- Email: [email protected]
Prostate Specific Antigen in the Diagnosis and Treatment of Adenocarcinoma of the Prostate. II. Radiation Treated Patients
0022-534 7/89 /1415-1084$02.00 /0 THE JOURNAL OF UROLOGY
Vol. 141, May Printed in U.S.A.
Copyright © 1989 by The Williams & Wilkins Co.
PROSTATE SPECIFIC ANTIGEN IN THE DIAGNOSIS AND TREATMENT OF ADENOCARCINOMA OF THE PROSTATE. III. RADIATION TREATED PATIENTS THOMAS A. STAMEY,* JOHN N. KABALIN
AND
MICHELLE FERRARI
From the Division of Urology, Stanford University Medical Center, Stanford, California
ABSTRACT
Serum prostate specific antigen was determined (Yang polyclonal radioimmunoassay) in 183 men after radiation therapy for adenocarcinoma of the prostate. A total of 163 men had received 7,000 rad external beam radiotherapy and 20 had been implanted with 125 iodine seeds. Only 11 per cent of these 183 patients had undetectable prostate specific antigen levels at a mean interval of 5 years since completion of radiotherapy. Prostate specific antigen levels after radiotherapy were directly related to initial clinical stage and Gleason score before treatment. Multiple prostate specific antigen determinations were performed with time in 124 of 183 patients. During year 1 after radiotherapy prostate specific antigen levels were decreasing in 82 per cent of the patients but only 8 per cent continued to decrease beyond year 1. Of 80 patients observed greater than 1 year after completion of radiotherapy 51 per cent had increasing values and 41 per cent had stable values. Increasing prostate specific antigen values after radiotherapy were correlated with progression to metastastic disease and residual cancer on prostate biopsy. Total serum acid phosphatase levels were poorly related to prostate specific antigen levels, were less effective in discriminating patients with metastatic disease and provided no additional information beyond that provided by prostate specific antigen. (J. Ural., 141: 1084-1087, 1989) Radiation therapy has been a popular form of treatment for prostate cancer. Actuarial survival rates have been reported for patients treated with linear accelerator supervoltage radiotherapy 1• 2 and the implantation of 125iodine (1 25 1) seeds. 3 However, for the evaluation of status free of disease most radiation therapists have relied upon digital rectal examination and enzymatic serum acid phosphatase measurements. With the recognition of the much greater sensitivity of prostate specific antigen (PSA) in the detection of intracapsular and extracapsular prostate cancer, 4 - 6 we have examined the concentrations and serial patterns of serum PSA in patients who have completed radiation therapy. METHODS
We determined the concentration of serum PSA in 183 patients at various intervals after completion of radiation therapy. Of these patients 163 received external beam radiotherapy with 7,000 rad to the prostate and 5,000 rad to the regional lymph node area during a 7-week period. 1• 2 Twenty patients were implanted with 1251 seeds at a mean total activity of 130 Gy. (range 80 to 200 Gy.). Mean patient age at the most recent serum PSA determination was 69 years. At initiation of radiotherapy cancer stage was assessed as stage A in 25 patients (21 stage A2), stage B in 81, stage C in 52 and stage Dl in 25. Of the 20 patients who received 1251 implantations 15 had clinical stages A and B disease. The average Gleason score on pre-radiotherapy biopsy was 6.6 and 6.3 for external beam and 1251 implants, respectively. For this analysis of PSA the 20 patients with 125 1 seed implants were included in the total 183 patients. Every post-radiotherapy patient seen at our institution who
had blood serum sent to the Stanford urology laboratories for a PSA determination between April 1985 and January 1988 formed the basis for this study, without any other selection bias except the exclusion of all patients who had received any form of anti-androgen therapy and any patient who had undergone transurethral or enucleation prostatectomy within 1 year of the first PSA. Ambulatory serum PSA, determined by the Yang polyclonal radioimmunoassay as in the previous studies,4-6 was obtained at intervals ranging from immediately after completion of radiotherapy up to 258 months (21½ years) after therapy. All assays were done in the Stanford urology laboratories. The mean interval between completion of radiotherapy and serum PSA determination was 61 months (approximately 5 years). The reported PSA values represent the most recent determination if more than 1 was available. Total serum acid phosphatase was determined in the Stanford clinical laboratories with the Bessey-Lowry enzymatic method, a technique used by these laboratories for 30 years. The range for normal men (2 standard deviations) is 0.13 to 0.64 Bessey-Lowry units. RESULTS
Of the 183 patients 11 per cent had undetectable PSA (less than 0.3 ng./ml.) 6 while 25 per cent were within the range of normal male levels (Oto 2.5 ng./ml. 4 ). Thus, 36 per cent of the patients had a recent PSA level within the normal range, with an average followup of 5 years after radiotherapy. However, another 36 per cent had elevated levels between 2.5 and 20 ng./ ml., while 28 per cent exceeded 20 ng./ml. with a mean PSA of 611 ng./ml. These elevations in PSA clearly were related to pre-radiation clinical stage; with mean followups of 51 to 89 months, the mean PSA levels for 25 stage Bl, 21 stage A2, 42 stage B2, 14 stage BS and 52 clinical stage C cancer patients were 7.8, 17, 54, 44 and 505 ng./ml., respectively. A total of 124 patients (107 who received external beam radiation and 17 who received 1251 seed implants) had multiple
Accepted for publication December 28, 1988. Supported in part by the Richard M. Lucas Cancer Foundation and the Flintridge Foundation. * Requests for reprints: Division of Urology, 8287, Stanford University Medical Center, 300 Pasteur Drive, Stanford, California 943055118. 1084
1085
PR0STATE SPECIFIC ANTIGEN
