| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Opinion |
1 Stanford University, Department of Urology S-287, Stanford University School of Medicine, Stanford, CA 94305-5118, E-mail tstamey{at}leland.stanford.edu
 |
Abstract |
|---|
 Top Abstract IntroductionReferences |
|---|
65 years of age is only 226 per 100 000 men. At least
40 000 of 100 000 men over age 65 (40%) have invasive prostate
cancer as judged by examination of prostates in 3- to 4-mm
step-sections. Thus, only 1 of every 177 men 65 years of age or older
(226 in 40 000) with invasive prostate cancer dies annually from his
cancer. Serum PSA between 2 and 10 µg/L is used almost universally as
an indication to biopsy the prostate. When 10–20 biopsies are commonly
taken, it is not surprising that 
40% of men are biopsy-positive for
prostate cancer. Despite this reliance on serum PSA as an indication
for biopsy, data at Stanford show no clinically useful relationship
between preoperative serum PSA (in the range 2–10 mg/L) and the volume
of Gleason grade 4/5 cancer or the volume of Gleason grades 3, 2, and 1
cancer, nor can we show any useful relationship of such preoperative
PSA concentrations (2–10 µg/L) to biochemical PSA failure rates
after radical prostatectomy. We urgently need a better serum marker for
prostate cancer. Because PSA biochemical failure rates after radical
prostatectomy are directly proportional to the amount of Gleason grade
4/5 cancer in the prostate, a serum marker of Gleason grade 4/5
carcinoma could be ideal. |
Introduction |
|---|
|
TopAbstractIntroductionReferences |
|---|
My concern with marketing these assays for use at the point of care is that it can send the wrong message to the clinician and the patient. To determine PSA "immediately after venipuncture" implies some kind of a hurry to know the PSA value to arrive at an immediate or early decision. To the contrary, what is actually needed in prostate cancer are a lot of thought, careful discussions and, often, second or even third opinions. Even the physiological variation of serum PSA is not inconsequential, and a second PSA value measured 2–3 weeks later without any examination of the prostate at either the first or second visit may be surprisingly different from the first value (2). More importantly, we must never forget that at least 40% of men over 50 years of age have histologically invasive prostate cancer (3)(4), but only a small fraction of those who have prostate cancer actually die from it.
It is easy to estimate how large the disparity is between histologic prevalence of invasive cancer and death from prostate cancer. The extraordinary incidence of prostate cancer is indisputable and increases with aging. It is best documented by examining serial sections of prostates from men who die unexpectedly in apparent good health and therefore require a coroner’s inquest. The earliest detailed study, to my knowledge, was that of L.M. Franks from London, England (3), who in 1954 found that 69 of 178 men (38%) over 50 years of age had invasive prostatic carcinoma when their whole prostates were sectioned at 4-mm intervals, a value that contrasted sharply with the 1.4% death rate for prostate cancer in the United Kingdom for men over 50 years of age (5). The elegant studies of Sakr et al. (6) on younger men in the 4th and 5th decades of life who were killed on the streets of Detroit showed invasive prostatic cancer in 27% and 34%, respectively; these studies argue strongly that prostate cancer begins in the 4th decade of life in a highly significant number of men. Almost as useful as autopsies in men requiring a coroner’s study are men who have their prostates removed not because of anything wrong with their prostate but because of invasive bladder cancer. In 139 consecutive such prostates at Stanford serially sectioned at 3-mm intervals, 40% had invasive prostate cancer at a mean age of 66 years (median, 65 years; range, 31–84 years) (4). Hirst and Bergman (7) reported an autopsy incidence of invasive carcinoma of 54% in men >80 years of age. For an additional 21 references on the autopsy incidence of prostate cancer, see Brawley et al. (8). The median age at diagnosis of prostate cancer in the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program is 71 years for whites and 69 years for blacks (8).
