Screening for Prostate Cancer: 2006 AUA Highlights
Mitchell H. Sokoloff, MD  

Introduction

It has been nearly a quarter of a century since the introduction of
prostate-specific antigen (PSA) serology as a screening and monitoring
biomarker for prostate cancer.[1] The PSA blood test has been a very
powerful tool for clinicians caring for men with prostate disease and
researchers developing improved treatments for prostate cancer. Still,
inherent weaknesses and inadequacies have made widespread acceptance of PSA
as a screening test very controversial.[2] Consequently, the validity of
PSA as a marker for prostate cancer screening and early detection has been
the focus of widespread investigations ever since its development. The
death knoll of PSA screening appeared imminent several years ago when Dr.
Thomas Stamey proclaimed that "PSA was dead as a screening test.[3]"
Despite this, PSA perseveres as a screening tool for prostate cancer. This
is due both to the dependence on it of an entire generation of urologists
as well as continued research focused on confirming its strengths and
creating new ways to interpret its values. As in past years, the utility of
PSA screening was an important topic at this year's American Urological
Association meeting.

Screening for Prostate Cancer With PSA: A Debate

Putting the controversy of PSA screening into perspective was a
"Point-Counterpoint" debate entitled "PSA Is a Valid Marker for Prostate
Cancer," which took place during the Tuesday morning Plenary Session.
Taking the "pro" side of the question was Dr. William Catalona from
Northwestern University, Chicago, Illinois, a pioneer of prostate cancer
diagnosis and treatment. The "con" side was argued by Dr. James Talcott, a
medical oncologist and outcomes researcher from Massachusetts General
Hospital, Boston, Massachusetts. Many of the arguments used in the debate
were supported by data presented at the meeting, and a selection of these
key abstracts is discussed after a summary of the debate.

Dr. Catalona framed his line of reasoning along the lines of published
research indicating that PSA, and, specifically, the median PSA that is
based on age, correlates with both the risk of developing cancer and the
risk of having aggressive cancer; moreover, PSA can predict pathologic
findings at the time of prostatectomy.[4,5] Using data from the Prostate
Cancer Prevention Trial (PCPT), he showed that as PSA increased in the
study population from 0 to 4.0 ng/mL, the risk for prostate cancer
increased from 7% to 27% and that the percentage of high-grade tumors in
those cancers increased from 13% to 25%, demonstrating that PSA is a
significant predictor of prostate cancer detection and high-grade disease
[6,7] Dr. Catalona also reiterated his work on PSA kinetics, in which he
showed that the overall rate of change in PSA is associated with a
diagnosis of prostate cancer and that the rate of change 1-2 years prior to
surgery correlates with aggressiveness of disease.[8] Dr. Catalona refuted
Dr. Stamey's claim that PSA no longer correlates with tumor size but,
rather, reflects the size of the prostate. Instead, Dr. Catalona proposed
that PSA is a marker of curable disease.

In his reply, Dr. Talcott did not address the latter points. Instead, he
focused on the question of whether PSA screening prolongs life. Dr.
Catalona used 2 arguments to indirectly support the benefits of screening
on survival: (1) stage migration reflects success in treatment, and (2)
falling prostate cancer mortality rates around the globe reflect the
success of screening programs. Dr. Talcott explained that both of these
arguments can be explained by lead-time bias (if diagnosed earlier,
screening makes survival and disease-free progression appear longer) and
length bias (cancers that are detected are slow-growing by nature, and more
likely to be clinically insignificant), which can make screening programs
appear to be successful. He emphasized that there has yet to be a
randomized trial in any population proving that PSA screening programs
prolong life. Using a number-needed-to-treat analysis, Dr. Talcott
explained that for clinically localized disease, 15 patients need to
undergo treatment in order to save one life. Taking the costs of treatment
(specifically the side effects of treatment) into consideration, Dr.
Talcott argued that the harm to those 15 men is considerable, and we must
be able to justify such harm (without definitive evidence of increased
survival) even if a single life may be prolonged.

