HHBM COLUMBIA LIBRARIES OFFSITE HEALTH SCIENCES STANDARD HX64121739 RC261 .B291914 Review of recent can RECAP Bashford, E.F. Review of Recent Cancer Research. 191u, intljeCttpoflettJgork Digitized by tine Internet Arciiive in 2010 witin funding from Open Knowledge Commons http://www.archive.org/details/reviewofrecentcaOObash PROCEEDINGS OF THE NEW YORK PATHOLOGICAL SOCIETY REVIEW OF RECENT CANCER RESEARCH BY DR. E. F. BASHFORD Director of the Laboratory, Imperial Cancer Research Fund, London MIDDLETON GOLDSMITH LECTURE I912 PRESS OF THE NEW ERA PRINTING COMPANr LANCASTER. PA. REVIEW OF RECENT CANCER RESEARCH BY DR. E. F. BASHFORD Director of the Laboratory, Imperial Cancer Researcii Fund, London MIDDLETON GOLDSMITH LECTURE I912 3^1 REMEW OE RECENT CAXXER RESEARCH^ Lecture I. Statistical and Comparative Ixvestigatioxs The recent advances made in our knowledge of cancer may be grouped together according as they have been obtained by statistical or by experimental methods. In this first lecture 1 shall direct your attention to the results which have been obtained by improving the national statistics of England instead of taking a separate cancer census as was done in Germany and other coun- tries, and to certain other statistical observations. In the second lecture I shall discuss the results of experiments. A few general remarks are necessary to introduce the statis- tical details. It is essential that the great frequency of cancer should be borne in mind, because this of itself explains why aggregations of cases are bound to occur. Cancer is so frequent in England that one out of every seven women and one out of less than e\'ery eleven men above the age of 35. ultimately dies of it. On this basis it is possible to calculate- how often one, two, three, or more, cases of cancer may be expected in a family. Of 100 families each of six members. 3 males and 3 females, no case of cancer would be expected in 51, one case would be ex- pected in 36, two cases in 11, and 3 or more cases in 2 families. Of 100 families of eight members, four women and four men. no case would occur in 41, one in 39, 2 in 16 and 3 or more in four families. In 100 families of ten members, five men and five W(jmen, only 33 would 1)e expected to escape altogether, whereas there would be one case in 39, two in 20, and 3 or more in eight families. Moreover, the disea.se is ten times as frequent at 70 ■ ^Middleton Goldsmith Lectures delivered before tlie Xcw York Patho- logical Society, October 2 and 4, ii>i_'. 2 Calculation made on the l)asis ni i in 8 and i in 11 respectively. 1 2 REVIEW OF RECENT CANCER RESEARCH as at 35 years of age; therefore similar calculations made for age periods other than 35 would show a corresponding greater liabil- ity than the a^•erages set forth above. Hence aggregations of cases of cancer to have any significance at all must occur with a frequency which removes them from all possibility of their being merely what would be expected owing to the great frequency of the disease. In the course of years it is quite natural that a number of cases should occur in a village, a street, a house, or even a family which has inhabited the same house for genera- tions, without the numbers necessarily meaning anything more than what was to be expected according to the theory of probabil- ities. This criticism applies without exception to all statements as to cancer houses, cancer streets, and cancer villages, which have been adduced as evidence that cancer is infective, although I do not wish to.be understood as denying that cancer or rather cer- tain forms of it occurring on different parts of the body, may, and actually do, vary in dififerent areas. As is apparent from the figures I gave of the relative liability of males and females above 35, the relative frequency of cancer is influenced very largely by the relative proportions of the two sexes and the respective numbers of each sex living at each age- period. If a curve be constructed of the mortality from an infective disease, c. g., tuberculosis, and regard be paid to the different age-periods, it is at once evident that the sacrifice of life is already high in the early years of life. After a fall the curve rises till about middle-age and thereafter falls to the end of life. The curve is very similar for the two sexes, although higher throughout for males, as also in the case of typhoid; but throw- ing the two sexes together into one curve does not materially modify it. If similar curves be constructed for cancer the appearances are markedly different. If both sexes are lumped together and age-periods are regarded, cancer is of no importance as a cause of death till 25-35 J but thereafter the curve rises progressively up to the end of life. If the sexes be considered apart the curves diverge widely, that for women being the higher. At birth can- cer is more common than at 10-15 years : this applies both to car- REVIEW OF RECENT CANCER RESEARCH O cinonia