beginning of the present century, the gross anatomical characters of
cancer and the main forms of the disease came to be more clearly
recognized. After the description of encephaloid cancer by Laennec[1]
in 1812, {531} and the first clear recognition of colloid cancer by
Otto[2] in 1816, these two forms of cancer took rank with scirrhus as
constituting the varieties of cancer of the stomach as well as of
cancer elsewhere. All that it was possible to accomplish in the
description of cancer of the stomach from a purely gross anatomical
point of view reached its culmination in the great pathological works
of Cruveilhier (1829-35) and of Carswell (1838), both of whom admirably
delineated several specimens of gastric cancer.
[Footnote 1: _Dict. des Sciences méd._, t. i. and t. xii., Paris,
1812-15.]
[Footnote 2: Otto, _Seltens Beobachtungen, etc._, 1816.]
During this period of active anatomical research the symptomatology of
gastric cancer was not neglected. The article on cancer by Bayle and
Cayol in the _Dictionnaire des Sciences médicales_, published in 1812,
shows how well the clinical history of gastric cancer was understood at
that period.
Cylindrical-celled epithelioma of the stomach could not be recognized
as a separate form of tumor until the application of the microscope to
the study and classification of tumors--an era introduced by Müller in
1838.[3] Cylindrical-celled epithelioma of the stomach was first
recognized by Reinhardt in 1851, was subsequently described by Bidder
and by Virchow, and received a full and accurate description from
Förster in 1858.[4]
[Footnote 3: _Ueber den feineren Ban, etc., der krankh. Geschwülste_,
Berlin, 1838.]
[Footnote 4: Reinhardt, _Annalen d. Charité_, ii. 1, 1851; Bidder,
_Müller's Archiv_, 1852, p. 178; Virchow, _Gaz. méd. de Paris_, April
7, 1855; Förster, _Virchow's Archiv_, Bd. 14, p. 91, 1858.]
Until the publication by Waldeyer[5] in 1867 of his memorable article
on the development of cancers, it was generally accepted that gastric
cancer originated in the submucous coat of the stomach, and that the
cells in the cancerous alveoli were derived from connective-tissue
cells. Waldeyer attempted to establish for the stomach his doctrine
that all cancers are of epithelial origin. In all varieties of gastric
cancer he believed that he could demonstrate the origin of the
cancer-cells from epithelial cells of the gastric tubules--a mode of
origin which had previously been advocated for cylindrical epithelioma
by Cornil[6] (1864). Waldeyer's view has met with marked favor since
its publication, but there are eminent pathologists who have not given
adherence to it in the exclusive form advocated by its author.
[Footnote 5: _Virchow's Archiv_, Bd. 41.]
[Footnote 6: _Journ. de l'Anat. et de la Phys._, 1864.]
It is somewhat remarkable that although in the early part of the
present century several monographs on gastric cancer appeared,[7] all
the more recent contributions to the subject are to be found in theses,
scattered journal articles, and text-books. Of the more recent careful
and extensive articles on cancer of the stomach, those of Lebert and of
Brinton are perhaps most worthy of mention.[8]
[Footnote 7: Chardel, Benech, Daniel, Germain, Prus, Sharpey, Barras,
etc.]
[Footnote 8: Lebert, _Die Krankheiten des Magens_, Tübingen, 1878;
Brinton, _Brit. and For. Med.-Chir. Rev._, 1857.]
ETIOLOGY.--The data for estimating the frequency of gastric cancer are
the clinical statistics of hospitals, series of recorded autopsies, and
mortuary registration reports.
Statistics with reference to this point based exclusively upon the
clinical material of hospitals have only relative value, as they do not
{532} represent in proper proportion both sexes, all ages, all classes
in life, and all diseases. Statistics based upon autopsies surpass all
others in certainty of diagnosis, but they possess in even greater
degree the defects urged against hospital statistics. Not all the fatal
cases in hospitals are examined post-mortem, and gastric cancer is
among the diseases most likely to receive such examination. Hence
estimates of frequency based exclusively upon autopsies are liable to
be excessive. Estimates from mortuary registration reports, and
therefore from the diagnoses given in death-certificates, rest
manifestly upon a very untrustworthy basis as regards diagnosis, but in
other respects they represent the ideal point of view, including, as
these reports do, all causes of death among all classes of persons. It
is evident that in all methods of estimating the frequency of gastric
cancer inhere important sources of error. In general, the larger the
number of cases upon which the estimates rest the less prominent are
the errors. Such estimates as we possess are to be regarded only as
approximate, and subject to revision.
