Cancer and Wisdom of the Body - Pernicious Cachexia

Despite intensive effort to cure breast cancer, treatment generally fails, as evidenced by the age adjusted mortality from breast cancer. For 60 years, breast cancer mortality remained virtually constant. As treatment failed to improve the life prospect of the average patient, it is based on false premises, e.g., Halsted's hypothesis, according to which the tumor is the only threat to the patient. Yet there is more to cancer than just the tumor. Two hallmarks of cancer, cachexia, and paraneoplasia, are usually ignored, since it is assumed that they are caused by the tumor. But, what if it is the other way around, and cancer is first of all a cachexia accompanied by a tumor? At least this could explain why in most cancers treatment fails.

Cancer is a chronic systemic disease with local manifestations. Like arteriosclerosis, that is also systemic and manifested solely by its local manifestations, e.g., stroke and myocardial infarction. In the same way as treatment of an ailing heart does not cure the underlying arteriosclerosis, tumor removal does not cure cancer, since being "metabolically" systemic.

It is proposed here that carcinogens, deplete a vital substance and induce a metabolic deficiency that ends in cachexia. In order to survive, the organism grows a protective organ, the tumor, that replenishes the missing substance. During pre-clinical phase of cancer, deficiency is slight and compensated even by a minute tumor. With time it gets worse and the tumor has to grow more and more in order to make up for the loss, causing pain and secondary damage to vital functions. The patient seeks help and the disease starts its clinical course. When deficiency worsens, the patient becomes cachectic and dies.

Such a metabolic relationship exists in pernicious anemia, that illustrates how a tumor might be protective. Cancer is viewed here as pernicious cachexia induced by the loss of a vital metabolite and compensated by the tumor. Until the discovery of the missing substance, treatment ought to preserve the tumor and alleviate its secondary manifestations.

2 An amazing statistic

Among the myriad statistical curves that describe the fate of cancer patients in the world, one stands out vividly, portraying a harsh reality (Fig. 1). Despite continuing effort to eradicate breast cancer upon detection, from 1930 to 1990, age adjusted mortality from breast cancer has remained virtually constant (1,2). Sixty years of continuous debates about the appropriate treatment of breast cancer did not improve the life prospect of the average patient. No other epidemiological phenomenon manifests such a constant behavior. All other statistics usually vary.
The curve highlights the failure of breast cancer treatment and its underlying theories, particularly that of W. S. Halsted (1852-1922). Accordingly, the tumor starts when a normal cell is transformed into a malignant. It proliferates faster than its normal neighbors and forms a small nodule. Initially it is localized, and when attaining a certain size, it spreads through lymphatic vessels into adjacent tissues. As tumor continues growing, some of its cells spread through blood vessels to remote organs where they settle as metastases. This hypothesis implies that tumor removal should cure the patient, yet sixty years of intensive effort to remove the tumor, did not change the biological outcome of the disease (Fig. 1). Obviously the hypothesis is wrong and should be modified. The subsequent arguments are based on observations made on breast cancer, and are true also for any cancer.

Fig. 1 Age adjusted mortality rate. Breast cancer in U.S. females.

3 Breast cancer progression

Tumor progression is described within two frameworks: biological and clinical The first describes solely the behavior of the breast tumor. The second, deals with clinical manifestations of breast cancer: Initially the tumor is a non invasive, "in situ carcinoma". It then proceeds through three invasive stages: 1. Localized, when tumor is confined to the breast, 2. Regional, when tumor cells enter regional lymph nodes, and 3. Distant, when tumor cells spread into remote organs. The disease starts when a normal cell becomes malignant. Initially the tumor is small and cannot be detected by available diagnostic tools, and the disease proceeds through its pre-clinical phase. On tumor detection, the disease starts its clinical phase. The two frameworks do not overlap. A tumor may reach its invasive stage during pre-clinical stage, long before diagnosis. In the year 1990, only 11.7% of breast tumors detected in the U.S. were non-invasive, 42.5% were localized, and 45.8% spread outside the breast (Table 1, (3, p.I.21)).

Table 1. Breast tumor progression

_____________________________________
Stage classification was adapted from (3, p. I.21). Stages 3 and 4 are pooled.
3.3% were unclassified (3, p. IV.17).

