Zeitschrift fiir Krebsforsehung 65, 30--36 (1962)

From the Department of Pathology (Head: Prof. D. F. CAP12ELL,M.D.) The University and Western Infirmary, Glasgow, Scotland

A Criticism of the Ilacmatogenous Theory of Cancer Metastasis By WILSON L B. ONUIGBO With 2 figures in the text I t is the purpose of the present paper to point out the various assumptions in the haematogenous theory of cancer metastasis, and to examine the grounds on which their validity m a y be challenged.

1. Visceral versus lymph nodal metastasis. WILLIS, WALTHE~, and others believe that carcinomas usually spread b y way of the lymphatics to the lymph nodes, but b y the blood to the viscera. Histology teaches however t h a t lymph nodes are nourished by arterial blood, and oncology teaches t h a t in cancer patients millions of cancer cells are probably released into the blood stream daily ( M o o ~ ) . Hence, metastasizing cells must inevitably be carried to the lymph nodes b y the blood stream. Consequently, unless it is proved t h a t lymph nodes differentially destroy blood-borne cancer cells but naturally nourish lymph-borne cancer cells, there is no valid ground for assuming t h a t cancer cells in the blood single out the viscera for attack whereas those in the lymph choose the lymph nodes. K the appearances are that most carcinomatous lymph node metastases are lymph-borne, then we must give consideration to the probability t h a t most visceral metastases are also lymph-borne. The prevailing concept of metastasis, which divorces the mechanism of spread to the lymph nodes from the mechanism of spread to the organs, is questionable. 2. Distant metastasis. According to CAMEgO~, for a tumour to spread to distant parts blood invasion is indispensable. However, does the location of a secondary turnout in a distant tissue necessarily mean t h a t the route of invasion must have been haematogenous ? Long ago, COATS (1883) taught t h a t cancer m a y be conveyed lymphogenously to the distant lymph nodes. Recently, both HAD~'I:ELD and SUT~E~LANI) included the lymph vessels among the channels responsible for distant metastasis. In his classical paper, which elucidated the nature of oat cell carcinoma of the bronchus, BARmAiD remarked on the occurrence of distant lymph node infiltration in this disease, the cells passing "downwards via the paraoesophageal to the coeliac lumbar and in some cases to the iliac groups." Se]~z~c]~g also commented on the frequency with which lymphatic permeation progressed as far as to the lower abdominal lymph nodes in carcinomas of the bronchus. We m a y conclude, therefore, t h a t the mere fact of distance can not be equated with blood transportation. 3. Distant metastasis without regional lymph node involvement. For over a century, interest has been shown in carcinomas which have spread to distant organs without apparently involving the regional lymph nodes of the organ from which they arose. This metastatic manifestation is thought to be a demonstration

A Criticism of the I-Iaematogenous Theory of Cancer Metastasis

31

of metastasis by way of the blood stream. However, we need to ask whether this pattern of metastasis is an incontrovertible proof of haematogenous spread. First, it is known that there arc natural channels which m a y be instrumental in causing lymph to circumvent particular groups of nodes (TOLDT). These collateral channels may bring about the metastasis of more distant groups of nodes in the absence of near node metastasis (PERI~OT). For example, the injection experiments of MEYER revealed the existence of uninterrupted lymph channels between the lungs and the upper abdominal organs and nodes. We may expect, in consequence, that in eases of lung cancer retrograde lymphatic spread m a y

Fig. I. The cut s~rface of the liver in a case of lung cancer. S~tperiorly, the parcnchyma is closely packed with whitish nod~les of tnmour; inferiorly, the deposits become sparser centrifugally

occur to the abdominal organs via these channels although the regional thoracic nodes are innocent of secondaries. Conversely, primary tmnours of the abdominal organs may spread directly to the lungs via these same channels without involving the regional abdominal nodes. Secondly, in lymphogenous metastasis much depends on the occurrence of preexisting lesions involving the lymphatic system. Years ago, COATS (1895) argued that blockage of the thoracic nodes by old infection may lead to lymph diversion to the abdominal nodes and that in such an individual cancer of the lung could spread directly to the abdominal nodes. Blockage of the nodes may be the result of inflammatory processes, healed tuberculosis or anthraeosis. Each of these pathological processes can lead to lymph diversion which in turn can lead to metastasis of distant tissues although contiguous nodes escape invasion, or suffer metastasis less grossly.

