- Email: [email protected]
Clinical use of Prostate Specific Antigen in Patients with Prostate Cancer
0022-534 7/89/1424-1011$02.00/0 Vol. 142, October
THE JOURNAL OF UROLOGY
Printed in U.S.A.
Copyright© 1989 by AMERICAN UROLOGICAL ASSOCIATION, INC.
CLINICAL USE OF PROSTATE SPECIFIC ANTIGEN IN PATIENTS WITH PROSTATE CANCER M'LISS A. HUDSON,* ROBERT R. BAHNSON
t,+ AND
WILLIAM J. CATALONA§
From the Division of Urologic Surgery, Washington University School of Medicine, St. Louis, Missouri
ABSTRACT
The clinical use of prostate specific antigen as a screening test for prostate cancer, as a preoperative determinant for staging of prostate cancer and to monitor response to therapy in prostatic cancer patients was evaluated in 168 men with benign prostatic hyperplasia and 231 men with prostate cancer. Only 3% of the men with benign prostatic hyperplasia had prostate specific antigen levels greater than 10 ng. per ml. compared to 44% of the men with proved prostate cancer. Preoperative prostate specific antigen levels increased with higher clinical stages of prostate cancer but there was substantial overlap among stages. Among patients with stage Al prostate cancer who were followed expectantly none had an elevated prostate specific antigen value or metastatic disease during a followup of 15 to 120 months. After radical prostatectomy serum prostate specific antigen values decreased to undetectable levels (less than 0.6 ng. per ml.) in 89% of the patients with organconfined disease, in 87% of those with microscopically positive margins only but in only 34% with seminal vesicles or lymph node involvement. Failure of the prostate specific antigen levels to decrease to the undetectable range after radical prostatectomy was associated with a greater likelihood of subsequent tumor recurrence. Only 3 of 18 patients (17%) treated with definitive radiation therapy had post-irradiation prostate specific antigen values of less than 0.6 ng. per ml., while in 39% the prostate specific antigens values remained greater than 4 ng. per ml. and in 4 of 18 (22%) the values were greater than 10 ng. per ml. Of patients with previously untreated stage D2 prostate cancer the mean pre-treatment prostate specific antigen value was 63.7 ng. per ml. compared to a post-hormonal therapy mean value of 31.1 ng. per ml. Of 32 patients treated with hormonal therapy 14 had stable disease, including 13 with prostate specific antigen levels of less than 10 ng. per ml. In contrast, 18 patients had progressive disease, of whom 16 had prostate specific antigen levels of more than 10 ng. per ml. We conclude that the serum prostate specific antigen assay is most useful clinically to monitor the response to therapy of prostate cancer patients. (J. Urol., 142: 1011-1017, 1989) Prostate specific antigen (PSA) is a glycoprotein with a molecular weight of approximately 35,000 daltons. It was described by Wang and associates in 1981 as an antigen associated only with prostatic tissue. 1 PSA has been localized to the cytoplasm of prostatic acinar cells, ductal epithelium and sem inal fluid. 2 PSA has not been found in any other normal human tissue nor is it associated with any malignancies other than prostate cancer. 2 • 3 Elevated serum levels of PSA have been associated with benign prostatic hyperplasia and prostatic carcinoma. Characterization studies have shown that PSA is composed of a single polypeptide chain of 240 amino acids. 4 ' 5 It is distinct biochemically and immunologically from prostatic acid phosphatase. PSA appears to be similar in composition to the serine proteases of the kallikrein family. Although the exact function of PSA is unknown, it has been suggested that PSA may be involved in liquefaction of the seminal coagulum.5 We assess the clinical usefulness of PSA as a serum tumor marker for diagnosis, staging, and monitoring of response to
therapy and disease progression in patients with known or suspected prostate cancer. METHODS
PSA assay. Serum PSA levels were determined using the Tandem-R PSAII immunoradiometric assay. 6 The normal range for serum PSA levefs recommended by the manufacturer is 0 to 4 ng,/ml. and the lower limit of detection is stated to be 0.15 ng,/ml. To set the lower limit of detection of this assay the manufacturer recommended the use of pooled female sera as a zero control. At our laboratory 10 pooled female sera duplicates were obtained from female volunteers for PSA controls and the mean PSA level plus or minus 3 standard deviations was determined. We also evaluated the lower limit of detection of PSA in this assay by performing serial dilutions of a known concentration (2 ng,/ml.) of PSA with pooled female sera or Hybritech Zero Diluent in 6 separate experiments. Patients. Serum PSA levels were determined in 168 men with benign prostatic hyperplasia, of whom 90 had histological con. firmation of adenomatous hyperplasia and 78 had bladder outlet obstructive symptoms with an enlarged, nonindurated prostate gland on digital rectal examination. A total of 14 men undergoing cystoprostatectomy for invasive bladder cancer also served as controls. Serum PSA levels were obtained preoperatively from 107 men with biopsy proved prostate cancer, organ confined tumors on rectal examination, and a negative bone
Accepted for publication April 5, 1989. * Recipient of American Urological Association American Foundation for Urologic Disease Scholarship 1988-1990. t Recipient of American Cancer Society Clinical Oncology Career Development Award 86-48. t Current address: Division of Urologic Surgery, University of Pittsburgh School of Medicine, Presbyterian University Hospital, Pittsburgh, Pennsylvania 15213. . § Requests for reprints: Division of Urologic Surgery, Washington University School of Medicine, 4960 Audubon Ave., St. Louis, Missouri 63110. 1011
II Hybritech, Inc., San Diego, California.
1012
HUDSON, BAHNSON AND CATALONA
scan and computerized tomography (CT) scan of the abdomen. PSA values were available in 174 men after radical prostatectomy. Of these patients 93 had a PSA level obtained within 6 months postoperatively, while in 81 the initial postoperative PSA values were obtained more than 6 months after radical prostatectomy, with a mean of 27.8 months (range 7 months to 11 years). Post-radiation therapy PSA levels also were determined in 18 patients who were irradiated for histologically proved clinically localized prostate cancer. Mean followup in these patients was 17.4 months (range 9 to 59 months). Of these 18 patients 7 had PSA levels obtained within 6 months of completion of radiation therapy. PSA values were available in 39 men with known metatastic prostate cancer. Mean followup in these patients was 26.5 months (range 8 to 48 months). Staging. Clinical staging of prostatic cancer was defined with a modification of the Whitmore-Jewett system: stage A-tumors not suspected on digital rectal examination but found on histological examination after simple prostatectomy (stage Al-well differentiated tumors confined to less than 5% of the specimen and stage A2-moderately or poorly differentiated tumors, or those with greater than 5% involvement of the specimen), stage B-tumors palpable on digital rectal examination and confined to the prostatic capsule (stage Bl-tumor limited to 1 lobe of the prostate and less than 2 cm. in diameter and stage B2-lesions involving both prostatic lobes or being greater than 2 cm. in diameter), clinical stage C-tumors that have extended beyond the prostatic capsule but without evidence of distant metastases and stage D2-patients with clinical evidence of distant metastases. Pathological staging for prostate cancer was defined by the following system. Pathological stages Al, A2, Bl and B2 tumors were defined as in the clinical staging system but were verified histologically. Organ-confined disease was defined as tumors that on histological examination showed neither evidence of extension through the prostatic capsule, extension to the bladder neck, urethral margins, fascia surrounding the seminal vesicles nor evidence of pelvic lymph node metastases. Microscopically positive margins on pathological examination (pathological stage Cl) refers to tumor at either the urethral or bladder neck margin, or extension through the prostatic capsule to the surgical margin but without seminal vesicle or pelvic lymph node involvement. Pathological stage C2 disease refers to patients with involvement of the seminal vesicles or fascia surrounding the seminal vesicles. Pathological stage Dl refers to patients with clinical stage A or B disease who at pelvic lymphadenectomy had histologically documented lymph node metastases. Extracapsular refers to patients with pathological stage C or D disease. All patients with local extracapsular tumor extension had either positive margins or seminal vesicle invasion. Tumor penetration through the prostatic capsule with negative surgical margins was not observed except in the area within the fascia surrounding the seminal vesicles, in which case it was considered as seminal vesicle invasion. In these patients there were no instances of capsular penetration of tumor covered by periprostatic fat or a neurovascular bundle with a negative surgical margin. Treatment options based on clinical stage included observation for patients with clinical stage Al disease who were more than 65 years old or had other serious medical problems, nervesparing radical prostatectomy for patients with stages Al, A2, Bl or B2 disease who had a life expectancy of 10 years and were in good health, external beam radiation therapy for patients with stage C prostate cancer or those with stages Al, A2, Bl or B2 disease who did not wish to undergo radical prostatectomy, and early or delayed hormonal therapy for patients with metastatic disease.
of detection of 0.2 ng./ml. and results were reported accordingly. In March 1988, because of observed fluctuations in repeated PSA assays in the low ranges of detectability, we conducted studies in which 3 PSA samples containing approximately 2.0 ng./ml. PSA were serially diluted with pooled female sera (fig. 1, A). PSA concentrations of O to 0.6 ng./ml. could not be distinguished from the zero control. Therefore, at that time our laboratory set the lower limit of detection for the assay as 0.6 ng./ml. Accordingly, much of the data reported in our study were obtained when the lowest possible value reported was less than 0.6 ng./ml. We performed subsequent studies in which the mean PSA value plus or minus 3 standard deviations for pooled female sera was determined. The mean value was 0.189 ± 0.318 ng./ml., and 3 standard deviations above the mean was 0.507 ng./ml., which is considerably higher than the 0.15 ng./ml. value reported by the manufacturer. More recently, we repeated the serial dilution studies after ·revision of the bead washing procedure and using the Hybritech Zero Diluent (pooled female serum treated to remove any PSA binding proteins). In these later studies the lower limit of detectability was demonstrated to be between 0.3 and 0.4 ng./ml. (fig. 1, B). For the purposes of this study the lower limit of detectability was 0.6 ng./ml. Benign prostatic hyperplasia patients. Of 168 men with benign prostatic hyperplasia (35 of 168) 21 % had PSA levels above the manufacturer's recommended upper limit of normal of 4 ng./ ml. (fig. 2). Only 3 patients (2%) had PSA levels greater than 10 ng./ml. PSA after cystoprostatectomy for bladder cancer. All 14 patients undergoing radical cystoprostatectomy for bladder cancer had postoperative PSA levels that decreased to the undetectable range. PSA and clinical stage. Of patients with clinical stage A or B prostate cancer 62% (64 of 103) had preoperative PSA values greater than 4 ng./ml. compared to 72% (31 of 43) of those with clinical stage C or D disease. Of these 146 men with 3.0
~;
Pooled Female Sero 2.5
• ••••___ ••• , .a .t:t···
2.0 1.5 1.0 0.5
E
~ C
3.0
~
/ ..a··-.,. ----~.,,,,,.:,:·...