In a rebuttal argument, Dr. Catalona made several arguments to show that
tumors diagnosed through screening programs are clinically significant,
using falling mortality rates as evidence. He concluded by proposing a
screening algorithm in which baseline PSA would be used for baseline risk
assessment and free/total PSA, PSA density, and PSA kinetics would be used
in follow-up for men with low baseline levels. He encouraged an intelligent
and rational application of PSA screening, which appeared to be accepted by
the audience.

The "Point-Counterpoint" debate effectively recapitulated contemporary
issues and controversies in PSA screening. Several of the arguments can be
further elucidated by reviewing key selected abstracts and research results
presented at the meeting.

PSA and Prostate Cancer Detection

Baseline PSA and the Risk for Prostate Cancer

The performance of PSA as a screening test for prostate cancer has been
called into question since its inception, especially for men with values <
10 ng/mL. Researchers at Stanford University, Stanford, California,
evaluated 999 first-time biopsy patients between the ages of 50 and 79
years with 12-core biopsies.[9] When men with normal digital rectal exams
(DREs) were studied, the positive predictive values (PPVs) for significant
cancer (> 3 mm) of a single baseline PSA test between 4.0 and 5.9 ng/mL
were 33%, 40%, and 48% for men aged 50-59, 60-69, and 70-79, respectively.
For the same groups of patients, the PPVs for high-grade cancer (Gleason
score > 6) were 14%, 22%, and 26%, respectively. The PPV increased
incrementally as baseline PSA values increased to values between 6.0 and
7.9 ng/mL and between 8.0 and 9.9 ng/mL, showing that PSA continues to be a
good marker for detecting prostate cancer. Investigators from Belfast,
Northern Ireland, had similar findings when they evaluated 68,364 men in
the Northern Ireland Cancer Registry with first-time screening PSA values
between 0.0 and 10 ng/mL.[10] Men were stratified according to baseline PSA
and followed for 5-10 years. Across all age groups, the risk of developing
cancer increased incrementally with baseline PSA; this was most significant
for a baseline PSA > 3.0 ng/mL. Lastly, in a study from Hamburg, Germany,
in which nearly 1000 men were treated with radical prostatectomy for PSA <
4.0 ng/mL and a normal DRE, 23% had cancer.[11] Of those with cancer, over
15% had Gleason scores > 7.

PSA Kinetics and the Risk for Prostate Cancer

PSA kinetics can be used to identify patients for biopsy. One multinational
study evaluated serial PSA measurements prospectively over 6 years in 353
men who eventually developed prostate cancer.[12] In men with cancer, PSA
velocity continuously increased in the 6 years prior to diagnosis, the most
significant increase being in the 2 years prior to diagnosis. Of note, this
study, as well as data from Loeb and associates,[13] suggest that the
cutoff for PSA velocity should be 0.5 ng/mL/year instead of 0.75
ng/mL/year. In another study with data from nearly 40,000 men screened for
prostate cancer in the Prostate, Lung, Colorectal, and Ovarian (PLCO)
Screening Trial, a PSA increase of 2.0 ng/mL in the year prior to diagnosis
helped identify men who had significant disease and would benefit from
diagnostic biopsy.[14]

Commentary: From the research presented at this year's AUA meeting, there is
little doubt that PSA continues to function adequately as a screening test
for detecting prostate cancer and for identifying patients for biopsy.
Clearly, it is paramount that physicians who perform PSA screening fully
inform their patients of the controversies surrounding prostate cancer
diagnosis and treatment, and that they maintain sound clinical acumen and
judgment when deciding who to screen and biopsy. The contemporary debate
should be drawn around how to proceed once the diagnosis of prostate cancer
is made, specifically, which patients need and would benefit from treatment
and which treatment is best for each individual patient. PSA is useful for
this last application because screening PSA values can be used to determine
risk stratification and select the most appropriate course of treatment.
PSA and Prostate Cancer Risk Stratification

