Unwarranted conclusions about treatment of oligometastatic prostate cancer

Many patients wonder, if they just have a couple of metastases, why can’t those be “zapped” by a few quick SBRT treatments and they can’t thereby be cured of their prostate cancer? Or, even if they can’t be cured, can’t the cancer’s progression be slowed down? To address those questions, we need to understand what is called the “natural history” of prostate cancer progression.

Even high-risk prostate cancer is quite a different sort of thing from metastatic prostate cancer. High-risk prostate cancer cells, for example those with a Gleason score of 5  +  5 = 10, are incapable of thriving outside the prostatic environment. At some point, they need to undergo a genetic transition called epithelial-to-mesenchymal transition (EMT), after which they can freely move throughout the body in the lymph, the blood, or the spaces around nerves, and plant themselves and accumulate in distant locations. Sometimes those microscopic metastases can circulate for a long time before planting themselves somewhere new. Sometimes they can plant themselves but do not proliferate appreciably for a long time. Sometimes they can alter the tissue environment in a new place (especially bone tissue) so it is more amenable to clumping and proliferation. Sometimes those cells get caught in lymph nodes (lymph nodes may be thought of as filters to catch cellular debris, including cancer cells) and proliferate there. All of these processes occur simultaneously.

Let’s try to gain an understanding of how many cancer cells are in systemic circulation at a given time. We have found that a count of five or more circulating tumor cells (CTCs) in each 7.5 ml of blood is associated with metastatic progression. (The prostate is also always shedding cells, healthy and cancerous, that are not capable of metastatic progression.) So, a 200 lb (91 kg) man with no detectable metastases and with a CTC count of 5, who has 6.5 liters of blood, will have at least 4,300 circulating tumor cells. In addition, there will be many thousands more lodged in and between tissues. Now, to be detectably metastatic with today’s best imaging technology, a clump of tumor cells must be at least 4 mm long. The cancer cell may be about 10 μm in  diameter, so there must be at least 200,000,000 of them (that’s 200 million of them) before the smallest metastasis becomes detectable. All of those cancer cells are constantly shedding and forming new daughter metastases elsewhere. Cancer cells may be circulating, clumping, and growing for a long time before they form a big enough clump to be detectable.

It should be clear that there is no possibility of a cure without systemic treatment. Currently, we have no systemic treatments that can cure metastatic prostate cancer on their own.

How long does it take to go from the first microscopic metastasis to the point where it is detectably metastatic? That’s impossible to know with any accuracy for a given individual. What we do know is that, on average, it takes 9 years from the time a man is biochemically recurrent after prostatectomy to the time when the first bone metastases are detected on a bone scan. That represents the accumulation of perhaps a billion cells in one place. It may be years more before the next bone metastasis is detected. Lymph node metastases are the most slowly progressing of all the kinds that prostate cancer causes. It is not unusual for many years to pass between new, detectable lymph node metastases. The new PET scans detect metastases much earlier, when the tumors are 80 percent smaller.

Now we can come back to the question of whether early detection and treatment of metastases can at least slow progression and increase survival. An [11C]choline PET/CT scan may be able to reliably detect metastases when the PSA is only about 2 ng/ml, rather than 20 ng/ml for a bone scan. The newer PSMA-based PET/CT scans may detect metastases even earlier, say at about 0.5 ng/ml. So, if any treatment is given when metastases are detected this early, and then we find that it takes a very long time — many years — to detect subsequent metastases, did the treatment really delay progression? This effect is called “lead-time bias.”

Adding to the confusion is the fact that those big clumps of detectable cancer cells are the source of much of the PSA. When those detected metastases are “zapped,” the cancer cells in them no longer secrete PSA and the cancer is controlled locally. We also know that old clumps of cancer are a rich source for new tumor cells. Is it possible that reducing at least that local source of metastatic cells will slow progression?

The only way to answer this question with any assurance is to conduct a randomized clinical trial. Some patients will get the treatment, in this case SBRT to the detected metastases, and the other patients will get standard of care — hormone therapy. Then we will be able to see how long it takes for new, distant metastases to be detected for the treated group as compared to the control group; and more importantly, did the treated group survive longer?

