Prostate cancer in younger men: an important new review

An article in this month’s issue of Nature Reviews: Urology has looked closely at a question of significant relevance to an important subset of our regular readers — Is prostate cancer in some way “different” in younger men?

This article by Salinas et al. has carefully examined the available data on the incidence and the known long-term outcomes of prostate cancer diagnosed very specifically in men of age 55 years and younger, and makes a series of important points:

  • As of 2012, approximately 10 percent of all cases of prostate cancer in the USA are being diagnosed in men of ≤ 55 years of age.
  • The median age of prostate cancer diagnosis decreased from 72 years in 1986 to 67 years in 2009 (primarily because of the increased role of PSA testing) — but “this shift does not account for the steep rise in the incidence of early onset prostate cancer.”
  • The data about long-term survival of men with early onset prostate cancer lack clarity — but there are at least strong suggestions that younger men have slightly worse long-term prognoses (for 10-year survival) than those diagnosed between the ages of 55 and 75.
  • In one study,
    • Men of 35 to 44 years of age diagnosed with AJCC Stage IV prostate cancer had a prostate cancer-specific mortality rate ~ 1.5 times higher than comparable men of 65 to 74 years of age.
    • Men of 35 to 44 years of age diagnosed with Gleason 8 – 10 disease had a prostate cancer-specific mortality rate ~ 1.4 times higher than comparable men of 65 to 74 years of age.

Salinas et al. go on to argue that:

The unexpectedly poor prognosis of advanced-stage early-onset prostate cancer supports the idea that a new clinical subtype might exist in the subset of men with early-onset prostate cancer.

This is not a completely new idea by any means. However, the paper by Salinas and her colleagues has helped to give focus to this possibility.

The paper addresses in some detail the possibility that there is a key genetic component to the risk for prostate cancer among younger males. As the authors note,

  • A man with a positive family history for prostate cancer is two or three times more likely to be diagnosed with prostate cancer than a man with no family history.
  • This risk is heightened among younger men with multiple affected relatives.
  • The effect of family history is not uniform across patient age groups.

Again, Salinas et al. go on to argue that

Despite the current limitations of genetic testing for prostate cancer, overall data suggest that genetic risk prediction is more useful in younger men than in older men.

Their basic conclusion is that early-onset prostate cancer provides a series of important research opportunities that deserve further scrutiny, for several reasons:

  • Careful study of early onset prostate cancer “could also lead to further understanding and identification” of key genetic and phenotypic data relevant to the initial development of prostate cancer in younger men.
  • Because the majority of younger men diagnosed with prostate cancer tend to have low-grade, organ-confined disease, these are potential long-term survivors who can help us to find ways to optimize quality of life over time.
  • Inheritance of prostate cancer susceptibility genes is more important in the development of prostate cancer in younger as compared to older patients, but may also be important in helping us to appreciate the precise roles of such genes for all patients.

The “New” Prostate Cancer InfoLink considers that this paper by Salinas et al. is an important informational resource for prostate cancer support group leaders and other prostate cancer educators, and we would encourage such individuals to request copies of the full text of this article from medical libraries or other resources.

Editorial note: The “New” Prostate Cancer InfoLink thanks Dr. Kathleen A. Cooney of the University of Michigan Medical School for kindly and promptly providing us with a full-text copy of this article. Dr. Cooney is the senior author of the original article on which our commentary is based.

11 Responses

  1. I’m having a hard time squaring the statement “but there are at least strong suggestions that younger men have slightly worse long-term prognoses (for 10-year survival) than those diagnosed between the ages of 55 and 75” with the findings of a recently published SEER database analysis that seems to show the opposite.

    Gandaglia et al. looked at 15-year prostate cancer-specific mortality (PCSM) rates and found they increased steadily with age at diagnosis: 2.3% in men ≤ 50 years; 3.4% in 51- to 60-year-olds; 4.6% in 61- to 70-year-olds; and 6.3% in men ≥ 71 years.