Once a diagnosis is made, PSA can be used in several ways to help patients
and their physicians understand the pathophysiology and natural history of
their disease and to select the best and most suitable therapy available.
PSA Kinetics and Gleason Score

Changes in PSA preceding biopsy are associated with the histopathologic
findings on biopsy. In the PLCO Screening Trial database of nearly 40,000
men screened for prostate cancer, a PSA increase of 2.0 ng/mL in the year
prior to diagnosis was associated with high-grade (Gleason score > 7)
disease on biopsy.[14] D'Amico and associates[15] had similar results,
demonstrating that a patient with a PSA velocity of > 1.8 ng/mL in the year
prior to diagnosis is significantly more likely to have a diagnosis of
Gleason score 8 and higher.

PSA and Clinically Significant Disease

Since the inception of PSA screening, there has been concern that many, if
not most, of the detected cancers are clinically insignificant. Researchers
from Germany evaluated a contemporary series of 349 consecutive
prostatectomies.[16] Only 9% of specimens qualified as being insignificant
(defined as < 0.5 cc3 in volume and containing no Gleason grade 4 or 5
disease). The study authors raised concerns that models designed to predict
insignificant disease may miss a substantial proportion of clinically
relevant tumors. When reviewing his database of 3331 prostatectomies, Dr.
Catalona looked for evidence of "unimportant disease" (< 0.5 cm3,
organ-confined, and no Gleason grade 4 or 5 disease), "insignificant
disease" (< 0.2 cm3, organ-confined, and no Gleason grade 4 or 5 disease),
and "underdiagnosed disease" (non-organ-confined: stage pT3 and above,
positive surgical margins, and/or lymph node involvement).[17] In the group
of men who were treated between 2001 and 2005, 32% had unimportant, 5% had
insignificant, and 29% had underdiagnosed disease. This is an important
rebuttal to the argument that length bias is the reason that screening
appears to be beneficial. By these calculations today, a man with prostate
cancer is more likely to be underdiagnosed (> 25%) than overdiagnosed (<
10%) through conventional detection by PSA and DRE.

Countering this argument, however, are 2 European studies, both using data
from the European Randomized Study on Screening for Prostate Cancer: The
first showed that 121 (49%) of 247 men with cT1 and cT2a disease were found
to have the equivalent of insignificant disease (< 0.5cm3 and Gleason score
< 6) at time of prostatectomy.[18] The second study showed that of 1014 men
diagnosed after screening, only 19 died of prostate cancer.[19] It is
important to note that this cohort is composed of mostly men with
well-differentiated disease; the median follow-up was only 5 years; and the
applicability to a typical American cohort is skewed.

PSA and Tumor Volume

Two years ago, Dr. Stamey wrote that "PSA was dead as a screening test" on
the basis of his contention that PSA no longer reflected the volume of
cancer, but rather the volume of the prostate, including predominantly
benign tissue.[3] Three papers presented at this year's AUA meeting refute
this claim. In 717 radical prostatectomy specimens from Walter Reed Army
Medical Center, Washington, DC, PSA correlated with tumor volume for glands
< 50 g in size.[20] Dr. Stamey's claim that PSA had lost its correlation to
tumor volume over the course of time was examined in a study of a group of
400 men treated with prostatectomy at M.D. Anderson Cancer Center, Houston,
Texas -- 200 of whom had been treated between 1991 and 1994, and 200
between 2000 and 2003. When results were compared, PSA was found to have
maintained its significant association with tumor volume.[21] Lastly,
researchers from The Cleveland Clinic, Cleveland, Ohio, evaluated 2067
radical prostatectomy specimens and showed that PSA remains predictive not
only of tumor volume, but also of the percentage of high-grade disease
(Gleason grade 4 and 5) found at time of prostatectomy.[22]

Commentary: Once a diagnosis of prostate cancer is made, preoperative PSA
values can help predict findings at the time of surgery. Treating
physicians mostly continue to consider the PSA values -- in addition to
other clinical parameters, such as clinical stage and Gleason score -- when
counseling a patient as to optimal management. This underscores how
important it is to focus our attention not on whether PSA screening is
effective, but as to how we can use PSA, in addition to the other clinical
parameters, to stratify patients once a diagnosis is made.
PSA and Disease-free Survival

