© 1990 S. Kargcr AG, Basel 0042-1138/90/0453-0153S2.75/0
Urol Int 1990;45:153-159
Prostate-Specific Antigen in Prostatic Carcinoma Carlos D. Jurincic, Harts U. Pixberg, Alexander Gasser. Karl-Friedrich Klippel Departments of Urology and Nuclear Medicine, General Hospital Celle, FRG
Key Words. Prostate-specific antigen • Prostatic carcinoma
Introduction Carcinoma of the prostate is the most common form of cancer in men over 65 years of age [1], and it is the second most frequent cause of cancer death in the Fed eral Republic of Germany. There are approximately 66,000 new cases of carcinoma of the prostate diagnosed each year, with 22,000 deaths annually in the USA [1], The incidence of prostatic cancer is steadily rising in Western industrial nations. We should emphasize the fact that the hospital-based clinical incidence of the whole population is expected to be far lower than the real incidence. The cell type is usually adenocarcinoma. Un fortunately an advanced stage of disease with dissemi nated metastases is diagnosed in the majority of patients. The different stages of cancer growth are correlated with the TNM classification. In the following we used two of the most common clinical classification systems to try to correlate the stage of disease and appropriate treatment, the American Uro logical Staging System and the TNM Pretreatment Clin ical Classification System. These systems of prostate cancer stages are not readily interconvertible, though this is often tried in order to compare results from insti
tution to institution and to give recommendations for adequate treatment at each stage of disease. As a result, reliable methods for early tumor detection and screening are sought. Besides digital rectal examina tion and biopsy of suspicious regions, the use of biologi cal markers has recently become of more clinical inter est. The ideal tumor marker in carcinoma of the prostate should give no or very few false-positive results among cases of benign prostatic hypertrophy (BPH), reflect tumor burden reliably, and should be a guide to appro priate treatment and follow-up so that growth of the neo plasm can be predicted before clinical evidence. Among the prostate-specific markers (table 1), pros tatic acid phosphatase (PAP), is known as a phosphohydrolase present in large quantities and is an exocrine secretion of the normal prostate. It has with an isoelec tric point of 4.4-5.5 and functions as a hydrolyzer of phosphorylcholine in human semen [2], PAP has been identified as a glycoprotein with a molecular weight of 100,000 daltons. Sites of possible enzymatic reactions have been determined and amino acid analysis has been performed [3]. With the development of antibodies to PAP, clinical use (EIA, RIA) of this test first found broad
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/6/2018 3:15:29 AM
Abstract. To evaluate the clinical and prognostic value of prostate-specific antigen (PSA) for the detection of tumor and tumor growth after therapy, 520 sera from 246 patients with prostatic carcinoma, 990 sera from patients with BPH, and 1,488 sera from patients with other urological diseases were analyzed. The values ranged from 0.1 to 1,828.9 ng/ml. 51 % of all values were about 2.5 ng/ml, and 76.8% of all values about 10 ng/ml. The commercial recommendation for the cutoff values is 2.5 ng/ml (IBL, FRG). In patients with benign prostatic hypertrophy this cutoff means 61% false-positive results, which makes the test highly sensitive but unspecific. In prostatic carcinoma patients this bor derline means a false-negative result in 9.75% (24 of 246). By determinating the cutoff at 10 ng/ml in our series, a false-negative result appeared in 14.6%. Therefore a plea is made for the 10-ng/ml cutoff. In follow-up studies a marked decline in PSA values after transurethral resection or antiandrogen therapy (orchiectomy/Zoladex® /ICI/flutamide, Essex). Generally, the greater the PSA levels the more advanced the stage of disease. These data suggest that PSA may be a useful adjuvant marker for monitoring tumor growth in patients with regionally confined tumor.
