I should know better. I really should. I’m referring, of course, to my having forgotten my usual avoidance of purely political posts yesterday. I’m beginning to remember why I so seldom blog about political matters in general and why I’ve never in two years discussed abortion on this blog in particular. I don’t know what came over me. Given all that, I think it’s high time for a straight science post, don’t you? After all, I could beat up on Dr. Egnor again or do another dichloroacetate post, but what would be the point? Dr, Egnor’s clearly an ideologue who is not likely to stop pushing “intelligent design,” and there hasn’t been anything new on the DCA front that I’m aware of. Time for some science!

Here’s a myth that I’ve been meaning to take on for a while, a belief that is prevalent among but by no means limited to “alternative medicine” aficionados, namely that operating on a cancer will lead it to “spread,” either by “exposing it to oxygen” or other unnamed processes. Unfortunately, for me, Sid Schwab beat me to it,. Fortunately for you, he did his usual great job. I’ll add my proverbial two cents after this quote, although you should by all means read Sid’s entire article:

I’ve been there: operating with the intention of doing a curative (or at least durably palliative) cancer operation, only to find, entirely unexpected, that the cancer is widely spread throughout the abdomen. Sometimes the only option is to close up, accomplishing nothing. And surgery which doesn’t do good is, by definition, bad. The effects of anesthesia, the demands of healing are, by themselves, adverse. So it’s reasonable to think that the angle of decline could be made steeper by an operation that produces nothing positive, in the context of already far-advanced cancer. But it’s not that surgery spread the cancer; nor that exposing it to air had an adverse effect. It’s that it was too far gone at the time of the operation. Sadly, it happens. Even when palliation is attempted, it’s not always as long-lasting as we’d have hoped. The stories, I think, are born of this.

I have two comments, one a quibble and one another a more substantive observation.

First, the quibble. Sid says:

When asked the question — which I was, frequently — I pointed out that everyone who’s ever been cured of cancer — and there are millions of them — began with an operation.

I think he means “a solid cancer.” After all, leukemias are cured without surgery all the time; chemotherapy alone does a good job on many of them. But that’s just me being picky. Sorry. Sometimes I just can’t resist it.

My second and hopefully much more substantive comment is that it has been observed sometime that, after extirpation of the primary tumor, distant metastases appear in short order, giving the appearance that definitive surgery with curative intent somehow “spread” the tumor. Although this observation in humans is mostly anecdotal and not well studied, Dr. Judah Folkman, the “grandfather of antiangiogenic therapy” hypothesized that in some cases the primary tumor secretes something that keeps microscopic metastases at bay. Indeed, using this hypothesis, he isolated the antiangiogenic peptides angiostatin and endostatin, among others, from mice bearing tumors. They did it by making the following observation for a mouse tumor called Lewis lung carcinoma:


The diagram above simply represents the observation that, when the tumor is left alone, mice develop only microscopic lung metastases. If, however, the primary cancer is removed, then many large lung metastases grow. In reality, this only happened with a particular strain of Lewis lung carcinoma with a very low metastatic potential, designated LLC-LM; other strains of LLC grow fast and metastasize fast. In any case, the results of the experiment above imply that the primary tumor is secreting something that suppresses the growth of microscopic metastases. After this, the Folkman group did what we in the biz like to call “brute force” science, collecting mouse urine and analyzing it for tumor suppressive activity until they purified a single 38 kDa peptide, which they designated angiostatin. This involved analyzing literally gallons of mouse urine. (Ah, the glamor of science!) Once they had a bunch of angiostatin on hand, they peformed the following experiment outlined here:


It turns out that treating the mice whose tumors were removed with purified angiostatin prevented the blossoming of lung metastases, just as the presence of the primary tumor did. Less than three years later, the Folkman group isolated endostatin using a similar brute force strategy.In the primary tumor, factors promoting angiogenesis far outweigh the angiostatin and endostatin secreted. Elsewhere in the body, because these microscopic metastases are so small, they do not generate enough pro-angiogenic factors; the balance is tipped in favor of the angiostatin or endostatin produced by the tumor.

So what are angiostatin and endostatin? Basically, angiostatin and endostatin are byproducts of these tumors secreting factors that digest the surrounding matrix, allowing the tumor to invade. Angiostatin, for instance, is the byproduct of the proteolytic cleavage of the plasminogen protein; endostatin is the byproduct of the cleavage of type XVIII collagen. Both angiostatin and endostatin demonstrate potent antiangiogenic activity, preventing tumors from inducing the ingrowth of blood vessels that would otherwise supply it with oxygen and nutrients. Because the diffusion of oxygen and nutrients can only supply cellular needs out to a diameter of less than 1 mm, antiangiogenic therapy in essence results in preventing tumors from growing to a size of more than 1mm. Better still, when administered to mice with preexisting tumors, antiangiogenic agents can result in their shrinking down to a tiny ball of cells, a state known as tumor dormancy, and they demonstrate no observable toxicity.