PSA assay. In our initial studies with the Tandem-R PSA assay we adopted the manufacturer's recommended lower limit
p'
<~1;. •· ---:e...···-~··' ".
A
•••
~~~--
'f-
I
I
Hybritech Zero Diluent
ct
U)
a.. 2.5 2.0 1.5 1.0 0.5
O
10
20
30
40
50
% PSA from RESULTS
.,., ·······
H
60
70
80
90
100
2 ng/ml
FIG. 1. Serial dilution of 3 PSA samples from known concentration of approximately 2 ng./ml. to lower limit of detection. A, diluent used was pooled female sera. B, after revision of bead washing procedure. Diluent used was Hybritech zero calibrator.
1013
CLU\TICAL tJSE OF PROSTA. TE SPECIFIC Al'>.JT1GEN IN PATIEl\ITS \VY.CH PROSTATE CAI~CER
100
••
iOO
e
• •
0
•
0
I •
O@
c
"®
IE
10.0
':?100
0
It:
--·
8"---,--
4.0
.Q
..,
OJ
H
"u
4.0
Gli!l~{j@
0.6
0e.@ee00
••
u 00 Q.
•
•• •"
0
:.
•
0 C
..
.•
H"
.. i•v..
@P
$~0"M0G~
H@
H
lb
LO
2
H
--;:-= • ...
• ©
1.0
• •
0 EUH)@$0®
06 01
BPH
A1
A2
81
62
C
D2
Clinictll Stage
Fw. 2. Preoperative PSA values in patients with benign prostatic hyperplasia (BPH), clinically localized prostate cancer or metastatic prostatic cancer.
prostate cancer 44% had PSA levels greater than 10 ngJml., including 36% (37 of 103) with clinical stages A or B disease (fig. 2).
There was a trend of increasing PSA levels with higher clinical stage of prostate cancer but considerable overlap was observed among the clinical stages. PSA and pathological stage. Preoperative PSA values could not be used to distinguish between patients found histologically to have organ-confined prostate cancer and those with extracapsular tumor extension (fig. 3). Higher PSA values did not always indicate extracapsular extension and lower levels did not always ensure that tumor was confined within the prostatic capsule. PSA radical prostatectomy. As mentioned previously, interval PSA levels of less than 0.6 ng./mL were eov.,w.,uc,g,u undetectable. By this criterion after radical prostatectomy 89% of the patients with organ-confined tumors had PSA levels that decreased to the undetectable range. Similarly, postoperative PSA levels decreased to less than 0.6 ng./ ml. in 87% of the patients with microscopically positive margins (capsular penetration with a positive margin, or tumor at the methral or bladder neck margin but with negative seminal vesicles) following radical prostatectomy (fig. 4). In patients with seminal vesicle (pathological stage C2) or lymph node (pathological stage Dl) involvement the results were markedly different. Only 34 % of the patients had PSA levels that decreased to less than 0.6 ng./ml. following radical prostatectomy. Of the patients with seminal vesicle or lymph node involvement 45% had postoperative PSA values greater than the upper limit of 4 ng./ml. and 7 of these 29 patients (24 %) had PSA values of more than 10 ng./ml. PSA and tumor recurrence. Of the post-radicalprostatectomy patients with PSA levels of greater than 10 ng./ml. 80% had recurrent disease. Conversely, to date only 1 patient with undetectable PSA values postoperatively has had recurrent disease (a local recurrence at the urethral anastomosis) following radical prostatectomy (fig. 4). The other 5 patients with
01-L-------------------Organ Confined
Extrocopsulor E)(lension
Pathologic Stage FIG. 3. Comparison of preoperative PSA values in patients with pathologically localized prostate cancer and those with extracapsular tumor extension on pathological examination of tumor. Lines indicate mean PSA values.
100
if
"'.,if
.,.
lO.O
E
'
0,
<::
~ U)
"
i
4.0
... .."°"
01.
...... IIHi!181lUlli
ef
;•
.,. 2
1.0 06
...
•=• awn, Bm!!ffl ~r
...
.,,_,
·-
v = Recurrenl disease
0.1 - - ' - - - - - , - - - - , - - - , - - - r - - - - - r - - - - - , - - - ~ -
C2 (+margins)
Pathologic Sto11e
(+SV)
D1 (+pelvic
nodu)
FIG. 4. Postoperative PSA values in relationship to postoperative pathological stage. Line attached to dot indicates patients who have recurrent tumor.
1014
HUDSON, BAHNSON AND CATALONA
recurrent disease shown as having undetectable PSA values had received adjuvant radiation or endocrine therapy for the recurrence before the PSA value was obtained. No patient with an undetectable PSA level following radical prostatectomy has had metastatic disease. PSA as an indicator of incomplete tumor excision. PSA levels were measured within 6 months after radical prostatectomy in 93 ofl 74 patients in this study. Of these 93 patients 4 7 had an undetectable PSA level (less than 0.6 ng./ml.) postoperatively and are without evidence of recurrent disease but they have been followed for less than 1 year, while 32 had an undetectable PSA value (less than 0.6 ng./ml.), are without evidence of recurrent disease and have been followed at least 12 months but not longer than 33 months. Seven of the remaining 14 patients had a PSA level of greater than 0.6 ng./ml., including 3 (43%) who had recurrent disease at less than 1 year, while 7 initially had an undetectable PSA value postoperatively that increased to greater than 0.6 ng./ml. during year 1 of followup. To date 1 of the latter 7 patients (14%) had a documented recurrence at 8 months. PSA as an indicator of current disease status. Of 81 patients with an initial PSA value measured more than 6 months after radical prostatectomy 48 have a level of less than 0.6 ng./ml. and are without evidence of recurrent disease. Among these 48 patients 18 have been followed for more than 1 year, 18 for more than 3 years and 12 for more than 5 years. Of the remaining 33 patients 16 have a PSA level of greater than 0.6 ng./ml. but they are without evidence of recurrent disease to date. Followup in these patients is more than 1 year in 6, more than 3 years in 9 and more than 5 years in 1. Of the final 17 patients with a documented recurrence 11 had a PSA level of greater than 0.6 ng./ml. Followup in these 11 patients was greater than 1 year in 5, greater than 3 years in 5 and greater than 5 years in 1. A total of 6 patients with recurrent disease had PSA levels of less than 0.6 ng./ml. but 5 had received either adjuvant radiation or endocrine therapy before measurement of the first PSA value. Prostatic acid phosphatase levels. All patients undergoing radical prostatectomy except 1 had a preoperative prostatic acid phosphatase level in the normal range. Postoperatively, prostatic acid phosphatase values remained in the normal range for all except this same patient. The patient with an elevated prostatic acid phosphatase level before and after radical prostatectomy was believed to have an elevated level on the basis of liver disease. The postoperative PSA value was undetectable and he had no evidence of recurrent tumor after 19 months. Only 4 patients have had an elevated prostatic acid phosphatase level after radical prostatectomy and each of these patients had a concomitantly elevated PSA value above 10 ng./ml. in the immediate postoperative period. Of these 4 patients 3 have had distant metastases. PSA and stage Al cancer patients managed conservatively. No patient followed expectantly with stage Al prostate cancer has had either an elevated PSA value or metastatic disease during followup. Of these 15 patients 5 have been followed for more than 8 years, 2 for more than 4 years and 8 for less than 2 years. The natural history of stage Al prostate cancer is such that disease progression would be unlikely to occur during this interval in the majority of these patients. PSA and patients treated with radiation therapy. Only 3 of 18 patients (17%) treated with primary radiation therapy who appear by standard clinical parameters to be free of recurrent tumor had post-radiation PSA values less than 0.6 ng./ml. Of the post-radiation PSA values 39% remained above 4 ng./ml. and 4 of 18 (22%) were greater than 10 ng./ml. during a mean followup of 17.4 months (range 9 to 59 months). None of these patients has had a clinical recurrence to date during the 2 years when PSA measurements have been available at our institution. Patients who failed radiation therapy before this period are classified with the stage D2 cancer patients.