The Iymphatics are, therefore, capable of conveying tumour cells to distant nodes with little or no involvement of the nearer nodes. The apparent escape of the regional nodes of an organ bearing a tumour does not warrant the inference that any distant deposit must have been blood-borne.

32

WiLso~ I. B. O~vmBo:

4. Discreteness o/ secondary growths. I n the opinion of WILLIS, discrete metastases within remote organs are almost a]ways blood-borne, but as early as 1831 ANDRAL described the occurrence of discrete deposits of tumour on the internal surface of the thoracic duct. Fig. 1 is from a case of lung cancer and shows striking invasion of the liver: most of the deposits are packed discretely in the upper parts and become sparse supero-inferiorly. This pattern is almost certainly not due to dispersal b y the blood stream but to centrifugal spread b y way of the lymph stream. I t is known t h a t some lymph vessels of the liver accompany the inferior vena cava to the posterior mediastinal lymph nodes. STriP,AT recognized the possibility of spherical deposits in the liver being of lymphogenous origin. There is need, then, for a wider recognition of the fact t h a t discrete deposits of tumour are not indubitably haematogenous but are undoubtedly consistent with lymphogenous metastasis. 5. Bilateral metastases. Over a century ago, WALS~E (1846) came to the conclusion t h a t one of the points in favour of the theory of blood-borne metastasis was the frequency with which paired organs were invaded bilaterally. We know t h a t anastomotic lymphatic vessels link up both halves of the body. We expect, consequently, t h a t lymphatic carriage of cancer cells must be able to bring about the development of bilateral metastases. I t is well known t h a t a growth localized to one half of the hypopharynx often involves the deep cervical nodes on the same side, but the nodes on the other side mayalso be attacked ;the bilateral metastases are then smaller on one side. This a s y m m e t r y of bilateral lymph node metastases has been noted in some other tumours (ScHuSTer, JOI~DAN and OHIO, STALEY and SCA~LON, SW~DLOW). I n m y view (O~vmno 1, 2), bilateral visceral metastases which are asymmetrical are an overt evidence of an underlying lymphogenous process. At least, bilaterality of invasion is not a hall-mark of haematogenous carriage of cancer. 6. Multiple metastases. A common concept is t h a t hacmatogcnous dissemination is more likely to produce a shower of emboli than a solitary one. We need, then, to examine the validity of the concept t h a t multiplicity of metastases is suggestive of blood-borne dissemination. Turning to lymphangitis carcinomatosa or lymphvessel carcinomatosis, we find t h a t this process " m a y involve a surprisingly wide a r e a " (WILLIS). If this form of lymphogenous dissemination, which is to the naked eye infiltrative, can be widespread, the probability is t h a t the purely embolic form of spread through the lymphatics would also be capable of yielding widespread and multiple metastases. Fig. 2 shows multiple lymph node metastases, which are distributed centrifugally, in a case of carcinoma of the right breast. If lymph node metastases can be discrete and cover a wide area, it is clear that, unless we argue t h a t lymph node metastases in such cases are haematogenous, mere multiplicity of secondary deposits is not indicative of spread by way of the blood stream. 7. Tumour cells in the blood. The impression has gained ground t h a t those tumour cells which have found their way into the blood stream must be those which lead to fatal results in most patients . In support of the view t h a t bloodborne tumour cells are all-important, it is pointed out t h a t tumour cells are found within the capillaries of distant organs (M. B. SCI~MIDT, S c ~ A I ~ ) . ttowevcr, we have to take into account a more striking fact, namely, the presence of tumour

A Criticism of the Haematogenous Theory of Cancer ~'[etastasis

33

cells within the lymphatics. Long ago COATS (1883) remarked that cancers show " p e c u l i a r " relations to the vessels of the lymphatic system. GAZAYERLI saw microscopic invasion of the lymphatics in every one of his necropsy cases of lung cancer, an observation with which I am in complete agreement. B n U ~ E R discovered cancer cells in the thoracic duct of 11 out of 53 unselected autopsy materials. I think that, if histology is any guide at all to ultimate metastasis, it is the lymphatic vessel rather than the blood vessel t h a t should take pride of place.