4 Cancer is a systemic disease

There is more to cancer than just the tumor. Two hallmarks of cancer, cachexia, and paraneoplasia, are usually ignored, since it is assumed that they are caused by the tumor. In some cases it appears as if cachexia and paraneoplasia accompany the tumor, yet usually weight loss does not correlate with the type of cancer and its duration, nor with the site or number of metastases (4). Weight loss is one of the earliest manifestations of malignancy (5), and cachexia can appear in patients with tumors that are less than 0.01% of the total body weight (6). Also paraneoplasia is unrelated to tumor size, location, or the degree of metastasis, and may antedate the discovery of the tumor by weeks, months, or even years (7). In spite of this, oncology maintains that tumor is the primary factor in cancer, and systemic effects are secondary. But, what if it is the other way around, and cancer is first of all a cachexia accompanied by the tumor? At least this could explain why in most cancers treatment fails.

Take for instance arteriosclerosis that is manifested by local phenomena, e.g., stroke and myocardial infarction, and yet is essentially systemic. The same could apply to cancer, which like arteriosclerosis is "metabolically" systemic, and presents itself also by local phenomena, e.g., tumor. In the same way as treatment of an ailing heart does not cure the underlying arteriosclerosis, tumor removal does not cure cancer.

The following epidemiological analysis reveals that this might be the case, and more:

5 The declining hazard rate of breast cancer

An outstanding phenomenon in cancer epidemiology lies hidden and buried in piles of statistical data that are published yearly. Despite that cancer is generally incurable from the time of its first clinical manifestation, along its entire course, the chances of the cancer patient continually improve (8,9). The chances of the patient are concisely represented by the cancer hazard rate m[c] that is defined as the probability per unit time, that an individual who has survived to the beginning of a given time interval, will die within that interval. Thus from the first clinical inception of cancer its hazard rate continually declines.

This phenomenon is most pronounced in patients who survive longer than three years. It was documented in breast cancer patients, e.g., in patients treated at the M.D.Anderson Hospital (10), in 3878 patients from Edinburgh (11,12), in 1141 patients from the Charity Hospital of Louisiana (13) and in 57,068 patients from Sweden (14). Some reports provide "conditional annual survival rates" instead, i.e. the chance that an individual who has survived until time "i" will live for another year. These reports bear witness to the same favorable outcome. In Norway the conditional survival of 14,731 patients with breast cancer improved from year to year (15), and so did the conditional survival of patients with other cancers, e.g. cervix, corpus, lung and prostate (16).
In all other chronic diseases as well as in the healthy population, the hazard rate continually rises (9).

Epidemiologists disregard this phenomenon ascribing it to a bias resulting from patient grouping. As some cancer patients will eventually be cured, while others continue dying, the group as a whole gets healthier. The same putative grouping bias should operate also in chronic non-cancerous diseases and in healthy individuals, yet their hazard rate continually rises.

6.1 An intuitive explanation of the statistical method

Figure 2 depicts hazard rates of patients with breast cancer. When studying the four curves please note the magnitude of the hazard when cancer was diagnosed, and its direction in subsequent years, whether it was rising or declining. In situ cancer starts with the lowest hazard that rises until the 12th year. The initial hazard of localized cancer is somewhat higher. Initially it rises and then remains constant. The hazard of regional cancer rises up to the third year, whereupon it continually declines. When patients with distant cancer are first diagnosed, their hazard is about 0.39, and it continually declines.

When diagnosed, a patient with distant disease, has a chance of 39% to die within a year. If surviving, her chance to die within the next year drops to 30%. By the 10th year, her chances to die within the subsequent year drop to about 10%. From year to year her chances to survive improve. Apparently her organism learns how to cope with the disease better and better. Unfortunately, her reserves do not suffice to cure herself, and she dies. In localized, and regional cancers, hazard rate starts to decline from the third year and onward. The longer they live, the better their chances to survive. A declining hazard is manifested only in cancer. In other chronic diseases, as well as in the healthy population, the hazard continually rises. In healthy individuals the hazard is much lower than in cancer, yet it rises. This trend is apparent in in-situ cancer (Fig. 2). You may remember that patients with in-situ cancer do not die from cancer, otherwise they would not have been grouped in this category. Their hazard thus follows that of non-cancerous diseases. In the statistical analysis. in situ cancer hazard, m[a], is confronted with the cancer hazard, m[c].

The observed hazard of a patient, m[o] is made of two components, hazard due to cancer, m[c], and due to other diseases, m[a], and, m[o]=m[c]+m[a]. In distant cancer all patients die from cancer, and m[c] >> m[a]. In other words the declining m[o] is due to cancer. Since in-situ cancer patients do not die from cancer, m[c] = 0, and their rising hazard is contributed by non cancerous diseases, m[o]=m[a]. Mathematically speaking, the derivative of a rising hazard is positive, and that of a declining hazard, negative. Which can be expressed as follows: m[c]' < 0 m[a]' > 0. A rigorous proof of these arguments is given in the subsequent sections.