Although the histologic prevalence of prostate cancer clearly increases
with increasing age from the 4th decade of life (27%) (6)
to the 9th decade of life (54%) (7), we need to compare the
histologic prevalence of invasive cancer to the SEER death rate from
prostate cancer in the US. Because 34% of men in the 5th decade have
invasive prostate cancer (6), I have taken a conservative
histologic prevalence of invasive prostate cancer for all men to be
40%. In the SEER database, 75% of prostate cancers are diagnosed in
men 65 years of age or older (8)(9). The SEER
mortality rates from prostate cancer are age-adjusted to the 1970 US
standard population and are available for the years 1993–1997 for all
ages, for those under 65, and for those 65 and over (9).
Because "all" ages and ages "under 65" include even males under
10 years of age, I have used only the SEER mortality data from prostate
cancer for men over 65 years, which is 226.2 deaths per 100 000 men,
or 0.2262 per 100 men, which represents an annual risk of death caused
by prostate cancer of 0.23% for this age group. Given a hypothetical
100 000 men in the >65 years of age group and a prevalence rate of
prostate cancer of 40% in this group, we would expect 40 000 men to
have prostate cancer. With an annual death rate of 226.2 per 100 000
men (incidence rate) in this age group, we can see that 226.2 of
40 000 men, or 1 in 177 men with histologic evidence of prostate
cancer is likely to die of his cancer each year. I recognize that it
may not be epidemiologically correct to compare age-adjusted death
rates to histologic prevalence of invasive prostate cancer, but I am on
safer ground using the SEER statistics only for men 65 years of age.
It is remarkable how few men 65 years of age in the US actually die
from prostate cancer, given the extraordinary histologic prevalence of
invasive prostate cancer. The numbers are easy to verify. The 226.2
deaths from prostate cancer per 100 000 men 65 years of age in 1995
is derived from the National Center for Health Statistics of the
CDC based in Hyattsville, MD. The last NCI publication of these
data showed that 34 475 men died of prostate cancer in 1995
(8). That the CDC number is approximately correct can be
confirmed by taking the 13 800 000 men in the US 65 years of age in
1995 (10) times the death rate from prostate cancer of
0.2262 per 100 men, which equals 31 216 deaths from prostate cancer,
which is close enough to the CDC count of 34 475 based on death
certificates.
Let us turn to recent evidence from Stanford as to just how good serum PSA is in reflecting the volume of prostate cancer and the probability of curing the cancer by radical prostatectomy or radiation therapy. To my knowledge, we presented the original evidence in 1987 that serum PSA was proportional to the volume of prostate cancer in our first 45 untreated radical prostatectomy specimens (11); the Pearson r2 x 100 for serum PSA vs cancer volume was only 35%, and even this correlation was driven mostly by the 58% of men who had serum PSA concentrations of 10 to >100 µg/L. We recently examined Pearson linear correlations for hundreds of men undergoing radical prostatectomy with a preoperative serum PSA of 2–22 µg/L; we chose 22 µg/L as the upper limit because in long-term follow-ups of men with peripheral zone cancers who were cured after radical prostatectomy, the highest preoperative serum PSA was 23 µg/L (12). The only two morphologic variables in the cancer that can be expected to reflect serum PSA concentration are the volume of Gleason grade 4/5 cancer (undifferentiated) and the volume of Gleason grades 1, 2, and 3 cancer (well-differentiated). Our 3-mm step-sections with estimates of the percentage of Gleason grade 4/5 cancers in all radical prostatectomies allows us to separate the largest (index) cancer volume into the volume (cm3) of Gleason grades 4/5 cancer and the volume of grades 3, 2, and 1 cancer. The r2 x 100 value for this relationship of serum PSA to 662 grade 4/5 cancers was 13%, and for 776 grade 3, 2, and 1 cancers, it was only 4%. These correlations are clearly much too low for PSA in the range of 2–22 µg/L to be used as a reliable marker for prostate cancer.