    They further note that “when considering patients affected by more aggressive disease, age was not significantly associated with higher risk of dying from PCa.” This seems to show that PCSM is not worse in younger men (or older men) diagnosed with high-risk disease.

    It may be that the phenomenon is unique to the 35- to 44-year-old subset diagnosed with virulent disease, a subset that is, thankfully, very small. I wonder, too, how many of those actually had a neuroendocrine prostate cancer.

  2. The sitemaster normally does a phenomenal job of condensing and summarizing papers whilst somehow reporting all important detail. And so I wonder if it was he or the paper’s author who might have nodded here.

    I see two obvious confounding factors not mentioned in the summary. Let’s imagine Mr. A and Mr. B, both newly diagnosed with prostate cancer. A is 40 years old and B is 70, putting both in the middle of the age ranges being compared.

    Factor #1: At age 40, A is far less likely than B to die in the next 10 years of stroke, Alzheimer’s, heart disease, lung cancer, etc. And as a 40-year-old has only a slightly greater incidence in other causes of death like accident or homicide. It therefore seems that A’s 10-year risk of prostate-cancer–specific mortality (PCSM10) would necessarily be greater than B’s.

    Factor #2. At age 40, A is vastly less likely than B to have had 10 years of PSA monitoring. Consequently, A’s disease is far more likely than B’s to have been diagnosed because of a primary symptom of prostate cancer rather than a lab test, or incidentally while investigating another prostate problem. PCSM10 is much higher for cases that have advanced to the point of being symptomatic at time of diagnosis. Restriction to Stage IV patients addresses this factor only to some extent.

    Perhaps both these confounding factors were addressed in the original paper in some way that was condensed down to the word “comparable” in the phrase “times higher than comparable men 65 to 74 years of age.” If so, then I would be curious to know how the 40-year-olds’ data was made comparable to the 70-year-olds’.

  3. Allen:

    I have brought the Gandaglia paper to Dr. Cooney’s attention (to see if she would like to make any comment on it) and I have also requested a copy of the full text of the Gandaglia paper.

    It is very possible that, even if Gandaglia et al. are correct, their data do not necessarily refute a primary premise of the paper by Salinas et al. (i.e., that by careful study of what triggers the onset of prostate cancer in younger men, we may be able to get a better handle on how to identify clinically significant forms of prostate cancer and distinguish them from indolent ones).

  4. Dear Paul:

    The “comparability” of the patients you are referring to would be discussed in the full text article of the original paper by Lin et al. to which Salinas and her colleagues are referring. I have not had the chance to read that article.

    All that I think that anyone is implying by “comparable” in the review by Salinas et al. is that there were no obvious outstanding reasons for why the patients would have been at any greater risk for prostate cancer-specific mortality.

    You are obviously correct in your identification of Factor 1 as affecting risk for prostate cancer-specific mortality. However, I am far less certain that Factor 2 is correct. In my experience the majority of men being diagnosed early are those who, for one reason or another (family history, race, etc.), start getting PSA tests early and are found to have an elevated PSA level. Diagnosis based on symptomatology in young men appears to be extremely rare.

    It is also worth noting that while you are theoretically correct that younger men at diagnosis should have a higher probability of 10- and 15-year prostate cancer-specific mortality than older ones, this appears to be refuted by the paper by Gandaglia et al. (mentioned by Allen).

    I suspect that we need to be cautious about all of these statistics, and I would note that Salinas et al. are careful to observe that the data about long-term survival of men with early onset prostate cancer lack clarity.

  5. I am curious if the paper analysed the disease demographics of younger men Dx with PCa; and, if so, whether there was a higher proportion of men with intermediate and high risk disease than in the overall newly diagnosed population.

    Anecdotally from my advocacy efforts, it seems like younger men have a more serious starting Dx.

  6. Rick:

    The paper didn’t get in to the details of demographics. However, those are well known and are available on other web sites such as the prostate cancer SEER fact sheet data.