Table 1. Biological markers in prostate carcinoma Prostate-specific markers Prostatic acid phosphatase (PAP) Prostatic-specific antigen (PSA) Nonspecific markers of prostate cancer Alkaline phosphatase (AP) Polyamines Creatine kinase (CK-BB) Lactate dehydrogenase (LDH isoenzyme)
Other markers CEA Ribonuclease Prostacyclin
response, but further experience with this marker has shown that general measurements for screening lead to an inadequately high number of false-positive results, with most of the really positive results proving to be stage-C or D carcinoma [4], The second specific marker is prostate-specific anti gen (PSA), which will be presented later on. Recently, Raynor et al. [5] presented another specific antigen of the prostate, called K.R-P8, which is different to PAP and PSA and localized within secretory vacuoles of prostatic cells with a possible role in tumor diagnosis and growth control. Further laboratory and clinical stud ies must be made until this marker is of clinical use. Among the nonspecific markers, there is alkaline phosphatase (AP), an enzyme with several isoenzymes. Since Bodansky [6] developed measurement procedures for the serum determination of AP in 1933, the enzyme level in serum was found to correlate well with the devel opment of bone metastases. In patients with ossary filialization, 91 % were generally found to have elevated AP values. Lately, levels of polyamines in the erythrocytes of a group of patients [7] with metastatic prostatic carci noma were found to reflect the tumor burden well and correlated positively with the response to therapy or pre dicted progression. As polyamines have been shown to be elevated in the serum and urine of patients with a variety of malignant diseases, their value is limited because of a lack of specificity [8]. Feld and Witte [9] reported the presence of creatine kinase-B B-isoenzyme in patients with prostatic carcino ma, noting that this isoenzyme is rarely found elevated in stage-D disease. Possible elevation of this enzyme is associated with tumor extension in other tissues, thus lacking specificity and selectivity. Denis and Prout [10] first measured the lactic dehy drogenase (LDH) isoenzyme in carcinoma of the pros tate. While several authors report special LDH ratio (LDH 5/1) patterns, which should be significant for PCA
[11], others found this test to be of no value in the initial clinical diagnosis [12]. Among other markers, carcinoembryonic antigen (CEA), ribonuclease, and prostacyclin should be men tioned. Guinan et al. [13] found no correlation of the increase or decrease of CEA with the stage of PCA. Recently, elevated levels of ribonuclease activity were found in 70% of patients with prostatic carcinoma by Chu et al. [14], Khan et al. [15] showed that persistently high levels of prostacyclin were associated with a poor prognosis of prostatic carcinoma. In 1979, Wang et al. [16] reported the isolation and purification of an antigen of prostatic tissue, found to have a molecular weight of about 33,000-36,000 daltons without subunits. Its carbohydrate portion is about 77%, the isoelectric point at 6.9. It shows mild activity as a protease at neutral pH, distinct from others known in physicochemical and catalytic properties [17]. The im munological characteristics are quite different from PAP or its subunits, therefore showing no cross-reaction. Hu man prostatic tissues contain microgram quantities of PSA per milligram protein with localization of this anti gen confined to the cytoplasm of prostate acinar and ductal epithelium. Significant amounts of PSA can be found in patients with BPH and prostatic carcinoma, primary or metastatic, while other tissues of non-pros tatic origin, either normal or cancerous, contained no detectable amount of PSA. Neither was PSA found in females [ 18]. Therefore, PSA seems to be cell-type and organ-site specific, i.e. a histiotypic product. Its biologi cal function is not yet known. Subsequently, PSA anti bodies have been produced in goats and rabbits, and immunoperoxidase techniques for the identification of PSA in tissues have been developed [19] establishing immunostaining methods and thus indicating the pros tatic origin of these tumors. To evaluate its potential value as a diagnostic/prognostic histological marker, several studies were made to demonstrate: (1) the presence of PSA in prostatic cancer, and (2) the correlation of staining intensity with the degree of tumor differentiation [20], Papsidero et al. [19] evaluated a large variety of dif ferent histological sections from different organs and malignancies, but found no positive results. Epstein et al. [21] reported a statistically significant correlation be tween the foci of poor immunoreactivity and progression of disease, establishing PSA as a valuable adjunct to histopathological grading, while Vernon and Williams [22] proved that the effects of therapy on immunostaining showed no significant difference before and after treat-
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/6/2018 3:15:29 AM
Jurincic/Pixberg/Gasser/Klippel
154
Prostate-Specific Antigen in Prostatic Carcinoma
155
ment. Variations of staining intensity were observed, but general staining remained. With the more recent development of monoclonal PSA antibodies, serum PSA determination for wide clin ical practice became possible. Wang et al. [ 16] confirmed the significance of PSA as a biological marker with regard to the diagnosis and prognosis by using an ELISA technique. They found markedly elevated PSA values in 68% of BPH, 63% of stage-A, 79% stage-B, 77% stage-C, and 86-92% in stage-D prostatic carcinoma. Similar results were obtained in a study by Kuriyama et al. [23] who showed a positive correlation of declining PSA val ues in response to treatment as well as an inverse rela tionship between pretreatment levels of PSA and patient survival time. Pontes et al. [24] found a good correlation between mean serum PSA levels and the surgical stage of disease, demonstrating PSA as a good indicator and predictor of disease progression.
Patients and Methods Fig. 1. Pretreatment PSA value distribution for all patients. Fig. 2. Distribution of pretreatment PSA values for patients with BPH and rate of false-positive results at different cutoff levels.
able at 2-8 °C; (4) sensitivity 0.1-0.2 ng/ml. and (5) specificity is high using monoclonal rabbit antibodies without cross-reaction to other enzymes.