From mortuary statistics Tanchou estimates the frequency of gastric
cancer as compared with that of all causes of death at 0.6 per cent.;
Virchow, at 1.9 per cent.; Wyss, at 2 per cent.; and D'Espine, at 2½
per cent.[9]
[Footnote 9: Tanchou, _Rech. sur le Traitement méd. des Tumeurs du
Sein_, Paris, 1844. These statistics, which are based upon an analysis
of 382,851 deaths in the department of the Seine, are necessarily
subject to sources of error, but they do not seem to me to deserve the
harsh criticisms of Lebert and others.
Virchow, _Verhandl. d. phys.-med. Gesellsch. Würzburg_, 1860, vol. x.
p. 49--analysis of 3390 deaths in Würzburg during the years 1852-55.
Wyss, quoted by Ebstein in _Volkmann's Samml. klin. Vorträge_, No.
87--analysis of 4800 deaths in Zurich from 1872-74.
D'Espine, _Echo médical_, 1858, vol. ii.--mortuary statistics of the
canton of Geneva, considered to be particularly accurate.]
In 8468 autopsies, chiefly from English hospitals, Brinton[10] found
gastric cancer recorded in 1 per cent. of the cases. Gussenbauer and
Von Winiwarter[11] found gastric cancer recorded in 1½ per cent. of the
61,287 autopsies in the Pathological Anatomical Institute of the Vienna
University. From an analysis of 11,175 autopsies in Prague, I find
gastric cancer in 3½ per cent. of the cases.[12]
[Footnote 10: _Loc. cit._]
[Footnote 11: _Arch. f. klin. Chirurg._, Bd. xix. p. 372.]
[Footnote 12: Statistics of Dittrich, Engel, Willigk, Wrany, and
Eppinger, in _Prager Vierteljahrschr._, vols. vii., viii., ix., x.,
xii., xiv., xxvii., l., xciv., xcix., and cxiv. Grünfeld found in 1150
autopsies in the general hospital for aged persons in Copenhagen 102
cancers of the stomach, or 9 per cent. (_Schmidt's Jahrb._, Bd. 198, p.
141).]
I have collected and analyzed with reference to this point the
statistics of death from all causes in the city of New York for the
fifteen years from 1868 to 1882, inclusive.[13] I find that of the
444,564 deaths during this period, cancer of the stomach was assigned
as the cause in 1548 cases and cancer of the liver in 867 cases.
Probably at least one-third of the primary cancers of the liver are to
be reckoned as gastric cancers. This would make the ratio of gastric
cancer to all causes of death about 0.4 per cent. This ratio becomes
about 1 per cent. (0.93) if only the deaths from twenty years of age
upward be taken: gastric cancer hardly ever occurs under that age. It
is probably fair to conclude that in New York not over 1 in 200 of the
deaths occurring at all ages and from all causes {533} is due to cancer
of the stomach, and that about 1 in 100 of the deaths from twenty years
of age upward is due to this cause.
[Footnote 13: These statistics are obtained from the records of the
Board of Health of the city of New York. These records are kept with
great care and system.]
The organs most frequently affected with primary cancer are the uterus
and stomach. In order to determine the relative frequency of cancer in
these situations, I have compiled the following table of statistics
from various sources:[14]
Primary Cancers. Stomach. Uterus.
------------------------- -------------- --------------
11,131 in Vienna 10 per cent. 31 per cent.
7,150 in New York 25.7 " " 24.2 " "
9,118 in Paris (Tanchou) 25.2 " " 32.8 " "
1,378 in Paris (Salle) 31.9 " " 32 " "
587 in Berlin 35.8 " " 25 " "
183 in Würzburg 34.9 " " 19 " "
1,046 in Prague 37.6 " " 33.3 " "
889 in Geneva 45 " " 15.6 " "
------ ---- ----
31,482 total 21.4 per cent. 29.5 per cent.
From this table it appears that in some collections of cases the uterus
is the most frequent seat of primary cancer, while in other collections
the stomach takes the first rank. If the sum-total of all the cases be
taken, the conclusion would be that about one-fifth of all primary
cancers are seated in the stomach, and somewhat less than one-third in
the uterus. Even if allowance be made for the apparently too low
percentage of cases of gastric cancer in the large Vienna
statistics,[15] I should still be inclined to place the uterus first in
the list of organs most frequently affected with primary cancer, and to
estimate the frequency of gastric cancer compared with that of primary
cancer elsewhere as not over 25 per cent.
[Footnote 14: Vienna cases: Gurlt, _Arch. f. klin. Chir._, Bd. xxv. p.