  7. PaulC,

    Unless I misunderstand your point, I don’t agree that Factor 1 is a confounder. Prostate cancer-specific mortality is calculated using a Kaplan-Meier plot, which censors out data such as deaths from other causes. What is compared is then the percentage of each group who have survived all other potential causes of mortality, but didn’t survive their prostate cancer.

    What I think you’re addressing in Factor 2 is commonly called “lead-time bias,” and I agree that it does present a problem in interpreting PCSM. Even if all of the prostate cancer in all age cohorts were detected by routine screening, the 60-year-old man who has just been diagnosed may have had prostate cancer screening for up to, say, 10 previous years, while the 40-year-old man has most likely only been screened beginning recently. Therefore, routine population screening means that the 60-year-old was probably diagnosed earlier in his disease progression than the 40-year-old, and all things being equal, the 60-year-old would have longer prostate cancer-specific survival. So lead-time bias would lead to longer prostate cancer-specific survival times in older men, which is what Salinas et al. observe, but Gandaglia et al. observe the opposite pattern. Lead-time bias rather than more virulent disease at younger age may explain the Salinas finding, but the Gandaglia finding of older-age more virulent disease is only accentuated by lead-time bias.


    Hi Allen,

    My interpretation is that Paul C and Sitemaster have suggested that the prominence of other causes of death may explain the slightly higher rate of prostate cancer specific mortality in younger men, and you have countered with the notion that other causes are censored out of the data set in the analysis. The table in the portion of the paper available at no cost as well as the figures in the article look very much like SEER survival data for 5 and 10 years (with 15 also available) that do not censor death from other causes but rather incorporate those deaths in the “relative survival” figures presented in SEER data. (For example, see <a href= presentation from the American Cancer Society. Can you think of an explanation for this?

  9. Allen:

    I have communicated with Dr. Cooney about the Gandaglia paper. She points out that that paper:

    “focuses on men undergoing prostatectomy therefore excluding men presenting with advanced and/or metastatic disease. We discussed this in our review; studies of men in RP series often show better outcomes for younger men likely due to fewer competing risks and perhaps earlier diagnoses of low grade disease? We included all men with prostate cancer similar to publications by Merrill and Bird as well as Lin and colleagues.”

    She also notes that, “The Gandaglia abstract also makes the point that in men with aggressive disease, age is not significantly associated with increased CSM.”

    I think it is safe to say that the Gandaglia et al. paper does not necessarily undermine the points made in the paper by Cooney and her colleagues in their review.

  10. Jim,

    I don’t have access to the full text, but the chart you’re referring to only speaks to relative overall survival, not cause-specific survival. The paper that the Sitemaster provided by Lin et al. makes clear that Kaplan-Meier plots were used.

    Sitemaster, thanks for that clarification from the author. So I would conclude that there is tragically a subset of young men 35-44 diagnosed with untreatable stage 4 disease who have an even worse prognosis than older men who are diagnosed with Stage IV disease. According to the Lin et al. paper (Table 1), 0.5% of all men 35-74 diagnosed with prostate cancer are in the 35-44 subset, and only 10.9% of that subset were diagnosed with Stage IV disease.

    I’m not sure what to make of this other than that genetic studies may be able to identify a rare genetic phenotype that is particularly lethal in younger men.

  11. Dear Allen:

    It is well appreciated that the number of men under 45 who are diagnosed with prostate cancer is relatively small and that the percentage of those men diagnosed with metastatic disease is also small. … This is a highly select group of men who almost certainly have some form of high-risk status (because of their genetics or possible their environmental exposure risks or both).

    However, this still doesn’t contradict the original premise of Cooney and her colleagues that greater focus on the diagnosis and treatment options for younger men in general (with diagnosis at any stage) may be informationally beneficial with respect to the diagnosis and treatment of men at all ages.

Leave a Reply

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

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

Google photo

You are commenting using your Google 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 )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: