The reality of the over diagnosis rate of prostate cancer.

Understanding the published numbers.
The value of screening for prostate cancer with PSA and DRE has been questioned for years. Even when the mortality rate from prostate cancer in the U.S. has experienced a 40% reduction since the commercialization of the test, experts rely on computer models to generate extraordinarily variable over diagnosis figures. The real test of how good a computer model really is depends on how results compare to reality. PCa over diagnosis published results have such variances that one, by necessity (and mental sanity), is driven to find out why this is so.

In doing so, one soon discovers that the mere definition of over diagnosis has a tremendous impact on results. Lead-time data and over diagnosis data are intrinsically combined in these equations to provide results that range from 15% to 93%.
Now, something has to be very wrong with either of these extremes. Which of these results resemble reality?

Let’s start by saying that generally PCa is a slow growing malignancy. Low growing is not equal to indolent. The natural history of untreated PCa is defined by a high mortality rate. Age at diagnosis is a significant factor affecting disease-specific mortality. Disease progression in untreated PCa is not without symptoms. There is a significant reduction in quality of life in disease progression. This has been ignored in favor of over diagnosis calculations using computer models.

By mostly ignoring the mortality reduction since the commercialization of the PSA test; repeatedly claiming that treatments do more harm than good; continually stating that early detection and treatments have not demonstrated any survival benefit has generated mixed signals about the value of PSA in reducing the risk of an advanced disease diagnosis. These have had long-term detrimental consequences in PCa advocacy.
Then there is the issue of how PCa over diagnosis rates are calculated. It depends on
definition:

1. Definition of over diagnosis by MacGregor M et al:
The detection of a prostate cancer that, when left untreated, would not cause death.
Estimate: 84%

2. Definition of over diagnosis by Zappa M et al:
Over diagnosis (detection of latent prostate carcinomas) estimated by the proportional excess of cancers detected by screening with respect to that expected in its absence.
Estimate: 51% or 93% for age 60 or 65 at entry.

3. Definition of over diagnosis by Etzioni R et al:
Over diagnosis was defined as the detection of prostate cancer through PSA testing that otherwise would not have been diagnosed within the patient’s lifetime.
Estimate: 15% in whites and 37% in blacks

These three definitions resulted in an incredible variance. In real life as more screening is applied at younger ages, the proportion of early disease with an insignificant character would tend to increase. Therefore an increase in diagnoses of insignificant PCa is possible.

This would then be a real-life rate of insignificant disease detected by screening with PSA and DRE. What is that reported rate? Are these reported rates close to the computer-generated rates? Here are some studies:

1. Ohori M et al:
Of 306 clinically detected tumors 9% were unimportant and 29% were advanced. In contrast, of the 90 cases of incidental cystoprostatectomy disease was either unimportant (78%) or curable (22%) and no tumor was advanced (p < 0.0005). Cancer detectable by digital rectal examination, ultrasonography or PSA was distributed similarly among the 3 groups. Impalpable cancer detected by PSA was less likely to be advanced (11%) than cancer detected by digital rectal examination (34%, p = 0.01) but no more likely to be unimportant (13% versus 8%). Of 29 tumors detected only by systematic biopsies because of an elevated PSA level only 4% were advanced, while 17% were unimportant

2. Graif T et al:
:A total of 2,126 men with clinical stage T1c prostate cancer were treated with
radical prostatectomy during 1 of the 3 periods 1989 to 1995, 1995 to 2001 and
2001 to 2005. The respective proportions of men with a tumor that met our
criteria for over diagnosis (0.5 cm3 or less, confined to the prostate with clear
surgical margins and no Gleason pattern 4 or 5) and under diagnosis (nonorgan
confined, pathological stage T3 or greater, or positive surgical margins) were
examined. RESULTS: The proportion of men with an over diagnosed tumor was 1.3% to
7.1%. The proportion with prostate cancer that was under diagnosed was 25% to
30%. An ancillary finding was that decreasing the prostate specific antigen
threshold for biopsy from 4.0 to 2.5 ng/ml in the screened population resulted in
a lower rate of under diagnosis from 30% to 26%, a higher rate of over diagnosis
from 1.3% to 7.1% and an increase in the 5-year progression-free survival rate
from 85% to 92%. Men who were 55 years or younger were significantly more likely
to meet our criteria for over diagnosed cancer. CONCLUSIONS: Under diagnosis of
prostate cancer continues to occur more frequently than over diagnosis. Lowering
the prostate specific antigen threshold for recommending biopsy to 2.5 ng/ml
resulted in a lower rate of under diagnosis and a higher progression-free
survival rate.

