Prostatic Cancer, Acid Phosphatase, Creatine Kinase-Bb and Race: A Prospective Study

0022-5347 /82/1284-0735$02.00/D Vol. 128, OctobeY

·~HZ JOURi'JAL OF UP..OLOG""'.{

Printed in U.S.A.

Copyright© 1982 by The Williams & Wilkins Co.

PROSTATIC CANCER, ACID PHOSPHATASE, CREATINE KINASE-BB AND RACE: A PROSPECTIVE STUDY WILLIAM R. FAIR, WARREN D. W. HESTON, DOV KADMON, DAVID B. CRANE, WILLIAM J. CATALONA, JACK H. LADENSON, JAY M. McDONALD, BURTON W. NOLL AND GLORIA HARVEY From the Division of Urology, Department of Surgery, and the Division of Laboratory Medicine, Departments of Pathology and Medicine, Washington University School of Medicine, and the Laboratory Service, John Cochran Veterans Administration Medical Center, St. Louis, Missouri

ABSTRACT

To examine the effectiveness ofp:rostatic acid phosphatase and creatine kinase-BB determinations in detecting prostatic cancer serum from 594 men more than 40 years old was assayed for prostatic acid phosphatase with the thymolphthalein monophosphate substrate and a radioimmunoassay kit. Creatine kinase-BB levels also were measured with a radioimmunoassay kit. Patients with benign prostatic hyperplasia had higher prostatic acid phosphatase levels than normal controls. Accordingly, to avoid a high incidence of false positives in patients with benign prostatic hyperplasia the 92.5 percentile level of the patients with benign prostatic hyperplasia (3.9 ng./ml.) was chosen as the upper limit of normal. With this critical value elevated prostatic acid phosphatase levels were observed in 6 per cent of the patients with clinical stage A disease, 8 per cent with stage B, 35 per cent with stage C and 68 per cent with stage D. The radioimmunoassay was no more effective than the enzymatic assay in detecting prostatic cancer. A correlation between prostatic acid phosphatase levels and patient race was observed, with 80 per cent of the black patients with extracapsular prostatic cancer having elevated prostatic acid phosphatase levels compared to 34 per cent of the white patients with similar stage disease. Creatine kinase-BB was elevated only in patients with advanced disease and was of little value in the diagnosis of prostatic cancer. Enzymatic determination of serum acid phosphatase was one of the earliest proposed procedures for identifying malignant disease. 1 Many modifications have been proposed to improve the specificity and sensitivity of the assay for prostatic acid phosphatase (PAP). 2 • 3 The most recent modification has been the measurement of enzyme concentration by immunologic techniques rather than measurement of enzymatic activity. 4 - 6 Preliminary results from laboratories using radioimmunoassay techniques to measure PAP raised the hope of increased detection of early, intracapsular prostatic cancer. 7 A comparison of results from different laboratories using this technique is hampered by the fact that different antibodies and techniques were used.6- 11 With the current availability of commercially obtainable reagents for radioimmunoassay, different laboratories can compare data using a standardized procedure and the same antibody. 8 ' 12- 14 Seminal plasma and the prostate have been found to contain high concentrations of the BB isoenzyme of creatine kinase (CK-BB), and the serum of patients with cancer of the prostate was reported to contain elevated CK-BB levels. 15 We examined the effectiveness of radioimmunoassay determinations for PAP and CK-BB in detecting prostatic cancer in men >40 years old. MATERIALS AND METHODS

Men ~40 years old had blood drawn by venipuncture upon admission to the hospital and before any other procedure or examination. The serum fraction was obtained by centrifugation and divided into 2 tubes. One tube was acid stabilized and the enzymatic assay was performed promptly with thymolphthalein Accepted for publication October 30, 1981. Read at annual meeting of American Urological Association, Boston, Massachusetts, May 10-14, 1981.