Patients treated with hormonal therapy. Of 16 patients with previously untreated stage D2 prostate cancer the mean pretreatment PSA value was 63. 7 ± 42.8 ng./ml. compared to a post-hormonal therapy mean value of 31.1 ± 14.2 ng./ml. Only 1 of 14 patients with apparent clinically inactive D2 prostate cancer as evidenced by stable changes on a bone scan, no bone pain and no voiding difficulties had a PSA level of greater than 10 ng./ml. Of 18 patients with clinically active metastatic prostate cancer treated with hormonal therapy 16 (88%) had PSA levels of greater than 10 ng./ml. These patients showed evidence of progressive disease as manifested by increasing bone pain, new changes on bone scan and/or difficulty voiding, while only 2 patients with progressive disease had PSA levels of less than 10 ng./ml. Among 9 patients in whom pre-treatment and post-treatment PSA levels were obtained the post-treatment PSA levels were obtained within 1 to 4 months after institution of hormonal therapy. Post-treatment levels decreased in 8 of the 9 patients (89%) (fig. 5). DISCUSSION
PSA as a screening test for prostate cancer. The use of serum tumor markers to detect prostate cancer has been a clinical goal since a correlation was first noted between elevated prostatic acid phosphatase levels and metastatic prostate cancer. 7 However, prostatic acid phosphatase elevations usually indicate the presence of metastatic disease and have not proved to be useful to detect localized prostate cancer. Because PSA has been shown to be more sensitive than prostatic acid phosphatase in following patients for development of progressive disease, hope was raised that PSA might prove to be a better marker for localized prostate cancer.s.-12 In most reports PSA levels were measured with 1 of 2 different assays, including a polyclonal antibody radioimmuno-
>100
PSA (ng/ml)
1.0
o Inactive disease
• Active disease
POST PRIOR TO HORMONAL THERAPY HORMONAL THERAPY FIG. 5. Pre-treatment and post-treatment PSA values in patients with clinical stage D2 prostate cancer treated with hormonal therapy. Lines connect pre-treatment and post-treatment values in same patient.
CLINICAL USE Of. PROSTATE SPECIFIC ANTIGEN. [N- PATIE:NTS ViITH PRU§Ti':,.TE CAl~ICER
Previous reports of positive PSA values in benign prostatic hyperplasia using the Tandem-R PSA immunoradiometric assay Pathologically Confirmed Benign Prostatic Hyperplasia Reference
Elevated PSA No./Total (%)
Myrtle and associates 6 • * Ferro and associates9 Ercole and associates8 Chan and associates 15 Oesterling and associates 17 Cooner and associates 18
7/352 (2) 13/40 (32.5) 10/357 (3) 0/69 (0) 7/72 (10) 8/197 (4)
PSA Value (ng./ml.) > 10 >10 ia':10 >12.4 > 10 >10
* Pathological confirmation of benign prostatic hyperplasia not stated.
assay (Pros-Check*) and a double monoclonal antibody assay (Tandem-R PSA assay). In the earliest studies investigators defined their own upper limits of normal for the assay used and based on their normal ranges, reported elevated PSA levels in ;rostate cancer patients and those with benign prostatic disease. These studies with various upper limits of normal have revealed that 20 to 86% of the patients with benign prostatic hyperplasia, and 41 to 98% wit~8 _;:';rious stages o~ pros!ate cancer have elevated PSA levels. · Each of these mvestigators concluded independently that PSA levels are not sufficiently specific to screen for prostate canc~r. . In our 168 patients with presumed bemgn prostat1c hyperplasia 21 % had a PSA level of greater than 4.0 ng./ml. as measured with the Tandem-R assay. However, only 3 of these patients (2%) had PSA levels of more than 10 ng./mL It sho1;1ld be appreciated that not all of our patients had an obstructive uropathic condition of sufficient severity to require surgical relief. Our results concur with those of Myrtle 6 and Ercole 8 and their associates who also reported that 2 to 3% of the surgically tested benign ~rostatic hyperplasia patients had preoperative PSA levels of more than 10 ng./ml. Review of previous reports confirms that PSA values measured by the Tandem-R assay in patients with pathologically confirmed benign prost~ti~ hyperplasia are less than 10 ng./ml. in the great maJonty (see table). e.s. 9, 15.17, 1s Of our patients believed to have clinically organ-confined prostate cancer, 66% had preoperative PSA levels of greater than 4.0 ng./ml. compared to 72% of those with extracapsul~r extension on rectal examination or documented metastatic disease. Over-all, of 146 men with biopsy proved prostate cancer 44% had PSA levels of greater than 10 ng./ml. (fig. 2). The observation that PSA levels of greater than 10 ng./ml. on the Tandem-R assay are associated more frequently with prostate carcinoma than benign prostatic hype_rplasia does not necessarily qualify PSA as an accurate screenmg test for prostate cancer. Statistical analyses by others have shown that the sensitivity and specificity of PSA are not sufficient to consider it an accurate screening test. 17 Nevertheless, PSA levels of greater than 10 ng./mL as measured in the Tandem-R assay are worrisome for prostate carcinoma. For example, Myrtle and associates reported that only 2% of 352 benign prostatic hyp~rplasia patients had PSA levels of more than 10 ng./ml., while PSA levels were greater than 10 ng./ml. in 30% of the stage A and 50% of the stage B prostate cancer patients. 6 In our patients 35% with clinically localized prostate cancer had a PSA value of greater than 10 ng./ml. Cooner and associates_ studied p~tients in whom the prostate gland was not sufficiently suspi cious on rectal examination to warrant performance of a biopsy and found that 10 of 18 (55 % ) with a PSA value of greater than 10 ng./ml. had prostate carcinoma. 18 It appears that additio~al studies to exclude the possibility of prostate cancer are advisable in patients with a PSA level of greater than 10 ng./ml. as measured by the Tandem-R assay. A PSA level of greater than
* Yang Laboratories, Bellvere, Washington.
1015
10 ng./mL should not be considered diagnostic of prostate cancer, since patients with a large, obstructing adenoma or prostatitis also may have PSA values above this level, Correlation between PSA value and stage of prostate cancer. In general, PSA values have been shown to increase with advancing clinical stages of prostate carcinoma. 8 • 1 2 · 14 Oesterling and associates studied 178 men with clinically confined prostate cancer undergoing radical prostatectomy and evaluated the ability of PSA to predict final pathological stage. 17 In their study preoperative PSA values were significantly higher in patients in whom the tumor penetrated through the prostatic capsule, and in those with seminal vesicle involvement and/or lymph node metastases. However, the diagnostic accuracy when one considers different upper limits of normal for PSA at 4.0, 10.0 and 16.0 ng./ml. ranged from 55 to 72% and was not sufficient to predict on an individual basis which patients had localized disease and which had extracapsular tumor extension. In our series also the mean PSA value increased with advancing clinical stage (fig. 2). Although we observed higher PSA values with more advanced stages of prostate cancer, our data also demonstrated considerable overlap in PSA values between patients with pathologically organ-confined disease and those with extracapsular tumor extension (fig. 3). Use of PSA values to monitor response to therapy in patients with prostate cancer. The greatest clinical value of PSA appears to be in monitoring response to therapy in patients undergoing treatment for prostate cancer. 13 Killian and associates suggested that PSA values may predict tumor recurrence before clinical symptoms appear in patients receiving curative therapy for organ-confined tumors. 14 Stamey and associates, using the Yang 19 assay for PSA in a series of 45 patients treated with radical prostatectomy for clinically confined prostate cancer reported that PSA routinely decreased to undetectable levels postoperatively. 10 In 6 of these patients serial postoperative PSA measurements appeared to be useful to detect residual tumor or early tumor recurrence, which was treated successfully with adjunctive radiation therapy in 5 of these 6 patients. Similarly, Chan and associates, using the Tandem-R assay, reported that in 67 of 73 patients (92%) undergoing radical prostatectomy for organ-confined tumor the PSA values decreased to undetectable levels. 15 Three patients with clinical tumor recurrence had PSA values above the normal reference range. The lower limit of detection of PSA in the Tandem-R assay is stated by the manufacturer to be 0.15 ng./ml. Special emphasis recently has been given to the lower limit of detection of the Tandem-R assay when evaluating post-radical prostatectomy patients. 8 • n. 12 · 1 5 • 17 In theory, these patients are expected to have undetectable PSA levels if all tumor has been resected. Thus, the presence of detectable PSA levels may have prognostic and therapeutic significance. However, there has been interlaboratory variation in the lower limit of detection of this assay. 11 • 12 • 15 · 17 The lower limit of detection for the Tandem-R PSA assay initially repmted by Chan and associates was 0.1 µg./1. 15 Further study by this group, cited Oesterling and associates, 17 showed a minimal detectable concentration of PSA at 0.07 ng./ml. with the Tandem-R assay. In both reports PSA decreased to undetectable levels in approximately 90% of the patients after radical prostatectomy. Stamey and associates, however, in a series of 50 post-radical prostatectomy patients who were presumed to have organ-confined disease, found that no patient had a postoperative PSA level measured with the Tandem-R assay of less than 0.2 ng./ml. and that values reported from their institution ranged from 0.4 to 1.2 ng./ml. 11 Lange and associates reported a lower limit of detection of PSA values using the Tandem-R assay as 0.5 ng./ml. 12 Our lower limit of detection has varied from 0.3 to 0.6 ng./ml. depending upon the bead washing procedure and the method used to establish the lower limit. At a recent conference on PSA sponsored by the Organ Sites
1016
HUDSON, BAHNSON AND CATALONA
Coordinating Center of the National Cancer Institute, clinical laboratory experts agreed that the most accurate means to establish the lower limit of detectability of the PSA assay is the limiting serial dilution method. For the PSA assay the diluent used should be the Zero Standard Diluent (pooled female serum with PSA binding proteins removed) rather than pooled female sera because pooled female sera may contain anti-PSA antibodies that could interfere with the assay. The other method to determine the lower limit of detectability involves establishing the mean plus 3 standard deviations of the zero standard. This latter method was stated to be somewhat less accurate than the dilution method but more commonly used in commercial assays. Early in our studies of PSA we adopted the manufacturer's recommendation for the lower limit of detectability of 0.2 ng./ ml. Later, when performing our own studies of limiting serial dilutions of pooled female sera, we found that values of less than 0.6 ng./ml. could not be distinguished from zero and, thus, for our patients the lower limit of detectability was considered to be less than 0.6 ng./ml. We have performed studies measuring the mean plus 3 standard deviations of pooled female sera and found the lower limit of detectability by this method to be 0.5 ng./ml. Finally, after consultation with the manufacturer, who advised a bead washing procedure using a commercial apparatus (Hybriwash), the serial dilution sensitivity experiment was repeated using the Zero Diluent from the Tandem-R PSA kit. An improvement in the lower limits of detectability of the assay to 0.3 to 0.4 ng./ml. PSA was noted. It should be emphasized that adherence to the specified wash procedure and performance of the procedure by the same medical technologist were important to maintain the precision and sensitivity of the assay. These results underscore the fact that minor technical details in the performance of the assay can account for appreciable differences in the lowest range of detectability. Accordingly, it seems advisable for each laboratory to establish its own lower limit of detectability. Important treatment decisions should not be based upon single assay results. Rather, a trend demonstrating a persistent elevation or an increase in serial PSA values provides a more meaningful basis for clinical decisions. In our series 89% of the patients with pathologically confirmed, organ-confined prostate cancer had PSA that decreased to undetectable levels (less than 0.6 ng./ml.). Of our patients with microscopically positive margins (pathological stage Cl) 87% also had undetectable PSA values postoperatively. None of these patients had clinical evidence of residual or recurrent tumor, although the followup is limited at 27.8 ± 18.1 months. Three patients with stage Cl disease had a local recurrence before a PSA value was obtained. These patients were treated with adjuvant radiation therapy, and they subsequently had an undetectable PSA value and are without evidence of disease to date. In contrast, PSA decreased to undetectable levels in only 34 % of our patients with seminal vesicle or lymph node involvement. Of these patients 45% had PSA levels of greater than 4.0 ng./ml. and 7 of 9 who had metastatic disease had PSA values of greater than 10 ng./ml. The other 2 patients had a value of more than 4 ng./ml. Of 18 patients with elevated PSA values 3 (17%) also had elevated prostatic acid phosphatase levels and 9 (50%) have had recurrent cancer to date. Thus, failure of postoperative PSA levels to decrease into the undetectable range after radical prostatectomy was associated with a greater likelihood of subsequent tumor recurrence. Among 108 patients treated with radiation therapy Stamey and associates10 found that PSA decreased to undetectable levels in only 13% and in only 35% did PSA levels return to the normal reference range. In our series 3 of 18 patients (17%) treated with radiation therapy who appear by standard clinical parameters to be free of recurrent tumor had a PSA value in the undetectable range and 7 of 18 (39%) had PSA levels of greater than 4 ng./ml. The significance of serial PSA values in
patients who receive primary radiation therapy requires further clarification but elevated PSA values are worrisome. Pre-treatment and post-treatment PSA values also have proved to be clinically useful in patients with known metastatic prostate cancer treated with hormonal therapy. 13 Ftirro and associates,9 and Ahmann and Schifman20 found PSA values to be more sensitive than prostatic acid phosphatase levels to monitor disease progression in patients with prostate cancer. Our data confirm that patients in clinical remission with hormonal therapy often showed a decrease in PSA values to the normal range, while those with further progression of disease after hormonal therapy again had markedly elevated PSA levels (fig. 5). A group of patients not often commented upon in the literature are those with stage Al prostate carcinoma who are managed expectantly. In our series among 15 such patients PSA values ranged from 0.6 to 3.5 ng./ml. With followup intervals of 15 to 120 months no patient in this group had an elevated PSA value nor has any patient had evidence of metastatic disease. Serial determination of PSA values in these patients also may prove to be clinically useful. Further clinical studies with long-term followup in larger numbers of patients will be required to define more completely the role of PSA as a tumor marker in prostate cancer patients.
REFERENCES
1. Wang, M. C., Papsidero, L. D., Kuriyama, M., Valenzuela, L.A., Murphy, G. P. and Chu, F. M.: Prostate antigen: a new potential marker for prostatic cancer. Prostate, 2: 89, 1981. 2. Nadji, M., Tabei, S. Z., Castro, A., Chu, T. M., Murphy, G. P., Wang, M. C. and Morales, A. R.: Prostate-specific antigen: an immunohistologic marker for prostatic neoplasms. Cancer, 48: 1229, 1981. 3. Kuriyama, M., Wang, M. C., Papsidero, L. D., Killian, C. S., Shimaro, T., Valenzuela, L., Nishiura, T., Murphy, G. P. and Chu, T. M.: Quantitation of prostate-specific antigen in serum by a sensitive enzyme immunoassay. Cancer Res., 40: 4658, 1980. 4. Watt, K. W. K., Lee, P. J., M'Timkulu, T., Chan, W. P. and Loor, R.: Human prostate-specific antigen: structural and functional similarity with serine proteases. Proc. Natl. Acad. Sci., 83: 3166, 1986. 5. Lilja, H.: A kallikrein-like serine protease in prostatic fluid cleaves the predominant seminal vesicle protein. J. Clin. Invest., 76: 1899, 1985. 6. Myrtle, J. F., Klimley, P. G., Ivor, L. P. and Bruni, j, F.: Clinical utility of prostate specific antigen (PSA) in the management of prostate cancer. Adv. Cancer Diagnostics, p. 1, 1986. 7. Gutman, A. B. and Gutman, E. B.: An "acid" phosphatase occurring in the serum of patients with metastasizing carcinoma of the prostate gland. J. Clin. Invest., 17: 473, 1938. 8. Ercole, C. J., Lange, P. H., Mathisen, M., Chiou, R. K., Reddy, P. K. and Vessella, R. L.: Prostatic specific antigen and prostatic acid phosphatase in the monitoring and staging of patients with prostatic cancer. J. Urol., 138: 1181, 1987. 9. Ferro, M. A., Barnes, I., Roberts, J. B. M. and Smith, P. J. B.: Tumour markers in prostatic carcinoma. A comparison of prostate-specific antigen with acid phosphatase. Brit. J. Urol., 60: 69, 1987. 10. Stamey, T. A., Yang, N., Hay, A. R., McNeal, J. E., Freiha, F. S. and Redwine, E.: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. New Engl. J. Med., 317: 909, 1987. 11. Stamey, T. A.: Editorial comment. J. Urol., 139: 764, 1988. 12. Lange, P. H., Ercole, C. J. and Vessella, R. L.: Serum prostatic specific antigen (PSA) in the management of patients after radical prostatectomy (RP). J. Urol., part 2, 139: 314A, abstract 607, 1988. 13. Kuriyama, M., Wang, M. C., Lee, C. L., Killian, C. S., Papsidero, L. D., Inaji, H., Loor, R. M., Lin, M. F., Nishiura, T., Slack, N. H., Murphy, G. P. and Chu, T. M.: Multiple marker evaluation
CLI!"'JICAL ·usE O:F' PROSTATE SPECIFIC AJ:¾JT!GEN I:t~ PAT!Ef'JTS V/ITH PROSTArI'E CAI"'1CER
in human prostate cancer with the use of tissue-specific antigens, J. NatL Cancer Inst., 68: 99, 1982. 14. Killian, C. S., Yang, N., Emrich, L. J., Vargas, F. P., Kuriyama, M., Wang, M. C., Slack, N. H., Papsidero, L. C., Murphy, G. P., Chu, T. M. and the Investigators of the National Prostatic Cancer Project: Prognostic importance of prostate-specific antigen for monitoring patients with stages B2 to Dl prostate cancer. Cancer Res., 45: 886, 1985. 15. Chan, D. W., Bruzek, D. J., Oesterling, J. E., Rock, R. C. and Walsh, P. C.: Prostate-specific antigen as a marker for prostatic cancer: a monoclonal and polyclonal immunoassay compared. Clin. Chem., 33: 1916, 1987. 16. Wang, T. Y. and Kawaguchi, T. P.: Preliminary evaluation of measurement of serum prostate-specific antigen level in detection of prostate cancer. Ann. Clin. Lab. Sci., 16: 461, 1986.
1017
17. Oesterling, J. E., Chan, D. W., Epstein, J. I., Kimball, A. W., Jr., Bruzek, D. J., Rock, R. C., Brendler, C. B. and Walsh, P. C.: Prostate specific antigen in the preoperative and postoperative evaluation of localized prostatic cancer treated with radical prostatectomy. J. Urol., 139: 766, 1988. 18. Cooner, W. H., Mosley, B. R., Rutherford, C. L., Jr., Beard, J. H., Pond, H. S., Bass, R. B., Jr. and Terry, W. J.: Clinical application of transrectal ultrasonography and prostate specific antigen in search for prostate cancer. J. Urol., 139: 758, 1988. 19. Yang, N.: Endogenous antibody to prostate-specific antigen in women. Letter to the Editor. Clin. Chem., 34: 647, 1988. 20. Ahmann, F. R. and Schifman, R. B.: Prospective comparison between serum monoclonal prostate specific antigen and acid phosphatase measurements in metastatic prostatic cancer. J. Urol., 137: 431, 1987.