and sarcoma and can be brought into connection with developmental anomalies in which complicated tumours, or sar- comata, more rarely carcinomata, have arisen. From the 20th year onwards the curve for each sex rises progressively, and like- wise both for carcinoma and sarcoma. The old teaching that sarcoma was a disease of youth is untenable; sarcoma is relatively more frequent in youth than is carcinoma, but both become more frequent in a parallel manner as life advances. The old teaching, now shown to be erroneous, was due to the statistical fallacy of generalising from hospital statistics which have no relation to the actual facts of population, its sex and age constitution. The cur\es in which all cases are thrown together, although they give valuable information, conceal very important facts which are only revealed when curves are constructed not only for each sex at the several age-periods, but also for each organ or site of the body. Then the greater frequency of cancer in women is seen to be due to cancer of the breast and uterus. It is also evident that the relation between age and mortality is different for the dift'erent organs but parallel for the same organ in the two sexes. This similarity is illustrated by the extreme example of the cancer of the breast so common in women, so rare in men. The curves are parallel and rise to the end of life in both sexes; the same obtains for ilie liver, skin, rectum, lip, and face. If dift'erent organs are compared, c. g., breast and uterus, the curve for the uterus attains only half the height of that for the Ijreast according to age, although there is a greater total mortality from cancer of the uterus than for any other single organ. This difference is in part explained because cancer of the uterus becomes a frequent cause of death, and therefore causes more deaths at earlier ages than does cancer of the breast. Nevertheless, the curves l)ring out real differences in the susceptibility to cancer at different ages in the two organs; whereas the curve for the breast rises to the end of life, that for the uterus flattens after 55 and falls progressively after 65. .Similar differences exist also for other organs; skin, lij), and face lia\e typical cur\-es, rising to the end <7 0l I2O0 IIOO 1000 900 $00 700 / 1 1 1 1 J \ 1 1 / iAI A LIVE h n ^ \ 1 ' \ 1900 iBoo 1700 1600 ISoo 14001 1300 1200 rlOO 1000 900 800 TOO 600 SOO 300 200 100 1600 1400 1200 1000 Boo 600 4.00 200 3eo 200 100 a iiiL lAI ml- \ 1 \\\ 1 iikis.J / 600 Soo 300 2O0 j! ! -1 : — - i ! R rfAst J- 1 ■ 1 too 7O0 eoo Soo 4-00 300 200 100 800 700 6oo Soo 400 500 200 100 o 40O 3oe 200 100 o Joo 20C IOC — c / \ / \J -■ ' _i rdJ '^1 mf / \ /' N ! /f \ ' J 1 1 h rr nil 1 ft 1 1/ 1 / 7 ^ V ^ / 1 .-, ^. 'r- - 1 ''1 / / y / ; V- T \r F // I IP ;/ V, / .*/ w / f / J / \ ''/ / '> f ,, / ; 11 1 , ikri IMI lOl TO^SfiijE ... V\ ' \ 1 1 1 n '' /iksAp* ATi rt i L_! / 1 1 U "^ II i ' ^' \ — ,>-^ '' ' \ 1 i 1 r^ \- BlAdDEjUae ,_ . 1 _iUREIl|fiA-^ 1 nni Th lAW, I i 1 1 1 y i i ! i : ^■■ ■■■^ 1 . ' 1 ■■•; ±. 1 ---^^>^ ' ' ■ ,-n.l','i— ^-^ 1 .. PXii— ^— 1 riAucs FenAiti Fio. 3. — E.N'fiLAND AND Wales. — Canccr. Age and sex mortality I mm cancer of various parts of the body. 1901-g. 8 REVIEW OF RECENT CANCER RESEARCH error of confounding death-rates with total deaths. It is quite true more people reach the cancer ages but the question is, Do a higher proportion of them die of cancer than did fifty years ago ? A further question is : Why has the increase been greater in men than women? RftTE t^ to 8 5 MILLIOn- 225 . 2/0 195 180 165 150 135 IZO 105 90 75 60 4-5 30 15 LIVERi GALL BLADE TR. UTERU; um-s^uf:g^^ ■ FECTlifT CVAR SKlfl CO . — Cancer oi various parts of the body; mor- tality at all ages, 1897-1910. Males. sidercd apart from the rest of the Ixxly. it is at once apparent that they have not shown so great an increase as the rest of the bodv. I'or long it was pleaded that llic expense would be nut of all proportion to the l)enelit which would follow from tabulating the several organs and sites of the l)ody sc[)aralcly ; 1)ul the luiglish 10 REVIEW OF RECENT CANCER RESEARCH national statistical office and the Imperial Cancer Research Fund have at last obtained this separation, and these data are available from 1 90 1 (really 1897) onwards for 50 different sites. The figures definitely prove how unequally the different parts of the body share in the increase in the number of deaths recorded. For some parts of the body there has been no increase at all dur- ing the past 13 years, c. g., the skin, liver and gall bladder in both sexes (Figs. 4 and 5) ; the uterus and ovary show little or no increase in contrast to the breast, for which there is an increase of 28 per cent. The greatest increase in both sexes is for the stomach and intestines, and it cannot be considered apart from the standstill of the figures for the liver, where primary cancer is known to be so rarely observed at autopsy. In this particular case it is quite obvious we have to do with better diagnosis and a more accurate certification of the causes of death whereby growths previously referred to the secondary site are now allo- cated to the primary site. The question of the importance of these factors has often been discussed, and the influence of the Analysis of the Cases of Neiv Groivths reported by tlie Hospital Authorities (Microscopical Examination has been made in all Cases), 1904-09, all Ages Malignant N ew Growths Wrongly Diagnosed Correctly Diagnosed Not Diagnosed as Cancer Accessible Inaccessible Intermediate 5,854 1. 