serum PSA determinations with time; the mean interval between consecutive PSA determinations was 9.1 months, with a range of 1 to 31 months. About a third of the patients had 2 consecutive determinations, a third had 3 and a third had 4 or more. Sixteen patients had clinical stage A (13 stage A2), 54 stage B, 34 stage C and 20 stage Dl disease at initiation of radiotherapy. The observed patterns of change in serial PSA values during years 1 and 2 and later are summarized in table 1. A total of 44 patients had multiple PSA determinations during year 1, 21 during year 2 and 59 after 2 years. During year 1 after completion of radiotherapy PSA was decreasing in 82 per cent (36 of 44) of the evaluable patients, including the majority of patients in all clinical disease stages. The mean rate of this decrease was 9. 7 ng./ml. per month. The absolute level of PSA was related to the pattern of change observed. PSA values were decreasing in two-thirds (24 of 36) of the patients with a PSA of 10 ng./ml. or less. Increasing PSA values were observed in an increasing number and percentage of patients with increasing absolute PSA levels. Of the 6 patients whose PSA elevated during the first 12 months 5 not only had the higher clinical stages (B3 or above) but they also had the highest PSA values (greater than 10 ng./ml.). Of the 80 patients with serial PSA values 12 months or more after radiotherapy (table 2) 51 per cent had increasing levels (evenly distributed across all stages) at a mean rate of increase of 28 ng./ml. per month. Only 6 of the 80 patients (8 per cent) continued to have decreasing levels more than 1 year following radiotherapy, and 4 of these 6 were evaluated during year 2. The remaining 2 patients with decreasing PSA were in years 3 and 5 after therapy, with no patient continuing to show decreases beyond 53 months. Additionally, 33 of the 80 evaluable patients (41 per cent) more than 1 year after receiving radiotherapy had stable PSA values (defined as variation less than 1 ng./ml.) during mean interval between PSA determinations of 12 months. The mean absolute PSA value in this stable group was 2.9 ng./ml., with 85 per cent (28 of 33) of these stable values being 2.5 ng./ml. or less (table 2). The data in table 2 also confirm that the 51 per cent of the patients with increasing PSA after 1 year of radiotherapy are heavily weighted toward those with higher serum PSA values (31 of 41 greater than 10 ng./ml.). Gleason scores 7 were available in 159 of the 183 patients at initiation of radiotherapy. Of 75 patients (mean patient age 68 years) with a Gleason score of 6 or less the mean post-radiotherapy PSA was 69 ng./ml., with a mean followup of 68 months. However, of 84 patients (mean age 69 years) with a Gleason score of 7 or greater the mean post-radiotherapy PSA was 268 ng./ml., with a mean followup of 39 months. Table 3 indicates that the patients with the higher Gleason scores before radiotherapy generally are those with the higher postradiotherapy PSA values, while the majority (38 versus 18) of patients with post-radiotherapy PSA values of 2.5 ng./mL or less had pre-radiotherapy Gleason scores of 6 or less. Total acid phosphatase serum levels, obtained at the same time as PSA, were available in 107 of the 183 patients (58 per cent) after radiotherapy. These values are tabulated and crossreferenced with PSA concentrations in tables 4 and 5. There is TABLE
1. Patterns of change in 124 patients with serial PSA values
after radiotherapy PSA Yrs. After Radiotherapy 1 2 >2 Totals
Decreasing No.(%) 36 4 2 42
(82)t (19) (3) (34)
Stable* No.(%) 2 7 26 35
Increasing No.(%)
(5) (33) (44) (28)
* Stable defined as variation of Jess than 1 ng./ml.
t Rate of decrease 9. 7 ng./ml. per month.
6 10 31 47
(14) (48) (53) (38)
Total No.(%) 44 21 59 124
(100) (100) (100) (100)
PSA levels in 80 patients with serial values obtained more than 12 months after radiotherapy
TABLE 2.
PSA Total No. Pts.
PSA {ng./ml.)
% Decreasing (6 pts.)
0.0 >0.0-2.5 >2.5-10 >10-20 >20 Totals
0 14 17 9 0
100 71 25 9 4
0 14 58 82 96
13 21 12 11 23
8
4lt
5it
80
% Stable'· (33 pts.)
% Increasing (41 pts.)
* Stable defined as variation of less than 1 ng./ml.
t Mean PSA 2.9 ng./ml.
t Rate of increase 28 ng./ml. per month. TABLE 3. Distribution of PSA values in 159 patients after radiotherapy compared to Gleason score at initiation of treatment
G Jeason Score PSA (ng./ml.)
;:i6 No.(%)
?;,7 No.(%)
12 26 14 8 15
5 13 20 13 33
0.0 >0.0-2.5 >2.5-10 >10-20 >20
(16) (35) (19) (11) (20)
(6) (15) (24) (15) (39)
TABLE 4. Comparison of mean total acid phosphatase and PSA levels in 107 patients after radiotherapy, broken down by clinical stage before treatment
Stage
No. Pts.
Al A2 Bl B2 B3
1 15 15 26
C
29 13
Dl
Mean Total Acid Phosphatase (Bessey-Lowry units)
Mean PSA {ng./ml.)
0.71 0.73 0.62 0.65 0.62 0.81 0.72
1.9 21.7 7.2 13.2 17.7 127.4 17.3
8
Normal range of total acid phosphatase 0.13 to 0.64 Bessey-Lowry units. Normal range of PSA Oto 2.5 ng./ml.
Distribution of total acid phosphatase levels after radiotherapy compared to PSA levels in 107 patients
TABLE 5.
PSA (ng./ml.)
Total No. Pts.
0.0 >0.0-2.5 >2.5-10 >10-20 >20 Totals
15 33 23 14 22 107
Total Acid Phosphatase (No. pts.) ;:i0,64
>0.64
9 23 18
10 5
6
8
6
7
15 42
65
Mean Total Acid Phosphatase (Bessey-Lowry units) 0.64 0.60 0.60 0.64 1.07
Normal range of total acid phosphatase 0.13 to 0.64 Bessey-Lowry units. Normal range of PSA Oto 2.5 ng./ml.
minimal correlation between PSA and total acid phosphatase values. The representative and important responses of PSA to radiotherapy are shown in figures 1 to 3. The patient in figure 1 is illustrative of the 82 per cent of those whose serum PSA decreases during year 1 of radiotherapy, even from rather substantial elevations (230 ng./ml.). Figure 2 also is representative of a common pattern, decreasing during year 1 to normal values but increasing steadily thereafter. Figure 3 represents radiation-resistant cancer, a slow decrease in PSA that never approaches normal and is followed by a rapid return to preradiation levels.