monophosphate as the substrate for the acid phosphatase determination. 16 The nonacid stabilized aliquots were frozen and stored at -20C. Frozen specimens later were thawed and analyzed by competitive binding double antibody radioimmunoassay with reagent kits.* PAP radioimmunoassay was performed on all specimens and CK-BB was determined only when sufficient serum remained after the PAP determination. Only patients with histologic verification of the presence of benign prostatic hyperplasia or prostatic cancer are included in this analysis. Table 1 lists the demographic characteristics of the study population. During the study (September 1979 to December 1980) 594 patients were enrolled. All patients had blood drawn for analyses when they were h.ospifaJized and before rectal examination. All had subsequent prostatectomy or prostatic biopsy so that histologic confirmation of the clinical diagnosis was available in all cases. There were 333 men with benign prostatic hyperplasia and 176 with prostatic cancer. Of the men with prostatic cancer 118 had had no prior treatment. Of the 58 previously treated with estrogen or radiation therapy most were rehospitalized for progressive disease and were considered treatment failures. Eighty-five additional men had other urologic disease in addition to prostatic pathology and were excluded from our analysis. The mean patient age of the total group was 67 years. The 333 patients with benign prostatic hyperplasia ranged from 45 to 90 years old, with a mean of 66 ± 9 years. This group of patients had an average weight of 171 ± 27 pounds, and 271 were white and 61 were black. In 1 individual race was not designated. Of these 333 patients 142 (42.6 per cent) were diagnosed on the basis of bilateral prostatic needle biopsies, taking multiple (4 to 12) cores of tissue. In the remaining 191 patients the diagnosis was made by routine histologic analysis of the tissue obtained at open prostatectomy or trans*

735

Mallinckrodt, Inc., St. Louis, Missouri.

736

FAIR AND ASSOCIATES

TABLE 1. Age, weight and racial characteristics of patients with benign prostatic hyperplasia or untreated carcinoma of the prostate

Age (yrs.)

Benign prostatic hyperplasia (333 cases) Prostatic cancer untreated (118 cases)

Mean* 66 ± 9

70 ± 8

Weight (pounds)

Race (%)t

Range

Av.*

Range

White

Black

45-90

171 ± 27

105-245

81.4

18.3

49-91

174 ± 31

110-312

75.4

23.7

* ± standard deviation. t One individual in the benign prostatic hypertrophy and 1 in the prostatic cancer groups were not designated as to race.

urethral resection of the prostate. The mean patient age of those with prostatic cancer was 70 ± 8 years, with a range of 49 to 91 years. This group of patients had an average weight of 174 ± 31 pounds, and 89 were white and 28 were black. In 1 case race was not designated. Benign prostatic hyperplasia constitutes the major coincidental prostatic disease in a population of men of this age who would have to be distinguished from patients with prostatic cancer. We found that patients with benign prostatic hyperplasia exhibited a more skewed distribution pattern for PAP than did a normal population. Previous work established empirically that the critical value that optimized the sensitivity and specificity of the PAP test corresponded to the upper 92.5 percentile of values observed in patients with benign prostatic hyperplasia. In the present series of patients with benign prostatic hyperplasia the upper 92.5 percentile of values was 3.9 ng./ml. Results above these values were considered to be elevated. The 92.5 percentile assay values in the 333 patients with histologically proved benign prostatic hyperplasia were 0.6 units per ml. for the enzymatic assay, 3.9 ng./ml. for the PAP test and 7.32 ng./ml. for the CK-BB test. Thus, we established and accepted a false positive rate of 7.5 per cent in a benign prostatic hyperplasia population. Surgical staging with pelvic lymph node dissection was performed in all cases of stages A and B prostatic cancer. Although not all patients with stage C disease had surgical confirmation of negative lymph nodes all had negative chest x-ray, excretory urograms, radioisotope bone scan and normal serum alkaline phosphatase. RESULTS

Acid phosphatase results. Total patients: Of the patients with prostatic cancer 59 of 176 (33. 7 per cent) had an elevated PAP by radioimmunoassay and 31.7 per cent (53 of 167) had an elevated PAP as determined by enzymatic assay. Untreated cancer patients: The cancer patients of primary interest were those with untreated disease. PAP values were elevated on radioimmunoassay in 29.7 per cent of the 118 individuals and by the enzymatic procedure in 27.4 per cent (table 2). Results according to cancer stage: Table 2 shows the values obtained in the patients grouped according to tumor stage of the disease. In patients with disease limited to the prostate (clinical stages A and B) only 2 of 48 (4.2 per cent) had an elevated serum acid phosphatase and only 4 of 51 (7.8 per cent) had an abnormal PAP radioimmunoassay. In those patients with extracapsular (clinical stages C and D) disease 27 of 61 (44.3 per cent) had an abnormal acid phosphatase value by enzymatic assay and 30 of 63 (47.6 per cent) were abnormal when measured by radioimmunoassay. Thus, there was no significant advantage for the radioimmunoassay over the enzymatic procedure in detecting prostatic cancer. Even in stage D disease >30 per cent of the patients had normal acid phosphatase values by enzymatic or radioimmunoassay techniques. Predictive value of PAP radioimmunoassay: In view of the