421--statistical analysis of 16,637 tumors observed in the three large
hospitals of Vienna from 1855 to 1878. New York cases: see preceding
foot-note. Paris cases: Tanchou, _op. cit._, and Salle, _Étiologie de
la Carcinose_, Thèse, Paris, 1877, p. 145 _et seq._--fatal cases in
Paris hospitals, 1861-63. Berlin cases: Lange, _Ueber den Magenkrebs_,
Inaug. Diss., Berlin, 1877--post-mortem material. Würzburg cases:
Virchow, _loc. cit._, and _Virchow's Archiv_, Bd. 27, p. 430. Prague
cases: reference given above--post-mortem material. Geneva cases:
D'Espine, _loc. cit._]
[Footnote 15: That this percentage is too low is apparent from the fact
that the number of cases of gastric cancer is only twice that of
primary cancer of liver in Gurlt's statistics.]
The liability to gastric cancer seems to be the same in both sexes. Of
2214 cases of gastric cancer which I have collected from hospital
statistics, and which were nearly all confirmed by autopsy, 1233 were
in males and 981 in females.[16] This makes the ratio of males to
females about 5 to 4. This difference is so slight that no importance
can be attached to it, especially in view of the fact that in most
hospitals the males are in excess of the females.
[Footnote 16: My statistics regarding sex are obtained from _Prager
Vierteljahrschr._, vols. xvii., l., xciv., xcix., cxiv.; Lange, _op.
cit._; Katzenellenbogen, _Beitr. zur Statistik d. Magencarcinoms_,
Jena, 1878; Leudet, _Bull. de l'Acad._, t. 29, p. 564; Gussenbauer and
V. Winiwarter, _loc. cit._; Lebert, _op. cit._; Habershon, _Diseases of
Abdomen_, Philada., 1879; and _Ann. d. Städt. Allg. Krankenh. zu
München_, Bd. i. and ii.
If to these accurate statistics be added collections of cases from
heterogeneous sources, including mortuary statistics (Brinton, Louis,
D'Espine, Virchow, Gurlt, Welch), there results a total of 5426 cases,
with 2843 males and 2583 females, the two sexes being more evenly
represented than in the more exact statistics given in the text. In
this collection of cases Gussenbauer and V. Winiwarter's cases only up
to the year 1855 are included, as the subsequent ones are doubtless in
great part included in Gurlt's statistics. According to Brinton,
gastric cancer is twice as frequent in males as in females.]
{534} The following table gives the age in 2038 cases of gastric cancer
obtained from trustworthy sources and arranged according to
decades:[17]
Age. Number of cases. Per cent.
------ ---------------- ---------
10-20. 2 0.1
20-30. 55 2.7
30-40. 271 13.3
40-50. 499 24.5
50-60. 620 30.4
60-70. 428 21
70-80. 140 6.85
80-90. 20 1
90-100. 2 0.1
Over 100. 1 0.05
From this analysis we may conclude that three-fourths of all gastric
cancers occur between forty and seventy years of age. The absolutely
largest number is found between fifty and sixty years, but, taking into
consideration the number of those living, the liability to gastric
cancer is as great between sixty and seventy years of age.
Nevertheless, the number of cases between thirty and forty years is
considerable, and the occurrence of gastric cancer even between twenty
and thirty is not so exceptional as is often represented, and is by no
means to be ignored. The liability to gastric cancer seems to lessen
after seventy years of age, but here the number of cases and the number
of those living are so small that it is hazardous to draw positive
conclusions.
[Footnote 17: The sources of the statistics for age are--Dittrich
(160), _Prager Vierteljahrschr._, vol. xvii.; D'Espine (117), _loc.
cit._; Virchow (63), _Virchow's Archiv_, Bd. 27, p. 429; Leudet (69),
_loc. cit._; Lange (147), _op. cit._; Katzenellenbogen (60), _op.
cit._; Gussenbauer and Von Winiwarter (493 cases up to 1855), _loc.
cit._; Lebert (314), _op. cit._; Habershon (76), _op. cit._; Gurlt
(455), _loc. cit._; _Trans. N.Y. Path. Soc._, vol. i. (41); and _Trans.
London Path. Soc._, vols. i.-xxxiv. (43). The results correspond
closely to those of the smaller statistics of Brinton and of Lebert.]
Cancer of the stomach in childhood is among the rarest of diseases.
Steiner and Neureutter[18] failed to find a single gastric cancer in
2000 autopsies on children. Cullingworth[19] has reported with
microscopical examination a case of cylindrical-celled epithelioma in a
male infant dying at the age of five weeks; it is probable that the
tumor was congenital. It is not certain whether Wilkinson's[20]
often-quoted case of congenital scirrhus of the pylorus in an infant
five weeks old was a cancer or an instance of simple hypertrophy.