3. Epstein JI et al:
Approximately 25% of radical prostatectomies performed for stage T1c disease show potentially insignificant prostate cancer. We previously reported the use of serum prostate specific antigen (PSA) density and needle biopsy findings to predict potentially insignificant cancer. We now evaluate whether using free/total serum PSA levels along with needle biopsy findings can better predict tumor significance. MATERIALS AND METHODS: We studied 163 radical prostatectomy specimens of stage T1c prostate cancer in which free/total serum PSA levels were determined. Free/total serum PSA levels were measured with Tandem-MP PSA assays. Insignificant prostate cancers were organ confined with tumor volumes less than 0.5 cc and Gleason score less than 7. Advanced tumors were either Gleason score 7 or greater, established extraprostatic extension with positive margins, or positive seminal vesicles or lymph nodes. Other cases were considered as moderate tumor. Moderate and advanced tumors were considered significant. RESULTS: Of the tumors 30.7% were insignificant, 49.7% moderate and 19.6% advanced. The best model to predict preoperatively insignificant tumor was a free/total PSA of 0.15 or greater and favorable needle biopsy findings (less than 3 cores involved, none of the cores with greater than 50% tumor involvement and Gleason score less than 7). Using this model of the 18 tumors predicted to be insignificant 17 were insignificant for a positive predictive value of 94.4%. Of the 145 cases that were predicted to be significant 112 were correctly predicted for a negative predictive value of 77.2%. There was only 1 tumor predicted to be insignificant which was classified as moderate. No tumor predicted to be insignificant was advanced. CONCLUSIONS: In conjunction with needle biopsy findings, free/total PSA levels accurately predict insignificant tumor in stage T1c disease.

4. Bartsch G et al:
In the Tyrol study, in which the mean age of screened men was <65 years, the estimate of over-diagnosis according to the criteria Epstein et al. was 8.7% .

It seems that from these real-life results, the degree of PCa over diagnosis is 1.3% to 25% as by the Epstein JI definition of disease insignificance. These results are based on cytoprostatectomy and radical prostatectomy actual cases. Given the variability of results by computer models as per the input data on age, incidence, definition, lead-time, population used and other variables it is more realistic to use actual surgical treatment (both TURP and RP) to identify insignificant prostate cancer to define over diagnosis.

One key factor in the whole discussion is to properly establish the real “indolent” nature of prostate cancer. Those saying that over diagnosis is extreme need to demonstrate how they know that those individual patients would have never been affected by clinical stages in their lifetimes once diagnosed. The calculation by computer models depend on so many variables that results can be questionable.

For example. Etzioni et al. reported:

The over diagnosis rates represent the proportion of screen-detected case
patients whose death by causes other than prostate cancer would have preceded
their date of clinical diagnosis and pertain to the modeled population, namely
white men aged 60–84 in 1988, with prostate cancer detected through PSA screening
between 1988 and 1998 inclusive.

Lead-time of 3 years: 17.38%

Lead-time of 5 years: 28.59%

Lead-time of 7 years: 39.24%

In other words, the estimated lead-time alone can have a 100%+ variance in such population when using a 3 or a 7 year lead-time in the calculation. Of all computer models, Etzioni seems to have the results that more closely resemble results of surgical methods of identifying insignificant disease. Still, we are dealing with a 15% to 25% of over diagnosis at the present time. In such cases a more judicious treatment/no treatment option could reduce immediate treatment for these men. To reduce or even recommend against testing with PSA seems a bit radical until a better marker is discovered. Men continue to be confused about the benefit of early detection given the opposing information provided to them.

I am grateful for the reduction in PCa mortality (40%) in the U.S, A., but
believe that this reduction could be greater if the current confusion
could be controlled. Rather than over diagnosis, I believe there still
is a high degree of advanced disease at diagnosis caused by the
controversy and that the current problem is under diagnosis and not over
diagnosis. Why has been the rate of diagnosis and treatment of
insignificant disease in the Tyrol experiment ignored? At 8.7% it
doesn’t fit the picture…

In the meantime while we argue, many men continue to die prematurely…and many unnecessarily

RalphV

Newporter,
Can you paste the link here?