Because of this unexpectedly poor (one might say disastrous)
correlation of both well-differentiated and undifferentiated cancer
volume with preoperative serum PSA in the range of 2–22 µg/L, we
also examined the relationship of preoperative serum PSA at 1-µg/L
intervals between 2 and 10 µg/L and at 10–11.9, 12–16.9, 17–21.9,
and >22 µg/L to biochemical PSA failure rates in men treated only by
radical prostatectomy. We used Wilcoxon statistics for comparison of
132 pairs, but used the Bonferroni correction for 0.05 statistical
significance (0.05/132 = 0.0004) to avoid falsely detecting
differences where none existed. As seen in Table 1
, Kaplan-Meier survival analysis for PSA failure rates was not
statistically different in men with preoperative PSA concentrations of
2–3 µg/L and higher preoperative PSA concentrations until PSA
reached 12 µg/L. At serum PSA concentrations of 12–17 µg/L and
certainly above 17 µg/L, there is no question in Table 1
that
preoperative serum PSA is reflecting Kaplan-Meier PSA cure rates. These
data argue that a man might be just as curable at a pretreatment serum
PSA of 12 µg/L as he is at 2–3 µg/L.
|
In confirmation of the poor discrimination of serum PSA values between 2 and 12 µg/L in predicting cure rates after radical prostatectomy, Shipley et al. (13) analyzed preradiation serum PSA concentrations in 1765 men treated only by a standardized external beam radiation therapy protocol. The authors used recursive partitioning analysis to find the lowest preradiation serum PSA associated with the best postradiation PSA cure rates. The lowest preradiation PSA was 9.2 µg/L, which the authors showed by bootstrap analysis did not give significantly different results from a value of 10 µg/L. Thus, for the two major treatment modalities for prostate cancer—radical prostatectomy and external beam radiation therapy—pretreatment serum PSA concentrations between 2 and 10 µg/L are not helpful in predicting therapeutic outcome.
It is highly likely that benign prostatic hyperplasia (BPH) is the
major contributor to serum PSA between 2 and 10 µg/L. The
contribution of BPH in the transition zone to serum PSA is most clearly
seen when all the BPH in the transition zone is removed by open
surgical enucleation for urinary obstruction in men who have negative
prostate biopsies. Although we first reported this observation in 1987
(11), eight cases in the last 10 years (Table 2
) with long-term follow-ups serve to emphasize this convincing
relationship. All of the men in Table 2
show striking decreases in
their serum PSA, often as much as 100-fold, a decrease that we now know
(but not in 1987) is maintained for years. Thus, BPH nodules exert a
powerful influence on serum PSA; in fact, the free/total PSA ratio (and
the complex/total PSA ratio, an indirect measure of free PSA) is
probably nothing more than an estimate of the amount of BPH
contributing to serum PSA, an estimate that can be made equally well by
measuring the size of the prostate, or better yet, the size of the
transition zone (in which BPH arises) by transrectal ultrasound
(14).
|
It is clear that we urgently need a better marker than PSA, regardless
of its molecular form. The ideal new marker would be one based on
Gleason grade 4/5 cancer. As seen in Fig. 1
, for each 10% increase in this undifferentiated form of
prostate cancer in the peripheral zone, 10% of men failed
biochemically on long-term follow-ups (15). Importantly,
although the index (largest) cancer volume in our Stanford database of
859 untreated radical prostatectomies for peripheral zone cancers has
gradually decreased from a mean and median of 4.77 and 3.13
cm3, respectively, in 1988–1991 to 2.54 and 2.01
cm3 in 1997–2000, 80% of all men continue to
have Gleason grade 4/5 cancer (mean of 35%, median of 20% in
1988–1991 vs mean of 33%, median of 20% in 1997–2000).
|
My major concern is that transrectal biopsy regimens have now become so efficient that their positive biopsy rates are virtually identical to the 40% histologic ("autopsy") prevalence of prostate cancer. Presti et al. (16), by adding only 2 far lateral biopsies at the base and mid-region on each side of the prostate to the standard 6 mid-lobe biopsies (a total of 10 biopsies), increased their positive biopsy rate from 34% to 42% in 483 consecutive men undergoing first-time biopsies. Others have recently recommended taking as many as 15–20 biopsies of the prostate in a "5-region technique"; in 119 biopsy sessions, the positive biopsy rate was 40% (17).
What is the solution to our dilemma that we are now detecting prostate
cancers that would usually be found only at autopsy in 3-mm
step-sections (40%) when only 1 of every 177 of these men 65
years of age will die annually of prostate cancer? The first solution
is to recognize the problem and to fully discuss it with our patients.