Kaulich[21] cites a case of colloid cancer affecting the stomach,
together with nearly all the abdominal organs, in a child a year and a
half old, but whether the growth in the stomach was primary or
secondary is not mentioned. The case which Widerhofer[22] has reported
as one of cancer of the stomach secondary to cancer of the
retro-peritoneal glands in an infant sixteen days old seems from the
description to be sarcoma. Scheffer[23] has reported a case of large
ulcerated encephaloid cancer of the fundus, involving the spleen, in a
boy fourteen years old. Jackson[24] has reported an interesting case of
encephaloid cancer in a boy fifteen years old in whom no evidence of
disease existed up {535} to ten weeks before death. These cases, which
are all that I have been able to find in children, are to be regarded
as pathological curiosities.[25]
[Footnote 18: _Prager Vierteljahrschr._, vol. lxxxix. p. 77.]
[Footnote 19: _British Med. Journ._, Aug. 25, 1877, p. 253.]
[Footnote 20: _London and Edinburgh Month. Journ. of Med._, 1841, vol.
i. p. 23.]
[Footnote 21: _Prager med. Wochenschr._, 1864, No. 34.]
[Footnote 22: _Jahrb. f. Kinderheilk. Alt. Reihe_, Bd. ii. Heft 4, p.
194.]
[Footnote 23: _Jahrb. f. Kinderheilk._, xv. p. 425, 1880.]
[Footnote 24: J. B. S. Jackson, _Extracts from the Records of the
Boston Society for Medical Improvement_, vol. v., Appendix, p. 109,
Boston, 1867.]
[Footnote 25: Mathien (_Du Cancer précoce de l'Estomac_, Paris, 1884)
has recently analyzed, chiefly from a clinical point of view, 27 cases
of gastric cancer occurring under thirty-four years of age. Of these, 3
were under twenty and 14 were between twenty and thirty years. He also
emphasizes the error of considering cancer of the stomach as
exclusively a disease of advanced life.]
Such statistics as we possess would make it appear that gastric cancer,
as well as cancer in general, is somewhat less common in the United
States than in the greater part of Europe.[26] These statistics,
however, are too inaccurate, and the problems involved in their
interpretation are too complex, to justify us in drawing any positive
conclusions as to this point. It is certain that cancer is not a rare
disease in the United States.
[Footnote 26: Of 1000 deaths in New York in 1882, 19.3 were from
cancer. The statistics on this point from some of the large European
cities are--Geneva, 53 deaths from cancer per mille; Frankfort, 47.6;
Copenhagen, 33.2; Christiania, 29; London, 28.7; Paris, 27; Edinburgh,
25.4; Berlin, 22.4; St. Petersburg, 15; Amsterdam, 12. These statistics
are obtained from the _Forty-fourth Annual Report of the
Registrar-General (for 1881)_, London, 1883; from _Preussische
Statistik_, Heft lxiii., Berlin, 1882; and from _Traité de la
Climatologie médicale_, Paris, 1877-80, by Lombard, in whose excellent
work will be found much information on this subject.
To judge from statistics in this country and in England, the death-rate
from cancer is undergoing a rapid annual increase. Whereas in New York
in 1868 this death-rate was only 12.6 per mille, in 1882 it was 19.3.
In England and Wales in 1858 the deaths from cancer per 1,000,000
persons living were 329, and in 1881 they were 520. It seems probable,
as suggested in the above report of the Registrar-General, that this
apparently increasing large death-rate is due to increased accuracy in
diagnosis. It may be also that decrease in infant mortality and
prolongation of life by improved sanitary regulations may account in
part for this increase. From this point of view Dunn makes the
paradoxical statement that the cancer-rate of a country may be accepted
as an index of its healthfulness (_Brit. Med. Journ._, 1883, i.).]
It is said on good authority that in Egypt and Turkey gastric cancer
and other forms of cancer are infrequent.[27] A similar infrequency has
been claimed for South America, the Indies, and in general for tropical
and subtropical countries; but all of these statements as to the
geographical distribution of cancer are to be accepted with great
reserve, as they do not rest upon sufficient statistical information.
[Footnote 27: Hirsch, _Handb. d. Historisch-geographische Pathologie_,
Bd. ii. p. 379, Erlangen, 1862-64.]
I have analyzed the frequency of gastric cancer among negroes upon a
basis of 7518 deaths among this race in New York, and I find the
proportion of deaths from this cause about one-third less than among
white persons.[28] It has been stated that cancer is an extremely rare
disease among negroes in Africa.[29] The admixture with white blood
makes it difficult to determine to what degree pure negroes in this
country are subject to cancer.