low sensitivity of the test in detecting intracapsular disease it is obvious that the acid phosphatase determination is of little value as a screening test in the early detection of prostatic cancer. However, the predictive value of either a positive or negative test in a population of patients admitted to a urology service with known or suspected prostatic disease is considerably better. The positive predictive value of all patients in the study is as follows: positive predictive value true positives 59 XlOO per cent: - XlOO per cent= 66 per cent .. all pos1t1ves 90 Excluding those patients known to have prostatic cancer before entrance into the study the predictive value of a positive test is 35 X 100, or 53 per cent. The negative predictive value in the same patient population is as follows: negative predictive value true negatives (387) X 100 per cent = 82 per cent all negatives (470) Acid phosphatase values in patients subjected to prostatic biopsy: Accurate PAP determinations theoretically could be of greatest potential value in the diagnosis of patients with a clinically suspicious prostate, that is the group of men in whom the clinician at least minimally suspected prostatic cancer and hospitalized the patient for a prostatic biopsy. In our study blood was drawn prospectively for acid phosphatase determinations in 203 patients hospitalized for needle biopsy of the prostate. The 92.5 percentile value in the 142 patients in whom the biopsy revealed no cancer was 2.57 ng./ml. Of the patients in whom the biopsy subsequently showed malignancy 24 of 61 (39.3 per cent) had elevated radioimmunoassay PAP values >2.57 ng./ml. and 23 of 59 (39 per cent) had increased levels when measured enzymatically. In this group of patients the positive predictive value of the PAP test was 69 per cent and the negative predictive value was 78 per cent. Racial differences: No statistically significant difference between black and white men with respect to PAP values existed in men with intracapsular disease, although the number of cases with positive values was low. However, black men with extracapsular disease had a statistically significant increased percentage of positive acid phosphatase values compared to white men (table 3). CK-BB determination. Over-all the 92.5 percentile level for CK-BB in 225 men with benign prostatic hyperplasia was 7.32 ng./ml. (table 4). In 84 men with documented prostatic cancer only 10 (11.9 per cent) had an elevated CK-BB level. Table 4 shows the correlation between CK-BB levels and tumor stage. Only 2 of 38 men (5.3 per cent) with intracapsular disease had an elevated value, while 8 of 45 (17.7 per cent) with extracap-

TABLE 2. Proportions of patients with untreated primary adenocarcinoma of the prostate having values greater than the upper 92.5 percentile of values found in patients with benign prostatic hyperplasia according to stage of disease*

Elevated Assays Stage:

Enzymatic P APt No.(%)

A

1/16

B

1/32 10/36 17/25 2/4 31/113

C

D Unstaged Totals

(6.2) (3.1) (27.8) (68.0) (50.0) (27.4)

Radioimmunoassay PAP No.(%) 1/16 3/35 13/37 17/26 1/4 35/118

(6.2) (8.6) (35.1) (65.4) (25.0) (29.7)

* The upper 92.5 percentile of enzymatic acid phosphatase values in 323 patients with benign prostatic hyperplasia was 0.60 units. The upper 92.5 percentile of PAP, radioimmunoassay values in 332 patients with benign prostatic hyperplasia was 3.90 ng./ml. t Not all sera were analyzed by the enzymatic procedure.

PROSTATIC CANCER, ACID PHOSPHATASE AND RACE

Percentage elevations of serum acid phosphatase and CKBB in black or white men with stage C or D untreated prostatic cancer

TABLE 3.

Enzymatic PAP No.(%)

Radioimmunoas~ say PAP No.(%)

Radioimmunoassay CK-BB No.(%)

17 /20 (85) 10/39 (27)

16/20 (80) 14/41 (34)

3/15 (20) 5/30 (17)

Black pts. White pts.

Difference between 2 racial groups is significant, at p 0.05 by the chi-square test.