Indirect evidence supporting a survival advantage from PSA screening was
presented in several studies.[23-25] In addition, there was an update from
Dr. Bartsch in regard to his seminal study of men in Tyrol, Austria,
originally published over 5 years ago.[26] With continued follow-up, there
has been considerable stage migration with an increased incidence of
organ-confined disease and, consequently, an increased number of men
undergoing successful surgical treatment for their disease. The
investigators believe that the increase in downstaging, localized disease,
and successful surgery is the probable cause of a nearly 50% drop in
mortality from prostate cancer.

Commentary: The lack of a definitive randomized trial means that we do not
know whether PSA screening can prolong survival. Still, indirect evidence
suggests that in populations in which prostate cancer screening is
available, earlier diagnosis of disease and stage migration have resulted
in reduced morbidity and mortality. Although still imperfect, PSA is a
powerful biomarker that can help identify men at risk for prostate cancer
and help guide treatment decision making.
Other Screening Modalities

PSA remains the best biomarker for prostate cancer. However, several
abstracts found that the PCA3 molecular assay may be beneficial for both
screening modalities.[27-30] Expression of the PCA3 gene is upregulated in
prostate cancer and can be detected in urinary sediment after DRE. A group
of investigators from California evaluated the PCA3 test in 147 men with
either treated cancer (post prostatectomy), benign prostatic hypertrophy,
untreated prostate cancer, and normal prostates; the last 3 were based on
transurethral ultrasonography (TRUS) biopsies.[30] There was near-complete
separation of the groups on the basis of only PCA3 results. With biopsy as
the reference method, the sensitivity was 62% and the specificity 82%.
These promising results indicate that PCA3 may be used in conjunction with
PSA serology for prostate cancer detection.[29,30] Investigators from The
Netherlands examined the results of PCA3 testing in 299 patients, 114 of
whom were diagnosed with prostate cancer and 48 who were treated with
surgery.[27] Although there was no association between PCA3 and pathologic
stage or Gleason grade, there was a correlation with prostate cancer
volume.

Commentary: With a better understanding of gene and protein expression in
prostate cancer as well as differences in gene and protein expression
between normal and benign tissue, new molecular assays will be developed to
aid in prostate cancer detection. It is too early to determine what role,
if any, PCA3 testing will have. Still, over the next few years expect
several new markers to be introduced and examined.
Conclusions

Without results from a definitive randomized trial, we cannot be sure
whether PSA screening truly prolongs the survival of men with prostate
cancer. Nonetheless, PSA remains a powerful biomarker and should be used in
the screening of men for prostate cancer when applied in a rational and
sensible manner. Moreover, PSA can be used to help guide men and their
physicians in deciding whether to proceed with therapy and to help choose
the optimal treatment course if treatment is indicated.

Although the concern of overtreatment is real, so must be the fear of
undertreatment and devastating outcomes if significant cancers are not
detected in young men. Perhaps it is the treatment that we should focus on
and not the need for biopsy. We are developing an improved understanding of
the pathophysiology of prostate cancer and are better at stratifying
patients for treatment. There is a growing acceptance of delayed treatment
in some men, and the role of active surveillance is increasing.[31-33] For
men who will benefit from treatment, improvements in therapy have
diminished the side effects and are making the diagnosis and treatment of
prostate cancer less devastating for the patient and his loved ones.

We should continue to use PSA for screening and early detection. Clinicians
who counsel and treat men with prostate cancer must be honorable about
applying the principles of screening and early detection in a rational
fashion and must provide their patients with informed consent. As we
continue to improve our diagnostic, prognostic, and therapeutic skills in
the future, men may want to undergo screening so that effective
stratification and treatment, if necessary, can be applied early and
effectively with minimal morbidity and even less mortality.