8. Regional distribution o/metastases. I t is generally thought t h a t cancers metastasize most commonly to the lungs and liver via the veins (CoMAN). The interpretation long placed on this observation has been t h a t regional spread oceurs b y way of the systemic veins to the lungs and b y the portal veins to the liver. I t is noteworthy, however, t h a t the observed distribution turns out in m a n y eases to be paradoxical. We m a y also consider the problem from other angles. On embryological grounds, the lymphatics of a part are closely associated with the veins. I t is not unlikely, therefore, that the regional disposition of the veins m a y be paralleled b y t h a t of the eymphaties. I n other words, the lymph vessels of the portal tract m a y be related to those of the liver, while those of the other areas m a y be linked with the lung. The apparent regional distribution of metastases to the liver and lungs m a y be explained on the basis t h a t these two organs have the richest lymphatic supply in the body. Years ago the old anatomists found it easiest to inject the lymphatics of the liver and lungs ; it m a y well be t h a t these are the organs to which tumour cells trayelling along the lymphatics would most easily arrive at. I t m a y be countered

Ft~. 2. Carcinomaof the breast with metastases in the thoracic a n d a b d o m i n a l nodes. T h e t h o r a x a n d t h e a b d o m e n were dissected s e p a r a t e l y a n d

then superimposed when photographed. Note

t h e m u l t i p l i c i t y of t h e i n v a d e d nodes, t h e i r t e n d e n c y to be larger on the r i g h t side, a n d the

topographical distribution of the larger nodes, d i m i n u t i o n in size being centrifugal

t h a t the presence of tumour cells in the portal veins in cases of carcinoma of an organ situated in the portal area is a clear indication of haematogenous metastasis. My own observations on extraportal primary tumours--principally lung cancer--convince me, however, t h a t the portal veins are always invaded when there are secondary deposits in the liver. Are we to say, therefore, t h a t in such cases the secondaries must have Z. Krebsforseh. Bd. 65

3

34

WILSONI. B. ONUIOBO:

arrived b y way of the portal vein ? I believe that we are dealing, as far as the liver is concerned, not with cause and effect but with associated phenomena. In my view, neither portal vein invasion nor the seeming regional distribution of metastases to the liver and lungs should be regarded as evidence that spread must have been haematogenous. 9. Intracranial metastases. Owing to the widespread belief that there are no lymphatics in the brain, when carcinomatous deposits are found in this organ, it is "assumed", as RICH put it, that metastasis must have been blood-borne. Let us consider the assumptive nature of this common concept. There are, it may be noted, modern workers who have accepted, with varying emphasis, lymphogenous metastasis to the central nervous system. Some authors believe that perineural lymphatics transport cancer cells to this part of the body (Earns) ; others accept lymphogenous dissemination to the coverings of the brain (GRAI~ and KARR) or to the brain itself. Although GLoBus and M~LTZEa and others raise the semantic objection that in brain we are not really dealing w i t h " t r u e " lymphatics, I think that what is relevant to the theory of cancer metastasis is the demonstration of the existence of lymphatic connections between the central nervous system and the rest of the body. My own work suggests that the topographical distribution of metastases in the brain is in accord with the conclusion that such connections exist (ONumBo). 10. The /requency o/ brain metastases in lung cancer. According to BAKER (1942), one of the factors suggesting that metastatic tumours of the central nervous system are blood-borne is that lung cancers frequently spread to this site. However, MEYmr and I~EAI~ have shown that, if the effect of selection were eliminated, the frequency with which breast cancer spreads to the brain would approximate that of lung cancer. The thyroid gland also furnishes secondary deposits in the brain commonly. The factor common to lung,- breast- and thyroid cancers is nearness to the brain. If we accept that a categorical statement about absence of lymphatics in the central nervous system is in all probability misleading, then we may conclude that lymphogenous spread, which tends to be centrifugal, also explains the frequency of brain metastases from supradiaphragmarie primary carcinomas. Summary and Conclusion The arguments used to support the haematogenous theory of carcinomatous metastasis are as follows: (1) the presence of secondary deposits in the parenchyma of organs, (2) the distance at which the deposits may be found, (3) the invasion of distant parts in the absence of secondary deposits in the regional nodes of the primary turnout, (4) the discreteness of the deposits, (5) the bilaterality of most metastases of paired organs, (6) the multiplicity of metastases, (7) the presence of turnout cells in blood vessels, (8) the tendency to regional distribution of metastases to the liver in tumours of the portal area and to the lungs in extraportal tumours, (9) the invasion of the brain which is an organ reputed to have no lymphatics, and (10) the frequency with which lung cancer metastasizes to the brain. A critical examination reveals the assumptive nature of each of these concepts. I t is shown that a lymphogenous theory of visceral metastasis is not invalidated by these assumptions. Perhaps a shift in emphasis from the