4. Proportions of patients with primary adenocarcinoma of the prostate having CK-BB values greater than the upper 92.5 percentile of values found in patients with benign prostatic hyperplasia according to stage of disease*

TABLE

Elevated Assays No.(%)

Stage

1/13 1/25 4/26 4/19 0/1

A B C D Unstaged Totals

(7.7) (4.0) (15.4) (21.1)

(0.0)

10/84 (11.9)

* The upper 92.5 percentile value in 225 patients with benign prostatic hyperplasia was 7.32 ng./ml.

Comparison of percentage elevations of PAP to other series of patients with prostatic cancer or benign prostatic hyperplasia

TABLE 5.

Benign Prostatic Hyperplasia

Foti and associates 7'* Bruce and associates8 '* Bruce and associates8 't Griffiths 1''t Bruce and associates8 ':j: Bruce and associates8 ':j: Current study:j:

6 3 11 9

27§

llll 7.5

Cancer Stage A

B

C

D

33 14 13 12 16 13 6

79 29

71 24 14 47 36 14 35

92

26 32 45 24 9

89 71 86 85 80 65

* Investigators' own antibody and procedure.

t Commercial PAP kit, New England Nuclear, Boston, Massachusetts. :j: Mallinckrodt, Inc., St. Louis, Missouri. § Bruce and associates' data based on the 97.5 percentile of the normal male population. The critical value used was 1.9 ng./ml. 8 II Bruce and associates' data revised by Bruce with the critical value increased to 2.5 ng./ml. 8 sular disease showed an elevation in CK-BB in the serum when measured the radioimmunoassay kit. DISCUSSION

The early hopes that serum PAP determinations by immunoassay would provide a screening tool for the early detection of prostatic cancer have still to be realized. Indeed, recent investigations with PAP radioimmunoassays have failed to demonstrate the high number of positives in cases of stages A and B cancer that were reported by Foti and associates 7 (table 5). One problem might be that many investigators used a normal population that had no relevance to the population in which PAP determinations are of most value, that is older men in whom there is a frequent occurrence of benign prostatic hyperplasia. In the literature supplied with the PAP radioimmunoassay there were 226 normal men 19 to 81 years old who had a normal bell-shaped distribution of PAP values. The mean serum level of PAP was 0.9 ng./ml. and no values > 2.0 ng./ml. were detected. However, the mean value for serum for individuals with histologically verified benign prostatic hyperplasia was reported to be 1.2 ng./ml. and the values were skewed to higher levels, with a number of values > 2.0 ng./ml. Our experience with 333 patients with histologically verified benign prostatic hyperplasia has been a mean serum value of l.65 ng./ml. and skewed distribution toward greater values. Because of the skewed distribution it was decided that a critical