Tony

How about we let the Moderators do their job, this place would be a mess without them.

I never claim to have all the answers, but there are a couple important relevant factors which have thus far been left out of this conversation...

First, as time has gone by in the PSA era, younger and younger men have been diagnosed with smaller and smaller cancers.  It used to be that men in the >75 age group dominated the SEER "rate of incidence by age", but by 1994 the 65-74 age group exceeded the older guys for the first time.  The <65 age shot up by 5x in that same timeframe.  Also, today, 82% of PC diagnosis is localized; in the early PSA era it was far below 50%.  So younger guys with less aggressive PC...of course survival is longer.

Another lesser known fact left out is that radical surgery in the >75 age group was common from 1987 (right after PSA introduction) to 1992 when the death rate peaked due to post operative heart attacks and pulmonary ebboli of men...many died on the operating table or shortly thereafter.  We've declined from that peak death rate, but some would say that rate was iatrogenic, and that we have merely returned to the pre-peak death rates.  Surgery on men in the >75 age groups is much less common today...the PSA testing has resulted in plenty of younger guys to fill the operating tables, and they can actually survive the procedure.

Over-diagnosis is not, in my view, nearly as big a problem as over-treatment.  We need to aggressively treat those who need aggressive treatment, and not aggressively treat those who have what may no longer even be called "cancer" if the name is changed (as it has been discussed in ASCO circles) to "IDLE" tumor. 

Casey,
No one has all the answers. To associate the reduction of PCa deaths to a reduction of surgical deaths in men older than 75 as the cause of an overall reduction in PCa deaths in the last 20 years is far fetched. The reported deaths after radical prostatectomy has never been high enough to mathematically have any impact. It would be proper when proposing such theory to provide sources. Talking of sources here is one from Canada that covers the issue:
“Overall, 53 men (0.5%) died, and 2,246 (20.4%) experienced one or more complications within 30 days of surgery. Age was associated with an increased risk of 30-day mortality, but the absolute risk of death was low even among men ages 70 to 79; 0.66% of men in this age group died within 30 days of surgery. The presence of comorbid conditions, particularly cardiovascular disease or stroke, was a stronger predictor than age of nearly all categories of complications.”
Source:
Alibhai SMH, Leach M, Tomlinson G, Krahn MD, Fleshner N, Holowaty E, et al. 30-Day Mortality and Major Complications After Radical Prostatectomy: Influence of Age and Comorbidity. J Natl Cancer Inst 2005;97:1525–32.

In your post, you say: “and not aggressively treat those who have what may no longer even be called "cancer" if the name is changed (as it has been discussed in ASCO circles) to "IDLE" tumor.”

Are you saying that they are treating patients without cancer? Be more explicit and clarify what you are saying. Provide sources so that we can understand.

RalphV

Casey,
I have not read Dr. Horan’s book. Is it possible for you that did to provide his arguments to support that hypothesis? Thanks.

RalphV

IIRC, I'd read that pathologists often assign a minimum Gleason of 6 to ANY Bx sample they receive - is that true?

What Purgatory said above about any cancer being cancer - we can attest to that. My wife had DCIS Dx - the cancer
treatment folks made light of it being "stage 0" breast cancer. Following surgery pathology determined a far greater
involvement that the docs believed - and nearly 'escaping' to become stage 2. Close call.

[DCIS = Ductal Cancer in Situ - USUALLY confined to certain areas of the breast]

Effectively, in 2005, the International Society of Urologic Pathologists recommended that ANY actual cancer found on biopsy should be graded as Gleason grade 3. This has since meant that newly diagnosed patients whose cancer has formally been confirmed by a biopsy now have a minimal Gleason score of 3 + 3 = 6.

This does not mean that GS less than 6 are not cancer. They are! Men continue to be diagnosed with GS 4 or 5 if the pathology is done on TURP chips. A little confusing, but true!

RalphV

Agree with Tim fully.

Are you saying that they are treating patients without cancer? Be more explicit and clarify what you are saying. Provide sources so that we can understand.

No, I'm not saying that (although this does, indeed, rarely happen)...but I referenced a previous discussion here at HW/PC that you probably didn't see or participate in, and are probably not aware of the national-level discussions on re-naming Gleason-6 PC with the term "IDLE" tumor.....so in hindsight, I can understand your confusion about my statement above. My apologies for being unclear.