THE JOURNAL OF UROLOGY
Printed in U.S.A.
Copyright© 1989 by AMERICAN UROLOGICAL ASSOCIATION, INC.
CLINICAL USE OF PROSTATE SPECIFIC ANTIGEN IN PATIENTS WITH PROSTATE CANCER M'LISS A. HUDSON,* ROBERT R. BAHNSON
t,+ AND
WILLIAM J. CATALONA§
From the Division of Urologic Surgery, Washington University School of Medicine, St. Louis, Missouri
ABSTRACT
The clinical use of prostate specific antigen as a screening test for prostate cancer, as a preoperative determinant for staging of prostate cancer and to monitor response to therapy in prostatic cancer patients was evaluated in 168 men with benign prostatic hyperplasia and 231 men with prostate cancer. Only 3% of the men with benign prostatic hyperplasia had prostate specific antigen levels greater than 10 ng. per ml. compared to 44% of the men with proved prostate cancer. Preoperative prostate specific antigen levels increased with higher clinical stages of prostate cancer but there was substantial overlap among stages. Among patients with stage Al prostate cancer who were followed expectantly none had an elevated prostate specific antigen value or metastatic disease during a followup of 15 to 120 months. After radical prostatectomy serum prostate specific antigen values decreased to undetectable levels (less than 0.6 ng. per ml.) in 89% of the patients with organconfined disease, in 87% of those with microscopically positive margins only but in only 34% with seminal vesicles or lymph node involvement. Failure of the prostate specific antigen levels to decrease to the undetectable range after radical prostatectomy was associated with a greater likelihood of subsequent tumor recurrence. Only 3 of 18 patients (17%) treated with definitive radiation therapy had post-irradiation prostate specific antigen values of less than 0.6 ng. per ml., while in 39% the prostate specific antigens values remained greater than 4 ng. per ml. and in 4 of 18 (22%) the values were greater than 10 ng. per ml. Of patients with previously untreated stage D2 prostate cancer the mean pre-treatment prostate specific antigen value was 63.7 ng. per ml. compared to a post-hormonal therapy mean value of 31.1 ng. per ml. Of 32 patients treated with hormonal therapy 14 had stable disease, including 13 with prostate specific antigen levels of less than 10 ng. per ml. In contrast, 18 patients had progressive disease, of whom 16 had prostate specific antigen levels of more than 10 ng. per ml. We conclude that the serum prostate specific antigen assay is most useful clinically to monitor the response to therapy of prostate cancer patients. (J. Urol., 142: 1011-1017, 1989) Prostate specific antigen (PSA) is a glycoprotein with a molecular weight of approximately 35,000 daltons. It was described by Wang and associates in 1981 as an antigen associated only with prostatic tissue. 1 PSA has been localized to the cytoplasm of prostatic acinar cells, ductal epithelium and sem inal fluid. 2 PSA has not been found in any other normal human tissue nor is it associated with any malignancies other than prostate cancer. 2 • 3 Elevated serum levels of PSA have been associated with benign prostatic hyperplasia and prostatic carcinoma. Characterization studies have shown that PSA is composed of a single polypeptide chain of 240 amino acids. 4 ' 5 It is distinct biochemically and immunologically from prostatic acid phosphatase. PSA appears to be similar in composition to the serine proteases of the kallikrein family. Although the exact function of PSA is unknown, it has been suggested that PSA may be involved in liquefaction of the seminal coagulum.5 We assess the clinical usefulness of PSA as a serum tumor marker for diagnosis, staging, and monitoring of response to
therapy and disease progression in patients with known or suspected prostate cancer. METHODS
PSA assay. Serum PSA levels were determined using the Tandem-R PSAII immunoradiometric assay. 6 The normal range for serum PSA levefs recommended by the manufacturer is 0 to 4 ng,/ml. and the lower limit of detection is stated to be 0.15 ng,/ml. To set the lower limit of detection of this assay the manufacturer recommended the use of pooled female sera as a zero control. At our laboratory 10 pooled female sera duplicates were obtained from female volunteers for PSA controls and the mean PSA level plus or minus 3 standard deviations was determined. We also evaluated the lower limit of detection of PSA in this assay by performing serial dilutions of a known concentration (2 ng,/ml.) of PSA with pooled female sera or Hybritech Zero Diluent in 6 separate experiments. Patients. Serum PSA levels were determined in 168 men with benign prostatic hyperplasia, of whom 90 had histological con. firmation of adenomatous hyperplasia and 78 had bladder outlet obstructive symptoms with an enlarged, nonindurated prostate gland on digital rectal examination. A total of 14 men undergoing cystoprostatectomy for invasive bladder cancer also served as controls. Serum PSA levels were obtained preoperatively from 107 men with biopsy proved prostate cancer, organ confined tumors on rectal examination, and a negative bone
Accepted for publication April 5, 1989. * Recipient of American Urological Association American Foundation for Urologic Disease Scholarship 1988-1990. t Recipient of American Cancer Society Clinical Oncology Career Development Award 86-48. t Current address: Division of Urologic Surgery, University of Pittsburgh School of Medicine, Presbyterian University Hospital, Pittsburgh, Pennsylvania 15213. . § Requests for reprints: Division of Urologic Surgery, Washington University School of Medicine, 4960 Audubon Ave., St. Louis, Missouri 63110. 1011
II Hybritech, Inc., San Diego, California.
1012
HUDSON, BAHNSON AND CATALONA
scan and computerized tomography (CT) scan of the abdomen. PSA values were available in 174 men after radical prostatectomy. Of these patients 93 had a PSA level obtained within 6 months postoperatively, while in 81 the initial postoperative PSA values were obtained more than 6 months after radical prostatectomy, with a mean of 27.8 months (range 7 months to 11 years). Post-radiation therapy PSA levels also were determined in 18 patients who were irradiated for histologically proved clinically localized prostate cancer. Mean followup in these patients was 17.4 months (range 9 to 59 months). Of these 18 patients 7 had PSA levels obtained within 6 months of completion of radiation therapy. PSA values were available in 39 men with known metatastic prostate cancer. Mean followup in these patients was 26.5 months (range 8 to 48 months). Staging. Clinical staging of prostatic cancer was defined with a modification of the Whitmore-Jewett system: stage A-tumors not suspected on digital rectal examination but found on histological examination after simple prostatectomy (stage Al-well differentiated tumors confined to less than 5% of the specimen and stage A2-moderately or poorly differentiated tumors, or those with greater than 5% involvement of the specimen), stage B-tumors palpable on digital rectal examination and confined to the prostatic capsule (stage Bl-tumor limited to 1 lobe of the prostate and less than 2 cm. in diameter and stage B2-lesions involving both prostatic lobes or being greater than 2 cm. in diameter), clinical stage C-tumors that have extended beyond the prostatic capsule but without evidence of distant metastases and stage D2-patients with clinical evidence of distant metastases. Pathological staging for prostate cancer was defined by the following system. Pathological stages Al, A2, Bl and B2 tumors were defined as in the clinical staging system but were verified histologically. Organ-confined disease was defined as tumors that on histological examination showed neither evidence of extension through the prostatic capsule, extension to the bladder neck, urethral margins, fascia surrounding the seminal vesicles nor evidence of pelvic lymph node metastases. Microscopically positive margins on pathological examination (pathological stage Cl) refers to tumor at either the urethral or bladder neck margin, or extension through the prostatic capsule to the surgical margin but without seminal vesicle or pelvic lymph node involvement. Pathological stage C2 disease refers to patients with involvement of the seminal vesicles or fascia surrounding the seminal vesicles. Pathological stage Dl refers to patients with clinical stage A or B disease who at pelvic lymphadenectomy had histologically documented lymph node metastases. Extracapsular refers to patients with pathological stage C or D disease. All patients with local extracapsular tumor extension had either positive margins or seminal vesicle invasion. Tumor penetration through the prostatic capsule with negative surgical margins was not observed except in the area within the fascia surrounding the seminal vesicles, in which case it was considered as seminal vesicle invasion. In these patients there were no instances of capsular penetration of tumor covered by periprostatic fat or a neurovascular bundle with a negative surgical margin. Treatment options based on clinical stage included observation for patients with clinical stage Al disease who were more than 65 years old or had other serious medical problems, nervesparing radical prostatectomy for patients with stages Al, A2, Bl or B2 disease who had a life expectancy of 10 years and were in good health, external beam radiation therapy for patients with stage C prostate cancer or those with stages Al, A2, Bl or B2 disease who did not wish to undergo radical prostatectomy, and early or delayed hormonal therapy for patients with metastatic disease.
of detection of 0.2 ng./ml. and results were reported accordingly. In March 1988, because of observed fluctuations in repeated PSA assays in the low ranges of detectability, we conducted studies in which 3 PSA samples containing approximately 2.0 ng./ml. PSA were serially diluted with pooled female sera (fig. 1, A). PSA concentrations of O to 0.6 ng./ml. could not be distinguished from the zero control. Therefore, at that time our laboratory set the lower limit of detection for the assay as 0.6 ng./ml. Accordingly, much of the data reported in our study were obtained when the lowest possible value reported was less than 0.6 ng./ml. We performed subsequent studies in which the mean PSA value plus or minus 3 standard deviations for pooled female sera was determined. The mean value was 0.189 ± 0.318 ng./ml., and 3 standard deviations above the mean was 0.507 ng./ml., which is considerably higher than the 0.15 ng./ml. value reported by the manufacturer. More recently, we repeated the serial dilution studies after ·revision of the bead washing procedure and using the Hybritech Zero Diluent (pooled female serum treated to remove any PSA binding proteins). In these later studies the lower limit of detectability was demonstrated to be between 0.3 and 0.4 ng./ml. (fig. 1, B). For the purposes of this study the lower limit of detectability was 0.6 ng./ml. Benign prostatic hyperplasia patients. Of 168 men with benign prostatic hyperplasia (35 of 168) 21 % had PSA levels above the manufacturer's recommended upper limit of normal of 4 ng./ ml. (fig. 2). Only 3 patients (2%) had PSA levels greater than 10 ng./ml. PSA after cystoprostatectomy for bladder cancer. All 14 patients undergoing radical cystoprostatectomy for bladder cancer had postoperative PSA levels that decreased to the undetectable range. PSA and clinical stage. Of patients with clinical stage A or B prostate cancer 62% (64 of 103) had preoperative PSA values greater than 4 ng./ml. compared to 72% (31 of 43) of those with clinical stage C or D disease. Of these 146 men with 3.0
~;
Pooled Female Sero 2.5
• ••••___ ••• , .a .t:t···
2.0 1.5 1.0 0.5
E
~ C
3.0
~
/ ..a··-.,. ----~.,,,,,.:,:·...