6.2 Material and Method

Computations were made from the SEER August 1993 public-use data files (17). The cases were diagnosed in 1973-91. The patients selected for this study were female Caucasian (report item number 9: Race Recode = 1). Cancers were staged into: in situ, localized, regional and distant (report item no 27: Historic Stage). Ages were grouped into five year interval groups (report item number 26: AgeRec). AgeRec= |Age/5|+1. File handling and statistical analysis was made with the JMP package version 3.1 (SAS Institute Inc. Cary, NC). The Kaplan-Meier (Product-Limit) analysis was used for computing cumulative survival and hazard rates. Patients with unknown survival time were excluded (report item number 28: 9999=unknown). The selected data set consisted of 198,623 patients (Table 2). Patients alive at cutoff date were censored, otherwise no censoring was done. A patient was included even if the cause of his death was unknown. The yearly hazard rate m = (Number that failed) / (Number at risk). The in-situ cancer hazard rate was subtracted from the hazard of clinical cancer. Log-hazard rates were then computed for different age groups (AgeRec), cancer stages (Stage) and years of survival (Years), ( 3 < Years < 16). These were then fitted by a least square model (Table 3).

Table 2 Number of patients and their mean age at diagnosis in each stage

Stage	Mean age	Standard deviation of age	N
In situ carcinoma	58.8	13.5	15,687
Localized	65.8	14.9	102,280
Regional	64.6	15.9	68,341
Distant	66.5	13.6	12,315
Total			198,623

Table 3 Parameter Estimates of the linear model:
LNhaz = intercept + b1*Stage + b2*AgeRec + b3*Years

Fig. 2 Hazard rates of breast cancer by stage.

6.3 Results

The data set consisted of 198,623 female Caucasian patients. With rising stage survival of breast cancer was poorer, and the hazard rate rose (Fig. 2). The hazard rate of in situ cancer started at 0.01. It climbed up till the 13th year whereupon it declined. The hazard rate of localized cancer started at 0.03, and till the third year it climbed up, whereupon it declined very moderately. In the figure the decline is not so obvious, but was statistically significant (Table 3). In regional cancers this bi-phasic trend was conspicuous. The initial hazard rate was 0.05. By year 3 it reached 0.09, and declined thereafter. The hazard rate of distant cancers was the highest, 0.39, and declined along the entire observation period.

6.3.1 Aging hazard m[a]

The subsequent analysis and discussion will be restricted to hazard rates observed between the years 4 and 15 (3 < Years < 16). Patients with in-situ cancer do not die from cancer, otherwise they would have been included in one of the other groups. Their hazard stands for death due to aging and non-cancerous diseases = m[a]. The other three stages represent clinical cancer. Their observed hazard m[o] consists of two components: Hazard due to cancer m[c] and hazard due to aging and non-cancerous diseases m[a]. m[o] = m[c]+ m[a]. In in-situ cancer m[o]=m[a]. Since all stages are affected by the same aging hazard m[a], the in-situ hazard may substitute for the aging hazards of clinical cancers, and m[o] = m[c] + m[in situ].

6.3.2 The cancer hazard declines with time, m[c]' < 0

In in-situ cancer the hazard rate continually rose. In the other three stages, the hazard rate in the interval 3 < Years < 16, declined. Since m[o] = m[c]+ m[in situ], and since m[in situ] rose, we may conclude that m[c] declined. In other words, in the time interval 3 < Years < 16 , the aging hazard rate derivative was positive m[a]' > 0 , and the cancer hazard rate derivative was negative m[c]' < 0). This statement was confirmed also by the least square model.
First, the in-situ hazard was subtracted from the observed clinical cancer hazards m[c] = m[o] - m[in situ]. The model was:

(1 =< Stage =< 3), (8 < AgeRec < 17), (3 < Years < 16) (Table 3). The slopes of AgeRec and Stage were positive, and the slope of Years, negative. This completes the proof of the opening hypothesis:

2. The derivative of the aging hazard is positive m[a]' > 0 , and the derivative of the cancer hazard is negative m[c]' < 0 . Although breast cancer is generally deadly, from the third year after its diagnosis and onward, it is manifested by a declining hazard.