A Criticism of the Haematogenous Theory of Cancer Metastasis

35

h a e m a t o g e n o u s t o the l y m p h o g e n o u s t h e o r y w o u l d resolve some presen~ difficulties a n d open new prospects. Zusammenfassung

Die A n n a h m e einer .corwiegend h/~matogenen M e t a s t a s i c r u n g m a l i g n e r Geschwiilste gr/indet sich auf folgcnde B e f u n d e : 1. Nachweis y o n M e t a s t a s e n in p a r e n e h y m a t 6 s e n Organen; 2. N a c h w e i s y o n F e r n m e t a s t a s e n ; 3. Nachweis y o n regionalen L y m p h k n o t e n f e r n m e t a s t a s e n bei F e h l e n y o n N a h m e t a s t a s e n ; 4. N a c h weis y o n G r u p p e n "con K l e i n m e t a s t a s e n im Versorgungsbezirk eines Gef/~Bes; 5. B i l a t e r a l i t / i t d e r M c t a s t a s e n in p a a r i g e n Organen; 6. Multiplizit/it der Metas t a s e n ; 7. N a e h w e i s v o n T u m o r z e l l e n im B l u r ; 8. U b e r w i e g e n y o n L e b e r m e t a s t a s e n bei Prim/~rgesehwiilsten im p o r t a l e n Einzugsgebiet, y o n L u n g e n m e t a s t a s e n bei P r i m a r g e s c h w i i l s t e n i m c a v a l e n E i n z u g s g e b i e t ; 9. Nachweis y o n G e h i r n m e t a stasen, einem Organ, y o n d e m a n g e n o m m e n wird, dal3 es keine Lymphgef/iBe bes i t z t u n d n i c h t d u r c h Lymphgef/~13e m i t den iibrigcn Organen v e r b u n d e n ist; 10. H/~ufigkeit von G e h i r n m e t a s t a s e n bei L n n g e n e a r e i n o m e n . Diese Befunde sind hie auf ihre B e w e i s k r a f t f/Jr den h g m a t o g e n e n U r s p r u n g y o n M e t a s t a s e n fiberpriift worden. Die A n n a h m e einer l y m p h o g e n e n Metas t a s i e r u n g in die v i s c e r a l e n Organe wird d u r c h diese Befunde n i e h t entkrgf~et, s o n d e r n im GegenteiI wiirde die A n n a h m e einer l y m p h o g e n e n M e t a s t a s i e r u n g fiir sog. ungewShnliche M e t a s t a s e f o r m e n die bessere Erkl/~rung abgeben. References