If you read this 2-month old thread, you will understand. Thread title (& link): If the term "cancer" weren't part of the diagnosis, would more people consider AS?

I am not against screening...absolutely not at all. I do see how the benefit of screening has been overstated by some, either deliberately to support a narrow point, or out of ignorance. The real situation is complex...with both favorable and unfavorable twists and turns listed here by both you and me. Unfortunately, even the very highly regarded Dr Catalona, whom I see as one of the top PC gurus and top surgeon (and is known as the "father of PSA screening"), has been publically called to the carpet for overstating the impact of screening. Maybe he was acting like an overprotective parent. I think the attention is misdirected on "overdiagnosis." Overtreatment is a much bigger problem than overdiagnosis.

tarhoosier said...
Ralph:

Do you mean that ANY cancer identified on biopsy is rated as a G3? Or do you mean that any SECONDARY cancer cells will be rated G3?
The first would mean that a 2+2 is now a 3+3, meaning that EVERY cancer identified is within treatment. The second simply ignores the impact of the second, lesser G score and considers the tumor to be G6.

That is correct Tarhoosier. Those are the recommendations of ISUP since 2005.
According to Mike Scott this has been accepted globally by pathologists.

RalphV

ralfinaz said...

Given the evidence provided by the natural history of untreated before the PSA era, one should be careful to assert that these GS 6 will never progress.

I haven't seen anyone assert this...have you?  That would be goofy.  This is the very premise for active surveillance...to actively monitoring the course of the disease with the expectation to intervene if the cancer progresses.  It will in some, and it won't in others.

ralfinaz said...

The current dilemma with active surveillance here in the US is that there is not enough support from the urology community. They are not united to benefit patients that can benefit from AS. Add to this that most men are uninformed about their risk of PCa and/or confused by the current debate between those that support PSA testing with those that oppose it.

Agreed.  But what we are increasingly seeing is that the heavy hitters in PC are coming out strongly in support of AS for the appropriate patients.  The small-time local guys are late to come on board...but this is the usual trend.

Casey said: "I haven't seen anyone assert this...have you? That would be goofy. This is the very premise for active surveillance...to actively monitoring the course of the disease with the expectation to intervene if the cancer progresses. It will in some, and it won't in others."

Casey,
The real problem is that a young age diagnosis presents a difficult situation. That progression can occur is a given. We all surviving PCa live with that even when treated. As more men are diagnosed with low-grade disease, the potential for treating those that might not progress fast enough increases. Separating those is difficult because the present diagnostics (Gleason grade based) are not sufficiently accurate enough to define the genetic characteristics that drive the variable progression rate of PCa.

I am not saying that every low-grade PCa diagnosed in younger men will progress and kill them if not treated. Far from it. What I am saying is that this concept of IDLE cancer is not remotely accurate. Some 30%+ of men on AS progress in the first 2 to 3 years. Ask yourself if there is so much over diagnosis/over treatment; if 80% are now diagnosed with low-grade disease, why so many current treatments fail?

What follows is evidence that untreated PCa can kill. Age at diagnosis along with the a progressive increase to a more poorly differentiated disease require attention. Ignoring this can diminish your quality of life and survival:

1. Ninety-one percent of patients in our cohort entered with Gleason 3 + 3 disease, and nearly all grade progression (38% of cohort) was to Gleason 3 + 4 disease. Follow-up was 3.6 years
Age,
<65 171 (53%)
>65 150 (47%)

Source: Carroll P et al. Active Surveillance for the Management of Prostate
Cancer in a Contemporary Cohort. CANCER June 15, 2008 / Volume 112 / Number 12

2. Johansson et al (2004) reported their recent update on a population-based cohort study with a mean observation period of 21 years.[6 ]In this study, 223 patients with early-stage, initially untreated prostatic cancer were observed.

Symptomatic patients with tumor progression received hormonal treatment
(orchiectomy or estrogens). Thirty-nine (17%) developed metastatic disease, with most cancers having an indolent course during the first 10-15 years. However, further follow-up at 15-20 years revealed a substantial decrease in cumulative progression-free survival (from 45% to 36%), survival without metastases (from 76.9% to 51.2%), and prostate cancer–specific survival (from 78.7% to 54.4%). Prostate cancer mortality increased from 15 deaths per 1000 person-years during the first 15 years to 44 deaths per 1000 person-years beyond 15 years of follow-up.