Results These patients were found to have: prostatic cancer in 246 cases, the remaining 274 determinations were serial samples; BPH in 990 cases, and other benign or malig nant urological diseases in 1,488 cases. For all these patients, the PSA value ranged between
Urol Int 1990;45:153-159
Prostate-Specific Antigen in Prostatic Carcinoma Carlos D. Jurincic, Harts U. Pixberg, Alexander Gasser. Karl-Friedrich Klippel Departments of Urology and Nuclear Medicine, General Hospital Celle, FRG
Key Words. Prostate-specific antigen • Prostatic carcinoma
Introduction Carcinoma of the prostate is the most common form of cancer in men over 65 years of age [1], and it is the second most frequent cause of cancer death in the Fed eral Republic of Germany. There are approximately 66,000 new cases of carcinoma of the prostate diagnosed each year, with 22,000 deaths annually in the USA [1], The incidence of prostatic cancer is steadily rising in Western industrial nations. We should emphasize the fact that the hospital-based clinical incidence of the whole population is expected to be far lower than the real incidence. The cell type is usually adenocarcinoma. Un fortunately an advanced stage of disease with dissemi nated metastases is diagnosed in the majority of patients. The different stages of cancer growth are correlated with the TNM classification. In the following we used two of the most common clinical classification systems to try to correlate the stage of disease and appropriate treatment, the American Uro logical Staging System and the TNM Pretreatment Clin ical Classification System. These systems of prostate cancer stages are not readily interconvertible, though this is often tried in order to compare results from insti
tution to institution and to give recommendations for adequate treatment at each stage of disease. As a result, reliable methods for early tumor detection and screening are sought. Besides digital rectal examina tion and biopsy of suspicious regions, the use of biologi cal markers has recently become of more clinical inter est. The ideal tumor marker in carcinoma of the prostate should give no or very few false-positive results among cases of benign prostatic hypertrophy (BPH), reflect tumor burden reliably, and should be a guide to appro priate treatment and follow-up so that growth of the neo plasm can be predicted before clinical evidence. Among the prostate-specific markers (table 1), pros tatic acid phosphatase (PAP), is known as a phosphohydrolase present in large quantities and is an exocrine secretion of the normal prostate. It has with an isoelec tric point of 4.4-5.5 and functions as a hydrolyzer of phosphorylcholine in human semen [2], PAP has been identified as a glycoprotein with a molecular weight of 100,000 daltons. Sites of possible enzymatic reactions have been determined and amino acid analysis has been performed [3]. With the development of antibodies to PAP, clinical use (EIA, RIA) of this test first found broad
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/6/2018 3:15:29 AM
Abstract. To evaluate the clinical and prognostic value of prostate-specific antigen (PSA) for the detection of tumor and tumor growth after therapy, 520 sera from 246 patients with prostatic carcinoma, 990 sera from patients with BPH, and 1,488 sera from patients with other urological diseases were analyzed. The values ranged from 0.1 to 1,828.9 ng/ml. 51 % of all values were about 2.5 ng/ml, and 76.8% of all values about 10 ng/ml. The commercial recommendation for the cutoff values is 2.5 ng/ml (IBL, FRG). In patients with benign prostatic hypertrophy this cutoff means 61% false-positive results, which makes the test highly sensitive but unspecific. In prostatic carcinoma patients this bor derline means a false-negative result in 9.75% (24 of 246). By determinating the cutoff at 10 ng/ml in our series, a false-negative result appeared in 14.6%. Therefore a plea is made for the 10-ng/ml cutoff. In follow-up studies a marked decline in PSA values after transurethral resection or antiandrogen therapy (orchiectomy/Zoladex® /ICI/flutamide, Essex). Generally, the greater the PSA levels the more advanced the stage of disease. These data suggest that PSA may be a useful adjuvant marker for monitoring tumor growth in patients with regionally confined tumor.
Table 1. Biological markers in prostate carcinoma Prostate-specific markers Prostatic acid phosphatase (PAP) Prostatic-specific antigen (PSA) Nonspecific markers of prostate cancer Alkaline phosphatase (AP) Polyamines Creatine kinase (CK-BB) Lactate dehydrogenase (LDH isoenzyme)
Other markers CEA Ribonuclease Prostacyclin
response, but further experience with this marker has shown that general measurements for screening lead to an inadequately high number of false-positive results, with most of the really positive results proving to be stage-C or D carcinoma [4], The second specific marker is prostate-specific anti gen (PSA), which will be presented later on. Recently, Raynor et al. [5] presented another specific antigen of the prostate, called K.R-P8, which is different to PAP and PSA and localized within secretory vacuoles of prostatic cells with a possible role in tumor diagnosis and growth control. Further laboratory and clinical stud ies must be made until this marker is of clinical use. Among the nonspecific markers, there is alkaline phosphatase (AP), an enzyme with several isoenzymes. Since Bodansky [6] developed measurement procedures for the serum determination of AP in 1933, the enzyme level in serum was found to correlate well with the devel opment of bone metastases. In patients with ossary filialization, 91 % were generally found to have elevated AP values. Lately, levels of polyamines in the erythrocytes of a group of patients [7] with metastatic prostatic carci noma were found to reflect the tumor burden well and correlated positively with the response to therapy or pre dicted progression. As polyamines have been shown to be elevated in the serum and urine of patients with a variety of malignant diseases, their value is limited because of a lack of specificity [8]. Feld and Witte [9] reported the presence of creatine kinase-B B-isoenzyme in patients with prostatic carcino ma, noting that this isoenzyme is rarely found elevated in stage-D disease. Possible elevation of this enzyme is associated with tumor extension in other tissues, thus lacking specificity and selectivity. Denis and Prout [10] first measured the lactic dehy drogenase (LDH) isoenzyme in carcinoma of the pros tate. While several authors report special LDH ratio (LDH 5/1) patterns, which should be significant for PCA
[11], others found this test to be of no value in the initial clinical diagnosis [12]. Among other markers, carcinoembryonic antigen (CEA), ribonuclease, and prostacyclin should be men tioned. Guinan et al. [13] found no correlation of the increase or decrease of CEA with the stage of PCA. Recently, elevated levels of ribonuclease activity were found in 70% of patients with prostatic carcinoma by Chu et al. [14], Khan et al. [15] showed that persistently high levels of prostacyclin were associated with a poor prognosis of prostatic carcinoma. In 1979, Wang et al. [16] reported the isolation and purification of an antigen of prostatic tissue, found to have a molecular weight of about 33,000-36,000 daltons without subunits. Its carbohydrate portion is about 77%, the isoelectric point at 6.9. It shows mild activity as a protease at neutral pH, distinct from others known in physicochemical and catalytic properties [17]. The im munological characteristics are quite different from PAP or its subunits, therefore showing no cross-reaction. Hu man prostatic tissues contain microgram quantities of PSA per milligram protein with localization of this anti gen confined to the cytoplasm of prostate acinar and ductal epithelium. Significant amounts of PSA can be found in patients with BPH and prostatic carcinoma, primary or metastatic, while other tissues of non-pros tatic origin, either normal or cancerous, contained no detectable amount of PSA. Neither was PSA found in females [ 18]. Therefore, PSA seems to be cell-type and organ-site specific, i.e. a histiotypic product. Its biologi cal function is not yet known. Subsequently, PSA anti bodies have been produced in goats and rabbits, and immunoperoxidase techniques for the identification of PSA in tissues have been developed [19] establishing immunostaining methods and thus indicating the pros tatic origin of these tumors. To evaluate its potential value as a diagnostic/prognostic histological marker, several studies were made to demonstrate: (1) the presence of PSA in prostatic cancer, and (2) the correlation of staining intensity with the degree of tumor differentiation [20], Papsidero et al. [19] evaluated a large variety of dif ferent histological sections from different organs and malignancies, but found no positive results. Epstein et al. [21] reported a statistically significant correlation be tween the foci of poor immunoreactivity and progression of disease, establishing PSA as a valuable adjunct to histopathological grading, while Vernon and Williams [22] proved that the effects of therapy on immunostaining showed no significant difference before and after treat-
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/6/2018 3:15:29 AM
Jurincic/Pixberg/Gasser/Klippel
154
Prostate-Specific Antigen in Prostatic Carcinoma
155
ment. Variations of staining intensity were observed, but general staining remained. With the more recent development of monoclonal PSA antibodies, serum PSA determination for wide clin ical practice became possible. Wang et al. [ 16] confirmed the significance of PSA as a biological marker with regard to the diagnosis and prognosis by using an ELISA technique. They found markedly elevated PSA values in 68% of BPH, 63% of stage-A, 79% stage-B, 77% stage-C, and 86-92% in stage-D prostatic carcinoma. Similar results were obtained in a study by Kuriyama et al. [23] who showed a positive correlation of declining PSA val ues in response to treatment as well as an inverse rela tionship between pretreatment levels of PSA and patient survival time. Pontes et al. [24] found a good correlation between mean serum PSA levels and the surgical stage of disease, demonstrating PSA as a good indicator and predictor of disease progression.
Patients and Methods Fig. 1. Pretreatment PSA value distribution for all patients. Fig. 2. Distribution of pretreatment PSA values for patients with BPH and rate of false-positive results at different cutoff levels.
able at 2-8 °C; (4) sensitivity 0.1-0.2 ng/ml. and (5) specificity is high using monoclonal rabbit antibodies without cross-reaction to other enzymes.
Results These patients were found to have: prostatic cancer in 246 cases, the remaining 274 determinations were serial samples; BPH in 990 cases, and other benign or malig nant urological diseases in 1,488 cases. For all these patients, the PSA value ranged between
