Can our track record in cancer research give us a hint of what we can expect in longevity research?


Can we look at the history of some leading diseases to get an indication of the potential breakthroughs in longevity research?

Research in the longevity field has dramatically grown in the last 3-4 years, and many are claiming we will realize great breakthroughs resulting in longer lives and more healthier years. And I for one hope that these breakthroughs do occur so there is reduced suffering, and enable people to live more healthier years, along with longer lives.

Aubrey de Grey, one of the leading drivers of longevity research, approach to longevity will be discussed in the concluding section of this piece in reference to the rest of this blog piece.

We could examine how are we doing in some of the leading causes of death over the last 50 years to maybe get some clues. The leading causes of death are led by the big 3, with the percentage of total deaths listed after the particular disease.

Heart diseases: 26.6%

Cancer: 22.8%

Cerebrovascular diseases (stroke): 5.9%

(all data is from the American Cancer Society and/or National Cancer Institute unless otherwise specified). Additionally, all data presented below in this piece is age adjusted rates, meaning we can compare across time despite different demographics from different eras.)

I am going to pick cancer to examine in more detail.  Cancer not only is the second leading cause of death, but also males have a 1 in 2 probability of developing cancer in their lifetime, while women are 1 in 3. So this disease affects most of us or someone we love. I have done a little bit of a background check to see how science has fared in treating this terrible disease over the last 50 years, or so.

(I am also reporting the cancer results of America, which according to recent data has the best cancer survival rates in the world, which means the conclusion might be worse for the rest of the world).

Now every week we are told of a new cure for cancer, a new breakthrough in cancer research that is going to save millions of lives. And we are also told that the survival rate of cancer X has improved ‘X’ percent in the last 10 years. Therefore, we must have huge decreases in cancer death rates over more than the last 50 years. In 1950 we hadn’t yet discovered DNA, not to mention sequenced the entire human genome, and started the whole biology revolution in research.

The cancer death rate (age adjusted) per 100,000 was:

In 1950: 193.9

In 2005: 183.8 (the most recent data set).

I am not even sure if this is a statistical difference? A 5.2% improvement in cancer death rate over the time of 55 years.

And possibly even more interesting if you look at the 2002 data and compare with 1950.

1950: 193.9

2002: 193.4

The cancer death rates are virtually the same for 1950 and 2002. How can there be no difference from 1950 to 2002? Is this not shocking to you – it is to me? What about all those new cures for cancer I have been reading about the last 30 years. And even the difference between 1950 and 2005 seem frighteningly small (5.2%).

So even if you skip all the basic advances we have made in molecular cancer research over the last 50 years, all the new treatments, etc, then just think of all our new advance screenings that we perform now, that we certainly did not do in 1950: pap test (papanicolaou test), PSA testing for prostate cancer, breast mammograms. Just the advanced screening should make a huge difference because all the cancer experts agree that the main key to treating cancer is catching it early. It could be argued that the small improvements we do see is all due to early diagnoses (screening and awareness), and nothing to do with actual better medical treatments, more research, new treatments, etc. Then why with the access of these new early screening tools, and far better public awareness why aren’t we seeing huge reductions in cancer death rates?

First we need to examine the leading types of cancers and which are the deadliest cancers. All the cancer groups when reporting this data put in the caveat of excluding non-melanoma skin cancers ( which is roughly 5 to 6 times higher incidence than breast or prostate, but very low death rate) so if we exclude then the leading new cases of cancer types are (again listing the big 3):

Males: ——————————————————–Females:

Prostate: 25%                                                                                    Breast: 26%

Lung and bronchus: 15%                                                                  Lung and bronchus: 14%

Colon and rectum (colorectal): 10%                                                  Colon and rectum: 10%

And the big 3 are also among the most deadliest, as I list the percentage of total cancer deaths by these three below.

Males: ——————————————————–Females:

Lung and bronchus: 31%                                                                   Lung and bronchus: 26%

Prostate: 10%                                                                                     Breast: 15%

Colon and rectum: 8%                                                                       Colon and rectum: 9%

A NY times graph probably gives you a more intuitive image of cancer types and deadliness (click on any of the graphs for a higher resolution image).

So to address cancer death rates one has to take into consideration the big 3, and in particular lung cancer because it has nearly 2 -3 times higher death rate than prostate or breast cancer.

One argument for the lack of large drops in the cancer death rate is a small but continuous increase in cancer incidence as depicted in the graph below (from 1975 onward).