7 The declining hazard is inversely correlated with the growing tumor

As cancer evolves the patient gradually deteriorates, he loses weight, develops paraneoplasia, and cachexia. These, obviously cannot account for the declining hazard and his improving chances. Attention should therefore be turned to the tumor. The declining hazard is inversely correlated with tumor growth. This is most pronounced in patients with distant disease. As metastases accumulate, the hazard declines. Once the tumor affects vital functions, hazard starts rising and the patient dies. Most tumors of patients with localized and regional cancer are clinically silent, manifested by a lump in the breast that does not cause pain. While in all other ailments pain and discomfort drive the patient to seek help, the cancer patient is not aware of her disease, and has to be screened. The hazard of these patients is depicted in Figure 2. As the disease progresses, the tumor grows and the hazard declines. The patient depends somehow on her tumor. More, the following analysis reveals that when the tumor is removed she gets worse.

8 Breast ablation is followed by a rising hazard rate

The hazard rate of patients with regional breast cancer is bi-phasic (Fig. 2). Up to the third year it rises, whereupon it declines. The hazard of localized breast cancer follows a similar pattern although less pronounced. All patients were treated when cancer was diagnosed, why should their condition deteriorate? Their fate was determined by their dependence on their tumors. Those who depended the most, died and the group hazard rose from 0.05 to 0.10. Patients who at that time of surgery had metastasis, were less dependent on their tumor, and their hazard rate declined. Patients with distant disease were entirely independent of their primary tumor, and their hazard declined despite treatment.

Tumor dependency explains why following surgery the condition of many patients aggravates, e.g., following colectomy for colon cancer, when the liver becomes seeded with metastases. In the past, this was attributed to tumor seeding induced by the surgical procedure. Yet the present analysis suggests a different explanation. As cancer progresses, organism and tumor maintain an equilibrium (homeostasis). When deprived from his primary tumor, the organism attempts to restore equilibrium, and induces micro-metastases to grow faster, until the missing tumor mass has been replenished.

Fig. 3 Cumulative survival of U.S. females with regional breast cancer

9 Long survival with micrometastasis

The cancer survival report No. 5 from the US National Institutes of Health describes the survival of 3369 patients with regional breast cancer diagnosed during 1950-1954 (18). Treatment consisted of mastectomy with or without irradiation. Their fate is depicted in Fig. 3. Initially the curve declines steeply, and gradually levels off. With time less and less women died, and indeed their hazard declined. Yet here we are interested in a different epidemiological feature. While 76% patients succumbed to the disease, 24% survived 20 years. Most of the time they were in remission and healthy. Since they eventually died from their disease, we can conclude retrospectively that all carried undetected micrometastases. Take for instance a woman that died after 20 years. Most of the time she was, and felt healthy, otherwise she would have been treated. During this prolonged remission her organism apparently "knew" how to live with, and adapt to micrometastasis. Let's examine the fate of 10% women that died in the period between 10-20 years. 34% survived 10 years, and 24% survived 20 years. 10% thus died from breast cancer between the year 10 and 20. Since living at least 10 years we may conclude that they carried hidden micro-metastases 10 years. For most of these years they were healthy otherwise they would have been treated. Micro-metastasis itself may not be harmful, and it might be even protective.

In the U.S., 33.6% of white females with breast cancer have regional disease (18 p. IV.17). The age adjusted incidence for all stages was 113.6/100,000. (19, p.117). 38/100,000 females carried micrometastases at least 10 years (= 0.336*113.6) which makes about 9,500 patients per year in the entire country. This estimate is extremely conservative. Actually, thousands of apparently healthy females carry micro-metastasis at least 10 years. Should they be poisoned with adjuvant chemotherapy? The nature of this adaptation to micro-metastasis is still unknown. Suppose medicine could harness it enabling the woman to live for another 20 years, cancer would turn into a benign disease. We may thus conclude that unless a metastasis impinges upon a vital organ, it is relatively harmless (20).

10 When to treat?

Even if one dislikes the idea that the patient might depend on her tumor, time has come to reconsider treatment objectives, and strategy. As long as a tumor or metastasis are clinically silent they should not be treated. From the time of its inception cancer is systemic, and a local cure is of no avail. Treatment is indicated only if a tumor causes discomfort, pain, or impinges on a vital function. In the past, mastectomy was a life saving procedure that protected the patient from infection and sepsis. Today we have time to wait and postpone the mutilating mastectomy, knowing that the patient's hazard continually declines. Treatment should be on an ad hoc basis. Some patients are anxious to get rid of the tumor which may be achieved in a gentler way than by mutilating mastectomy. Above all, they ought to be told the truth, that they carry a wide spread disease, but their chances continually improve.

Since apparently cancer is not yet curable, we ought to aim at prolonged remission. Exactly as done in all other chronic diseases. We never promise to cure a patient with an heart ailment, saying to him: "Your pump is somewhat damaged, yet with proper exercise you might participate even in the 'cardiac marathon'". Why not encourage the cancer patient saying to him: "You got cancer, it may be generalized, yet with proper . . . . you will be able to run for president". Here we are stuck, since all we know is to cut and poison. The declining hazard indicates that the patient is mobilizing her defense, and we ought to learn how to help her in this endeavor.

