The problem I see with studies such as the BJC one is the fact that in
press releases, the true nature of the study is misinterpreted and what
comes to light is that treatments for early disease (as discovered by
screening with PSA) do not provide a survival benefit (and invariably
affect the QOL of the patient). The model does not directly account for
age and Gleason score, which are important covariates in modelling
survival from prostate cancer.

To resolve this the authors went to the med lit to obtain some survival
data in men treated conservatively. They used the Albertsen P et al
study which monitored the survival of 767 men from the Connecticut
Tumor Registry with localized prostate cancer diagnosed between 1971
and 1984, treated conservatively and followed up for a median of 24
years. (note that this data is pre-PSA data). Based on this and other
mathematical manipulations (involving lead-time and overdetection) they
did their calculations and came out with the study results.

Reading the conclusions of the study by Parker C et al does clarify the
issue. This is a modeling exercise. In other words, it is a
mathematical calculation that depending on the data input predicts
results. The authors, rightfully so, caution on interpreting results
and yet what is published in the press is that treatment is
ineffective. This is nothing new in the confused world of PCa. Just one
more step to make the issue more confusing.

Why then the authors ignored the history of the natural course of
untreated prostate cancer? It has been almost ten years since
significant data was published in Sweden. Age and tumor grade
(differentiation) had a significant impact in prostate cancer
mortality. Why is the reality of actual data not considered when
developing a mathematical model? Read on:

In a retrospective study 6890 patients with prostate cancer from
theNorth Swedish Cancer Register were analyzed according to cancer
specific survival. Prostate cancer mortality was 40% in patients with
well-differentiated cancers, 54% in patients with
moderate-differentiated prostate cancer and 72% in men with
low-differentiated prostate cancer. Prostate cancer mortality was 80%
in men younger than 60 years, 63% in men 60-69 years old, 53% in men
70-79 years old and 49% in men older than 80 years.

Source:
Damber JE, Gronberg H.[Mortality due to prostatic carcinoma in northern
Sweden]
Urologe A. 1996 Nov;35(6):443-5
PMID: 9064879 [PubMed - indexed for MEDLINE]

The reality is that at this point the decision to treat or not should
belong to the patient. Data such as reported by the Swedish study above
should enter in the decision making process of current patients.

Experts that tell us that there is overtreatment cannot distinguish (at
this point in time) who benefits and who doesn't with any degree of
significant clarity or even with how much overtreatment exists. Figures
are bounced off the wall from 80% down to low teens. There is a
recognized and significant amount of understaging in the interpretation
of biopsy results. Based on this and on the fact that since the
inception of PCa screening, disease-specific mortality is being
reduced, until further clarified, screening should be encouraged rather
than discouraged and educated patients should decide what to do for
their own benefit. Patient education to the risk of prostate cancer is
the only avenue presently available.

This is one more example on how to manipulate data to fit purpose. In
this case to proclaim the indolence of PCa. This while mortality in the
UK is increasing...

Source:
Vercelli M, Quaglia A, Marani E, Parodi S.
Prostate cancer incidence and mortality trends among elderly and adult
Europeans. Crit Rev Oncol Hematol. 2000 Aug;35(2):133-44.
PMID: 10936470 [PubMed - indexed for MEDLINE]

RalphV

Most men diagnosed at age 55-60 would be expected to live considerably
longer than 15 years if they were to live out their statistically
projected years.  So 99% odds of making it 15 years doesn't provide
much useful information for these men.  Two studies have been presented
at recent AUA meetings.  Tewari (last year) and Albertsen (this year)
have shown survival advantages for men treated by RP as opposed to
observation.  One of the keys is that this survival advantage takes 8
or so years to emerge.  Therefor, studies with median follow-ups less
than this time will show no survival advantage.  Albertsen shows that
low-risk men who elect observation have a risk of death from PCa that
is 3.8 greater than that for their RP-treated counterparts.  These
longer-term studies are just starting to appear, hopefully
peer-reviewed papers will follow...ron

American Urological Association Annual Meeting
May 20 - 25, 2006
Atlanta, Georgia, USA

Publishing #: 652
Presentation Title: Ten Year Outcomes Following Treatment for
Clinically Localized Prostate Cancer: A Population Based Study
Category: 43 Localized
Author Block: Peter C. Albertsen*, Farmington, CT; James A. Hanley,
Montreal, PQCanada; David F. Penson, Los Angeles, CA; Judith Fine,
Farmington, CT

Introduction and Objective: No data from randomized trials are
available to compare treatment outcomes among men diagnosed with
localized prostate cancer as a result of screening. A retrospective,
population-based outcomes analysis of men diagnosed in Connecticut with
localized prostate cancer between 1990 and 1992 was performed to
estimate prostate cancer specific survival and all cause survival
following surgery (n=806), radiation (n=703) or observation (n=114).

Methods: Analyses were conducted using an intention to treat
perspective. Two approaches were utilized: a proportional hazards model
and a classification system separating patients into low, intermediate
and high risk disease. The proportional hazards model included Gleason
score, pre-treatment PSA, clinical stage, age at diagnosis and
co-morbidities. The classification system utilized Gleason score,
pre-treatment PSA and clinical stage.

Results: After an average follow up of 11.8 years, 11% of the cohort
have died from prostate cancer, 4% from other cancers, and 23% from
non-cancer causes. Patients undergoing surgery tended to be younger and
have a more favorable distribution of histology and lower pre-treatment
PSA values when compared to patients undergoing radiation. Patients
electing observation tended to be older and had a more favorable
profile. After adjusting for differences in patient characteristics,
the men undergoing surgery had consistently better cause-specific
survival when compared to men undergoing radiation or observation.
Survival differences for men with low risk disease did not become
apparent until 8 years following diagnosis, for men with intermediate
disease, about 4 years following diagnosis and men with high risk
disease, almost immediately. The risk of death from prostate cancer for
men undergoing radiation versus surgery was 3.2, 2.5 and 2.2 times
greater for low, moderate and high risk disease respectively. The risk
of death from prostate cancer for men undergoing observation versus
surgery was 3.8, 2.3, and 3.3 times greater for men with low, moderate
and high risk disease respectively. There was no difference in
cause-specific survival between men receiving radiation and observation
although there may be a small trend in favor of radiation for men with
high risk disease.

Conclusions: Patients undergoing surgery for clinically localized
prostate cancer appear to have a survival advantage that increases in
magnitude over ten years when compared to men electing either radiation
or observation.