The blue line is for male cancer incidence (per 100,000 age adjusted), the gray bottom line is female rate, and the red line represents both sexes. You see a general increase, but with a unusual blip in the male and combined sex lines between 1990 and 1993.

So let us look at individual trends in males and females as depicted below.


The red line represents prostate cancer incidence rate (per 100,000), while the blue and green are lung and colorectal respectively, which are decreasing.

You can see a large increase in prostate (red) cancer incidence, which then then falls back down before remaining relatively constant. The prostate cancer incidence mirror the overall cancer rates in males (blue line) in the previous graph. This suggest that new technology such as PSA screening and doctor and public awareness was catching more cancers that previously determined. A newscientist article reports that in 2002 1/3 of PSA detection of cancer in older males would not actually  cause symptoms in their lifetime, while the majority of these subjects would still  undergo cancer therapy. And there is an ongoing debate on the accuracy and usefulness of PSA testing.


The purple color represents breast cancer which is increasing slightly, along with the blue line – lung cancer, while the colon and rectum cancer (green line) incidences are trending downwards.

Therefore, it appears that the overall increase of cancer incidence over the years is largely attributed to the rise of prostate and breast cancer in the respective sexes. There is an open argument, particularly for prostate cancer, that some of this increase might be false positives. Either way the increase is not that large and mainly centered around prostate and breast cancer (there is also an increase in lung cancer in females – see below).

(I am leaving out discussion on possible environmental causes of the increased incidences of cancers for this blog piece).

But since lung cancer has a relatively high incidence and the most deadly of the most prominent cancers it plays a large role in the overall cancer statistics so we should examine this form of cancer in more detail.

For this I offer these graphs that go far back as 1930 and report cancer death rates for multiple cancers.



What these graphs displays in an increase in lung cancer for males until around 1993 then a decline, while female lung cancer rates rise about the same year then levels off. Therefore, behavioral trends in humans, such as smoking might be playing an important role and we should explore this possibility.

Tobacco use:

The black is the tobacco use, and the blue depicts cancer death rates from lung cancer for males, with the red representing the same for females. Tobacco use peaks about 1962 and the lung cancer death rate for males peak in 1990. So we can see a 25-30 year delay in tobacco use and death by lung cancer, especially for males.

Now we know that females in general took up smoking at a later time period than males and the death rate from lung cancer are offset in the sexes so we should look at the sexes separately. And if we look at data for the separate the sexes for smoking prevalence since 1965 as depicted below we can learn a few more things.

From this graph we can see that since 1965 there is a decrease in smoking prevalence in both sexes. Therefore, given the 25-30 year delay from smoking behavior changes and cancer death rates from lung cancer we would expect to see a drop in both sexes starting at say 1995 onwards. We do see this for males but have yet to see it in the female population.

But overall from 1992 we see a drop from 90 to 70 (rate per 100,000 age adjusted) in males and steady rate of around 40 for females in death rate from lung cancers. And since lung cancer has such a large influence on the overall cancer death rates (due to high incidence and high death rate) that this overall (combined sexes) drop in lung cancer death rates could largely account for the small overall drop we see in cancer death rates.

An equal argument can be made that the 1950 death rates are relatively low because the 25-30 year lag had not shown up in the rising rate of smoking that occurred from 1925 to 1950. And the 1950 cancer rates are largely determined from the level of smoking in 1920-25. And this is a valid argument, but it also points to the fact that behavioral changes have far larger influence on cancer death rates than our medical advances.

So one way to approach the potential confounding variable of tobacco use is to eliminate it from our analysis.

Now if you go back to the most telling graph of cancer death rates starting from 1930 and examine all the other cancers we see that for males a large decrease in stomach cancer death rate (most likely due to changes in diet in food preservation) and a smaller decrease for liver cancer, no real changes in prostate or colorectal cancer death rates (using the eyeball test) and an increase in death rate for leukemia and pancreas cancers. The news is better for females as we observe a decrease in stomach, uterus, colorectal and a smaller decrease in breast cancer death rates, and a increase in ovary and pancreas cancer.

But the bottom line is over 70 years it is hard to argue we are not much further ahead in overall cancer death rates that seem disproportional to the advances we have made in biological science and more specially cancer research.


Cancer is the second leading cause of death in humans, so it appears to me that if we are going to make any significant improvement in longevity and/or more healthy years we would have to make large strides in reducing cancer death rates (but see below).

However, despite all the vast sums of money and time spent on the fundamental basic biology of cancer, and all the clinical studies we have made very little, if any, real progress in cancer death rates over the last 50-70 years. Additionally, think of all the advances we have made in screening, awareness and early diagnoses of cancer over the last 50 years. They could easily account for the little improvements we might observe, but then again you would ask why not better? Then you have the weekly announcements of new cancer treatments that we have seen for the last 30 odd years – where are the results?
It is actually hard to believe the lack of results.

