[Help] [Aide] [Up]

Science Tribune - Article - August 1996


An analysis of the state of cancer research

Jean-Claude Salomon

CNRS, BP 8, 94801 Villejuif, France
E-mail : salomon@infobiogen.fr.


Research into cancer along the lines that have been pursued over the last thirty years has not given the expected results despite regular exultatory media announcements. This article questions whether the prevailing emphasis on technology and on the study of specific mechanisms of action is the only way of viewing cancer research. Cancer is not a disease that can be reduced to just the malignant tumour. It is associated with biological dysregulations and clinical manifestations ('paraneoplasic syndromes') that also require study if the true causes of death are to be established. The author suggests a more global, all-embracing, approach to cancer research.

Appraising 'progress'

If we compare present figures for cancer mortality rates with those of the early fifties, it is clear that there has been no substantial reduction in mortality except in the case of children and young adults (1) (2) (3) (4).

No-one seriously disputes these poor results but they both surprise and shock the lay-public, scientists and physicians. The public is accustomed to hearing about triumphant medical achievements. This sharply contrasts, however, with continuing demands made by public authorities and media-oriented charities who clamour for more funds : "Your money can help us. We can do wonders ! Give us more money and we will do even better". Scientists and physicians, themselves, have become caught up in the whirlwind of technological 'progress' and publication of positive results at all costs. Even they are surprised to realize that the efforts expended over the last 35 years have not achieved the advances in cancer research and treatment that they expected and longed for.

Do not get me wrong. I enthusiastically applaud the accomplishments that have been attained so far. Almost 50% of cancers in children and young adults are now cured whereas, before the 1960s, nearly all these patients would have died. But I do not forget the 50% who still die and for whom we have to do better. I therefore ask myself whether there is a relationship between progress made and the quality and effectiveness of the scientific research which preceded, accompanied and followed this progress. I want to question current reasoning, which may not be at the source of all progress as is so often thought, but may prove to be a collectively shared error in appreciation. I do not wish to cast doubt on the good faith of most triumphalists but to peer into the source of their convictions.

Where do the dominant opinions originate and what are the stakes involved ?

The importance of technology

In medicine, as in all other economic sectors, there exists a strong manufacturing body which exerts a powerful influence on medical research. It is difficult to resist the seduction of technological innovation. University academics, who were for so long resistant to the lures of industry, have at last capitulated and succumbed to "modernisn". The message goes something like this: "The answers to unsolved problems must lie and will lie in technology. In order to explore the human body, we manufacture ever-better chemical tools, devices, scanners, instruments, etc.. to access the deepest organs and cavities at a minimum cost, with least difficulty and maximum efficiency." We now have computer-assisted drug design and biotechnology-derived drugs.

Despite the fascinations of technology, it has yet to be proved that technology markedly improves man's health, that it adds to his life expectancy, brings benefits that are shared by mankind, helps preserve his habitat, and improves his nutrition and hygiene. Living standards need to be maintained or bettered during our technical explosion in medicine even if they are not directly related to technical developments.

The prevalent cancer research culture

It is by going to the very root of things, by ignoring all external pressures - the media and social pressures - that we can radically reappraise the function of cancer research. But first of all, we must establish a reliable system of evaluation. For a number of reasons - and until proof of the contrary - cancer mortality rate and population age categories are the most reliable indicators of progress. Neither the survival rate from the time of diagnosis (which is longer when small tumours are detected early) nor the rate of recovery (which is better when not only cancers in progression but latent cancers never provoking ill-health are considered) constitute good criteria.

Present-day biology is divided into two cultures. One culture is dominated by molecular and cellular theories. It is based on the acquisition of many simple, irrefutable, detailed scientific data that, for instance, relate to the specificity of the structure/function relationships that govern the phenomena under study. The other culture is sometimes described as phenomenological and reaffirms, often defensively, a truism, namely, that the whole is different from the sum of its parts and that accumulating facts is no substitute for understanding the complexity of physiological or pathological phenomena. Such a truism is not only irrefutable but also incantatory and cannot contribute to medical progress.

Trapped between these two visions lies the "hic et nunc" (here and now) which can only be termed a crisis situation that needs to be examined by epistemologists, historians, scientists and physicians. We possess an abundance of techniques and data but lack theories that can integrate the different levels which fracture the field into higher-order hierarchies. These levels extend from molecule to cell, to tissue, to organ, to the individual, and to the population as a whole.

Researchers everywhere are constantly put under pressure to produce results and publications in a world that is dominated foremost by an Anglo-Saxon culture that tends to privilege the technical/industrial complex. The public, goaded by the media, perverts medical research even further by its love of hero-worship and the "Star System".

A fresh introduction to an empirical view of cancer

The disease 'cancer', consisting of one or several tumours, can progress in only three ways :

1) Towards spontaneous healing. This is exceptional.
2) Towards cure by therapy which is more likely in developed countries where about 45% of diagnosed cancers are cured.
3) Towards death despite medical action.

We shall consider point number 3 only because it is the most frequent situation and the one that justifies the need for converging all scientific effort. It elicits two surprising observations :

First, although there are very many studies by biologists on the early stages of malignant transformation, there are virtually no studies on death from cancer. This is as if, biologically speaking, the terminal phase were of no interest and unable to reveal anything of importance. Yet, systematic autopsies performed a century ago -and also much more recently - have revealed that 80 to 85% of deaths are not linked to the growth of the primary tumour and/or to metastases. The scientific study of the conditions of cancer deaths is a huge "terra incognita" .