References

  1. Han M, Gann PH, Catalona WJ. Prostate-specific antigen and screening
for prostate cancer. Med Clin North Am. 2004;88:245-265. Abstract
  2. Sirovich BE, Schwartz LM, Woloshin S. Screening men for prostate and
colorectal cancer in the United States: does practice reflect the evidence?
JAMA. 2003;289:1414-1420.
  3. Stamey TA, Caldwell M, McNeal JE, Nolley R, Hemenez M, Downs J. The
prostate specific antigen era in the United States is over for prostate
cancer: what happened in the last 20 years? J Urol. 2004;172:1297-1301.
  4. Loeb S, Roehl KA, Antenor JA, Catalona WJ, Suarez BK, Nadler RB.
Baseline prostate-specific antigen compared with median prostate-specific
antigen for age group as predictor of prostate cancer risk in men younger
than 60 years old. Urology. 2006;67:316-320. Abstract
  5. Loeb S, Gonzalez CM, Roehl KA, et al. Pathological characteristics of
prostate cancer detected through prostate specific antigen based screening.
J Urol. 2006;175:902-906. Abstract
  6. Thompson IM, Ankerst DP, Chi C, et al. Assessing prostate cancer risk:
results from the Prostate Cancer Prevention Trial. J Natl Cancer Inst.
2006;98:529-534. Abstract
  7. Thompson IM, Pauler DK, Goodman PJ, et al. Prevalence of prostate
cancer among men with a prostate-specific antigen level < or = 4.0 ng per
milliliter. N Engl J Med. 2004;350:2239-2246. Abstract
  8. D'Amico AV, Chen MH, Roehl KA, Catalona WJ. Identifying patients at
risk for significant versus clinically insignificant postoperative
prostate-specific antigen failure. J Clin Oncol. 2005;23:4975-4979.
Abstract
  9. Presti J, McNeal J, Hsu T, Brooks J, Gill H. Performance of PSA levels
between 4 and 10 ng/mL in the era of extended biopsy schemes. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 468.
 10. Connolly D, Black A, Nambirajan T, Murray L, Gavin A, Keane P. Initial
PSA levels and the long-term risk of prostate cancer. Program and abstracts
of the American Urological Association 2006 Annual Meeting; May 20-25,
2006; Atlanta, Georgia. Abstract 469.
 11. Makarov D, Humphreys E, Mangold L, et al. Pathologic outcomes and
biochemical progression in men with T1C prostate cancer undergoing radical
prostatectomy with PSA 2.6-4.0 ng/mL versus PSA 4.1-6.0 ng/mL. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 474.
 12. Berger A, Deibl M, Klocker H. PSA-velocity increases continuously in
the years prior to diagnosis of prostate cancer. Program and abstracts of
the American Urological Association 2006 Annual Meeting; May 20-25, 2006;
Atlanta, Georgia. Abstract 485.
 13. Loeb S, Roehl K, Graif T, Viprakasit D, Catalona W, Nadler R. PSA
velocity threshold for predicting prostate cancer in young men with PSA < 4
ng/mL. Program and abstracts of the American Urological Association 2006
Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 471.
 14. Grubb R, Levin D, Pinsky P. Relationship of PSA velocity and Gleason
score in the PLCO Cancer Screening Trial. Program and abstracts of the
American Urological Association 2006 Annual Meeting; May 20-25, 2006;
Atlanta, Georgia. Abstract 522.
 15. D'Amico A, Cullen J, Chen Y, McLeod D. Prostate-specific antigen
velocity and the odds of identifying prostate cancer at biopsy. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 523.
 16. Alschibaja M, Massmann J, Funk A, van Randenborgh H, Hartung R, Paul
R. Clinically insignificant prostate cancer -- can we reliable predict low
volume and low grade prostate cancer preoperatively? Program and abstracts
of the American Urological Association 2006 Annual Meeting; May 20-25,
2006; Atlanta, Georgia. Abstract 475.
 17. Graif T, Yu X, Loeb S. Underdiagnosis and overdiagnosis of prostate
cancer. Program and abstracts of the American Urological Association 2006
Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 476.
 18. Roobol M, Steyerberg E, Kattan M, de Koning H, Schröder F. Prediction
of indolent prostate cancer: validation and updating of a prognostic