PSA assay. In our initial studies with the Tandem-R PSA assay we adopted the manufacturer's recommended lower limit
p'
<~1;. •· ---:e...···-~··' ".
A
•••
~~~--
'f-
I
I
Hybritech Zero Diluent
ct
U)
a.. 2.5 2.0 1.5 1.0 0.5
O
10
20
30
40
50
% PSA from RESULTS
.,., ·······
H
60
70
80
90
100
2 ng/ml
FIG. 1. Serial dilution of 3 PSA samples from known concentration of approximately 2 ng./ml. to lower limit of detection. A, diluent used was pooled female sera. B, after revision of bead washing procedure. Diluent used was Hybritech zero calibrator.
1013
CLU\TICAL tJSE OF PROSTA. TE SPECIFIC Al'>.JT1GEN IN PATIEl\ITS \VY.CH PROSTATE CAI~CER
100
••
iOO
e
• •
0
•
0
I •
O@
c
"®
IE
10.0
':?100
0
It:
--·
8"---,--
4.0
.Q
..,
OJ
H
"u
4.0
Gli!l~{j@
0.6
0e.@ee00
••
u 00 Q.
•
•• •"
0
:.
•
0 C
..
.•
H"
.. i•v..
@P
$~0"M0G~
H@
H
lb
LO
2
H
--;:-= • ...
• ©
1.0
• •
0 EUH)@$0®
06 01
BPH
A1
A2
81
62
C
D2
Clinictll Stage
Fw. 2. Preoperative PSA values in patients with benign prostatic hyperplasia (BPH), clinically localized prostate cancer or metastatic prostatic cancer.
prostate cancer 44% had PSA levels greater than 10 ngJml., including 36% (37 of 103) with clinical stages A or B disease (fig. 2).
There was a trend of increasing PSA levels with higher clinical stage of prostate cancer but considerable overlap was observed among the clinical stages. PSA and pathological stage. Preoperative PSA values could not be used to distinguish between patients found histologically to have organ-confined prostate cancer and those with extracapsular tumor extension (fig. 3). Higher PSA values did not always indicate extracapsular extension and lower levels did not always ensure that tumor was confined within the prostatic capsule. PSA radical prostatectomy. As mentioned previously, interval PSA levels of less than 0.6 ng./mL were eov.,w.,uc,g,u undetectable. By this criterion after radical prostatectomy 89% of the patients with organ-confined tumors had PSA levels that decreased to the undetectable range. Similarly, postoperative PSA levels decreased to less than 0.6 ng./ ml. in 87% of the patients with microscopically positive margins (capsular penetration with a positive margin, or tumor at the methral or bladder neck margin but with negative seminal vesicles) following radical prostatectomy (fig. 4). In patients with seminal vesicle (pathological stage C2) or lymph node (pathological stage Dl) involvement the results were markedly different. Only 34 % of the patients had PSA levels that decreased to less than 0.6 ng./ml. following radical prostatectomy. Of the patients with seminal vesicle or lymph node involvement 45% had postoperative PSA values greater than the upper limit of 4 ng./ml. and 7 of these 29 patients (24 %) had PSA values of more than 10 ng./ml. PSA and tumor recurrence. Of the post-radicalprostatectomy patients with PSA levels of greater than 10 ng./ml. 80% had recurrent disease. Conversely, to date only 1 patient with undetectable PSA values postoperatively has had recurrent disease (a local recurrence at the urethral anastomosis) following radical prostatectomy (fig. 4). The other 5 patients with
01-L-------------------Organ Confined
Extrocopsulor E)(lension
Pathologic Stage FIG. 3. Comparison of preoperative PSA values in patients with pathologically localized prostate cancer and those with extracapsular tumor extension on pathological examination of tumor. Lines indicate mean PSA values.
100
if
"'.,if
.,.
lO.O
E
'
0,
<::
~ U)
"
i
4.0
... .."°"
01.
...... IIHi!181lUlli
ef
;•
.,. 2
1.0 06
...
•=• awn, Bm!!ffl ~r
...
.,,_,
·-
v = Recurrenl disease
0.1 - - ' - - - - - , - - - - , - - - , - - - r - - - - - r - - - - - , - - - ~ -
C2 (+margins)
Pathologic Sto11e
(+SV)
D1 (+pelvic
nodu)
FIG. 4. Postoperative PSA values in relationship to postoperative pathological stage. Line attached to dot indicates patients who have recurrent tumor.
1014
HUDSON, BAHNSON AND CATALONA
recurrent disease shown as having undetectable PSA values had received adjuvant radiation or endocrine therapy for the recurrence before the PSA value was obtained. No patient with an undetectable PSA level following radical prostatectomy has had metastatic disease. PSA as an indicator of incomplete tumor excision. PSA levels were measured within 6 months after radical prostatectomy in 93 ofl 74 patients in this study. Of these 93 patients 4 7 had an undetectable PSA level (less than 0.6 ng./ml.) postoperatively and are without evidence of recurrent disease but they have been followed for less than 1 year, while 32 had an undetectable PSA value (less than 0.6 ng./ml.), are without evidence of recurrent disease and have been followed at least 12 months but not longer than 33 months. Seven of the remaining 14 patients had a PSA level of greater than 0.6 ng./ml., including 3 (43%) who had recurrent disease at less than 1 year, while 7 initially had an undetectable PSA value postoperatively that increased to greater than 0.6 ng./ml. during year 1 of followup. To date 1 of the latter 7 patients (14%) had a documented recurrence at 8 months. PSA as an indicator of current disease status. Of 81 patients with an initial PSA value measured more than 6 months after radical prostatectomy 48 have a level of less than 0.6 ng./ml. and are without evidence of recurrent disease. Among these 48 patients 18 have been followed for more than 1 year, 18 for more than 3 years and 12 for more than 5 years. Of the remaining 33 patients 16 have a PSA level of greater than 0.6 ng./ml. but they are without evidence of recurrent disease to date. Followup in these patients is more than 1 year in 6, more than 3 years in 9 and more than 5 years in 1. Of the final 17 patients with a documented recurrence 11 had a PSA level of greater than 0.6 ng./ml. Followup in these 11 patients was greater than 1 year in 5, greater than 3 years in 5 and greater than 5 years in 1. A total of 6 patients with recurrent disease had PSA levels of less than 0.6 ng./ml. but 5 had received either adjuvant radiation or endocrine therapy before measurement of the first PSA value. Prostatic acid phosphatase levels. All patients undergoing radical prostatectomy except 1 had a preoperative prostatic acid phosphatase level in the normal range. Postoperatively, prostatic acid phosphatase values remained in the normal range for all except this same patient. The patient with an elevated prostatic acid phosphatase level before and after radical prostatectomy was believed to have an elevated level on the basis of liver disease. The postoperative PSA value was undetectable and he had no evidence of recurrent tumor after 19 months. Only 4 patients have had an elevated prostatic acid phosphatase level after radical prostatectomy and each of these patients had a concomitantly elevated PSA value above 10 ng./ml. in the immediate postoperative period. Of these 4 patients 3 have had distant metastases. PSA and stage Al cancer patients managed conservatively. No patient followed expectantly with stage Al prostate cancer has had either an elevated PSA value or metastatic disease during followup. Of these 15 patients 5 have been followed for more than 8 years, 2 for more than 4 years and 8 for less than 2 years. The natural history of stage Al prostate cancer is such that disease progression would be unlikely to occur during this interval in the majority of these patients. PSA and patients treated with radiation therapy. Only 3 of 18 patients (17%) treated with primary radiation therapy who appear by standard clinical parameters to be free of recurrent tumor had post-radiation PSA values less than 0.6 ng./ml. Of the post-radiation PSA values 39% remained above 4 ng./ml. and 4 of 18 (22%) were greater than 10 ng./ml. during a mean followup of 17.4 months (range 9 to 59 months). None of these patients has had a clinical recurrence to date during the 2 years when PSA measurements have been available at our institution. Patients who failed radiation therapy before this period are classified with the stage D2 cancer patients.