555 1,322 S67 945 289 488 IS9 160 Total 8,731 1,801 757 Accessible Inaccessible Intermediate Accessible 91.1% 62.3% 82.0% 93.0% 90.7% 89.2% 8.9% 37.8% 18.0% 7.0% 9.3% 10.8% Intermediate accuracy of diagnosis may be illustrated by the results of a com- parison of the clinical diagnoses with the pathological and his- tological findings for 13,000 cases treated in London hospitals. The circumstances under which these diagnoses were made are much more favoural)le than obtain among the population in REVIEW OF RECENT CANCER RESEARCH 11 general, and it should be noted that the increase in deaths affects mainly those internal parts of the body where diagnosis, as the abo\e table shows, is dit^cult, or formerly was impossible before the introduction of autopsies, the great advances of surgery and histolog}-. These factors account in part, if not wholly, for the greater increase in men, when all cases of cancer are con- sidered together, for in them cancer is more often internal. They must be given due weight even if the almost equal import- ance of the improvement in certification of causes of death, and in the accuracy of statistics of population as a whole and of improved statistical methods is taken full account of. For some parts of the body no increase has taken place, for other parts the increase can be explained, if not wholly, at anv rate, largelv. by the foregoing considerations. ])ut for other sites, c. g., the tongue in man and the breast in women, the explanation is not so easy. Comparative investigations have shown how the relative anatomical distribution of cancer in the body is modified by other diseases c. g., Biharziosis. or by the practice of peculiar customs l)y aboriginal races, and there is ever}- justification for conclud- ing that the adopti(jn of these customs in America or in England would at once become responsible for a large increase of cancer in sites in which it is rare or does not occur at all. Hence the possibility of variations in the relative proportions of cases of cancer occurring in different parts of the body or in different areas even in the same country, from time to time, must be ad- mitted. In any case it is still too soon to conclude that for all parts of the body the accuracy of diagnosis and of certification of causes of death had increased, but cancer itself had not in- creased for any single site or organ, although this exi)lanati(m certainly holds for the mouse. In 1896 Tivingood described in America two cases of carcinoma in a mouse, but tweU'c vears ago not a single case had been recorded in F.ngland or in Ger- many. To-day, about 1,000 cases each ha\"e been recorded from Khrlich's laboratory in Frankfort, and the Imperial Cancer Re- search Fund in London. \\'hy? Sim])ly because the disease has been looked for and found where previously it was unsus- pected. The same remarks apply in my personal experience to 12 REVIEW OF RECENT CANCER RESEARCH the COW, which is very hable to several forms of cancer. It may be asserted that it is quite wrong to talk about an increase of cancer in general, and to frighten the public unduly by extrava- gant statements based upon figures unworthy of the name of sta- tistics. I am sorry to say this criticism applies with force to the United States where the registration of births and deaths is so imperfect and the difficulties bound up with the large floating population partly inseparable from the vast amount of immigra- tion such that, combined, they render any comparison between the past and the present or between different States impossible. My remarks will have made it evident that the question of the increase of cancer cannot be discussed in general terms on the basis of lumping all cases together and stating them as so many dcatJis per i,ooo or 100,000 of population, no regard being taken of age and sex, which is the only method available for the United States. I am sorry to say it is possible for me to draw more reliable conclusions as to the incidence of cancer in the nati\'es of some British colonies than in the population of New York. In considering the real or apparent increase of cancer, it is necessary to regard the different parts of the body separately for other reasons, since it has been asserted, especially in Germany, that its most anxious aspect was that the increase affects mainly the younger ages. This is