737

value could not be based on the assumption of a normal distribution. The critical value that optimized the sensitivity and specificity of the assay of the PAP radio immunoassay test was found to correspond to the upper 92.5 percentile of values in patients with benign prostatic hyperplasia (3.9 ng./ml.). Similarly, the critical values for the enzymatic PAP and radioimmunoassay CK-BB also were set at the 92.5 percentile of their respective values observed in patients with benign prostatic hyperplasia. In patients with untreated prostatic cancer the mean serum value of PAP was 6.9 ng./ml. and the values range from O to >40 ng./ml. We did not find a significant difference between the radioimmunoassay and enzymatic method in detecting elevated acid phosphatase values in this population of patients. It appears with this radioimmunoassay kit as well as others currently marketed that no substantial clinical advantage is gained by the radioimmunoassay method when compared to an enzymatic assay using an appropriate substrate, such as thymolphthalein monophosphate. However, it is possible that other radioimmunoassay kits may yield better results in the future. In the group with benign prostatic hyperplasia we found individuals with values in the upper ranges of those found with prostatic cancer. These were found more often in individuals who were to have prostatic surgery than those hospitalized for biopsy only. This fact is reflected in the decrease in the 92.5 percentile for the group with benign prostatic hyperplasia as a whole from 3.9 to 2.57 ng./ml. for the biopsy group. This probably reflects the trend toward urinary obstructive symptoms in the surgical grnup and is consistent with the observation of others that urinary retention is associated with increased serum levels of acid phosphatase, and suggests that PAP determinations should be interpreted with caution in patients with severe obstructive symptoms. Another parameter that may affect the outcome of the incidence of increased PAP levels in prostatic cancer patients is the number of black men in the study. It has been our experience to date that black men have a higher incidence of PAP elevations, as well as a greater value of PAP, compared to white men. This may be because black men present with more bulky disease than white men or it may represent a difference in the biology of the tumor between these racial groups. It has been reported in epidemiological studies that prostatic cancer in American black men appears to behave more aggressively. 17 The differential trend of PAP elevations in our studies is consistent with the epidemiological findings. Radioimmunoassay determinations of PAP represent a technical innovation in the methodology for detecting this enzyme in serum. A major advantage for this method is that the immunologic reactivity is more stable than the enzymatic activand less care is required in handling serum specimens. We not found this immunoassay to offer a clinical advantage over the enzymatic procedure within the population at risk. Furthermore, we have no elevations in patients with intracapsular disease that differ from the percentage elevations found in patients with benign prostatic hyperplasia. As is seen in table 5 this has been the experience of most of the more recent investigations with immunoassay procedures. Even with this technical advance in detection it appears that the elevations in stages A and B prostatic cancer are either the result of understaging owing to micrometastases not found with staging lymphadenectomy or a reflection of a skewed distribution, such as that found in patients with benign prostatic hyperplasia. It also may be that the elevation in benign prostatic hyperplasia when not due to urinary retention may be an indication of the presence of undetected prostatic cancer, with a high malignant potential. 8 These possibilities will .have to be addressed on followup studies of these individuals. Even in stage D prostatic cancer elevations are not always highly elevated, which limits the usefulness of acid phosphatase as a marker for the presence

738

FAIR AND ASSOCIATES

of prostatic cancer or for estimating tumor burden. PAP radioimmunoassay is definitely not a screening method for the general population. One may ask whether PAP determination by radioimmunoassay could be useful in screening patients admitted to the urologic service when the incidence of prostatic cancer would presumably be much greater. In our series the positive predictive value is only 66 per cent if all patients are included. If those known to have malignancy at the time of hospitalization are excluded the positive predictive value decreases to only 53 per cent. It is clear that this method would not be useful as a screening test even in a population enriched in its proportion of patients with prostatic cancer. As currently accepted PAP will be useful in following response to therapy in those cases in which it is elevated but other markers need to be found that will be of greater use in predicting the malignant potential of prostatic cancer. CK-BB does not appear to be as useful a marker as PAP for prostatic cancer. A similar conclusion was reached recently by Zweig and Van Steirteghem. 18 However, Silverman and associates believe that antibodies raised against CK-BB purified from the prostate may be more specific than those raised against brain CK-BB. 19 Whether such will be the case remains to be determined. It is possible that CK-BB may provide a means to follow response to therapy in those patients in whom PAP levels are not elevated. It should be emphasized that antibodies raised against high molecular weight proteins of varying degrees of purity will result in heterogeneous antibody production. Recent advances in cell biology have resulted in cell fusion techniques that make it possible to produce large quantities of monoclonal antibody of defined and unchanging specificity. 20 Hopefully, this will result in an array of specific antibodies against features unique to cancerous as well as to normal prostatic tissue, which will enable us to detect and treat better prostatic cancer. 21 • 22 REFERENCES 1. Gutman, A. B. and Gutman, E. B.: An acid phosphatase occurring

2. 3. 4. 5. 6.

in the serum of patients with metastasizing carcinoma of the prostate gland. J. Clin. Invest., 17: 473, 1938. Yam, L. T.: Clinical significance of the human acid phosphatase: a review. Amer. J. Med., 56: 604, 1974. Mercer, D. W.: Separation of tissue and serum acid phosphatase isoenzymes by ion-exchange column chromatography. Clin. Chem., 23: 653, 1977. Shulman, S., Mamrod, L., Gonder, M. J. and Soanes, W. A.: The detection of prostatic acid phosphatase by antibody reactions in gel diffusion. J. Immunol., 93: 474, 1964. Milisauskas, V. and Rose, N. R.: Immunochemical quantitation of prostatic acid phosphatase. Clin. Chem., 18: 1529, 1972. Moncure, C. W., Johnston, C. L., Jr., Smith, M. J. V. and Koontz, W. W., Jr.: Immunological and histochemical evaluation of marrow aspirates in patients with prostatic carcinoma. J. Urol., 108: 609, 1972.