The second solution is to recognize the serious limitations of serum
PSA <12 µg/L in reflecting the histologic volume and grade of the
cancer as well as the potential cure rates by radical prostatectomy or
irradiation. There clearly is an urgent need for a better serum marker
than PSA, including all of its molecular forms. A marker proportional
to the volume of Gleason grade 4/5 cancer would be ideal (Fig. 1
), but
that will require knowing the gene expression characteristics of grade
4/5 cancer as well as some good luck that some of the related proteins
reach the serum in proportion to the volume of grade 4/5 cancer.
It will not be easy. However difficult the task, the statement by Brawley et al. (8) in November 1998 that "Medicine’s ability to diagnose prostate cancer has improved profoundly and has outpaced medicine’s ability to predict the biological behavior and the true clinical significance of diagnosed tumors" should serve as the understatement for the new millennium. Until we are able to predict the biological behavior of prostate cancer, I think it is difficult to justify a "point of care" rapid assay for serum PSA.
|
References |
|---|
|
TopAbstractIntroductionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  R. W. Veltri, M. C. Miller, S. Isharwal, C. Marlow, D. V. Makarov, and A. W. Partin Prediction of Prostate-Specific Antigen Recurrence in Men with Long-term Follow-up Postprostatectomy Using Quantitative Nuclear Morphometry Cancer Epidemiol. Biomarkers Prev., January 1, 2008; 17(1): 102 - 110. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. F. Shariat, M. W. Kattan, E. Traxel, B. Andrews, K. Zhu, T. M. Wheeler, and K. M. Slawin Association of Pre- and Postoperative Plasma Levels of Transforming Growth Factor {beta}1 and Interleukin 6 and Its Soluble Receptor with Prostate Cancer Progression Clin. Cancer Res., March 15, 2004; 10(6): 1992 - 1999. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. W. Kattan, S. F. Shariat, B. Andrews, K. Zhu, E. Canto, K. Matsumoto, M. Muramoto, P. T. Scardino, M. Ohori, T. M. Wheeler, et al. The Addition of Interleukin-6 Soluble Receptor and Transforming Growth Factor Beta1 Improves a Preoperative Nomogram for Predicting Biochemical Progression in Patients With Clinically Localized Prostate Cancer J. Clin. Oncol., October 1, 2003; 21(19): 3573 - 3579. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Thompson, R. J. Leach, B. H. Pollock, and S. L. Naylor Prostate Cancer and Prostate-Specific Antigen: The More We Know, the Less We Understand J Natl Cancer Inst, July 16, 2003; 95(14): 1027 - 1028. [Full Text] [PDF]
|
|
|
|
|
|
S. D. Cramer, B.-L. Chang, A. Rao, G. A. Hawkins, S. L. Zheng, W. N. Wade, R. T. Cooke, L. N. Thomas, E. R. Bleecker, W. J. Catalona, et al. Association Between Genetic Polymorphisms in the Prostate-Specific Antigen Gene Promoter and Serum Prostate-Specific Antigen Levels J Natl Cancer Inst, July 16, 2003; 95(14): 1044 - 1053. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Jain, A. Lam, I. Vivanco, M. F. Carey, and R. E. Reiter Identification of an Androgen-Dependent Enhancer within the Prostate Stem Cell Antigen Gene Mol. Endocrinol., October 1, 2002; 16(10): 2323 - 2337. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B.-L. Adam, Y. Qu, J. W. Davis, M. D. Ward, M. A. Clements, L. H. Cazares, O. J. Semmes, P. F. Schellhammer, Y. Yasui, Z. Feng, et al. Serum Protein Fingerprinting Coupled with a Pattern-matching Algorithm Distinguishes Prostate Cancer from Benign Prostate Hyperplasia and Healthy Men Cancer Res., July 1, 2002; 62(13): 3609 - 3614. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. F. Shariat, M. Shalev, A. Menesses-Diaz, I. Y. Kim, M. W. Kattan, T. M. Wheeler, and K. M. Slawin Preoperative Plasma Levels of Transforming Growth Factor Beta1 (TGF-{beta}1) Strongly Predict Progression in Patients Undergoing Radical Prostatectomy J. Clin. Oncol., June 1, 2001; 19(11): 2856 - 2864. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