1086
STAMEY AND ASSOCIATES
PSA (ng/ml) 230
PSA (ng/ml)
\
100
100 49.8 -~
42.2
27.7
\
~15.3
12.8
"\
10
;·
· ~0. 6 -~4
10
•,\
5.4
2.2
•
0
20
# MONTHS FOLLOWING RADIOTHERAPY FIG. 1. PSA levels in 71-year-old man with clinical stage B2 nodule,5 Gleason score 7 (3 plus 4) on biopsy. Pre-radiation PSA level was 230 ng./ml. Computerized tomography and pedal lymphangiograms failed to show any lymph nodes. Within 2 weeks of completing 5,000 rad to pelvis and 7,000 rad to prostate, serum PSA decreased to 27.7 ng./ml., 5 months later it was 2.2 ng./ml. and 15 months later PSA was 1.5 ng./ ml. PSA normal range O to 2.5 ng./ml.
PSA (ng/ml)
100
13.5 10
20
0
# MONTHS FOLLOWING RADIOTHERAPY
FIG. 2. Findings in 75-year-old man with clinical stage C prostate cancer,5 Gleason score 7 (3 plus 4) on biopsy. PSA was 13.5 ng./ml. before receiving 7,000 rad to prostate. Six months after completing radiotherapy PSA was 1.6 ng./ml. but level gradually increased in next 13 months and 13 months after 12. 7 entry it increased to 57 ng./ml. PSA normal range O to 2.5 ng./ml.
DISCUSSION
With the Yang polyclonal radioimmunoassay we obtained blood serum for PSA from 183 patients at a mean interval of 61 months after completion of radiotherapy. Each patient had been treated only with irradiation; patients who had been placed on anti-androgen therapy at any time or who had
0
20 # MONTHS FOLLOWING RADIOTHERAPY
FIG. 3. Findings in 78-year-old man with clinical stage C prostate cancer, Gleason score 9 (4 plus 5). Patient received 7,000 rad immediately after serum PSA of 49.8 ng./ml. was obtained. PSA gradually decreased to low of 8.4 ng./ml. 15 months after radiation therapy but then increased 5-fold to 42.2 ng./ml. during next 5 months. PSA normal range Oto 2.5 ng./ml.
undergone transurethral resection of the prostate within 1 year of the first PSA determination were excluded from this analysis. Only 11 per cent of these 183 patients had undetectable PSA levels but an additional 25 per cent were within the normal range (2.5 ng./ml. or less). The remaining 64 per cent of the patients all had abnormal elevations of serum PSA. PSA levels after radiotherapy were related directly to initial clinical stage and Gleason score before treatment. Multiple PSA determinations were performed with time in 124 of 183 patients (table 1). We have established that PSA levels decreased in 82 per cent of irradiated patients during the first 12 months after therapy but only 6 of 80 patients (8 per cent) continued to have decreasing levels beyond year 1. It is of some interest that this decrement in PSA during year 1 occurred regardless of the clinical stage at initiation of therapy and that the mean rate of decrease was 9.7 ng./ml. per month. Of more importance are the patterns of change in 80 patients observed greater than 1 year after completion of radiotherapy (table 2). Of the patients 51 per cent had increasing PSA values, occurring at all clinical stages with a mean rate of increase of 28 ng./ml. per month. These patients with increasing values are distributed towards those with the higher final concentrations of PSA, while the 41 per cent of the 80 patients with stable PSA (PSA values that vary less than 1 ng./ml., 85 per cent of which were less than 2.5 ng./ml.) are clearly distributed towards the lower PSA concentrations (mean 2.9 ng./ml.). Increasing PSA values after radiotherapy correlated with progression to metastatic disease and residual cancer on prostate biopsy. In 27 of 183 irradiated patients we documented progression to clinical stage D2 and we obtained multiple PSA values on 16 (59 per cent). All 16 patients had increasing PSA values before progression to clinical stage D2. More convincingly, we biopsied 9 of the 41 patients (22 per cent) with increasing PSA values 12 months after completion of therapy. Despite the observation that 8 of these 9 prostates were palpably flat, symmetrical and normal on digital rectal examination before biopsy, all 9 were positive for prostate cancer. While elevated levels as well as increasing PSA concentra-
1087
PROSTATE SPECIFIC ANTIGEN
tions with time strongly suggest failure of radiation therapy to cure prostate cancer, positive biopsies are the best evidence for failure. 8- 11 We recently showed that multiple, ultrasoundguided, transrectal biopsies can be performed safely, painlessly and quickly with minimal morbidity in an outpatient setting. 12 Random systematic biopsies can be used to obtain 1.5 cm. cores from apex to base on both sides of the prostate and seminal vesicles. 13 With this technique asymptomatic, post-irradiation patients (mean 5 years) with palpably normal prostates can have substantial residual cancer in the presence of a normal PSA. 14 Thus, while increasing PSA concentrations after radiotherapy surely imply failure to cure the cancer, normal serum concentrations of PSA, even when accompanied by a flat or palpably normal prostate, cannot be relied upon to exclude the presence of residual cancer. It is important to emphasize that irradiation of the prostate drastically changes the significance of the absolute serum PSA value and that it cannot be interpreted by comparing postradiotherapy values to the database derived from untreated patients. 4 •5 For example, the lowest PSA values we have seen in untreated patients with documented metastasis to bone (mean PSA 563 ng./ml.) 5 were in 2 or 3 rare patients in the 30 to 40 ng./ml. range. By contrast, of 18 post-radiotherapy patients who have progression to positive bone scans and in whom a PSA value was obtained within 3 months of that scan 4 had serum PSA concentrations of 8, 10, 19 and 20 ng./ml. Thus, it appears that metastases can occur at low serum levels of PSA after irradiation to the prostate and pelvis. This finding suggests that the bulk of the tumor burden in untreated stage D2 disease still is within the radiated prostate and pelvis. In a unique study, Mahan and associates examined prostate cancer biopsy specimens from the same group of 27 patients before and after radiation therapy. 