certainly not the case in England and I ha^■e not found reliable evidence that it is so in Germany. The accompan3nng tables show quite distinctly that in England it is the higher age-periods that are chiefly affected. This fact is not so remarkable if the relati\e proportions of persons dying of "old age" is compared with the relative proportion of deaths from cancer to deaths from all causes. Each amounted to 65 per cent, in the year 1900. Although the present generation looks after the aged better than in years gone by, there is vet room for great improvement which, as in the past so also in the future, will lead to deaths l)eing transferred to an accurate cause of death, in many cases to cancer. Up till now I have confined my remarks mainly to England but all the reservations and criticisms made in discussing the real REVIEW OF RECENT CANCER RESEARCH 13 or apparent increase of cancer in that country apply with added force to all other countries. In Greece, Portugal. Turkey, and Russia, no statistics exist. For France. Denmark, Sweden, Roumania, and Bulgaria they are compiled only for the towns. Death-rates per i,ooo Persons Living, iSSi-igog Countries (Arranged in Order of Crude Rates in 1901-5) Switzerland The Netherlands . . . . England and Wales . . Scotland Austria Victoria Ireland New Zealand South Australia Prussia New South Wales . . . Belgium Queensland Tasmania Italy Ontario, Province of. Western Australia . . . Spain Hungary Servia Crude Rates Corrected Rates 1881- 1886- 1891- 1896- I90I- 1901- 1906 1907 1908 1909 1885 1890 i89r -1900 1905 1905 1.03 1. 14 1.22 1.27 1.30 1. 10 1. 12 1.06 I. II 0.60 0.70 0.81 0.92 0.97 0.85 0.88 0.89 .0.90 0.92 0.90 0-55 0.63 0.71 0.80 0.86 0.86 0.92 0.91 0.95 0.54 0.62 0.69 0.77 0.84 0.83 0.94 0.94 0.94 — 0.44 0.50 0.59 0.69 0.74 0.69 0.73 0.72 0.72 — 0.45 0.53 0.62 0.69 0.74 0.76 0.77 0.82 0.81 0.82 0.38 0.43 0.49 0.58 0.69 0.56 0.64 0.62 0.62 0.65 0.30 0.42 0.52 0.59 0.67 0.75 0.79 0.82 0.79 0.82 0.32 0.39 0.48 0.56 0.67 0.76 0.84 0.80 0.77 0.86 0.34 0.41 0.50 0.57 0.6s 0.64 0.69 0.72 0.73 0.74 0.27 0.36 0.43 0.54 0.64 0.80 0.85 0.87 0.84 0.90 — — — — 0.582 0.49 0.49 0.51 0.54 — 0.25 0.27 jo.34 0.44 0.57 0.79 0.76 0.90 0.71 0.83 — 0.49 0.49 0.5s 0.56 0.68 0.63 0.77 0.82 0.82 — 0.43' 0.44 0.51 0.5s 0.45 0.51 0.50 0.53 0.53 0.21 0.29 ? 0.44' 0.52 — — — — — 0.33 0.41 0.31 0.31 0.45 0.74 0.98 0.83 0.86 1.09 — — — — 0.44 0.38 0.41 0.41 0.44 0.44 — , — — o.30> 0.39 0.38 0.39 0.41 0.42 0.43 — — 0.061 0.08 O.IO — — — — — Therefore only the crudest comparison on the basis of the num- ber of deaths per looo li\'ing is ])(issil)le, and it is pn)hable that this comparison as set out in the accompanying ta])le is valueless. In all probability the figures in this table illustrate simply the degree of the development of the statistics of the several coun- tries and do not reveal real differences in the incidence of cancer. Read from above, downwards, they indicate the increasing worth- lessness of the statistics from Switzerland, where a compulsory medical inspection of the dead obtains, to Ser\ia where they probably have no value at all. l\ead from left to right, they in- dicate mainlv the impro\-ement in the statistics of each Cf^untry. The figures for tlie Tnited Slates ha\e been .already referred to. Allhougli comparisons l)(.-l\\een l""nropcan cnnniiirs are use- ^ Four years. - 3 years. 14 REVIEW OF RECENT CANCER RESEARCH less, the study of cancer in widely removed peoples has had results of importance. It has been definitely proved that the idea that cancer is rare or does not occur at all in certain parts of the earth is false, c. g., Egypt, India, Japan. In Japan there are 30,000 to 40,000 deaths a year, though the inhabitants are mainly rice eaters, and there is no discoverable difference between the parts of that country where fish enters more largely into the diet. In India, cancer is by no means rare either in vegetarian or other castes. Of course no conclusions as to the relative frequency of cancer is possible, but it has been ascertained that diet, climate, and race, if not of no importance at all, are of little moment in comparison with chronic irritation. Here again it is not by lumping all cases of cancer together, but by separately consider- ing the several sites and organs, that advances have been made. Through chronically irritating certain parts of the body by prac- tising certain exotic native customs the dwellers in widely re- moved parts of the earth have performed unintentional experi- ments of the highest value, in that they have thereby changed the anatomical distribution of cancer as it is known in the bodies of Europeans. Epithelioma of the skin of the abdomen, prac- tically unknown in Europe, is very common in Kashmir, where the Kangri or fire-basket filled with burning charcoal is carried round the abdomen and leads to repeated burning. Carcinoma of the mouth, so rare in women although not rare in men in Europe, is as common in