(I am sure cancer researchers and lobbyist could sell a far brighter picture that I have just presented – but I have just presented their own data at face value – though they might to do a better job at interpretation.)

Now, with this all said we have made some progress in heart disease from 1950 to the current time, though if you look at the limited amount of research from 1900 we might have not had any improvements from 1900 to  the year 2000 (suggesting lifestyle and environmental factors). But there is no argument that we have a longer lifespan now compared to 1950, but this is largely due to reduction in death during childbirth (for both child and mother) and reduction in death in childhood. Vaccination for common childhood diseases and antibiotics have probably had the biggest contribution to the increase in lifespan, and it is open to debate if we have made any gains on the other end of the lifespan spectrum.

Aubrey de Grey, also points out the lack of success of the single disease paradigm (discussed in these video links: via PIMM , TED talk, and this google talk). His argument is we have to get at the core of the problem, but isn’t this similar to the molecular cancer researchers approach? Specifically, Aubrey argues that the gerontology approach that tries to stop the damage, by usually altering the metabolism, or the geriatric approach which tries to reduce or halt the pathology that results from the initial damage might not be the best approach. He believes an engineering approach concentrating on the damage might be a better approach (see his video lectures).

I think Aubrey would argue that the cancer results over the last 50 – 70 years can not tell us much about the possibilities of his approach to longevity. But I still think that if we are going to have fundamental advances in longevity we are still going to have an appropriate level of molecular biological manipulation that is similar in principle as treating cancer (to the point that we reduce cancer death rates).

I wonder if we can not made huge strides in cancer death rates then what are the hopes for meaningful progress in longevity research?

Now the argument that usually occurs at this point is that only now do we know enough at the biological level to come up with real breakthroughs. But this is the exact same line we have been told for at least the 30 years of my memories. The new breakthrough, the new cure is heralded yesterday, today, tomorrow – but then 1, 5, or 10 years later nothing has changed (based on the cancer results).

I hope that this time we really are at the threshold of fundamental breakthroughs that result in reduced death rates from cancer, but also all other diseases. But if you want me to believe you that this time is different from all the other times – show me the results.

5 comments for “Can our track record in cancer research give us a hint of what we can expect in longevity research?

  1. August 14, 2008 at 3:09 am

    “it is open to debate if we have made any gains on the other end of the lifespan spectrum.”

    – Really? Can you present data on this please, for example the distribution of ages that people born in, say, 1880 died at?

    I suspect that you will find that improvements have been made at the “old” end of life.

    It seems to me (and again, we need data) that the average 50 year old of today has it a lot better than the average 50 year old of 1920 or 1950.

    On way of combating the negative outlook of this analysis is to think of the question of fighting aging (or not) as a lottery. If we choose to fight aging, we live for a very long time with probability p, and die as normal with probability (1-p). If we choose not to fight aging, we die for certain.

    An important question to ask ourselves is: what value of p are we willing to tolerate before we give up?

    Personally, I’d say that I’d give up bothering at about 1/10000.

  2. Ward
    August 14, 2008 at 8:31 am

    thanks for the comment and good point. I haven’t found the data regarding the other end of the spectrum of the lifespan yet. Combined with the apparent lack of results with cancer progress, and with very real gains in survival in childhood I ‘wondered’ (maybe I should have used that word instead of saying it is still an open debate if we have made gains at the far end of life spectrum. Maybe one of the readers will point to a blog piece or a data source so we can examine this question.
    Don’t get me wrong I am a strong proponent of longevity research and try to personally make the correct choices in my personal life (based on current research) to maximize my lifespan, but I just happened to look at the cancer data and was surprised by what appears to be not very encouraging results – so I thought I should point out the facts as best as I can.
    ps I like you blog and will have to read more of it.

  3. August 14, 2008 at 4:12 pm

    ah, thanks – I should have a few more “deep” blog posts out in the next week, hope you like them.

    re: “not very encouraging results”, i have had another thought… I may do a short blog response…

  4. February 6, 2009 at 12:06 pm

    would you also have a graph concerning the cancer incidence in aging population, as in age and the increased incidence of cancer?

  5. Ward
    February 6, 2009 at 8:16 pm


    I believe I do have some graphs on the rising incidence of cancer as we age – but it might take me a couple days to get to it because of schedule – but I will post it when I get a chance.

    thanks for your question.