7. Foti, A. G., Cooper, J. F., Herschman, H. and Malvaez, R. R.: Detection of prostatic cancer by solid phase radioimmunoassay of serum prostatic acid phosphatase. New Engl. J. Med., 297: 1357, 1977.

8. Bruce, A. W., Mahan, D. E., Sullivan, L. D. and Goldenberg, L.: The significance of prostatic acid phosphatase in adenocarcinoma of the prostate. J. Urol., 125: 357, 1981. 9. Murphy, G. P., Chu, T. M. and Karr, J.P.: Prostatic acid phosphatase-the developing experience. Clin. Biochem., 12: 226, 1979. 10. Choe, B. K., Pontes, J. E., Dong, M. K. and Rose, N. R.: Doubleantibody immunoenzyme assay for human prostatic acid phosphatase. Clin. Chem., 26: 1854, 1980. 11. Vihko, P., Sajanti, E., Janne, 0., Peltonen, L. and Vihko, R.: Serum prostate-specific acid phosphatase: development and validation of a specific radioimmunoassay. Clin. Chem., 24: 1915, 1978. 12. Griffiths, J.C.: Prostate-specific acid phosphatase: re-evaluation of radioimmunoassay in diagnosing prostatic disease. Clin. Chem., 26: 433, 1980. 13. Quinones, G. R., Rohner, T. J., Jr., Drago, J. R. and Demers, L. M.: Will prostatic acid phosphatase determination by radioimmunoassay increase the diagnosis of early prostatic cancer. J. Urol., 125: 361, 1981. 14. Brody, J. P., Savory, J., Sturgill, B. C., Lehman, M. R. and Wills, M. R.: Prostatic acid phosphatase as measured by two radioimmunoassay kits in the detection of prostatic adenocarcinoma. Clin. Chem., 27: 605, 1981. 15. Silverman, L. M., Dermer, G. B., Zweig, M. H., Van Steirteghem, A. C. and Tokes, Z. A.: Creatine kinase BB: a new tumorassociated marker. Clin. Chem., 25: 1432, 1979. 16. Ladenson, J. H., McDonald, J.M., Renoe, B. W. and Michael, J. M.: Acid phosphatase and prostatic carcinoma. Clin. Chem., 24: 129, 1978. 17. Hutchison, G. B.: Incidence and etiology of prostate cancer. Urology, suppl., 17: 4, 1981. 18. Zweig, M. H. and Van Steirteghem, A. C.: Assessment by radioim-

munoassay of serum creatine kinase BB (CK-BB) as a tumor marker: studies in patients with various cancers and a comparison of CK-BB concentrations to prostate acid phosphatase concentrations. J. Natl. Cancer Inst., 66: 859, 1981. 19. Silverman, L. M., Chapman, J. F., Jones, M. E., Dermer, G. B., Pullano, T. and Tokes, Z. A.: Creatine kinase BB and other markers of prostatic carcinoma. Prostate, 2: 109, 1981. 20. Milstein, C.: Monoclonal antibodies. Sci. Amer., 243: 66, 1980. 21. Bolmer, S. D. L. and Davidson, E. A.: Preparation and properties of a glycoprotein associated with malignancy. Biochemistry, 20: 1047, 1981. 22. Lee, C.-L., Jou, Y.-J., Kirdani, R., Murphy, G. P. and Chu, T. M.:

Monoclonal antibodies to human prostatic acid phosphatase. Fed. Proc., 40: 1595, 1981. EDITORIAL COMMENT This is an excellent study, confirming that immunoassays for PAP and CK-BB are of no value in screening populations for prostate cancer (even when such surveys are limited to men in the prostate cancer age group). Although the authors have selected arbitrarily the 92.5 percentile levels of these enzymes in patients with benign prostatic hyperplasia as the critical value determining an elevation, a reasonable attempt has been made to verify the absence of malignancy by histologic sectioning of the resected specimen or multiple random needle biopsies. This contribution emphasizes that the more costly immunoassays for PAP yield no substantial clinical advantage compared to standard enzymatic assays, which are less expensive. It remains to be seen whether immunoassays using monoclonal antisera will yield a degree of specificity sufficient to warrant the routine use of these more expensive laboratory studies. C. A. 0.