women in certain parts of India as in men. In China rice is eaten very hot by the men. who are served first In- the women. The latter get the rice cold ; they escape the carcinoma of the oesophagus so common in the men. In the regions of the Upper Nile cases of melanotic sarcoma are not uncommon on the sole of the foot and always subsec^uent upon the injury due to the entrance of a thorn. The fullest biological importance is given to these observations by corresponding ob- servations on animals ; for example in India, where cattle are harnessed by the right horn, epithelioma develops in consequence at the root of that horn but ne\-er at the root of the left one. In parts of England and Ireland cancer is very common in the liver of the cow, alwavs associated with cirrhosis of that oroan. REVIEW OF RECENT CANCER RESEARCH 15 Thus comparative investigations have given a new and enhanced importance to the forms of chronic irritation — kipus-scar, burn- scar, bilharzia, etc. — long known, or new, hke the X-rays. Like the older forms of irritations these newer, or hitherto neglected forms of irritation, have nothing in common, unless it be argued they permit the entrance of a ubiquitous cancer parasite. In considering the importance of irritation in Europe and America again, it is necessary to consider the different parts of the body separately. In the case of the different parts of the intestinal canal, the curves of relative frequency cross at the stomach in both sexes ; above the stomach cancer is more common in the male, below it more common in the female, both in the national statistics and in hospital statistics of England. This circumstance may not be without relation to the different habits of the two sexes, the male irritating the upper half of the canal by smoking, alcohol, gulping his food, etc., more than the female, who is more prone to the chronic irritation of constipation. Moreover, the unwillingness or inability of women to nurse chil- dren may not be unconnected with the increase in the number of deaths recorded from cancer of the breast. The increase for cancer of the tongue in men and the stomach and intestines in both sexes, therefore, should, perhaps, not be dismissed as due merely to improved diagnosis and certification of the causes of death. The importance of considering different sites apart can also be argued on a comparative basis. In surveying the incidence of cancer in the vertebrate kingdom, one has been struck by the fact that certain forms of cancer appear to preponderate in differ- ent classes. It is. of course, obvious that the incidence of cancer in re])rcsentati\es of the diff'erent zoological classes must differ, since, c. c/., structures peculiar to mammals are absent in other vertel)rates. Ihit if we consider the mammalia themselves, it api)ears probal)lc that some species are very liable to forms of cancer from which others, even nearly allied, are relatively or altogether exempt, as illustrated, r. m lime to time in Inniors of the same strain, lliat a person ignorant (•f its historv would certainlv have held thev were reallv distinct Lecture II. Experimental Study of Tumor Cells Up to the present, tumours of mammals when transplanted into other individuals have grown progressive!}' only in other animals of the same species. An inoculation of a tumour from another species does not alter the suitahility of an animal for the subse- quent transplantation of a tumour of its own species. The claim has been made that resistance could be produced in this latter way; but our very extensive observations show that the constant result is as stated, whereas the occasional and apparent protection obtained after inoculation of strange tumour is ex- plicable by accidental circumstances. If the animals are ill from any cause, c. g., from a too large inoculation dose, Sepsis, enteritis, etc., they are rendered less suitable for transplantation, and an appearance of resistance is produced. This point is of impor- tance because the apparent protection after inoculation of a tumour of a strange species has been used as evidence that the tumours of different species, especially the sarcomata, have some- thing in common in the nature of a parasitic etiology. As a matter of fact the inoculation of mice with transplantable rabbit or rat sarcoma has no effect upon a later inoculation of mouse sarcoma. Thus, the specificity of cancer is ]:)ro\-ed in two ways: by absence of power to grow progressively in a strange species and by failure of a strange tumour to induce resistance to homol- ogous inoculation. The specific character of tumours is, how- ever, brought out even more clearly ])y the induction of resistance by means of the normal tissues of the same species, but not of a strange species. Even in the same species biological differences between tumours can be detected by these methods. If a tumour be inoculated which takes in loo per cent, and all the resulting tumours soon