Taken together, these data suggest that, although most prostate cancers diagnosed at an early stage have an indolent course, local tumor progression and aggressive metastatic disease may develop in the long term. In addition, these findings would support early radical treatment, notably among patients with an estimated life expectancy exceeding 15 years.

Source: Johansson JE, Andren O, Andersson SO, Dickman PW, Holmberg L,
Magnuson A, et al. Natural history of early, localized prostate cancer. JAMA. Jun 9 2004;291(22):2713-9.

3. The results indicate that cancer specific mortality in patients with localized prostate cancer increases steadily over time and is approximately 50% after 15 years. Localized prostate cancer is a slow-growing but progressive neoplastic disease. When diagnosed in a man with a longer life expectancy it should be handled as such. The study population was composed of 6514 patients diagnosed with prostate cancer during 1971to 1987 in northern Sweden.

Source: Hugosson J, Aus G. Natural course of localized prostate cancer. a personal view with a review of published papers. Anticancer Res. 1997 May-Jun;17(3A):1441-8.

4. RESULTS: about 85% of these patients died during the 7 to 23 years of follow-up, and the prostate cancer-specific mortality was estimated to be 55%. Age at diagnosis was found to be a strong predictor of prostate cancer death. Patients diagnosed before the age of 60 had an 80% risk of dying of prostate cancer, whereas those over 80 years of age at diagnosis had less than a 50% risk of prostate cancer-related death.

Source: Grönberg H, Damber L, Jonson H, Damber JE. Prostate cancer mortality in Northern Sweden, with special reference to tumor grade and patient age.Urology. 1997 Mar;49(3):374-8.

5. The long-term outcome in patients with prostate cancer treated with palliative intent was examined in two populations from Göteborg, Sweden. The results showed a prostate-cancer-related mortality of 62%. The cumulative mortality increased over time, indicating that prostate cancer may be a slow-growing tumour, but that patients were at considerable risk for disease progression and eventual death. Dying from prostate cancer was associated with a long hospital stay and frequent demands for palliative treatments such as TURP, radiation and procedures due to upper-urinary tract obstruction. In a subpopulation of patients who survived for more than 10 years, the cancer-related mortality was surprisinglyhigh,62% after noncurative treatment. Even if the patients were diagnosed before the PSA era, the above findings should be taken into account when advising patients with prostate cancer about therapy if they have a long life expectancy.

Source: Aus G, Hugosson J. [Non-curative treatment of prostate carcinoma. Outcome in Göteborg]. Urologe A. 1996 Nov;35(6):449-53.

6. Dr. B. Tribukait and coworkers in Sweden have supported the multi-step
progression of prostate cancer with their work on successive needle aspirations of prostate tumors. They evaluated the yearly rate of mutation accumulation as determined by DNA ploidy measurements. In each cell of their bodies, except their germ cells, humans have 23 pairs of chromosomes that dictate the person’s genetic makeup. Normal cells containing 23 pairs of chromosomes are said to be diploid. Dr. Tribukait was able to show that, as time goes by and mutations accumulate, there is
a loss or gain of chromosomes in prostate tumors. This process changes the cells from diploid to aneuploid. Aneuploid cells are cells containing an abnormal number of chromosomes. There is a close relationship between DNA ploidy and Gleason Grades. Higher Gleason Grades are mostly aneuploid and tend to metastasize or invade local tissues more readily.

Source: Adolfsson J, Tribukait B.Evaluation of tumor progression by repeated fine needle biopsies in prostate adenocarcinoma: modal deoxyribonucleic acid value and cytological differentiation. J Urol. 1990 Dec;144(6):1408-10.

Abstract:
Repeated fine needle aspiration biopsies of the prostate were taken during a period of 24 months or more from 84 patients with untreated prostate cancer. Serial followup regarding modal deoxyribonucleic acid values and cytological differentiation of the tumor cells was possible in 72 and 78 patients, respectively. During followup the modal deoxyribonucleic acid values in the tumor cells changed towards an increased aneuploidy in 17 patients and the cytological differentiation decreased in 18. These findings of a change in modal deoxyribonucleic acid values and cytological differentiation of prostate cancer cells during the course of untreated patients support the concept of a gradual dedifferentiation of prostate cancer.
PMID: 2231936

RalphV