I,,!. J Radmrrm OnwIo~v tJ,ol Phv Vol Printed in the USA All rights reserved
19. pp. 1383-1388
Cupynght
0360.3016/YO $3.W + .Xl cc 1990 Pergamon Press plc
??Original Contribution
THE ROLE OF SERUM PROSTATIC ACID PHOSPHATASE IN THE MANAGEMENT OF ADENOCARCINOMA OF THE PROSTATE WITH RADIOTHERAPY JEFFREY
C.
CARLTON,
Department
M.D.,
GUNAR
K.
ZAGARS,
M.D.
AND
MARY
of Clinical Radiotherapy, The University of Texas, M. D. Anderson 15 15 Holcombe Boulevard. Houston, TX 77030
J. OSWALD,
B.S.
Cancer Center,
Between 1974 and 1983,472 patients with clinically-staged adenocarcinoma of the prostate treated by radiotherapy had baseline and follow-up prostatic acid phosphatase (SPAP) measurements by the enzymatic Roy method. The mean pretreatment SPAP was higher in Stage C (0.65 mIU/ml) than in combined Stages A2/B (0.43 mIU/ml), (p < 0.05). Likewise, the incidence of elevated SPAP (>0.8 mIU/ml) was also higher in Stage C (12%) than in Stages A2/B (3%), (p < 0.01). Only 3 of 113 patients in Stages A2/B had an elevated SPAP and all three remain disease-free. In Stage C elevated SPAP was an adverse prognostic factor, and patients with a normal SPAP fared worse if their value was in the upper half of normal (>0.4 mIU/ml) rather than in the lower half (10.4 mIU/ml). However, in Stage C, tumor grade was found to correlate with initial SPAP, so that the higher the grade, the higher was the mean SPAP and the greater was the incidence of elevated SPAP. When stratified for grade, the prognostic significance of low-normal versus high-normal SPAP in Stage C was lost. An elevated SPAP was, however, an independent adverse prognostic factor for patients with intermediate and high grade tumors. Following radiotherapy, mean SPAP values fell significantly within l-3 months. For patients with initially normal SPAP, this fall was of no prognostic significance. In 80% of the patients with baseline elevation of SPAP, the values normalized following treatment and the relapse rate in these patients was 51%, which was still higher than the relapse rate of patients with initially normal SPAP (33%) (p < 0.05) but was lower than the 89% relapse rate in patients whose postradiation SPAP did not normalize (p < 0.05). Pretreatment SPAP was of independent prognostic significance for only 6% of the study population and therefore has quite limited usefulness in the management of this disease. SPAP decreases following radiotherapy, but this is of prognostic significance only for the small group of patients with elevated pretreatment values. Carcinoma
of prostate,
Radiotherapy,
Prostatic
acid phosphatase.
in 11 of 15 patients with metastatic adenocarcinoma of the prostate (11). Shortly thereafter, investigators demonstrated a decrease in previously elevated acid phosphatase values following hormonal manipulation paralleling the clinical response of metastatic disease ( 14). Since that time, the measurement of serum activity of acid phosphatase has become well established as a useful monitor for response of metastatic prostate cancer to systemic therapy (18). With the development of assay methods more specific for the prostatic component of serum acid phosphatase, investigators have attempted to define more clearly the role of this marker in the management of early prostatic cancer. However. despite the long use of more specific enzymatic assays for
serum prostatic acid phosphatase (SPAP) and the more recent introduction of radioimmunoassay techniques, a clear role for SPAP in early prostate cancer has not been adequately defined. The significance of an elevated SPAP in patients with tumor clinically confined to the prostate and periprostatic tissues remains controversial. Surgical series report metastatic nodal disease at pelvic lymphadenectomy in 50 to 100% of patients with preoperative clinically localized disease and elevated SPAP (8. 10, 15. 26). This finding, along with an observed high relapse rate for clinicallystaged patients with apparently localized disease but elevated SPAP, has prompted some investigators to conclude that such patients are not curable and should be offered hormonal management (1. 26). Whitesel et ul. recommended a new metastatic staging category, Do, for such patients (26).
Reprint requests to: Gunar K. Zagars, M.D. ilcknclwled~ment-Supported in part by grant CA-06294 awarded by the National Cancer Institute. U.S. Department of
Health and Human Services, and the American Clinical Fellowship Award. Accepted for publication 2 I June IWO.