undergo spontaneous healing, the animals will then be found to be completely protected against a re-inoculation of 20 REVIEW OF RECENT CANCER RESEARCH 21 the same tumour. Against other tumour strains there may be complete or incomplete protection, or, indeed, there may be no pro- tection at all. In this connection there appears to be some dif- ference between carcinoma and sarcoma, and although the sub- ject requires further investigation it seems that sarcoma protects better against carcinoma than the latter does against sarcoma. Thus, there is a degree of immunity which is common to all tumours of a species, a pan-immunity in Ehrlich's sense, and degrees of immunity which are specific. The pan-immunity de- 1 44E. Lmbrya Zlkin 1 umar LanTral 9> Fig. 7.1 Similarity of resistance produced by tumuur and liy normal tissue respectively. I)ends on the properties of the tumours as tissues of the species ; the specific differences in the tumours will be considered later. The high degree of protection produced by normal tissue is most remarkable, as the accompanying curves from W'oglom's experi- ments show. The two cur\es run parallel and the coni[)arison 1 Reprinted hy courtesy of the Jminuil of lixf'criiiiciitiil M i'Jiciiu'. 22 REVIEW OF RECENT CANCER RESEARCH established between tumour and normal tissues shows that the resistance induced by tumour is not due to any foreign agency. ^Mlen normal tissue is inoculated, necrotic tissue and other irreg- ularities insepara1)le from tumour are avoided; therefore, too much weight must not be laid upon the fact that normal tissue in- duces a higher degree of resistance than does tumour tissue. Nevertheless, this difference may have some etiological signifi- cance and will be referred to again. In order to induce resistance the normal and tumour cells must be ali\'e. If their vitality be destroyed in any way, e. g., by heat, chemical means, mechanical crushing, or radium, the power to induce resistance is entirely lost. Unable to separate this property from the life of the cells, we are in a position simi- lar to that obtaining before the ferment action of yeast was separated from the living cells. The resistant change is spread throughout the body, and state- ments to the effect that a subcutaneous inoculation will not pro- tect against a later inoculation into an internal organ are erron- eous. The protection is just as effective there, and, indeed, can be shown to be active even in the blood. If tumour emboli be induced in the lungs of mice immunised with normal or tumour tissue, the emboli do not develop into tumours as in normal ani- mals, but remain confined within the vessel walls and degenerate. It is evident that complete protection can be conferred against the introduction of the cancer cell from without, so that were cancer spread in this way or communicated from individual to individual as has been alleged in so-called cancer a deux, we should already be alile to confer exemption on the communitv. The foregoing results have reference only to testing animals by an inoculation. Once a tumour has established itself the resist- ance has much less effect, except in the case of tumours which in- duce resistance to their own growth and in consequence undergo spontaneous healing. .Vn animal's own tissues do not induce any re- sistance against tumour inoculation. Animals already bearing trans])lanted tumours can l)e immunised against a second inocula- tion, so, also, mice with spontaneous tumours can be protected against the inoculation of a transplantable tumour; but in the REVIEW OF RECENT CANCER RESEARCH 23 latter case the resistance is without any effect upon an inocula- tion of an animal with its own spontaneous tumour. The prog- ress of spontaneous tumours and the formation of metastases is in no way affected by any of these methods. ]\Iice which have been kept immunised for months will develop spontaneous tu- mours of their own, which fact also speaks for the endogenous origin of cancer. These facts discredit all claims to cure cancer b}' means of autolytic products, vaccines, or immune sera. The importance of the resistance that can be induced against inocula- tion is purely biological, enabling us to obtain information on some hitherto hidden qualities of the tumour cell and its relation to the organism, and leading to the inevitable conclusion that every case of cancer is a problem individual to the person in whom it develops. They have no therapeutic value as yet and it is not apparent if they ever will have any. Furthermore, they throw no light on the great frecjuency of cancer. Other observations to which I have already referred have explained the relative frecjuency, in different human races, of certain forms of cancer associated with chronic irritation. Sim- ilar conditions were also noted for animals and the fact also alluded to that cancer has special predilections for certain mam- malian organs which are exempted in other species. I pass