15 Although all 27 cancers in their series stained positive for acid phosphatase before irradiation 3 failed to stain and 3 showed only equivocal staining after irradiation. If acid phosphatase production is absent ifi some irradiated prostate cancers, then it is logical to assume that PSA similarly may be absent or diminished in a minority of irradiated cancers. This might explain the unusually low PSA values sometimes observed following radiation therapy despite demonstrable residual cancer and even distant metastases. Much remains to be learned regarding the exact nature of the biological effect of radiation on prostate cancer in general, and on the known biochemical markers for prostate cancer in specific. Whatever the explanation, a high index of suspicion must be maintained for disease that may be much more advanced than the serum PSA might indicate in any patient whose prostate has been irradiated. It is interesting, however, that once the concentration of serum PSA increased to more than 50 ng./ml. in the postirradiation patient (16 of 23 patients in table 2) subsequent increases occurred at a mean rate of 138 ng./ml. per month, which is the same as we observed in all 16 untreated stage D2 cancer patients whose initial PSA also was more than 50 ng./ ml.5 Radiation therapists have used for their definition of no evidence of disease not only a flat, smooth, nonirregular prostate on digital rectal examination but also a normal total serum acid phosphatase. 1 • 2 However, there is almost no correlation between total serum acid phosphatase and PSA concentrations. In fact, in 4 consecutive ranges of increasing PSA values (table 5) mean total acid phosphatase remained constant between 0.60 and 0.64 (normal range for total acid phosphatase 0.13 to 0.64 Bessey-Lowry units), except for patients with the highest PSA levels (greater than 20 ng./ml.), which was the only group with an elevated mean total acid phosphatase. Additionally, of the 27 men in whom we documented progression to stage D2 after radiotherapy, we obtained total acid phosphatase values at the time of PSA in 9. As noted previously, all 9 men had elevated PSA levels, with a minimum value of 8.2 ng./ml.
(greater than 3 times normal). However, 2 of these 9 patients had normal total acid phosphatase levels. We must conclude that not only is PSA of superior use in comparison to total acid phosphatase to monitor the post-radiotherapy patient but that total acid phosphatase adds no useful information beyond that provided by serum PSA. Finally, we believe that serial PSA values, especially when combined with multiple, transrectal, ultrasound guided 1.5 cm. cores (prostate mapping), 12- 1• will allow an infinitely more accurate assessment of residual cancer 18 months after completion of radiotherapy. Our observations that 53 per cent of the men greater than 2 years after radiotherapy have increasing levels of PSA and that ultrasound guided, systematic biopsies of even palpably normal prostates frequently are positive in the presence of normal serum levels of PSA 14 strongly suggest that there is a higher probability of persistent cancer after radiotherapy than generally is recognized. It also is likely that the 25 per cent 15-year survival rates reported for stages B2 and B3 nodules' come from the 44 per cent of men with stable serum levels of PSA, 85 per cent of whom have PSA concentrations of O to 2.5 ng./ ml. REFERENCES
1. Bagshaw, M. A., Cox, R. S. and Ray, G. R.: Status of radiation treatment of prostate cancer at Stanford University. Natl. Cancer Inst. Monogr., 7: 47, 1988. 2. Bagshaw, M.A., Ray, G. R. and Cox, R. S.: Selecting initial therapy for prostate cancer: radiation therapy perspective. Cancer, 60: 521, 1987. 3. Whitmore, W. F., Jr., Hilaris, B., Batata, M., Sogani, P., Herr, H. and Morse, M.: Interstitial radiation: short-term palliation or curative therapy? Urology, suppl. 2, 25: 24, 1985. 4. Stamey, T. A., Yang, N., Hay, A. R., McNeal, J.E., Freiha, F. S. and Redwine, E. R.: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. New Engl. J. Med., 317: 909, 1987. 5. Stamey, T. A. and Kabalin, J. N.: Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. I. Untreated patients. J. Urol., 141: 1070, 1989. 6. Stamey, T. A., Kabalin, J. N., McNeal, J. E., Johnstone, I. M., Freiha, F., Redwine, E. A. and Yang, N.: Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. II. Radical prostatectomy patients. J. Urol., 141: 1076, 1989. 7. Gleason, D. F.: Histologic grading and clinical staging of prostatic carcinoma. In: Urologic Pathology. Edited by M. Tannenbaum. Philadelphia: Lea & Febiger, part II, chapt. 9, pp. 171-198, 1977. 8. Freiha, F. S. and Bagshaw, M. A.: Carcinoma of the prostate: results of post-irradiation biopsy. Prostate, 5: 19, 1984. 9. Scardino, P. T., Frankel, J.M., Wheeler, T. M., Meacham, R. B., Hoffman, G. S., Seale, C., Wilbanks, J. H., Easley, J. and Carlton, C. E., Jr.: The prognostic significance of post-irradiation biopsy results in patients with prostatic cancer. J. Urol., 135: 510, 1986. 10. Kiesling, V. J., McAninch, J. W., Goebel, J. L. and Agee, R. E.: External beam radiotherapy for adenocarcinoma of the prostate: a clinical followup. J. Urol., 124: 851, 1980. 11. Leach, G. E., Cooper, J. F., Kagan, A. R., Snyder, R. and Forsythe, A.: Radiotherapy for prostatic carcinoma: post-irradiation prostatic biopsy and recurrence patterns with long-term followup. J. Urol., 128: 505, 1982. 12. Hodge, K. K., McNeal, J.E. and Stamey, T. A.: Ultrasound-guided transrectal core biopsies of the palpably abnormal prostate. J. Urol., in press. 13. Hodge, K. K., McNeal, J. E., Terris, M. K. and Stamey, T. A.: Random-systematic versus directed ultrasound-guided transrectal core biopsies of the prostate. J. Urol., in press. 14. Kabalin, J. N., Hodge, K. K., McNeal, J. E., Freiha, F. S. and Stamey, T. A.: Identification of residual cancer in the prostate following radiation therapy: role of transrectal ultrasound guided biopsy and prostate specific antigen. J. Urol., in press. 15. Mahan, D. E., Bruce, A. W., Manley, P. N. and Franchi, L.: Immunohistochemical evaluation of prostatic carcinoma before and after radiotherapy. J. Urol., 124: 488, 1980.