Triggiani et al. retrospectively report on patients at several centers in Italy (for some reason, most of these studies have been done in Italy) who had three or fewer detected metastases treated with SBRT:

  • About 100 patients with a recurrence after primary treatment with metastases detected by [11C]choline PET/CT scan (the oligo-recurrent group)
  • 41 castration-resistant patients with metastases detected by bone scan/CT (the oligo-CRPC group)

After a median of 20 to 23 months of follow-up, distant progression-free survival was:

  • 43 percent after 2 years for the oligo-recurrent group
  • 22 percent after 2 years for the oligo-CRPC group

The authors conclude that:

Stereotactic body radiotherapy seems to be a useful treatment both for oligo-recurrent and oligo-CRPC.

We are now ready to understand why this is an unwarranted conclusion. There is no way to know, based on the data they provided, whether the treatment was “useful” or not. We have no way of knowing what the distant progression-free survival would have been had they not received the SBRT treatment. Inexplicably, several groups from Italy also reached such unwarranted conclusions.

In fact, in a meta-analysis with longer-running follow-up data, Ost et al. (commented on here) found that for oligo-recurrent patients, distant progression-free survival was:

  • 31 percent after 3 years, and
  • Only 15 percent after 5 years

In other words, the vast majority (85 percent) of men with SBRT-treated oligometastatic recurrence had detectably relapsed within 5 years. Given the lead-time bias and the slow rate of detectable early progression anyway, it is impossible to say that the radiation treatment accomplished anything. Until we have some proof, patients should approach metastatic treatment for anything but palliative purposes with caution. There is currently no evidence, none, that treatment of metastases has any effect on survival.

In spite of the lack of evidence, if a radiation oncologist looking at the patient’s anatomy finds metastatic radiation to be safe, then there is little reason other than cost to abstain from it. However, a patient is taking a survival risk if he puts off hormone therapy in order to find metastases, especially in light of early evidence from the TOAD study.

Treatment of pelvic lymph nodes is a special case. If a patient is able to detect any metastatic pelvic lymph nodes, and he is convinced that he should have treatment at all, he should consider treatment of the entire pelvic lymph node field rather than isolated pelvic lymph nodes. One has to treat what one can’t see as well as what one can see; again, provided that it is safe to do so. Safety may be questionable because of anatomy, lack of visceral fat, history of bowel inflammation, and previous pelvic radiation. The evidence for efficacy is mixed. Some retrospective data analyses (Rusthoven et al.Abdollah et al.Jegadeesh et al.) found a survival benefit, while some did not (Kaplan et al. and Johnstone et al.). These retrospective studies are notoriously confounded by selection bias (i.e., the patients who got the therapy were the most likely to improve anyway). We await the outcomes of the randomized clinical trials before we have a more definitive answer.

Editorial note: This commentary was written by Allen Edel for The “New” Prostate Cancer InfoLink.

8 Responses

  1. Interesting article. You may like to take a look at the data we published from STAMPEDE looking at outcomes in node-postive, M0 patients where we permitted, but did not mandate, pelvic radiation therapy (RT). About 50% got pelvic RT and the hazard ratio for failure-free survival was around 0.5 in favour of RT. Not randomised of course but a multivariate analysis of a large pragmatic trial.

    Reference is James ND, Spears MR, Clarke NW, Dearnaley DP, Mason MD, Parker CC, et al. Failure-free survival and radiotherapy in patients with newly diagnosed nonmetastatic prostate cancer: data from patients in the control arm of the STAMPEDE trial. JAMA Oncol. 2016;2(3):348-357.

  2. Absolutely excellent analysis Allen! This seems to put a knife in the oligometatstatic theory. I guess the point of these new more accurate PET/CT scans is to find out if metastases are only in pelvic lymph nodes so they can be radiated. I guess you’re saying that radiation of any other areas of metastases is unfounded except to relieve pain?

  3. Readers should appreciate, in the context of patients being treated in the control arm of STAMPEDE trial mentioned above by Prof. James, that all those patients were receiving androgen deprivation therapy (ADT) with adjuvant radiation therapy as an electable option. No one was receiving radiation therapy without ADT.

  4. Bob:

    I am sure that Allen will chime in from the West Coast once he is awake, but in the interim I think that what he is saying is that radiation of visible, oligometastatic tumors is a reasonable option in carefully selected patients, but only when the patients are also getting systemic therapy with ADT. This would correlate with the data Prof. James refers to as well. However, we are still going to need more specific, prospective, randomized trials to actually be able to provide clear demonstration of whether the combination of RT + ADT compared to ADT alone can have real therapeutic impact in men with oligometastatic prostate cancer.

  5. I agree completely with the Sitemaster’s take on it. I’m really not trying to stick knives in anything. I am totally open to the hypothesis that significant cytoreduction with surgery or radiation plus systemic therapy to kill off many of the distant and circulating tumor cells may reduce the cancer burden to the extent that the immune system (stimulated by those therapies) might be able to clean up a lot more (the abscopal effect). Systemic therapies may include ADT, chemotherapy, and nuclear medicines. The abscopal effect may be enhanced by immune stimulants given at the same time. We have a lot to learn about all this, and large, long-running randomized clinical trials (RCTs) are the only way to do this.

    It’s also worth mentioning that the largest such RCT (see this link) is currently recruiting at Royal Marsden Hospital in London, which also participated in Dr. James’ landmark STAMPEDE trial. Another large RCT is recruiting at the Sir Mortimer B. Davis-Jewish General Hospital in Montreal. Hopefully, they will be large enough and run long enough to detect survival differences. The only RCT I’m aware of in the US is a small, brief one at Johns Hopkins (see this link).

  6. Length of tumor needed for imaging: 4 mm or 3 mm?

    Thank you Allen for an excellent informative and thought provoking article. I’ll be filing this to keep all the good information.

    I look forward to making a number of comments, but I’ll start with the size of the tumor for detection, hoping that you or others may chime in on this issue.

    The above article stated: “Now, to be detectably metastatic with today’s best imaging technology, a clump of tumor cells must be at least 4 mm long.” Or 3 mm?

    Of course the difference between 4 mm and 3 mm is relatively pretty big in percentage terms and potentially important in dealing with metastases, depending on the overall situation. I’ve heard radiologists talk of 3 mm (and even sometimes, iffy, 2 mm) as the smallest for detection. One was Dr. Stephen Bravo who was researching the now defunct (due safety) feraheme USPIO scan (which I had). Dr. Margolis is the other.

    At the 2016 Conference on Prostate Cancer during early September in Los Angeles, during Q&A in the Ask the Experts Roundtable Discussion panel led by Dr. Mark Scholz, MD, Dr. Scholz asked the very well published prostate cancer researcher and radiologist Dr. Dan Margolis, MD (formerly UCLA, now Weill Cornell), what the smallest tumor was that an mpMRI could identify. He replied that “our spatial resolution allows us to see tumors on the order of 3 mm.” He was not talking about mpMRI of the prostate, but rather about mpMRI protocols for imaging locally and regionally, basically, the bones and lymph nodes of the abdomen and pelvis which area serves as a sentinel for whether there are more distant metastases. (PCRI Conference Set, Disc 3 minutes 11-12.) While he said it was often hard to tell how suspicious a tumor is when it is that small, some of the mpMRI sub-scans and other nuclear scans like the C-11 PET/CT scans can add metabolic and other clues.

  7. Jelle Berentz claims to be able to find lymph node metastases as small as 2 mm with a Combidex-enhanced MRI. It is silly to discuss whether it is 2 mm, 3 mm, or 4 mm — that misses the point entirely. The point is that there are hundreds of millions of cancer cells in any of those tumors.

  8. Oligometastatic Prostate Cancer: Size matters?

    The article above clearly shows that even small tumors have awesomely high numbers of cancer cells. Who of us is not impressed by that? But the unanswered question is the role that the size of the metastatic tumor plays in seeding additional metastases and curtailing survival.

    Looking at this in reverse with an extreme example, theoretically it would take a minimum of just two metastatic cancer cells to spread further: one to hold the fort while the other goes out and colonizes at another site. If that held true for every metastasis, there would seem to be virtually no oligometastatic cancer as the cancer would spread widely from the start. Yet we see that oligometastatic cancer does exist, at least at the now observable level where the length is 3-4 mm or more.