Indeed Folkman did another cool experiment in which he administered endostatin at high doses, stopped the drug and allowed the tumors to start to grow back, and then started the endostatin again. Eventually, the tumors didn’t grow back anymore:


The tumor is still there; it’s just been rendered dormant. Indeed, the small lump of tumor cells can be induced to grow again by various manipulations.

Sadly, humans aren’t mice (at least with respect to tumor models), and antiangiogenic therapy, although it has shown promise in prolonging life when combined with chemotherapy for patients with liver metastases from colorectal cancer and some other tumors, is not the “miracle cure” that would turn cancer into a manageable chronic disease. And, although there is evidence that tumors can exist in a dormant state in humans for a long time, we haven’t found a way, angiogenesis inhibitors or no angiogenesis inhbitors, to induce tumor dormancy in an established human tumor. All of this is, of course, one reason why I’ve been so vociferous in cautioning the hypesters on the Internet who think that dichloroacetate will be a cure for cancer, just based on experiments in which it caused impressive tumor growth inhibition in rats. Anyone who’s been in cancer research for a while has seen this story before.

But let’s get back to the launching off point for this post. Does any of this mean that there’s any truth to the contention that operating on tumors somehow results in its spread? Not really. For one thing, there’s little evidence in humans that this happens. (Once again, humans are not mice.) For another thing, the presence of microscopic metastases is the very rationale for adjuvant chemotherapy given after surgery for many cancers; the chemotherapy “mops up” the microscopic tumor deposits that might otherwise grow into macroscopic metastases. For another thing, even if removing the primary tumor in humans actually did “release” micrometastases from growth inhibition in many human cancers, the option of leaving the primary tumor in place is not a good one. It will keep growing and ultimately cause complications, and eventually the metastases will escape growth inhibition anyway. Also, if metastases pop up after surgery, they were almost surely there before the surgery, too small to be detected on preoperative imaging studies, and not “spread” from the primary tumor by the surgery or by oxygen hitting the tumor. As is so often the case, correlation does not necessarily equal causation. Finally, it is possible to seed tumor cells by surgery to nearby areas. (It is not possible to seed tumors to distant organs that way, because this is purely a mechanical spreading of tumor cells.) Recurrences of colorectal cancer showing up at the port site after laparoscopic surgery were once a big concern early in the history of the procedure until data was obtained to show that the rate is actually quite low. In any case, attention to careful surgical technique goes a long way to prevent tumor seeding.

In the end, as Sid says, the only modality that has any chance of curing most solid tumors is complete surgical extirpation of the primary tumor with a healthy margin of normal tissue around it. (One exception is anal cancer, where radiation and chemotherapy alone, a.k.a. the Nigro protocol, can be curative.) After that, radiation and chemotherapy are just icing on the cake that decrease the rate of recurrence. Without surgery, there would be precious few survivors of any solid malignancy, and it would be foolhardy to deny oneself a chance at a surgical cure based on what is, in essence, theoretical possibility for which the evidence is quite sketchy. Thus, when patients ask me the question that Sid described, I answer that there is no good evidence that surgery causes tumors to spread and lots of good evidence that surgery is the best treatment.


  1. #1 Mark
    February 27, 2007

    I wonder if dogs would exhibit some of the same results as mice. I have had three dogs die from cancer. Two were from osteosarcomas that metastasized to the lungs. One was from some other type of tumor which initially appeared to be a benign lipoma. I now have a dog with rather large lipomas which are apparently benign but which I have an irrational fear of having removed. I guess some of that is natural, even if irrational. After all, if you become nauseated after eating a particular food, you will be disinclined to eat that food even if it had nothing to do with the nausea.

  2. #2 Ahistoricality
    February 27, 2007

    Nicely done.

  3. #3 Arakasi
    February 27, 2007

    After reading your posts on DCA over the last week or so, I would not be surprised if someone out there was selling mouse urine “for veterinary purposes only” online.