11 Tumor as protective organ

It is proposed here that carcinogens, deplete a vital substance, inducing a metabolic deficiency that ends in cachexia. In order to survive, the organism mobilizes a protective organ, the tumor, that replenishes the missing substance. During pre-clinical phase of cancer, deficiency is slight and compensated by a minute tumor. With time it gets worse and tumor has to grow more and more in order to make up for the loss, causing pain and secondary damage to vital functions. The patient seeks help and the disease starts its clinical course. When deficiency intensifies, the patient dies in a state of decompensation, known as crisis or relapse.

There is a disease called pernicious anemia that illustrates how a tumor might be protective. It is triggered by a "carcinogen" preventing the entry of vitamin B12 into the body. During its pre-clinical phase, that lasts about two years, the patient is healthy. The clinical phase starts with anemia and "paraneoplasia", known as combined degeneration of the spinal cord and brain. The bone marrow displays "neoplastic" features, e.g., hyperplasia, maturation arrest, and ineffective erythropoiesis, that were regarded in the past as "pseudo-leukemia" (21). These are protective means by the bone marrow that keep the patient alive. With time deficiency deepens more and more until reaching the state of decompensation whereupon the patient dies.

Clinical (phases)
1. Inception
	Pre-clinical phase
2. Detection
	Clinical phase

	Intercept	b1	b2	b3	r-square	N
Parameter estimates	-5.8	0.59	0.173	-0.024	0.72	279
Prob > \|t\|	0.00005	0.00005	0.00005	0.00060
Effect test
Prob > F	-	0.00005	0.00005	0.00060

G. Zajicek

1 Summary

Fig. 1 Age adjusted mortality rate. Breast cancer in U.S. females.

3 Breast cancer progression

Table 1. Breast tumor progression

Biological (stages)

Clinical (phases)

4 Cancer is a systemic disease

5 The declining hazard rate of breast cancer

6.1 An intuitive explanation of the statistical method

6.2 Material and Method

Table 2 Number of patients and their mean age at diagnosis in each stage

Fig. 2 Hazard rates of breast cancer by stage.

6.3 Results

6.3.1 Aging hazard m[a]

6.3.2 The cancer hazard declines with time, m[c]' < 0

7 The declining hazard is inversely correlated with the growing tumor

8 Breast ablation is followed by a rising hazard rate

Fig. 3 Cumulative survival of U.S. females with regional breast cancer

9 Long survival with micrometastasis

10 When to treat?

11 Tumor as protective organ

Cancer is viewed here as pernicious cachexia induced by the loss of a vital metabolite and compensated by the tumor. Until the discovery of the missing substance, treatment ought to preserve the tumor and alleviate its secondary manifestations (22-26).

12 References

Farewell my breast

Biological (stages)
Non-invasive	In situ carcinoma	11.7%
Invasive	1. Localized	42.5%
	2. Regional	33.6%
	3. Distant	09.0%

Physician		Philosopher
Click on your icon to follow your trail

G. Zajicek

1 Summary

Fig. 1 Age adjusted mortality rate. Breast cancer in U.S. females.

3 Breast cancer progression

Table 1. Breast tumor progression

Biological (stages)

Clinical (phases)

4 Cancer is a systemic disease

5 The declining hazard rate of breast cancer

6.1 An intuitive explanation of the statistical method

6.2 Material and Method

Table 2 Number of patients and their mean age at diagnosis in each stage

Fig. 2 Hazard rates of breast cancer by stage.

6.3 Results

6.3.1 Aging hazard m[a]

6.3.2 The cancer hazard declines with time, m[c]' < 0

7 The declining hazard is inversely correlated with the growing tumor

8 Breast ablation is followed by a rising hazard rate

Fig. 3 Cumulative survival of U.S. females with regional breast cancer

9 Long survival with micrometastasis

10 When to treat?

11 Tumor as protective organ

Cancer is viewed here as pernicious cachexia induced by the loss of a vital metabolite and compensated by the tumor. Until the discovery of the missing substance, treatment ought to preserve the tumor and alleviate its secondary manifestations (22-26). 12 References

Farewell my breast

Cancer is viewed here as pernicious cachexia induced by the loss of a vital metabolite and compensated by the tumor. Until the discovery of the missing substance, treatment ought to preserve the tumor and alleviate its secondary manifestations (22-26).

12 References