    This fact suggests that more than size is at work in enabling the spread of metastases, and this is not a new concept of course. It’s been known for a long time that some cancer cells are unable to spread while others can (as Allen mentioned). As just one example, cadherin and catenin levels play a role in determining whether some cells can become metastatic, breaking away from the base group. As a fairly closely related example, my own case illustrates the situation of a lot of cancer but arguably no detectable spread, again an illustration of a lot of “size” but minimal or no distant metastatic spread. I was diagnosed with a PSA of 113.6, a Gleason of 4 + 3 = 7, all cores positive with most 100% cancer plus perineural invasion, stage III with likely seminal vesicle invasion, and a “rock hard prostate” -– an expression I heard from at least two urologists. The size was described as “mildly enlarged,” making it probably 35 to 45 cc. Based on the size and other evidence, there had to have been an extremely high number of cancer cells, with some of them outside the capsule. That would seem a perfect formula for metastasis, which is what respected doctors at three institutions, including Johns Hopkins and the City of Hope, were convinced was the case. However, both a Tc bone scan and a CT scan -– both admittedly fairly insensitive — were negative, and a ProstaScint showed just one “suspicious” lesion in an unlikely location for a sole detectable metastasis -– all this in early 2000. (My current PSA since radiation in 2013, supported by ADT, has never exceeded < 0.05, with fingers crossed for another good outcome from pending test.)

    Some leading doctors, perhaps many who are exploring the oligometastatic prostate cancer phenomenon, envision metastatic cancer in two basic varieties: the explosive kind that quickly produces many metastases, and a kind that tends not to spread metastases quickly; rather than considering an imaging finding of just a few cancers as a matter of catching the spread early, they see that five, or fewer, metastases (especially bone metastases, with three or fewer a more favorable indicator) is a likely indicator of a much milder form of metastasis, at least in terms of its potential for spreading.

    On the other hand, Dr. Eugene Kwon, MD, famous for his C-11 choline PET/CT scanning use at the Mayo Clinic in Rochester, Minnesota, recently stated that “the vast majority of [metastatic] prostate cancer that we identify [with the C-11 choline PET/CT scan] is in the form of oligometastatic disease, which means 5 or less metastases within the body.” (September PCRI/Us Too 2016 Conference on Prostate Cancer, Los Angeles, Disc. 2, 7:24–8:06) His evidence is from the 1,500 C-11 choline PET/CT scans for prostate cancer done each year at the Mayo Clinic in Rochester, apparently many being done as a result of low but significant PSAs after primary treatment. Later (Disc 2, 13:36) he mentioned the findings in Mayo’s paper by Sobol and colleagues on their map of post RP relapse sites:

    — 30.4% local recurrence only (in the “fossa”);
    — 53.6% metastatic disease only; and
    — 16.0% local plus metastases.

    Thus, about 70% of patients getting scanned for relapses had metastatic disease, and these would constitute the base for the “vast majority” referred to by Dr. Kwon who had oligometastatic disease.

    While hardly a randomized trial, this study is giving us a highly disciplined look at early metastatic disease. At first impression it seems that metastatic prostate cancer may take a while -– perhaps for dangerous mutations — before it becomes “explosive.”

    Dr. Kwon used controlling dandelions as an analogy to illustrate his view of the natural history of metastatic prostate cancer. My interpretation is that after weed killer (primary therapy) wipes out the main sources of dandelion seeds, the gardener uses imaging (our eyes) to spot the few remaining small plants that are still growing, and surgically removes them with a trowel or applies weed killer. That solves almost all the problem, but a few plants that were too small to see shortly reveal themselves and also get the gardener’s “therapy.” Perhaps another round or two –- after days, which correspond to many months or some years for many of us — is required before the grass is dandelion free, and follow-up surveillance is needed to catch plants that have seeded after a windy trip from nearby lawns. The end result is a lawn that is doing fine.

    Dr. Kwon reported on several patients, as examples of wiping out metastatic disease, who were enjoying extended times free of disease and therapy after having their metastatic disease wiped out (follow-ups of 2, 4, 5, and 5 years). He did not give us a statistical overview of their successes in wiping out apparent metastatic disease, but his examples are a proof of the principle that the “no evidence of disease” time after being freed from metastatic disease can be substantially extended, by years. Even when there is a subsequent recurrence, further rounds of imaging and targeted therapy, perhaps coupled with systemic therapy, can further extend survival. Moreover, as so many of us have experienced, added time means an ever-increasing likelihood of improved disease treatment and management technology.

    From a patient’s viewpoint, identifying and treating oligometastatic prostate cancer can have a highly meaningful impact and is definitely a worthwhile accomplishment!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: