Evaluation of PSA, Free PSA, PSMA, and Totaland Bone Alkaline Phosphatase Levels

Compared to Bone Scans in the Management ofPatients With Metastatic Prostate Cancer

Gerald P. Murphy,1* Michael J. Troychak,2 Oliver E. Cobb,3 Victoria A. Bowes,1Richard J. Kenny,1 Robert J. Barren III,1 Gerald M. Kenny,1 Haakon Ragde,1

Eric H. Holmes,4 and Robert L. Wolfert5

1Cancer Research Division, Pacific Northwest Cancer Foundation, Northwest Hospital,Seattle, Washington

2Department of Nuclear Medicine, Northwest Hospital, Seattle, Washington3Department of Urology, Northwest Hospital, Seattle, Washington

4Department of Cell Surface Biochemistry, Pacific Northwest Cancer Foundation, NorthwestHospital, Seattle, Washington

5Research Division, Hybritech Incorporated, San Diego, California

BACKGROUND. Metastatic prostate cancer clinical evaluation is difficult. A revaluation ofnew prostate markers with regard to bone scans was performed.METHODS. Serial markers, including bone alkaline phosphatase (BAP), total alkaline phos-phatase (TAP), prostate-specific antigen, total (PSA) and free (fPSA), and prostate-specificmembrane antigen (PSMA), were obtained in patients under evaluation and treatment forpossible or known metastatic prostate cancer. These were correlated with bone scan results(BSR).RESULTS. Seventy patients were observed from mid-October 1996–January 1997, duringwhich time 171 serum samples were obtained and correlated with semiquantitative bone scanstatus. PSA and fPSA provided some correlation with BAP and BSR, but only at high levels(>16–50 ng/ml). Receiver-operating curve (ROC) analysis demonstrated that BAP and TAPhad a significant discriminating ability for positive and negative bone scans (>.78), comparedto PSMA, PSA, and fPSA. However, percent BAP and TAP only correlated with BSR at a levelabove six lesions. As the lesions detected by BSR increased, the correlation increased.CONCLUSIONS. BAP is a valuable marker for clinical response evaluations to use in theserial follow-up of patients with metastatic prostate cancer, and correlates well with the bonescan as the number of lesions increase to >6. PSA or fPSA show comparable results, but onlyat high levels (>16–50 ng/ml). Prostate 33:141–146, 1997. © 1997 Wiley-Liss, Inc.

KEY WORDS: bone alkaline phosphatase; PSA; free PSA; PSMA

INTRODUCTION

The evaluation of metastatic prostate cancer in pa-tients with hormone-sensitive or -resistant disease hasdepended largely on the ability in the 1970–1985 pe-riod to evaluate serial bone scans, as few patients hadso-called measurable disease, or on an elevation in theavailable serum marker, acid phosphatase [1,2]. TheNational Prostate Cancer Project (NPCP) devoted con-siderable efforts to defining and attempting to quan-

tify bone scans to assist in the evaluation of these pa-tients [1,2]. In fact, based on these efforts a satisfactorysemiquantitative method for assessing the extent ofdisease in such patients was developed [3]. At the

*Correspondence to: Gerald P. Murphy, M.D., D.Sc., Director, Re-search Division, Pacific Northwest Cancer Foundation, NorthwestHospital, 120 Northgate Plaza, Suite 205, Seattle, WA 98125.Received 12 February 1997; Accepted 15 April 1997

The Prostate 33:141–146 (1997)

© 1997 Wiley-Liss, Inc.

same time, the NPCP was also evaluating the possi-bility that a quantitative association could be esti-mated between tumor extent and response to therapy,as judged by serial serum bone alkaline phosphataseactivity [4]. Early results available at the time sug-gested that bone alkaline phosphatase activity, com-pared to total alkaline phosphatase activity, could beuseful for initial and follow-up evaluation of such pa-tients [4,5]. The response criteria of the NPCP chieflydepended on serum markers and bone scan evalua-tion [6]. This current report on the evaluation of boneand prostate serum markers was prompted by recentpublications on the availability of a kit assay for bonealkaline phosphatase as well as on the availability ofother new serum prostate markers, including prostate-specific antigen (PSA), free prostate-specific antigen(fPSA), and prostate-specific membrane antigen(PSMA) [7–9]. PSMA assays, as well as PSA assays,have been under evaluation at our center and form thebasis for this evaluation of their effectiveness in theinitial evaluation and follow-up of patients suspectedof having metastatic prostatic cancer, especially to thebone [9–11]. We herein report our current evaluation.

MATERIALS AND METHODS

Seventy patients, age range 47–87 years, suspectedor known to have metastatic prostate cancer at initialdiagnosis and during treatment, were evaluated at theNorthwest Hospital, Pacific Northwest Cancer Foun-dation, from mid-October 1996–January 1997 in a se-quential fashion, to attempt to assess the relationshipbetween initial diagnosis, treatment, and a short-termfollow-up with regard to a comparison between theresponse of the patient to various prostate cancermarkers (total and free PSA, PSMA, bone alkalinephosphatase, and total alkaline phosphatase) withquantitative assessments of bone scans performedwithin 30 days of the serum and clinical assessments.NPCP criteria of clinical response were then utilized[1–6,9–12]. Bone scans were classified as 0, M0 > nolesions; 1, M1 = 0–6 lesions; 2, M2 > 6–20 lesions; 3, M3> 20 lesions; and 4, M4 = SuperScan [3]. The NPCPresponse criteria have been modified as follows: acidphosphatase has been replaced with total PSA withthe requirement for a response to be a >50% reduction,over two measurements at least 2 weeks apart [12].

The bone scans were performed in a similar mannerto that reported, and graded in the described semi-quantitative system by one of us, without knowledgeof clinical status [1,2,13]. Bone scans performed priorto the observation period were also available for com-parison.

The assay methodology for PSMA, PSA, and fPSAas performed in our center has been previously de-

scribed [10]. Total alkaline phosphatase was measuredby Northwest Hospital Clinical Laboratory using theTechnicon DAX System (Bayer Corp., Diagnostic Di-vision, Laguna Hills, CA), which has a reported nor-mal range of 25–125 U/l [13]. Bone alkaline phospha-tase was measured using the Hybritech Tandem-ROstase Immunoradiometric Assay (Hybritech Incor-porated, San Diego, CA). This test has a range of 2–120ng/ml, a within- and between-assay variation of <7%and <9%, respectively, a crossreactivity with liver al-kaline phosphatase of <15% (information supplied bythe manufacturer), and a normal range of 3–19 ng/ml.Briefly, 0.100 ml of standard, control, or serumsamples were mixed with 0.100 ml tracer antibody(radiolabeled anti-skeletal alkaline phosphatase) for15 sec. Then, anti-skeletal alkaline phosphatase-coatedbeads were added and incubated without rotation at2–8°C for 19 hr (±2 hr). The beads were then washedthree times and counted using a Wallac 1470 WizardGamma Counter (Wallac Oy, Turku, Finland) [7].

Receiver-operating curves (ROC) depicting rates ofsensitivity by 1-specificity for each marker comparedto positive or negative bone scans were performed,utilizing current methodology [14]. An area under theROC >0.75 is currently felt to be a significantly [14].

Statistical analysis with a variety of tests was com-pleted using conventional analysis techniques [15–17].

RESULTS

Table I describes the generally hormone-refractoryand nonresponsive nature of this 70-patient popula-tion with regard to NPCP criteria of clinical response.The more advanced the bone scan score, the fewerresponses observed. A number of patients had had aprior radical prostatectomy, which affected in part thepostoperative total and free PSA levels. This entiregroup of patients was compared to the others reportedin terms of the findings of positive or negative bonescans, and was found to be similar (pairwise Pearsoncorrelation coefficient values >.95) [7,8,15–19].

Table II describes the average values of the 70 pa-tients for the individual 2–3-serum measurements ob-tained during the test period with regard to bone scancategory. There was a progressive elevation in totalalkaline phosphatase (TAP) as well as bone alkalinephosphatase (BAP). PSA values, total or free, fluctu-ated without direct regard to the number of bone me-tastases, or individual response category. Variations inPSA or free PSA in relationship to a present or absentprostate were less of a factor in these patients than theresponse status in relation to the bone scan score.There were no clinical responders in bone scan scorelevel 3 or 4 in this series (Table II). Several markerswere significant (P < .0001), i.e., comparing BAP, PSA,

142 Murphy et al.

and fPSA to TAP. The cutoff value in using BAP orPSA to determine a relationship to the bone scanyielded a value of BAP equal to 18 ng/ml, and of PSAequal to 16 ng/ml. Using these values in tandemyielded a sensitivity of 0.565 and a specificity of 0.659.Table III describes the individual serum values for to-tal alkaline phosphatase, bone alkaline phosphatase,PSA, free PSA, and PSMA with regard to the quanti-fied NPCP bone scan response criteria. TAP, BAP, andpercent of bone AP show a progressive correlation.PSA or free PSA did not. PSMA levels, except in bonescan category 4 (3 patients), were all elevated.

Figure 1 depicts the ROC values for each markershown in relation to the positive bone scans. Bonealkaline phosphatase, total alkaline phosphatase, freePSA, and total PSA were all significant (P >.7500).

DISCUSSION

This report, as summarized in Figure 2, essentiallydemonstrates that BAP significantly correlates withPSA level as a quantitative assessment of the bone

scan only after the PSA level is >50 ng/ml. The sameis true for fPSA. As noted previously, the cutoff valuesfor BAP and PSA are actually lower (18 and 16 ng/ml,respectively). However, grouping the values as shownin Figure 2 provides a better means of depicting theassociation. This is in contrast to earlier reports thatcollectively grouped all bone scans vs. total PSA re-sults, based only on initial evaluation [18,19]. As notedpreviously, this is not due to patient selection bias, asour group composition was similar to those reportedpreviously [7,8,18,19]. Total alkaline phosphatase orbone alkaline phosphatase, based on ROC cutoff val-ues (>.75), were discriminatory beyond >6 lesions inthe quantified bone scan series (Tables II and III, Fig.1). PSA or free PSA were not of similar value (Fig. 1,Tables II and III). PSMA, which reflects a clonal resis-tance category present in all these individuals withmostly hormone-refractory disease, was generally el-evated (Tables II and III) [9–11].

At present, there is difficulty in clinically evaluatingthese individual patients with metastatic prostate can-cer. The relationship of bone alkaline phosphatase tototal alkaline phosphatase in a serial fashion as shown

TABLE 1. Category of Clinical Responses in Patients DuringObservation Period

Bone scan score 0 1 2 3 4 All

Clinical response status by number of samplesProgression 67 13 15 16 11 122Partial response 20 9 13 0 0 42Stable (no change) 7 0 0 0 0 7All 94 22 28 16 11 171Clinical response status by number of patientsProgression 31 4 5 5 3 48Partial response 12 3 3 0 0 18Stable (no change) 4 0 0 0 0 4All 47 7 8 5 3 70

Patients with intact prostate29/18 2/5 7/1 3/2 3/0 44/26

Patients postradical prostatectomy

TABLE II. Mean Average of Serum Prostate Markers Obtained During Observation Period*

ScoreNumber

of patients TAP BAP % BAP PSA fPSA % fPSA%

complexed PSMA

0 47 81.594 12.247 15.3 26.155 10.433 9.2 84.9 0.26371 7 75.869 12.918 17.0 8.394 1.189 9.8 83.0 0.27282 8 132.58 31.537 21.8 231.24 43.870 17.4 82.6 0.30443 5 425.56 152.71 32.1 1282.49 195.79 19.7 80.3 0.35724 3 971.85 378.98 27.8 544.97 99.956 18.3 81.7 0.1944All 70

*Means calculated on a per patient basis. Normal range TAP, 25–125; BAP, 3–19; PSMA, .08—.211.

Metastatic Prostate Cancer Evaluation 143

in this study appears to offer an improvement in thissituation. As before and even today, these patientsas we see them have few to no instances of other-wise measurable disease. On the other hand, thenewly available ProstaScintt Scan (Cytogen CYT-356;CYTOGEN Corporation, Princeton, NJ) has demon-strated the ability to increase the detection of post-prostatectomy recurrences, as well as extraprostatic

soft-tissue lesions [20,21]. Based on its recent FDA ap-proval, this test should provide an additional means toevaluate such patients before and during treatment.These results are valid for our patients under the con-ditions of our current studies. Different results may beobtained elsewhere. We are interested in how theseavailable prostate markers correlate with serial bonescans at other centers.

TABLE III. Prostate Serum Markers Related to Bone Scan Score

Bone scan score 0 1 2 3 4 All

Total alkaline phosphataseNumber of samples 94 22 28 16 11 171Mean 78.766 79.409 133.79 371.88 1029.00 338.57Standard deviation 21.713 31.207 57.704 378.61 975.17 409.32Range 30–133 42–144 62–278 117–1,364 238–2,783 30–2,78395% confidence limits (±) 4.389 13.04 21.374 185.52 576.27 61.350

Bone alkaline phosphataseNumber of samples 94 22 28 16 11 171Mean 12.378 13.733 32.370 124.38 404.29 117.43Standard deviation 5.729 6.694 19.679 150.33 489.46 207.57Range 2.273–31.691 1.946–25.100 6.348–77.770 11.121–453.21 19.419–1190 1.946–119095% confidence limits (±) 1.158 2.797 7.289 73.661 289.24 31.111

% bone alkaline phosphataseNumber of samples 94 22 28 16 11 171Mean 15.9 17.1 22.6 28.7 27.6 22.380Standard deviation 6.0 5.5 6.6 18.8 17.4 6.639Range 2.5–41.0 4.4–26.1 10.2–32.4 6.5–86.1 6.9–53.4 2.5–86.195% confidence limits (±) 1.2 2.3 2.4 9.2 10.3 0.9951

PSANumber of samples 94 22 28 16 11 171Mean 22.787 10.200 226.06 2157.17 564.88 596.22Standard deviation 65.165 12.607 331.13 2330.0 452.07 963.17Range 0.000–471.44 0.000–32.798 0.113–890.39 20.672–6299.8 144.23–1290.64 0.000–6299.895% confidence limits (±) 13.173 5.268 122.65 1141.66 267.15 144.36

fPSANumber of samples 94 22 28 16 11 171Mean 7.247 1.491 41.159 341.46 106.65 99.601Standard deviation 35.047 1.938 59.140 384.23 94.47 154.27Range 0.000–280.7 0.000–5.505 0.000–180.7 3.234–1035.3 23.38–255.0 0.000–1035.395% confidence limits (±) 7.085 0.8099 21.905 188.27 55.830 23.122

PSMANumber of samples 94 22 28 16 11 171Mean 0.2714 0.2750 0.3190 0.3730 0.2004 0.2878Standard deviation 0.1236 0.1162 0.1617 0.1110 0.0624 0.1150Range 0.0359–0.5451 0.0999–0.5207 0.0202–0.5677 0.1212–0.5628 0.0619–0.2662 0.0202–0.567795% confidence limits (±) 0.0250 0.0486 0.0599 0.0542 0.0369 0.0172

144 Murphy et al.

REFERENCES

1. Slack NH, Karr JP, Chu TM, Murphy GP: An assessment of bonescans for monitoring osseous metastases in patients beingtreated for prostate carcinoma. Prostate 1980;1:259–270.

2. Pollen JJ, Gerber K, Ashburn WL, Schmidt JD: Nuclear boneimaging in metastatic cancer of the prostate. Cancer 1981;47:2585–2594.

3. Soloway MS, Hardeman SW, Hickey D, Raymond J, Todd B,Soloway S, Moinuddin M: Stratification of patients with meta-static prostate cancer based on extent of disease on initial bonescan. Cancer 1988;61:195–202.

4. Killian CS, Vargas FP, Pontes EJ, Beckley S, Slack NH, MurphyGP, Chu TM, Investigators of the National Prostatic CancerProject’s Cooperative Clinical Trials: The use of serum isoen-zymes of alkaline and acid phosphatase as possible quantitativemarkers of tumor load in prostate cancer. Prostate 1981;2:187–206.

5. Wajsman Z, Chu TM, Bross D, Saroff J, Murphy GP, JohnsonDE, Scott WW, Gibbons RP, Prout GR, Schmidt JD: Clinicalsignificance of serum alkaline phosphatase isoenzyme levels inadvanced prostatic carcinoma. J Urol 1978;119:244–246.

6. Slack N, Murphy G: Criteria for evaluating patient responses totreatment modalities for prostate cancer. Urol Clin North Am1984;11:337–342.

7. Cooper EH, Whelan P, Purves D: Bone alkaline phosphataseand prostate-specific antigen in the monitoring of prostate can-cer. Prostate 1994;25:236–242.

8. Morote J, Lorente JA, Encabo G: Prostate carcinoma staging:Clinical utility of bone alkaline phosphatase in addition to pros-tate specific antigen. Cancer 1996;78:2374–2378.

9. Murphy GP, Ragde H, Kenny G, Barren R, Erickson S, Tjoa B,

Fig. 2. Relationship between PSA levels, bone alkaline phospha-tase, and bone scan results.

Fig. 1. ROC values for each markerare shown in relation to a positive bonescan. Bone alkaline phosphatase (BAP),total alkaline phosphatase (TAP), freePSA, and total PSA were all significant(P > 0.7500 value) [12].

Metastatic Prostate Cancer Evaluation 145

Boynton A, Holmes E, Gilbaugh J, Douglas T: Comparison ofprostate specific membrane antigen and prostate specific anti-gen levels in prostatic cancer patients. Anticancer Res 1995;15:1473–1480.

10. Murphy GP, Barren RJ, Erickson SJ, Bowes VA, Wolfert RL,Bartsch G, Klocker H, Pointner J, Reissigl A, McLeod DG, Doug-las T, Morgan T, Kenny GM, Ragde H, Boynton AL, Holmes EH:Evaluation and comparison of two new prostatic cancer mark-ers: Free PSA (F-PSA) and prostate specific membrane antigen(PSMA). Cancer 1996;78:809–818.

11. Morgan TO, McLeod DG, Leifer ES, Moul JW, Murphy GP:Prospective use of free PSA to avoid repeat prostate biopsies inmen with elevated total PSA. Prostate [Supp 1] 1996;7:58–63.

12. Murphy G, Tjoa B, Ragde H, Kenny G, Boynton A: Phase Iclinical trial: T-cell therapy for prostate cancer using autologousdendritic cells pulsed with HLA-A0201-specific peptides fromprostate-specific membrane antigen. Prostate 1996;29:371–380.

13. Alkaline Phosphatase Study Group, Committee on Standards ofthe AACC, Subcommittee on Enzymes, Tiez W, Burtis C, ErvinK, Petitelere CJ, Rinker AD, Zygowics E: Progress in the devel-opment of a recommended method for alkaline phosphataseactivity measurements (abstract). Clin Chem 1980;7:1023.

14. Metz, Charles: Basic principles of ROC analysis. Semin NuclMed 1978;8:283–298.

15. Mehta CR, Patel NR: A network algorithm for performing Fish-er’s exact test in r × c contingency tables. J Am Stat Assoc 1983;78:427–434.

16. Hayes WL: ‘‘Statistics,’’ 3rd ed. New York: Holt Rinehart &Winston, 1981.

17. Hosmer DW, Lemeshow S: ‘‘Applied Logistic Regression,’’ NewYork: John Wiley & Sons, 1989.

18. Chybowski FM, Larson KJJ, Bergstralh EJ, Oesterling JE: Pre-dicting radionuclide bone scan findings in patients with newlydiagnosed, untreated prostate cancer: Prostate specific antigenis superior to all other clinical parameters. J Urol 1991;145:313–318.

19. Wolff JM, Ittel T, Boeckmann W, Reinike T, Habib FK, Jakse G:Skeletal alkaline phosphatase in the metastatic workup of pa-tients with prostate cancer. Eur Urol 1996;30:302–306.

20. Murphy GP: Radioscintiscanning of prostate cancer. Cancer1995;75:1819–1822.

21. Wynant GE, Murphy GP, Horoszewicz JS, Neal CE, Collier BD,Mitchell E, Purnell G, Tyson I, Heal A, Abdel-Nabi H, Winzel-berg G: Immunoscintigraphy of prostatic cancer: Preliminaryresults with 111In-labeled monoclonal antibody 7E11-C5.3 (CYT-356). Prostate 1991;18:229–241.

146 Murphy et al.