nomogram. Program and abstracts of the American Urological Association 2006
Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 1392.
 19. Roemeling S, Vis A, Roobol M, van der Kwast T, Schröder F. Death from
screen-detected prostate cancer: the demographics at diagnosis. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 1472.
 20. Furusato B, Rosner I, Osborn D, et al. The relationship of
preoperative PSA levels to prostatic weight and tumor size. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 863.
 21. Ochiai A, Vakar-Lopez F, Troncoso P, Babaian R. The relationship
between PSA and tumor volume persists in current era. Program and abstracts
of the American Urological Association 2006 Annual Meeting; May 20-25,
2006; Atlanta, Georgia. Abstract 867.
 22. Figler B, Reuther A, Dhar N, Magi-Galluzzi C, Zhou M, Klein E. PSA is
still predictive of tumor volume and grade late in the PSA era. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 1473.
 23. Desireddi V, Roehl K, Loeb S. Progression-free survival rates for
radical prostatectomy in the PSA era. Program and abstracts of the American
Urological Association 2006 Annual Meeting; May 20-25, 2006; Atlanta,
Georgia. Abstract 650.
 24. Stephenson A, Eastham J, Kattan M, Bianco F, Dotan Z, Scardino P.
Long-term risk of metastatic progression of prostate cancer following
radical prostatectomy. Program and abstracts of the American Urological
Association 2006 Annual Meeting; May 20-25, 2006; Atlanta, Georgia.
Abstract 651.
 25. Albertsen P, Hanley J, Penson D, Fine J. Ten-year outcomes following
treatment for clinically localized prostate cancer: a population based
study. Program and abstracts of the American Urological Association 2006
Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 652.
 26. Bartsch G, Horninger W, Oberaigner W, et al. Tyrolean screening study:
update 2005 -- stage migration and decrease of mortality. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 473.
 27. van Gils M, Hessels D, Hulsbergen-van de Kaa C, et al. Detailed
analysis of histopathological parameters and PCA3 test results. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 537.
 28. Fradet Y, Groskopf J, Walker S, et al. Protoype Aptima PCA3 molecular
urine test: development of a method to aid in the diagnosis of prostate
cancer. Program and abstracts of the American Urological Association 2006
Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 538.
 29. Hessels D, van Gils M, Witjes J, et al. The prognostic value of PCA3
gene-based analysis of urine sediments after extended digital rectal
examination. Program and abstracts of the American Urological Association
2006 Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 539.
 30. Marks L, Aubin S, Deras I, et al. Specificity of the APTIMA PCA3 Assay
for Prostate Cancer. Program and abstracts of the American Urological
Association 2006 Annual Meeting; May 20-25, 2006; Atlanta, Georgia.
Abstract 540.
 31. Hamdy F, Lane A, Neal D, Mason M, Donovan J. The ProtecT (Prostate
testing for cancer and Treatment) study: evaluating screening and treatment
of clinically localised prostate cancer in the UK. Program and abstracts of
the American Urological Association 2006 Annual Meeting; May 20-25, 2006;
Atlanta, Georgia. Abstract 653.
 32. Warlick C, Trock B, Landis P, Epstein J, Carter H. Pathological
outcomes are similar for men in an expectant management program undergoing
delayed surgical intervention compared to those undergoing immediate
intervention. Program and abstracts of the American Urological Association
2006 Annual Meeting; May 20-25, 2006; Atlanta, Georgia. Abstract 654.
 33. Klotz L, Loblaw A, Holden L, Morton G, Nam R, Danjoux C. Active
surveillance with selective delayed intervention for favorable risk
prostate cancer: Updated experience with 500 patients. Program and
abstracts of the American Urological Association 2006 Annual Meeting; May
20-25, 2006; Atlanta, Georgia. Abstract 655.