INTRODUCTION In 1938 Gutman of serum
and Gutman
reported
finding
elevation
acid phosphatase
1383
Cancer Society
1384
1. J. Radiation Oncology 0 Biology 0 Physics
With the advent of more specific assay techniques, the significance of high-normal SPAP versus low-normal SPAP has received attention. Several series have reported a worse prognosis for patients with a pretreatment SPAP in the upper half of the normal range compared to those with values in the lower half (2, 6, 19, 23). However, the relationship of such SPAP values to other prognostic variables such as grade and stage has not been adequately defined. Moreover. changes in SPAP following local-regional therapy have not been well-documented and the potential prognostic importance of early post-therapy SPAP has not been defined. Radiotherapy is the best established curative treatment modality for Stage C prostate cancer and is experiencing an increasing role in the management of Stage A2 and B disease. This retrospective review of 472 patients with Stages A2, B, and C prostatic cancer treated with external beam radiotherapy (EBRT) alone was performed to define better the role of SPAP in the radiotherapeutic management of this disease.
MATERIALS
AND
METHODS
This series includes all patients with clinical Stage A2, B, and C prostatic adenocarcinoma who were treated definitively with external beam radiotherapy at U.T. M. D. Anderson Cancer Center (UT MDACC) between 1974 and 1983, who had no hormonal manipulation prior to relapse, and who had pretreatment and follow-up SPAP measurements performed by the Roy et al. method (22). Four hundred seventy-two patients fulfilled all these criteria and are the subject of this report. Stage A2 disease was defined as multifocal or diffuse carcinoma not evident on rectal palpation of the prostate. Stage B lesions were palpable without palpatory evidence of extension beyond the prostate, whereas Stage C disease was defined as tumor extension beyond the prostatic capsule on digital examination. Stage C included disease extension to the bladder, to the seminal vesicles, or to the pelvic sidewall. Of the 472 patients, 46 were staged as having A2 disease, 67 as having Stage B, and 359 as having Stage C. The mean and median ages for the population were 65 and 66 years, respectively, with a range from 44 to 80 years of age. None of the patients had evidence of lymphatic or hematogenous metastases by staging evaluation prior to start of radiotherapy. Radionuclide bone scans were performed in 388 of the cases (82%). Bipedal lymphangiography was performed in 388 of the cases (82%), and no patient underwent staging lymphadenectomy. All patients had a histological confirmation of prostatic adenocarcinoma by pathologists at our institution. Histologic grading by the M. D. Anderson system, the MDA grade (4). was available for 390 patients (83%). All prostatic acid phosphatase measurements were determined at our institution by enzymatic assay using thymolphthalein monophosphate as a substrate according to
December 1990. Volume 19. Number 6
the method described by Roy et ul. (23). Normal range has been determined to include values up to 0.80 mIU/ml. External beam radiotherapy was delivered using high energy photon beams ( 18-25 MV) according to techniques described previously (27, 28). Following treatment, most patients were seen regularly at U.T. M.D.A.C.C. for follow-up evaluation. The remaining patients were followed by their referring urologists, and follow-up information was obtained from these physicians or from correspondence with the patient. In general, follow-up evaluations were performed every 3 months for 2 to 3 years, then every 6 months up to 5 years, and then annually thereafter. Routine follow-up evaluation consisted of a digital rectal exam, pelvic and lumbar spine radiograph, routine blood studies, and serum SPAP. Radionuclide bone scans were routinely performed annually or more often when clinically indicated. The duration of follow-up for the 338 patients alive at the time of analysis ranged from 7 to 148 months with a mean of 5.6 years and a median of 5.1 years. Only eight patients were followed less than 3 years because of loss of contact. Datu analNs rnethou’s Disease-free survival (DFS) curves were calculated using the actuarial Berkson and Gage method (3). These curves represent the probability that a patient is free of evidence of disease on the assumption he is alive at the time in question. This does not represent the probability that the patient is both alive and free from disease (27,28). Diseasefree survival was selected as an endpoint in this study, as overall survival is influenced heavily by intercurrent disease in this elderly population. A limited analysis using survival as an endpoint revealed no major departures from the conclusions derived from DFS analysis. An increasing SPAP or a change from a normal to an elevated value was not considered an adequate reason to score a patient as a treatment failure. Such patients were considered disease-free until there was radiographic, bone scan, or-histologic evidence of relapse. Local control was defined as a prostate gland persistently normal by palpation. even in the presence of obstructive symptoms if transurethral resection of the prostate (TURP) revealed no cancer. To allow for the slow regression of the prostatic lesion after EBRT, we adopted the following convention. All patients who achieved complete clinical regression of their primary tumor before beginning hormonal management for metastatic disease were scored as being in remission at the time EBRT began. Those patients who never achieved complete clinical regression were scored as local failures at the commencement of EBRT. Tests of significance between actuarial curves were done with the log-rank statistic (7), and tests of significance for differences between proportions were done with the Chisquare statistic or the one-sided Fisher exact test. Statistical
Prostatic acid phosphatase 0 J. C. CARLTON elal
comparisons of means were performed using T-tests by the pooled-variance estimate method. All follow-up times were calculated from the start of EBRT.