The second surprise is the almost total "oncocentrism" (vision centered on tumours derived from the Greek word 'once' for tumour) of contemporary 'cancerology'. The terms "malignant tumour" and "cancer" are not synonymous, nor are "oncogenesis" and "carcinogenesis" or "oncology" and "cancerology". The confusion is just as common among biologists as clinicians. Oncocentrism gives a view of cancers that is as debatable as the use of the word geocentricism when applied to the universe before the Copernican revolution. There is an empty space full of objects, apart from tumours and their causes, calling for scientific study.

- Paraneoplasic syndromes and biological dysregulation

The clinical turmoil of the disease is a greater reality than the tumour alone. "Paraneoplasic syndromes" (to underline their laterality with 'central' tumours) are clinical expressions of the symptoms and signs of the turmoil. For example, endocrine syndromes, that often appear in a variety of lung cancer cases and during disease progression, accompany tumour development but are not directly related to tumour appearance, increase in tumour volume, or to the invasion of healthy tissue by tumour tissues.

At the biological level, the disease can be an expression of genetic dysregulation. Cells - once derepressed through malignancy - become empowered with the capacity of producing hormones, growth factors, mediators of immune or nervous response, receptors, etc ... in places where they should normally not occur. At times, the molecules that are synthesized do not have the normal structure of the corresponding molecules and may thus display one (or several) partially agonist and/or antagonist effect(s).

- Cancer = tumour(s) + paraneoplastic syndrome

We shall focus on paraneoplasic syndromes as examples of manifestations ignored by current cancer research which focusses exclusively on one major postulate : that a thorough analysis of tumour cells and of their genomes will provide the keys to mastering cancer(s). We shall postulate that all malignant tumours are associated either with clinical paraneoplasic syndromes (physical or functional signs; morphological manifestations) or with an infraclinical biological syndrome that are the major determinants of the fatal outcome of most lethal cancers. The postulate is plausible and its testing may provide useful information on the death of patients with cancer and, by extension, on the biology of cancers in general.

Paraneoplasic syndromes can be better analyzed today because we have the knowledge and techniques for relating clinical observations to the production or consumption of measurable quantities of biological substances by tumour tissues. There is a mobile equilibrium between the tumour, the paraneoplastic syndrome, and the organism.

If clinical and biological paraneoplasic syndromes are correlated to tumour growth, we formulate two complementary hypotheses:

1) Extratumoral infraclinical manifestations exist as soon as the primary tumour appears and can even precede its appearance.

2) The association between tumours/paratumoral phenomena is bidirectional : If tumour growth affects the occurrence of paraneoplasic syndromes, paraneoplasic syndromes affect tumour growth. As shown by the relationships between immune function, tumorigenesis and tumour growth, this is probably not always true.

It has been known for some time that some tumours are associated with immunodepression which is partly relieved when the tumour is removed or reduced and that immunodepression favours tumour growth. It is also known that immune depression (genetic, acquired or secondary to treatment) is a determining factor in cancers and leukemia. This leads to two nonexclusive concepts :

- The paraneoplasic syndrome determined by tumour growth favours tumour growth.

- The paraneoplasic syndrome precedes or coincides with oncogenesis (tumour genesis) which it determines.

Are tumour/paraneoplasic syndrome inseparable ? Are they in equilibrium ? There is no answer to these questions but it is likely that, depending upon cancer phase and progression rate, they move to and from a state of equilibrium. Furthermore, therapy currently directed solely against tumours may increase the imbalance and accelerate disease progression.

Two tumours belonging to the same anatomical category may have very different paraneoplasic syndromes; this may help explain the unpredictable evolution of some cancers. For example, cancers of the pancreas have a brief survival rate, even after radical surgery, but a few cases progress only slowly. The evolutionary gaps between tumours at a similar stage and of similar form depend perhaps on different paraneoplasic syndromes.

Proposals for future research into neoplasic syndromes

In view of the probable clinical, evolutionary and biological importance of paraneoplasic syndromes, we propose that research into these syndromes should be an indispensable complement to research on tumour tissues and cells. This research could involve :

- a descriptive study not restricted to very advanced cancers;
- a methodical study of biological and clinical semeiology to infer syntax and grammatical rules with relevance to pathology;
- therapeutic studies with combinations of biological molecules made acceptable by administration of small doses (search for synergy among substances).


Scientists are unanimous in recognizing the great complexity of living organisms and of their diseases. They also agree that exploring a sick man piece by piece - crumb by crumb - is totally inadequate. A sick man needs to be considered as a global entity, as a complex system. Over the last 20 years, the systemic approach, Chaos theory and the theory of fractals have contributed substantially to the understanding of other complex natural systems. How much longer must one wait for their application to research in cancerology, not for curiosity's sake but as a must ?


1. Bailar JC III, Smith EM. Progress against cancer? N Engl J Med 314, 1226-12232, 1986.

2. Becker N, Smith EM, Wahrendorf J. Time trends in cancer mortality in the Federal Republic of Germany: progress against cancer?Int J Cancer 43, 245-249, 1989.

3. Extramural Committee. Measurement of progress against cancer. J National Cancer Institute 82, 825-835, 1990.

4. Davis DL, Hoel D, Fox J, Lopez A. International trends in cancer mortality in France, West Germany, Italy, England and Wales, and the USA. Lancet 336, 474-481, 1990.

Further reading

Salomon JC. Le tissu déchiré. Propos sur la diversité des cancers. Editions Seuil, Paris, 1991.

Salomon JC. The somatic code.The Cancer J 2, 169-172, 1988.