Vol. 141, May Printed in U.S.A.
Copyright © 1989 by The Williams & Wilkins Co.
PROSTATE SPECIFIC ANTIGEN IN THE DIAGNOSIS AND TREATMENT OF ADENOCARCINOMA OF THE PROSTATE. III. RADIATION TREATED PATIENTS THOMAS A. STAMEY,* JOHN N. KABALIN
AND
MICHELLE FERRARI
From the Division of Urology, Stanford University Medical Center, Stanford, California
ABSTRACT
Serum prostate specific antigen was determined (Yang polyclonal radioimmunoassay) in 183 men after radiation therapy for adenocarcinoma of the prostate. A total of 163 men had received 7,000 rad external beam radiotherapy and 20 had been implanted with 125 iodine seeds. Only 11 per cent of these 183 patients had undetectable prostate specific antigen levels at a mean interval of 5 years since completion of radiotherapy. Prostate specific antigen levels after radiotherapy were directly related to initial clinical stage and Gleason score before treatment. Multiple prostate specific antigen determinations were performed with time in 124 of 183 patients. During year 1 after radiotherapy prostate specific antigen levels were decreasing in 82 per cent of the patients but only 8 per cent continued to decrease beyond year 1. Of 80 patients observed greater than 1 year after completion of radiotherapy 51 per cent had increasing values and 41 per cent had stable values. Increasing prostate specific antigen values after radiotherapy were correlated with progression to metastastic disease and residual cancer on prostate biopsy. Total serum acid phosphatase levels were poorly related to prostate specific antigen levels, were less effective in discriminating patients with metastatic disease and provided no additional information beyond that provided by prostate specific antigen. (J. Ural., 141: 1084-1087, 1989) Radiation therapy has been a popular form of treatment for prostate cancer. Actuarial survival rates have been reported for patients treated with linear accelerator supervoltage radiotherapy 1• 2 and the implantation of 125iodine (1 25 1) seeds. 3 However, for the evaluation of status free of disease most radiation therapists have relied upon digital rectal examination and enzymatic serum acid phosphatase measurements. With the recognition of the much greater sensitivity of prostate specific antigen (PSA) in the detection of intracapsular and extracapsular prostate cancer, 4 - 6 we have examined the concentrations and serial patterns of serum PSA in patients who have completed radiation therapy. METHODS
We determined the concentration of serum PSA in 183 patients at various intervals after completion of radiation therapy. Of these patients 163 received external beam radiotherapy with 7,000 rad to the prostate and 5,000 rad to the regional lymph node area during a 7-week period. 1• 2 Twenty patients were implanted with 1251 seeds at a mean total activity of 130 Gy. (range 80 to 200 Gy.). Mean patient age at the most recent serum PSA determination was 69 years. At initiation of radiotherapy cancer stage was assessed as stage A in 25 patients (21 stage A2), stage B in 81, stage C in 52 and stage Dl in 25. Of the 20 patients who received 1251 implantations 15 had clinical stages A and B disease. The average Gleason score on pre-radiotherapy biopsy was 6.6 and 6.3 for external beam and 1251 implants, respectively. For this analysis of PSA the 20 patients with 125 1 seed implants were included in the total 183 patients. Every post-radiotherapy patient seen at our institution who
had blood serum sent to the Stanford urology laboratories for a PSA determination between April 1985 and January 1988 formed the basis for this study, without any other selection bias except the exclusion of all patients who had received any form of anti-androgen therapy and any patient who had undergone transurethral or enucleation prostatectomy within 1 year of the first PSA. Ambulatory serum PSA, determined by the Yang polyclonal radioimmunoassay as in the previous studies,4-6 was obtained at intervals ranging from immediately after completion of radiotherapy up to 258 months (21½ years) after therapy. All assays were done in the Stanford urology laboratories. The mean interval between completion of radiotherapy and serum PSA determination was 61 months (approximately 5 years). The reported PSA values represent the most recent determination if more than 1 was available. Total serum acid phosphatase was determined in the Stanford clinical laboratories with the Bessey-Lowry enzymatic method, a technique used by these laboratories for 30 years. The range for normal men (2 standard deviations) is 0.13 to 0.64 Bessey-Lowry units. RESULTS
Of the 183 patients 11 per cent had undetectable PSA (less than 0.3 ng./ml.) 6 while 25 per cent were within the range of normal male levels (Oto 2.5 ng./ml. 4 ). Thus, 36 per cent of the patients had a recent PSA level within the normal range, with an average followup of 5 years after radiotherapy. However, another 36 per cent had elevated levels between 2.5 and 20 ng./ ml., while 28 per cent exceeded 20 ng./ml. with a mean PSA of 611 ng./ml. These elevations in PSA clearly were related to pre-radiation clinical stage; with mean followups of 51 to 89 months, the mean PSA levels for 25 stage Bl, 21 stage A2, 42 stage B2, 14 stage BS and 52 clinical stage C cancer patients were 7.8, 17, 54, 44 and 505 ng./ml., respectively. A total of 124 patients (107 who received external beam radiation and 17 who received 1251 seed implants) had multiple
Accepted for publication December 28, 1988. Supported in part by the Richard M. Lucas Cancer Foundation and the Flintridge Foundation. * Requests for reprints: Division of Urology, 8287, Stanford University Medical Center, 300 Pasteur Drive, Stanford, California 943055118. 1084
1085
PR0STATE SPECIFIC ANTIGEN
serum PSA determinations with time; the mean interval between consecutive PSA determinations was 9.1 months, with a range of 1 to 31 months. About a third of the patients had 2 consecutive determinations, a third had 3 and a third had 4 or more. Sixteen patients had clinical stage A (13 stage A2), 54 stage B, 34 stage C and 20 stage Dl disease at initiation of radiotherapy. The observed patterns of change in serial PSA values during years 1 and 2 and later are summarized in table 1. A total of 44 patients had multiple PSA determinations during year 1, 21 during year 2 and 59 after 2 years. During year 1 after completion of radiotherapy PSA was decreasing in 82 per cent (36 of 44) of the evaluable patients, including the majority of patients in all clinical disease stages. The mean rate of this decrease was 9. 7 ng./ml. per month. The absolute level of PSA was related to the pattern of change observed. PSA values were decreasing in two-thirds (24 of 36) of the patients with a PSA of 10 ng./ml. or less. Increasing PSA values were observed in an increasing number and percentage of patients with increasing absolute PSA levels. Of the 6 patients whose PSA elevated during the first 12 months 5 not only had the higher clinical stages (B3 or above) but they also had the highest PSA values (greater than 10 ng./ml.). Of the 80 patients with serial PSA values 12 months or more after radiotherapy (table 2) 51 per cent had increasing levels (evenly distributed across all stages) at a mean rate of increase of 28 ng./ml. per month. Only 6 of the 80 patients (8 per cent) continued to have decreasing levels more than 1 year following radiotherapy, and 4 of these 6 were evaluated during year 2. The remaining 2 patients with decreasing PSA were in years 3 and 5 after therapy, with no patient continuing to show decreases beyond 53 months. Additionally, 33 of the 80 evaluable patients (41 per cent) more than 1 year after receiving radiotherapy had stable PSA values (defined as variation less than 1 ng./ml.) during mean interval between PSA determinations of 12 months. The mean absolute PSA value in this stable group was 2.9 ng./ml., with 85 per cent (28 of 33) of these stable values being 2.5 ng./ml. or less (table 2). The data in table 2 also confirm that the 51 per cent of the patients with increasing PSA after 1 year of radiotherapy are heavily weighted toward those with higher serum PSA values (31 of 41 greater than 10 ng./ml.). Gleason scores 7 were available in 159 of the 183 patients at initiation of radiotherapy. Of 75 patients (mean patient age 68 years) with a Gleason score of 6 or less the mean post-radiotherapy PSA was 69 ng./ml., with a mean followup of 68 months. However, of 84 patients (mean age 69 years) with a Gleason score of 7 or greater the mean post-radiotherapy PSA was 268 ng./ml., with a mean followup of 39 months. Table 3 indicates that the patients with the higher Gleason scores before radiotherapy generally are those with the higher postradiotherapy PSA values, while the majority (38 versus 18) of patients with post-radiotherapy PSA values of 2.5 ng./mL or less had pre-radiotherapy Gleason scores of 6 or less. Total acid phosphatase serum levels, obtained at the same time as PSA, were available in 107 of the 183 patients (58 per cent) after radiotherapy. These values are tabulated and crossreferenced with PSA concentrations in tables 4 and 5. There is TABLE
1. Patterns of change in 124 patients with serial PSA values
after radiotherapy PSA Yrs. After Radiotherapy 1 2 >2 Totals
Decreasing No.(%) 36 4 2 42
(82)t (19) (3) (34)
Stable* No.(%) 2 7 26 35
Increasing No.(%)
(5) (33) (44) (28)
* Stable defined as variation of Jess than 1 ng./ml.
t Rate of decrease 9. 7 ng./ml. per month.
6 10 31 47
(14) (48) (53) (38)
Total No.(%) 44 21 59 124
(100) (100) (100) (100)
PSA levels in 80 patients with serial values obtained more than 12 months after radiotherapy
TABLE 2.
PSA Total No. Pts.
PSA {ng./ml.)
% Decreasing (6 pts.)
0.0 >0.0-2.5 >2.5-10 >10-20 >20 Totals
0 14 17 9 0
100 71 25 9 4
0 14 58 82 96
13 21 12 11 23
8
4lt
5it
80
% Stable'· (33 pts.)
% Increasing (41 pts.)
* Stable defined as variation of less than 1 ng./ml.
t Mean PSA 2.9 ng./ml.
t Rate of increase 28 ng./ml. per month. TABLE 3. Distribution of PSA values in 159 patients after radiotherapy compared to Gleason score at initiation of treatment
G Jeason Score PSA (ng./ml.)
;:i6 No.(%)
?;,7 No.(%)
12 26 14 8 15
5 13 20 13 33
0.0 >0.0-2.5 >2.5-10 >10-20 >20
(16) (35) (19) (11) (20)
(6) (15) (24) (15) (39)
TABLE 4. Comparison of mean total acid phosphatase and PSA levels in 107 patients after radiotherapy, broken down by clinical stage before treatment
Stage
No. Pts.
Al A2 Bl B2 B3
1 15 15 26
C
29 13
Dl
Mean Total Acid Phosphatase (Bessey-Lowry units)
Mean PSA {ng./ml.)
0.71 0.73 0.62 0.65 0.62 0.81 0.72
1.9 21.7 7.2 13.2 17.7 127.4 17.3
8
Normal range of total acid phosphatase 0.13 to 0.64 Bessey-Lowry units. Normal range of PSA Oto 2.5 ng./ml.
Distribution of total acid phosphatase levels after radiotherapy compared to PSA levels in 107 patients
TABLE 5.
PSA (ng./ml.)
Total No. Pts.
0.0 >0.0-2.5 >2.5-10 >10-20 >20 Totals
15 33 23 14 22 107
Total Acid Phosphatase (No. pts.) ;:i0,64
>0.64
9 23 18
10 5
6
8
6
7
15 42
65
Mean Total Acid Phosphatase (Bessey-Lowry units) 0.64 0.60 0.60 0.64 1.07
Normal range of total acid phosphatase 0.13 to 0.64 Bessey-Lowry units. Normal range of PSA Oto 2.5 ng./ml.
minimal correlation between PSA and total acid phosphatase values. The representative and important responses of PSA to radiotherapy are shown in figures 1 to 3. The patient in figure 1 is illustrative of the 82 per cent of those whose serum PSA decreases during year 1 of radiotherapy, even from rather substantial elevations (230 ng./ml.). Figure 2 also is representative of a common pattern, decreasing during year 1 to normal values but increasing steadily thereafter. Figure 3 represents radiation-resistant cancer, a slow decrease in PSA that never approaches normal and is followed by a rapid return to preradiation levels.
1086
STAMEY AND ASSOCIATES
PSA (ng/ml) 230
PSA (ng/ml)
\
100
100 49.8 -~
42.2
27.7
\
~15.3
12.8
"\
10
;·
· ~0. 6 -~4
10
•,\
5.4
2.2
•
0
20
# MONTHS FOLLOWING RADIOTHERAPY FIG. 1. PSA levels in 71-year-old man with clinical stage B2 nodule,5 Gleason score 7 (3 plus 4) on biopsy. Pre-radiation PSA level was 230 ng./ml. Computerized tomography and pedal lymphangiograms failed to show any lymph nodes. Within 2 weeks of completing 5,000 rad to pelvis and 7,000 rad to prostate, serum PSA decreased to 27.7 ng./ml., 5 months later it was 2.2 ng./ml. and 15 months later PSA was 1.5 ng./ ml. PSA normal range O to 2.5 ng./ml.
PSA (ng/ml)
100
13.5 10
20
0
# MONTHS FOLLOWING RADIOTHERAPY
FIG. 2. Findings in 75-year-old man with clinical stage C prostate cancer,5 Gleason score 7 (3 plus 4) on biopsy. PSA was 13.5 ng./ml. before receiving 7,000 rad to prostate. Six months after completing radiotherapy PSA was 1.6 ng./ml. but level gradually increased in next 13 months and 13 months after 12. 7 entry it increased to 57 ng./ml. PSA normal range O to 2.5 ng./ml.
DISCUSSION
With the Yang polyclonal radioimmunoassay we obtained blood serum for PSA from 183 patients at a mean interval of 61 months after completion of radiotherapy. Each patient had been treated only with irradiation; patients who had been placed on anti-androgen therapy at any time or who had
0
20 # MONTHS FOLLOWING RADIOTHERAPY
FIG. 3. Findings in 78-year-old man with clinical stage C prostate cancer, Gleason score 9 (4 plus 5). Patient received 7,000 rad immediately after serum PSA of 49.8 ng./ml. was obtained. PSA gradually decreased to low of 8.4 ng./ml. 15 months after radiation therapy but then increased 5-fold to 42.2 ng./ml. during next 5 months. PSA normal range Oto 2.5 ng./ml.
undergone transurethral resection of the prostate within 1 year of the first PSA determination were excluded from this analysis. Only 11 per cent of these 183 patients had undetectable PSA levels but an additional 25 per cent were within the normal range (2.5 ng./ml. or less). The remaining 64 per cent of the patients all had abnormal elevations of serum PSA. PSA levels after radiotherapy were related directly to initial clinical stage and Gleason score before treatment. Multiple PSA determinations were performed with time in 124 of 183 patients (table 1). We have established that PSA levels decreased in 82 per cent of irradiated patients during the first 12 months after therapy but only 6 of 80 patients (8 per cent) continued to have decreasing levels beyond year 1. It is of some interest that this decrement in PSA during year 1 occurred regardless of the clinical stage at initiation of therapy and that the mean rate of decrease was 9.7 ng./ml. per month. Of more importance are the patterns of change in 80 patients observed greater than 1 year after completion of radiotherapy (table 2). Of the patients 51 per cent had increasing PSA values, occurring at all clinical stages with a mean rate of increase of 28 ng./ml. per month. These patients with increasing values are distributed towards those with the higher final concentrations of PSA, while the 41 per cent of the 80 patients with stable PSA (PSA values that vary less than 1 ng./ml., 85 per cent of which were less than 2.5 ng./ml.) are clearly distributed towards the lower PSA concentrations (mean 2.9 ng./ml.). Increasing PSA values after radiotherapy correlated with progression to metastatic disease and residual cancer on prostate biopsy. In 27 of 183 irradiated patients we documented progression to clinical stage D2 and we obtained multiple PSA values on 16 (59 per cent). All 16 patients had increasing PSA values before progression to clinical stage D2. More convincingly, we biopsied 9 of the 41 patients (22 per cent) with increasing PSA values 12 months after completion of therapy. Despite the observation that 8 of these 9 prostates were palpably flat, symmetrical and normal on digital rectal examination before biopsy, all 9 were positive for prostate cancer. While elevated levels as well as increasing PSA concentra-
1087
PROSTATE SPECIFIC ANTIGEN
tions with time strongly suggest failure of radiation therapy to cure prostate cancer, positive biopsies are the best evidence for failure. 8- 11 We recently showed that multiple, ultrasoundguided, transrectal biopsies can be performed safely, painlessly and quickly with minimal morbidity in an outpatient setting. 12 Random systematic biopsies can be used to obtain 1.5 cm. cores from apex to base on both sides of the prostate and seminal vesicles. 13 With this technique asymptomatic, post-irradiation patients (mean 5 years) with palpably normal prostates can have substantial residual cancer in the presence of a normal PSA. 14 Thus, while increasing PSA concentrations after radiotherapy surely imply failure to cure the cancer, normal serum concentrations of PSA, even when accompanied by a flat or palpably normal prostate, cannot be relied upon to exclude the presence of residual cancer. It is important to emphasize that irradiation of the prostate drastically changes the significance of the absolute serum PSA value and that it cannot be interpreted by comparing postradiotherapy values to the database derived from untreated patients. 4 •5 For example, the lowest PSA values we have seen in untreated patients with documented metastasis to bone (mean PSA 563 ng./ml.) 5 were in 2 or 3 rare patients in the 30 to 40 ng./ml. range. By contrast, of 18 post-radiotherapy patients who have progression to positive bone scans and in whom a PSA value was obtained within 3 months of that scan 4 had serum PSA concentrations of 8, 10, 19 and 20 ng./ml. Thus, it appears that metastases can occur at low serum levels of PSA after irradiation to the prostate and pelvis. This finding suggests that the bulk of the tumor burden in untreated stage D2 disease still is within the radiated prostate and pelvis. In a unique study, Mahan and associates examined prostate cancer biopsy specimens from the same group of 27 patients before and after radiation therapy. 15 Although all 27 cancers in their series stained positive for acid phosphatase before irradiation 3 failed to stain and 3 showed only equivocal staining after irradiation. If acid phosphatase production is absent ifi some irradiated prostate cancers, then it is logical to assume that PSA similarly may be absent or diminished in a minority of irradiated cancers. This might explain the unusually low PSA values sometimes observed following radiation therapy despite demonstrable residual cancer and even distant metastases. Much remains to be learned regarding the exact nature of the biological effect of radiation on prostate cancer in general, and on the known biochemical markers for prostate cancer in specific. Whatever the explanation, a high index of suspicion must be maintained for disease that may be much more advanced than the serum PSA might indicate in any patient whose prostate has been irradiated. It is interesting, however, that once the concentration of serum PSA increased to more than 50 ng./ml. in the postirradiation patient (16 of 23 patients in table 2) subsequent increases occurred at a mean rate of 138 ng./ml. per month, which is the same as we observed in all 16 untreated stage D2 cancer patients whose initial PSA also was more than 50 ng./ ml.5 Radiation therapists have used for their definition of no evidence of disease not only a flat, smooth, nonirregular prostate on digital rectal examination but also a normal total serum acid phosphatase. 1 • 2 However, there is almost no correlation between total serum acid phosphatase and PSA concentrations. In fact, in 4 consecutive ranges of increasing PSA values (table 5) mean total acid phosphatase remained constant between 0.60 and 0.64 (normal range for total acid phosphatase 0.13 to 0.64 Bessey-Lowry units), except for patients with the highest PSA levels (greater than 20 ng./ml.), which was the only group with an elevated mean total acid phosphatase. Additionally, of the 27 men in whom we documented progression to stage D2 after radiotherapy, we obtained total acid phosphatase values at the time of PSA in 9. As noted previously, all 9 men had elevated PSA levels, with a minimum value of 8.2 ng./ml.