ANDRAL,G.: A treatise on pathological anatomy, vol. 2, pp. 443--444. Dublin: Hedges and Smith 1831. BA~ER, A. B. : Metastatic tumors of the central nervous system. Arch. Path. 84, 495 (1942). BARNARD,W. G.: The nature of the "oat-celled sarcoma" of the mediastinum. J. Path. Bart. 29, 241 (1926). BRUNNE~, U. : Die Entstehung lymphogener Metastasen im Ductus thoracicus. Virchows Arch. path. Anat. 833, 241 (1960). CAMERON,G. 1~.: The liver as a site and source of cancer. Brit. med. J. 1954 I, 347. CoaTs, J.: A manual of pathology, pp. 177 and 222. London: Longman 1883. - - A specimen of cancer of the lung with enormous deposits in the liver. Trans. Glasgow Path. Clin. See. 6, 18 (1895). CelIA)r, D. 1R.: Mechanisms responsible for the origin and distribution of blood-borne tumor metastases: a review. Cancer Res. 18, 397 (1953). EARLE,K. M.: Metastatic brain tumors. Dis. herr. Syst. 16, 86 (1955). GAZAYE~LI,M. E. : Ph.D. Thesis, University of Edinburgh 1935. GLOB~rs, J. B., and T. MELTZER: Metastatic tumors of the brain. Arch. Neurol. Psychiat. (Chic.) 48, 163 (1942). GRAIN, G. O., and J. P. KAI~R:Diffuse leptomeningeaI careinomatosis. Neurology (Minneap.) 5, 706 (1955). HADr'IELD, G. : The dormant cancer cell. Brit. med. J. 1954 II, 607. JORDAN, R.T., and S. Omwo: Comparative pathological changes in AKI~ mice following intranasal instillation of tumor cells and streptococci. Cancer l%es. 18, 47 (1958). ~EYE~, K. K. : Direct lymphatic connections from the lower lobes of the lung to the abdomen. J. thorac. Surg. 35, 726 (1958). MEYER, P.C., and T. G. REAH: Secondary neoplasms of the central nervous system and meninges. Brit. J. Cancer 7, 438 (1953). MOORE, G. E.: The spread of cancer. Gastroenterology 83, 313 (1957). ONUmB0, W. I. B. : The spread of lung cancer to the brain. Brit. g. Tuberc..52, 141 (1958). - - Patterns of metastasis in lung cancer: a review. Cancer Res. 21, 1077 (1961a). - - Ph.D. Thesis, University of London 1961b. 3*

36

WILSONI. B. ONUmBO: A Criticism of the ttaematogenous Theory of Cancer Metastasis

P~RgOT, J. W. : Advances in the anatomical investigation of the lymphatic system with special reference to some pathological applications. Edinb. med. J. 61, 50 (1954). l~rc~, G. J. : The distribution of metastatic tumors in the cerebrum. Arch. Neurol. Psyehiat. (Chic.) 23, 742 (1930). SC~Ara~R, t~.: IJlber den Untergang von Tumorzellen in der mensehliehen und tierisehen Leber. Verb. Dtseh. Ges., 31. Tagg 1938, S. 463--468. SC~MIDT, M. B. : Die Verbreitungswege der Carcinome und die Beziehung genera]isierter Sarkome zu den leuk~mischen Neubildungen. Jena: Gustav Fischer ]903. SC~VSTER, N. H.: Primary carcinoma of the lung. J. State Med. 87, 278 (1929). SPEncEr, H. : P h . D . Thesis, University of London 1954. STALmr, C. J., and E . F . SCA~LON: Bilateral radical neck dissection. Amer. J. Surg. 98, 851 (1959). STI~RAT, J. H. : M. D. Thesis, University of Glasgow 1945. SV~HERL~D, R.: Cancer. The significance of delay, p. 41. London: Butterworth 1960. Sw~DLow, M. : Malignant tumors of the head and neck. Amer. J. Surg. 98, 858 (1959). TOLDT, C. : An atlas of human anatomy, vol. 2, p. 717. New York: Macmillan 1928. WALS~E, W. H.: The nature and treatment of cancer, pp. 106 and 327. London: Taylor and Walton 1846. W~TKE~, H. El: Krebsmetastascn. Basel: Benno Schwabe & Co. 1948. WILLIs, R. A.: The spread of tumours in the human body, p. 25ff. London: Butterworth 1952. Dr. Wmso~ I. B. O~vm~o, University and Western Infirmary, Department of Pathology, Glasgow, Scotland