  4. #4 Sid Schwab
    February 27, 2007

    Elegant. Eleganter by far than my mostly off-the-cuff post. Shows the difference between an academic and an amateur. And yeah, I wish I’d said “solid.” I am aware of the “truthiness” of what I wrote, in the sense that it’s not as entirely simple as I implied. My hope was to dispel the superstition, and figured a little rounding off was ok. Meanwhile I really like your ability (not unlike PZ, I’d say) to compile scientific detail in a readable and relevant — not to mention enjoyable — manner. Even, dare I say, the occasional (not so occasional over at PZ’s place) political rant.

  5. #5 qalmlea
    February 27, 2007

    Mildly off-topic, but I just ran across this and had to share: “But what is the PRECISE AND SOLE CAUSE OF CANCER? CANCER OCCURS WHEN WE EAT OUR OWN FLESH. (Two very simple examples of eating our own flesh are biting our cuticles or biting our lips.) ”

    Found here: http://www.prattle.net/archives/002140.html

  6. #6 Dianne
    February 27, 2007

    When asked the question — which I was, frequently — I pointed out that everyone who’s ever been cured of cancer — and there are millions of them — began with an operation.

    Fussy addendum: A small percentage of small cell lung carcinomas can be cured with XRT/chemo as well, without surgical involvement. And beginning with the operation isn’t always the best way: neo-adjuvant chemotherapy is best in some cases. But you certainly need a surgeon to get a cure in the vast majority of cases.

  7. #7 Sid Schwab
    February 27, 2007

    I could have said “had” an operation, instead of “began with.” Still, the point is discernible. Can you begin treating small cell without a tissue diagnosis? Sputum, I suppose. I’d include a bronch with bx or a needle aspirate as “surgery” in the broadest sense, at least in the context of tumor manipulation. Bx for rectal ca too, for that matter. (The next sentence, as I recall, was that to dx it you need tissue, which requires some sort of procedure….) Guess I should stick to mowing my lawn.

  8. #8 Nat
    February 27, 2007

    Thank you for writing this. It’s an excellent example of the process of scientific thinking.

  9. #9 Torbjörn Larsson
    February 27, 2007

    not so occasional over at PZ’s place

    Perhaps he gets rant money?

  10. #10 Dianne
    February 27, 2007

    Perhaps he gets rant money?


  11. #11 Calli Arcale
    February 27, 2007

    Regarding exposure of tumor cells to oxygen, isn’t this a pretty illogical argument? After all, they’ve all been continually supplied with oxygen (via the bloodstream) or they’d be dead and the cancer wouldn’t be such a problem.

  12. #12 Andrew Dodds
    February 28, 2007


    I think the whole surgery=spread idea is based on a layman’s understanding of metastatis – it is usually explained as ‘cells breaking off the primary and lodging randomly in the body’.

    If this were true, then the idea of surgery spreading tumor cells around randomly and hence ‘seeding’ metastases dosen’t sound that far fetched. However, I believe that more recent research shows that the process is far more complex, with sites downstream of the cancer being ‘prepared’ hormonally before any cancer cells lodge.
    Which is why there are preferred sites of metastatis for different primaries; initially only a few organs will contain suitable receptors to organise a site.

  13. #13 Shalini
    February 28, 2007

    Great post. Thanks.

  14. #14 Faithful Reader
    February 28, 2007

    As someone whose father and mother are now five-year survivors or (respectively) colon cancer and breast cancer, I am always interested in your posts about cancer.

    But I just had to say that ‘Judah Folkman’ is an absolutely wonderful, resonant name!

  15. #15 Ben
    February 28, 2007

    Good call Dianne. I was going to suggest the same addendum myself!

  16. #16 epador
    February 28, 2007

    1) Sniff, sniff. I always thought a bone marrow biopsy and aspiration was an operation. So I used to think our leukemics had surgery before treatment too. [lol]

    2) Sorry I don’t have the references, but a decade or so ago I remember data showing tumor growth spurts in either breast or ovarian cancer that occurred post op and could only be abrogated by pre-operative chemotherapy. I always worked on the assumption that the growth factors secreted as a response to healing/injury stimulated the tumor cells. That was my explanation for sudden appearance of mets appearing post surgery.

    3) Tumor cells can indeed be spread hematogenously, not locally, at surgery, at least in the case of renal tumors involving the renal vein – you gotta close that off or big clumps can hit the lungs during the resection.

  17. #17 Darby
    March 1, 2007

    Has much work been done on the effects of preliminary tests like punch biopsies in aggravating and/or accelerating tumorgenesis? As mentioned above, it seems like some of the reactions that could occur in nonstandard tissues (a cystitic prostate, for example) might produce a cancer where one wouldn’t have occured otherwise. I know that some rare cancers have been linked to tissue traumas, so it doesn’t seem that far-fetched…

New comments have been temporarily disabled. Please check back soon.