell types of lung eancer.OntJ. Cancer, 15, 440. j I