now to consider the results of the prolonged propaga- tion of tumours of the mouse, wliich is peculiarly liable to can- cer of the mamma. This is a Ijiological problem of as much im- portance as the great liability of the human female to cancer of the Ijreast. In the case of the mouse, Haaland has shown how common is chronic inllammation in the mamma of old females and has pointed out one cause. \\'/.., the presence of nematodes, tliough il does not follow lliat this is the only cause. The chronic inllammation is associated w itli hy|)ertrophic nodules and adenoma- tous and cancerous changes in the ei)ithelium in all combinations. A similar combination occurs in the liver of cows, where cirrhosis, h}'pertropln'c nodules, adenoma, and c.nrcinoma — witli extensi\e metastases in the lungs and lymph glands — are frequently com- l)ined. Tiie cirrhosis is here associated with the presence of li\er-llukes in the l)ile duels, 1)ut it is 1)\- n^ means certain tliai the 24 REVIEW OF RECENT CANCER RESEARCH cirrhosis is not due to some chemical substance absorbed from the food. Thus, the more comparative the study of chronic irrita- tion becomes, the more it increases in importance as a mediate or indirect cause of certain forms of cancer. It is ahnost certain that an irritant effective in one species of mammal will not be an eft'ective /. c, indirect, etiological factor in another species. The forms of irritation are manifold; they have nothing in common beyond their causing cell injury and their induction of cell prolif- eration. When tumours are transplanted the cell proliferation is simply prolonged artificially, and in reality by again causing cellular trauma we are continuing the experiment which had occurred naturally. This prolonged propagation of tumours has been criticised as of no etiological value, but I shall show you how greatly it has added to knowledge, and that by observing the behaviour of tumour cells over a prolonged period certain inferences may be drawn as to how the cancer cells were likely to behave before they were placed under experimental condi- tions. The longing expressed by Goodhart in 1875: "What a subject for Darwin would be the cells of a cancer if only they were tangible; how the immortal pigeon would be completely eclipsed, while the hungry pathologist would be filled with food, if we could observe the variation of tumours under judicious cultivation!" has now been gratified, if cancer has not yet found its Darwin. For more than ten years the behaviour of a large number of transplantable tumour .strains has been carefully noted. The observations fall mainly into two groups, those relating to mor- phology and those bearing on the phenomena of growth. It is im- possible to consider here all the factors concerned in the first transference of a spontaneous tumour to normal animals. It is a true transplantation, the daughter tumours developing from the tumour cells, whilst the host supplies the connective tissue and vascular scaffolding. A certain selection is effected of those cells adapting themselves to the strange conditions or able to live under them. The soil acts, as it were, as a sieve allowing cer- tain cells to pass, selecting only some that are capable of growth. At the next transplantation these have increased in number, and REVIEW OF RECENT CANCER RESEARCH 25 by the third or fourth passage the percentage of takes reaches a level which has remained constant for years for the majority of our tumours, when grown in as many parallel sister strains as possible. Most of the tumour strains show as yet unexplained and apparently periodic fluctuations from time to time. There are, however, some remarkable exceptions to constant behaviour and to them special reference will be made. AMien a tumour is transplanted it undergoes entire histolog- ical disorganisation. This process has been repeated over and over again every se^■en or ten days, or after longer intervals, for years. Assuming that the only common factor for all the irri- tants associated with cancer is their power to induce proliferation, the process is as analogous as is possible to the chronic prolifer- ation which precedes the development of cancer, always bearing in mind that it is the read}- made tumcjur or cancer cell which is being studied. Xutwithstanding the fact that we are not produc- ing the cancer cell at will, but are merely studying its behaviour — since it is clearly demonstrated that the cancer cell is a genealog- ical descendant of a normal cell — we are justified in assuming that the propagated cancer cell will behave similarly to its geneal- ogical ascendants in its host of origin provided, of course, that we largely neglect the possible influence of strange environment. After the disorganisation caused by transplantation, it is nec- essary to give time for a tumour to resume its normal haljits. Therefore, it is advisal)le to study not merely the material pre- ser\-ed at the time of each