[Footnote 28: According to the Ninth Census Report of the United
States, in the census year 1870 the deaths from cancer among white
persons were 13.7 per mille, and among colored persons only 5.7 per
mille; but it is well known that the registration returns upon which
the vital statistics in these reports are based are very incomplete and
unsatisfactory.]
[Footnote 29: Bordier, _La Geographie médicale_, Paris, 1884, p. 464.
Livingstone speaks of the infrequency of cancer among the negroes in
Africa.]
The question as to what rôle is played by heredity in the causation of
gastric cancer belongs to the etiological study of cancer in general.
Probably in about 14 per cent. of the cases of cancer it can be
determined that other members of the family are or have been affected
with the disease.[30] {536} The influence of inheritance, therefore, is
apparent only in a comparatively small minority of the cases. As
suggested long ago by Matthew Baillie, this hereditary influence is
better interpreted as in favor of a local predisposition (embryonic
abnormality?) in the organ or part affected than in favor of the
inheritance of a cancerous diathesis. It has been claimed by D'Espine,
Paget, and others that cancer develops at an earlier age when there is
a family history of the disease than when such history is absent.
[Footnote 30: This statement is based upon the collection of 1744 cases
of cancer analyzed with reference to this question. Of these, a family
history of cancer was determined in 243 cases. The cases are obtained
from statistics of Paget and Baker, Sibley, Moore, Cooke, Lebert,
Lafond, Hess, Leichtenstern, Von Winiwarter, and Oldekop. There is
extraordinary variation in the conclusions of different observers upon
this point. Velpeau asserted that he could trace hereditary taint in 1
in 3 cancerous subjects; Paget, in 1 in 4; Cripps, in 1 in 28. My
conclusions agree with those obtained at the London Cancer Hospital
(Cooke, _On Cancer_, p. 11, London, 1865).
The most remarkable instance of inherited cancer on record is reported
by Broca (_Traité des Tumeurs_, vol. i. p. 151, Paris, 1866): 15 out of
26 descendants over thirty years of age of a woman who died in 1788 of
cancer of the breast were likewise affected with cancer. As is well
known, Napoleon the First, his father, and his sister died of cancer of
the stomach.]
It may be considered established that cancer sometimes develops in a
simple ulcer of the stomach, either open or cicatrized. It is most
likely to develop in large and deep ulcers with thickened edges, where
complete closure by cicatrization is very difficult or impossible. It
is difficult to prove anatomically that a gastric cancer has developed
from an ulcer, and hence such statements as that of Eppinger, that in
11.4 per cent. of cancers of the stomach this mode of development
existed, are of no especial value.[31] No etiological importance can be
attached to the occasional association of cancer with open or
cicatrized simple ulcers in different parts of the same stomach. Of the
comparatively few cases in which strict anatomical proof has been
brought of the origin of cancer in simple gastric ulcer, probably the
most carefully investigated and conclusive is one studied and reported
by Hauser.[32] It is, however, by no means proven that Hauser's view is
correct, that cancer develops from the atypical epithelial growths
often to be found in the cicatricial tissue of gastric ulcer. In a few
instances both the clinical history and the anatomical appearances
speak decisively for the development of cancer in a simple gastric
ulcer;[33] and the establishment of this fact is of clinical
importance.
[Footnote 31: _Prager Vierteljahrschr._, vol. cxiv.]
[Footnote 32: _Das chronische Maqengeschwür_, Leipzig, 1883, p. 61. See
also Heitler, "Entwicklung von Krebs auf narbigen Grunde in Magen,"
_Wien. med. Wochenschr._, 1883, p. 961. It seems to me that at present
there is a tendency to exaggerate the frequency with which cancer
develops from gastric ulcer.]
[Footnote 33: A particularly satisfactory case of this kind is reported
by Lebert, _op. cit._, p. 503.]
Many other factors in the causation of gastric cancer have been
alleged, but without proof of their efficacy. This is true of chronic
gastritis, which was once thought to be an important cause of gastric
cancer, and is even recently admitted by Leube to be of influence.[34]
Certainly the majority of cases of cancer of the stomach are not
preceded by symptoms of chronic gastritis. Although in a few instances
gastric cancer has followed an injury in the region of the stomach,
there is no reason to suppose that this was more than a coincidence.
[Footnote 34: In _Ziemssen's Handb. d. spec. Path. u. Therap._, Bd.
vii. p. 134, Leipzig, 1878.]