Patients treated with hormonal therapy. Of 16 patients with previously untreated stage D2 prostate cancer the mean pretreatment PSA value was 63. 7 ± 42.8 ng./ml. compared to a post-hormonal therapy mean value of 31.1 ± 14.2 ng./ml. Only 1 of 14 patients with apparent clinically inactive D2 prostate cancer as evidenced by stable changes on a bone scan, no bone pain and no voiding difficulties had a PSA level of greater than 10 ng./ml. Of 18 patients with clinically active metastatic prostate cancer treated with hormonal therapy 16 (88%) had PSA levels of greater than 10 ng./ml. These patients showed evidence of progressive disease as manifested by increasing bone pain, new changes on bone scan and/or difficulty voiding, while only 2 patients with progressive disease had PSA levels of less than 10 ng./ml. Among 9 patients in whom pre-treatment and post-treatment PSA levels were obtained the post-treatment PSA levels were obtained within 1 to 4 months after institution of hormonal therapy. Post-treatment levels decreased in 8 of the 9 patients (89%) (fig. 5). DISCUSSION
PSA as a screening test for prostate cancer. The use of serum tumor markers to detect prostate cancer has been a clinical goal since a correlation was first noted between elevated prostatic acid phosphatase levels and metastatic prostate cancer. 7 However, prostatic acid phosphatase elevations usually indicate the presence of metastatic disease and have not proved to be useful to detect localized prostate cancer. Because PSA has been shown to be more sensitive than prostatic acid phosphatase in following patients for development of progressive disease, hope was raised that PSA might prove to be a better marker for localized prostate cancer.s.-12 In most reports PSA levels were measured with 1 of 2 different assays, including a polyclonal antibody radioimmuno-
>100
PSA (ng/ml)
1.0
o Inactive disease
• Active disease
POST PRIOR TO HORMONAL THERAPY HORMONAL THERAPY FIG. 5. Pre-treatment and post-treatment PSA values in patients with clinical stage D2 prostate cancer treated with hormonal therapy. Lines connect pre-treatment and post-treatment values in same patient.
CLINICAL USE Of. PROSTATE SPECIFIC ANTIGEN. [N- PATIE:NTS ViITH PRU§Ti':,.TE CAl~ICER
Previous reports of positive PSA values in benign prostatic hyperplasia using the Tandem-R PSA immunoradiometric assay Pathologically Confirmed Benign Prostatic Hyperplasia Reference
Elevated PSA No./Total (%)
Myrtle and associates 6 • * Ferro and associates9 Ercole and associates8 Chan and associates 15 Oesterling and associates 17 Cooner and associates 18
7/352 (2) 13/40 (32.5) 10/357 (3) 0/69 (0) 7/72 (10) 8/197 (4)
PSA Value (ng./ml.) > 10 >10 ia':10 >12.4 > 10 >10
* Pathological confirmation of benign prostatic hyperplasia not stated.
assay (Pros-Check*) and a double monoclonal antibody assay (Tandem-R PSA assay). In the earliest studies investigators defined their own upper limits of normal for the assay used and based on their normal ranges, reported elevated PSA levels in ;rostate cancer patients and those with benign prostatic disease. These studies with various upper limits of normal have revealed that 20 to 86% of the patients with benign prostatic hyperplasia, and 41 to 98% wit~8 _;:';rious stages o~ pros!ate cancer have elevated PSA levels. · Each of these mvestigators concluded independently that PSA levels are not sufficiently specific to screen for prostate canc~r. . In our 168 patients with presumed bemgn prostat1c hyperplasia 21 % had a PSA level of greater than 4.0 ng./ml. as measured with the Tandem-R assay. However, only 3 of these patients (2%) had PSA levels of more than 10 ng./mL It sho1;1ld be appreciated that not all of our patients had an obstructive uropathic condition of sufficient severity to require surgical relief. Our results concur with those of Myrtle 6 and Ercole 8 and their associates who also reported that 2 to 3% of the surgically tested benign ~rostatic hyperplasia patients had preoperative PSA levels of more than 10 ng./ml. Review of previous reports confirms that PSA values measured by the Tandem-R assay in patients with pathologically confirmed benign prost~ti~ hyperplasia are less than 10 ng./ml. in the great maJonty (see table). e.s. 9, 15.17, 1s Of our patients believed to have clinically organ-confined prostate cancer, 66% had preoperative PSA levels of greater than 4.0 ng./ml. compared to 72% of those with extracapsul~r extension on rectal examination or documented metastatic disease. Over-all, of 146 men with biopsy proved prostate cancer 44% had PSA levels of greater than 10 ng./ml. (fig. 2). The observation that PSA levels of greater than 10 ng./ml. on the Tandem-R assay are associated more frequently with prostate carcinoma than benign prostatic hype_rplasia does not necessarily qualify PSA as an accurate screenmg test for prostate cancer. Statistical analyses by others have shown that the sensitivity and specificity of PSA are not sufficient to consider it an accurate screening test. 17 Nevertheless, PSA levels of greater than 10 ng./mL as measured in the Tandem-R assay are worrisome for prostate carcinoma. For example, Myrtle and associates reported that only 2% of 352 benign prostatic hyp~rplasia patients had PSA levels of more than 10 ng./ml., while PSA levels were greater than 10 ng./ml. in 30% of the stage A and 50% of the stage B prostate cancer patients. 6 In our patients 35% with clinically localized prostate cancer had a PSA value of greater than 10 ng./ml. Cooner and associates_ studied p~tients in whom the prostate gland was not sufficiently suspi cious on rectal examination to warrant performance of a biopsy and found that 10 of 18 (55 % ) with a PSA value of greater than 10 ng./ml. had prostate carcinoma. 18 It appears that additio~al studies to exclude the possibility of prostate cancer are advisable in patients with a PSA level of greater than 10 ng./ml. as measured by the Tandem-R assay. A PSA level of greater than
* Yang Laboratories, Bellvere, Washington.
1015
10 ng./mL should not be considered diagnostic of prostate cancer, since patients with a large, obstructing adenoma or prostatitis also may have PSA values above this level, Correlation between PSA value and stage of prostate cancer. In general, PSA values have been shown to increase with advancing clinical stages of prostate carcinoma. 8 • 1 2 · 14 Oesterling and associates studied 178 men with clinically confined prostate cancer undergoing radical prostatectomy and evaluated the ability of PSA to predict final pathological stage. 17 In their study preoperative PSA values were significantly higher in patients in whom the tumor penetrated through the prostatic capsule, and in those with seminal vesicle involvement and/or lymph node metastases. However, the diagnostic accuracy when one considers different upper limits of normal for PSA at 4.0, 10.0 and 16.0 ng./ml. ranged from 55 to 72% and was not sufficient to predict on an individual basis which patients had localized disease and which had extracapsular tumor extension. In our series also the mean PSA value increased with advancing clinical stage (fig. 2). Although we observed higher PSA values with more advanced stages of prostate cancer, our data also demonstrated considerable overlap in PSA values between patients with pathologically organ-confined disease and those with extracapsular tumor extension (fig. 3). Use of PSA values to monitor response to therapy in patients with prostate cancer. The greatest clinical value of PSA appears to be in monitoring response to therapy in patients undergoing treatment for prostate cancer. 13 Killian and associates suggested that PSA values may predict tumor recurrence before clinical symptoms appear in patients receiving curative therapy for organ-confined tumors. 14 Stamey and associates, using the Yang 19 assay for PSA in a series of 45 patients treated with radical prostatectomy for clinically confined prostate cancer reported that PSA routinely decreased to undetectable levels postoperatively. 10 In 6 of these patients serial postoperative PSA measurements appeared to be useful to detect residual tumor or early tumor recurrence, which was treated successfully with adjunctive radiation therapy in 5 of these 6 patients. Similarly, Chan and associates, using the Tandem-R assay, reported that in 67 of 73 patients (92%) undergoing radical prostatectomy for organ-confined tumor the PSA values decreased to undetectable levels. 15 Three patients with clinical tumor recurrence had PSA values above the normal reference range. The lower limit of detection of PSA in the Tandem-R assay is stated by the manufacturer to be 0.15 ng./ml. Special emphasis recently has been given to the lower limit of detection of the Tandem-R assay when evaluating post-radical prostatectomy patients. 8 • n. 12 · 1 5 • 17 In theory, these patients are expected to have undetectable PSA levels if all tumor has been resected. Thus, the presence of detectable PSA levels may have prognostic and therapeutic significance. However, there has been interlaboratory variation in the lower limit of detection of this assay. 11 • 12 • 15 · 17 The lower limit of detection for the Tandem-R PSA assay initially repmted by Chan and associates was 0.1 µg./1. 15 Further study by this group, cited Oesterling and associates, 17 showed a minimal detectable concentration of PSA at 0.07 ng./ml. with the Tandem-R assay. In both reports PSA decreased to undetectable levels in approximately 90% of the patients after radical prostatectomy. Stamey and associates, however, in a series of 50 post-radical prostatectomy patients who were presumed to have organ-confined disease, found that no patient had a postoperative PSA level measured with the Tandem-R assay of less than 0.2 ng./ml. and that values reported from their institution ranged from 0.4 to 1.2 ng./ml. 11 Lange and associates reported a lower limit of detection of PSA values using the Tandem-R assay as 0.5 ng./ml. 12 Our lower limit of detection has varied from 0.3 to 0.6 ng./ml. depending upon the bead washing procedure and the method used to establish the lower limit. At a recent conference on PSA sponsored by the Organ Sites