(greater than 3 times normal). However, 2 of these 9 patients had normal total acid phosphatase levels. We must conclude that not only is PSA of superior use in comparison to total acid phosphatase to monitor the post-radiotherapy patient but that total acid phosphatase adds no useful information beyond that provided by serum PSA. Finally, we believe that serial PSA values, especially when combined with multiple, transrectal, ultrasound guided 1.5 cm. cores (prostate mapping), 12- 1• will allow an infinitely more accurate assessment of residual cancer 18 months after completion of radiotherapy. Our observations that 53 per cent of the men greater than 2 years after radiotherapy have increasing levels of PSA and that ultrasound guided, systematic biopsies of even palpably normal prostates frequently are positive in the presence of normal serum levels of PSA 14 strongly suggest that there is a higher probability of persistent cancer after radiotherapy than generally is recognized. It also is likely that the 25 per cent 15-year survival rates reported for stages B2 and B3 nodules' come from the 44 per cent of men with stable serum levels of PSA, 85 per cent of whom have PSA concentrations of O to 2.5 ng./ ml. REFERENCES
1. Bagshaw, M. A., Cox, R. S. and Ray, G. R.: Status of radiation treatment of prostate cancer at Stanford University. Natl. Cancer Inst. Monogr., 7: 47, 1988. 2. Bagshaw, M.A., Ray, G. R. and Cox, R. S.: Selecting initial therapy for prostate cancer: radiation therapy perspective. Cancer, 60: 521, 1987. 3. Whitmore, W. F., Jr., Hilaris, B., Batata, M., Sogani, P., Herr, H. and Morse, M.: Interstitial radiation: short-term palliation or curative therapy? Urology, suppl. 2, 25: 24, 1985. 4. Stamey, T. A., Yang, N., Hay, A. R., McNeal, J.E., Freiha, F. S. and Redwine, E. R.: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. New Engl. J. Med., 317: 909, 1987. 5. Stamey, T. A. and Kabalin, J. N.: Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. I. Untreated patients. J. Urol., 141: 1070, 1989. 6. Stamey, T. A., Kabalin, J. N., McNeal, J. E., Johnstone, I. M., Freiha, F., Redwine, E. A. and Yang, N.: Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. II. Radical prostatectomy patients. J. Urol., 141: 1076, 1989. 7. Gleason, D. F.: Histologic grading and clinical staging of prostatic carcinoma. In: Urologic Pathology. Edited by M. Tannenbaum. Philadelphia: Lea & Febiger, part II, chapt. 9, pp. 171-198, 1977. 8. Freiha, F. S. and Bagshaw, M. A.: Carcinoma of the prostate: results of post-irradiation biopsy. Prostate, 5: 19, 1984. 9. Scardino, P. T., Frankel, J.M., Wheeler, T. M., Meacham, R. B., Hoffman, G. S., Seale, C., Wilbanks, J. H., Easley, J. and Carlton, C. E., Jr.: The prognostic significance of post-irradiation biopsy results in patients with prostatic cancer. J. Urol., 135: 510, 1986. 10. Kiesling, V. J., McAninch, J. W., Goebel, J. L. and Agee, R. E.: External beam radiotherapy for adenocarcinoma of the prostate: a clinical followup. J. Urol., 124: 851, 1980. 11. Leach, G. E., Cooper, J. F., Kagan, A. R., Snyder, R. and Forsythe, A.: Radiotherapy for prostatic carcinoma: post-irradiation prostatic biopsy and recurrence patterns with long-term followup. J. Urol., 128: 505, 1982. 12. Hodge, K. K., McNeal, J.E. and Stamey, T. A.: Ultrasound-guided transrectal core biopsies of the palpably abnormal prostate. J. Urol., in press. 13. Hodge, K. K., McNeal, J. E., Terris, M. K. and Stamey, T. A.: Random-systematic versus directed ultrasound-guided transrectal core biopsies of the prostate. J. Urol., in press. 14. Kabalin, J. N., Hodge, K. K., McNeal, J. E., Freiha, F. S. and Stamey, T. A.: Identification of residual cancer in the prostate following radiation therapy: role of transrectal ultrasound guided biopsy and prostate specific antigen. J. Urol., in press. 15. Mahan, D. E., Bruce, A. W., Manley, P. N. and Franchi, L.: Immunohistochemical evaluation of prostatic carcinoma before and after radiotherapy. J. Urol., 124: 488, 1980.