transplantation. Init also tuniour> that have grown for some time. It is then found that the char- acteristic histological structure is reproduced as a rule, it may l)e very rapidly, or it may be only after prolonged growth. Tliis characteristic diO'ercnliation occurs in si)ite of the fact that the connective tissue and vascular scaffolding has to be reproduced anew by eacli fresli liost, taking i)lace with e(jual regularity in adenoma, cystic papillifcroiis carcinoma, and solid carcinoma of the mamma, as well as in squamous-cellcd carcinoma and in ade- noma of the sebaceous and preputial glands. The latter, notwith- standing vears of ])ropagation, still re|)roduces a structure whicli cannr)t l)e distinguislied under tlie microscojK' from tlic normal 26 REVIEW OF RECENT CANCER RESEARCH preputial gland. The conclusion clearly follows that tumours which histologically are "benign" can be propagated for the same length of time as those which are devoid of all differentiation and are in current terminology histologically "malignant." Experi- ment therefore supports clinical experience in its inability to establish a sharp division between benign tumours — for which a parasitic etiology is not assumed — and the malignant new growths for which a parasitic etiology is asserted to be essential. Other more subtle features revealed by microscopical tech- nique may be retained unaltered, c. g., tumours which form gly- cogen in large droplets have continued to do so for years, from the spontaneous tumour onwards. A tumour strain which caused a sarcomatous transformation of the connective tissue scaffolding supplied by the new hosts has continued for six years to effect this change in every mouse in which it has grown for 60 days. In 1905, attention was drawn to the fact that the propagated tumours derived from a single parenchyma, viz. the mammary epithelium, varied so much in microscopical appearances that a person ignorant of their life-history would declare them to be of distinct primary origin. This early observation led to the ex- tensive propagation of a large number of tumour-strains of sepa- rate primary origin, as many as 86 strains being kept in propa- gation at one time, many of them in several parallel sister strains. The net result is that the histological differences present at the outset of propagation of tumours arising from the same paren- chyma, remain in many cases constant, and are, therefore, to be re- garded as true variations of relative constancy. The further con- clusion may be drawn that the change which occurs in the trans- formation from normal to the cancer cell deeply affects the cell mechanism, and not only in one way, but in several ways. Although the constancy with which strains maintain their his- tological features is most noteworthy the exceptions to this gen- eral rule are perhaps of even greater biological significance. Two s(|uamous celled carcinomata have grown for years and have always shown complete differentiation; but another one behaved in this way only up to the ninth transplantation and REVIEW OF RECENT CANCER RESEARCH 27 has grown for years without any reappearance of keratinisation. A fourth tumour of the same kind has also behaved in the latter wav, and some adenomatous tumours have likewise l!)St all evidence of acinous differentiation. There are adenomatous tumours which have exhibited a most remarkable histological change; after transplantation there is the usual undifferentiated cell mass, but areas of spindle cells appear later, indistinguish- al)le histologically from spindle celled sarcoma, which later differ- entiate again into typical acini and reproduce the picture of the mother tumour. This spindle celled appearance may also become constant, and only in the very oldest tumours may there l)e found evidence of the tendency to acinous dift'erentiation. Sim- ilar morphological changes have also been observed among the propagable sarcomata. An osteo-chondro-sarcoma has been grown in two parallel strains. One shows a great tendency to necrosis and is soft to the touch, the other feels hard and forms much collagen. Both strains have lost all tendency to form either cartilage or bone, and cannot be identified with the mother tumour except from their life histories. As a last example an adeno-carcinoma which has been grown in i8 parallel sister strains may be referred to. It illustrates alike the constancy and the variability of the histological structure and other i)rop- erties of the epithelium. I'or three years some of the strains of this tumour caused irregularly sarcomatous transformation in the stroma, and the transformation was to spindle-, round-, and p()l\"mi)rph<)us-cclled sarcoma. Other sister strains ])os- sessed no such property but have grown as pure adeno-carcinoma or as pure solid carcinoma. Now, after several years, all the strains have lost the power to produce sarcoma. One of these strains, altliougli remaining a pure cpilhclial lum