1016
HUDSON, BAHNSON AND CATALONA
Coordinating Center of the National Cancer Institute, clinical laboratory experts agreed that the most accurate means to establish the lower limit of detectability of the PSA assay is the limiting serial dilution method. For the PSA assay the diluent used should be the Zero Standard Diluent (pooled female serum with PSA binding proteins removed) rather than pooled female sera because pooled female sera may contain anti-PSA antibodies that could interfere with the assay. The other method to determine the lower limit of detectability involves establishing the mean plus 3 standard deviations of the zero standard. This latter method was stated to be somewhat less accurate than the dilution method but more commonly used in commercial assays. Early in our studies of PSA we adopted the manufacturer's recommendation for the lower limit of detectability of 0.2 ng./ ml. Later, when performing our own studies of limiting serial dilutions of pooled female sera, we found that values of less than 0.6 ng./ml. could not be distinguished from zero and, thus, for our patients the lower limit of detectability was considered to be less than 0.6 ng./ml. We have performed studies measuring the mean plus 3 standard deviations of pooled female sera and found the lower limit of detectability by this method to be 0.5 ng./ml. Finally, after consultation with the manufacturer, who advised a bead washing procedure using a commercial apparatus (Hybriwash), the serial dilution sensitivity experiment was repeated using the Zero Diluent from the Tandem-R PSA kit. An improvement in the lower limits of detectability of the assay to 0.3 to 0.4 ng./ml. PSA was noted. It should be emphasized that adherence to the specified wash procedure and performance of the procedure by the same medical technologist were important to maintain the precision and sensitivity of the assay. These results underscore the fact that minor technical details in the performance of the assay can account for appreciable differences in the lowest range of detectability. Accordingly, it seems advisable for each laboratory to establish its own lower limit of detectability. Important treatment decisions should not be based upon single assay results. Rather, a trend demonstrating a persistent elevation or an increase in serial PSA values provides a more meaningful basis for clinical decisions. In our series 89% of the patients with pathologically confirmed, organ-confined prostate cancer had PSA that decreased to undetectable levels (less than 0.6 ng./ml.). Of our patients with microscopically positive margins (pathological stage Cl) 87% also had undetectable PSA values postoperatively. None of these patients had clinical evidence of residual or recurrent tumor, although the followup is limited at 27.8 ± 18.1 months. Three patients with stage Cl disease had a local recurrence before a PSA value was obtained. These patients were treated with adjuvant radiation therapy, and they subsequently had an undetectable PSA value and are without evidence of disease to date. In contrast, PSA decreased to undetectable levels in only 34 % of our patients with seminal vesicle or lymph node involvement. Of these patients 45% had PSA levels of greater than 4.0 ng./ml. and 7 of 9 who had metastatic disease had PSA values of greater than 10 ng./ml. The other 2 patients had a value of more than 4 ng./ml. Of 18 patients with elevated PSA values 3 (17%) also had elevated prostatic acid phosphatase levels and 9 (50%) have had recurrent cancer to date. Thus, failure of postoperative PSA levels to decrease into the undetectable range after radical prostatectomy was associated with a greater likelihood of subsequent tumor recurrence. Among 108 patients treated with radiation therapy Stamey and associates10 found that PSA decreased to undetectable levels in only 13% and in only 35% did PSA levels return to the normal reference range. In our series 3 of 18 patients (17%) treated with radiation therapy who appear by standard clinical parameters to be free of recurrent tumor had a PSA value in the undetectable range and 7 of 18 (39%) had PSA levels of greater than 4 ng./ml. The significance of serial PSA values in
patients who receive primary radiation therapy requires further clarification but elevated PSA values are worrisome. Pre-treatment and post-treatment PSA values also have proved to be clinically useful in patients with known metastatic prostate cancer treated with hormonal therapy. 13 Ftirro and associates,9 and Ahmann and Schifman20 found PSA values to be more sensitive than prostatic acid phosphatase levels to monitor disease progression in patients with prostate cancer. Our data confirm that patients in clinical remission with hormonal therapy often showed a decrease in PSA values to the normal range, while those with further progression of disease after hormonal therapy again had markedly elevated PSA levels (fig. 5). A group of patients not often commented upon in the literature are those with stage Al prostate carcinoma who are managed expectantly. In our series among 15 such patients PSA values ranged from 0.6 to 3.5 ng./ml. With followup intervals of 15 to 120 months no patient in this group had an elevated PSA value nor has any patient had evidence of metastatic disease. Serial determination of PSA values in these patients also may prove to be clinically useful. Further clinical studies with long-term followup in larger numbers of patients will be required to define more completely the role of PSA as a tumor marker in prostate cancer patients.
REFERENCES
1. Wang, M. C., Papsidero, L. D., Kuriyama, M., Valenzuela, L.A., Murphy, G. P. and Chu, F. M.: Prostate antigen: a new potential marker for prostatic cancer. Prostate, 2: 89, 1981. 2. Nadji, M., Tabei, S. Z., Castro, A., Chu, T. M., Murphy, G. P., Wang, M. C. and Morales, A. R.: Prostate-specific antigen: an immunohistologic marker for prostatic neoplasms. Cancer, 48: 1229, 1981. 3. Kuriyama, M., Wang, M. C., Papsidero, L. D., Killian, C. S., Shimaro, T., Valenzuela, L., Nishiura, T., Murphy, G. P. and Chu, T. M.: Quantitation of prostate-specific antigen in serum by a sensitive enzyme immunoassay. Cancer Res., 40: 4658, 1980. 4. Watt, K. W. K., Lee, P. J., M'Timkulu, T., Chan, W. P. and Loor, R.: Human prostate-specific antigen: structural and functional similarity with serine proteases. Proc. Natl. Acad. Sci., 83: 3166, 1986. 5. Lilja, H.: A kallikrein-like serine protease in prostatic fluid cleaves the predominant seminal vesicle protein. J. Clin. Invest., 76: 1899, 1985. 6. Myrtle, J. F., Klimley, P. G., Ivor, L. P. and Bruni, j, F.: Clinical utility of prostate specific antigen (PSA) in the management of prostate cancer. Adv. Cancer Diagnostics, p. 1, 1986. 7. Gutman, A. B. and Gutman, E. B.: An "acid" phosphatase occurring in the serum of patients with metastasizing carcinoma of the prostate gland. J. Clin. Invest., 17: 473, 1938. 8. Ercole, C. J., Lange, P. H., Mathisen, M., Chiou, R. K., Reddy, P. K. and Vessella, R. L.: Prostatic specific antigen and prostatic acid phosphatase in the monitoring and staging of patients with prostatic cancer. J. Urol., 138: 1181, 1987. 9. Ferro, M. A., Barnes, I., Roberts, J. B. M. and Smith, P. J. B.: Tumour markers in prostatic carcinoma. A comparison of prostate-specific antigen with acid phosphatase. Brit. J. Urol., 60: 69, 1987. 10. Stamey, T. A., Yang, N., Hay, A. R., McNeal, J. E., Freiha, F. S. and Redwine, E.: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. New Engl. J. Med., 317: 909, 1987. 11. Stamey, T. A.: Editorial comment. J. Urol., 139: 764, 1988. 12. Lange, P. H., Ercole, C. J. and Vessella, R. L.: Serum prostatic specific antigen (PSA) in the management of patients after radical prostatectomy (RP). J. Urol., part 2, 139: 314A, abstract 607, 1988. 13. Kuriyama, M., Wang, M. C., Lee, C. L., Killian, C. S., Papsidero, L. D., Inaji, H., Loor, R. M., Lin, M. F., Nishiura, T., Slack, N. H., Murphy, G. P. and Chu, T. M.: Multiple marker evaluation
CLI!"'JICAL ·usE O:F' PROSTATE SPECIFIC AJ:¾JT!GEN I:t~ PAT!Ef'JTS V/ITH PROSTArI'E CAI"'1CER
in human prostate cancer with the use of tissue-specific antigens, J. NatL Cancer Inst., 68: 99, 1982. 14. Killian, C. S., Yang, N., Emrich, L. J., Vargas, F. P., Kuriyama, M., Wang, M. C., Slack, N. H., Papsidero, L. C., Murphy, G. P., Chu, T. M. and the Investigators of the National Prostatic Cancer Project: Prognostic importance of prostate-specific antigen for monitoring patients with stages B2 to Dl prostate cancer. Cancer Res., 45: 886, 1985. 15. Chan, D. W., Bruzek, D. J., Oesterling, J. E., Rock, R. C. and Walsh, P. C.: Prostate-specific antigen as a marker for prostatic cancer: a monoclonal and polyclonal immunoassay compared. Clin. Chem., 33: 1916, 1987. 16. Wang, T. Y. and Kawaguchi, T. P.: Preliminary evaluation of measurement of serum prostate-specific antigen level in detection of prostate cancer. Ann. Clin. Lab. Sci., 16: 461, 1986.
1017
17. Oesterling, J. E., Chan, D. W., Epstein, J. I., Kimball, A. W., Jr., Bruzek, D. J., Rock, R. C., Brendler, C. B. and Walsh, P. C.: Prostate specific antigen in the preoperative and postoperative evaluation of localized prostatic cancer treated with radical prostatectomy. J. Urol., 139: 766, 1988. 18. Cooner, W. H., Mosley, B. R., Rutherford, C. L., Jr., Beard, J. H., Pond, H. S., Bass, R. B., Jr. and Terry, W. J.: Clinical application of transrectal ultrasonography and prostate specific antigen in search for prostate cancer. J. Urol., 139: 758, 1988. 19. Yang, N.: Endogenous antibody to prostate-specific antigen in women. Letter to the Editor. Clin. Chem., 34: 647, 1988. 20. Ahmann, F. R. and Schifman, R. B.: Prospective comparison between serum monoclonal prostate specific antigen and acid phosphatase measurements in metastatic prostatic cancer. J. Urol., 137: 431, 1987.