RESULTS A histogram displaying the frequency distribution of pretreatment SPAP values is shown in Figure I. For values within the normal range, the distribution is approximately Gaussian. Four hundred twenty-five patients had pretreatment SPAP values within the normal range (O-0.80 mIU/ml), and 47 patients had elevated values. In the 47 patients with elevated SPAP. the values ranged from 0.82 to 27.32 mIU/ml, and the median elevated SPAP value was 1.33 mIU/ml. The incidence of elevated SPAP by stage was as follows: Stage A2. 0 of 46; Stage B. 3 of 67 (4%); Stage C 44 of 359 ( 12%). The incidence of elevated SPAP in Stage C disease was significantly greater than the incidence of elevated SPAP in Stages A2 and B combined (p = ~0.0 1). In keeping with this finding, the mean SPAP value in all patients with Stage C disease (0.65 mIU/ml) was significantly higher than the mean value of (0.43 mIU/ ml) in patients with Stages A2 or B (p = ~0.05). Pretreatment SPAP was of no prognostic significance in patients with Stages A2 or B disease since the three patients with elevated values ( 1.09 to 3.83 mIU/ml) have remained continuously disease-free. The 5-year DFS rate for patients with Stages A2 or B was 9 1% (89% for A2, 93% for B). For Stage C disease, however, pretreatment SPAP was a significant prognostic factor. Figure 2 shows that the 5year DFS rate was 59% for those patients with normal SPAP value and 38%) for those with an elevated value (p = ~0.02). Of the 44 patients in stage C with elevated SPAP, 38 (86%) were treated with fields that encompassed the pelvic lymph nodes. whereas 25 1 of 3 15 patients (80%) with normal SPAP received elective pelvic node treatment. Thus, the extent of the treatment fields was approximately in the two patient groups. Moreover. even comparable
__
SPAP Elevated
1
20
No
_.......
yes
Oc----1 0
1
Time
Fig. 2. Disease-free disease: pretreatment I 0.02).
48
24
12
80
in Monk
survival curves for patients with Stage C SPAP in normal range versus elevated (p
for patients whose SPAP was normal, we found that values in the upper half of the normal range (0.80 2 SPAP > 0.4) were associated with a significantly poorer 5 year DFS rate (53%) than values in the lower half of the normal range (DFS 65%) (p = ~0.01) (Fig. 3). No such correlation was found for Stages A2 and B. Within the Stage C group. initial SPAP was found to correlate with tumor grade so that the higher the grade, the higher the SPAP value and the greater was the incidence of elevated SPAP values (see Fig. 4). For M. D. Anderson grade 4 lesions, 8/33 (24%) had elevated SPAP values as compared with only 6/89 (7%) elevated SPAP in patients with M. D. Anderson grade 1 (p = 0.017). Similarly. an SPAP I 0.4 was found in 55189 (62%) patients with grade 1 lesions, whereas only lo/33 (30%) patients with grade 4 lesions had such low SPAP values (p = 13
(472
Oi
0
I
40
12
T21e
80
in MonZs
Fig. 3. Disease-free survival curves of patients with Stage C disease and normal pretreatment SPAP: SPAP in lower half of normal range versus SPAP in upper half of normal range (p 4 0.01).
1386
I. J. Radiation Oncology 0 Biology 0 Physics SPAP Value 0 -
MDA 1
0.4
MDA 2-3
.41 -
0.8
> 0.8
MDA 4
Grade
Fig. 4. Correlation between pretreatment SPAP and M. D. Anderson grade for patients with Stage C disease (p < 0.05).
SPAP category, prognosis worsened as grade increased. However, for low grade lesions (MDA grade l), SPAP was of no prognostic significance. For intermediate (MDA grades 2 and 3 combined) and high-grade lesions (MDA grade 4), SPAP values within the normal range had no prognostic significance. For intermediate and high-grade lesions, elevated SPAP, however, appears to predict independently for a worse outcome. For grades 2 and 3, the difference in DFS at 5 years was statistically significant (p =
19. pp. 1383-1388
Cupynght
0360.3016/YO $3.W + .Xl cc 1990 Pergamon Press plc
??Original Contribution
THE ROLE OF SERUM PROSTATIC ACID PHOSPHATASE IN THE MANAGEMENT OF ADENOCARCINOMA OF THE PROSTATE WITH RADIOTHERAPY JEFFREY
C.
CARLTON,
Department
M.D.,
GUNAR
K.
ZAGARS,
M.D.
AND
MARY
of Clinical Radiotherapy, The University of Texas, M. D. Anderson 15 15 Holcombe Boulevard. Houston, TX 77030
J. OSWALD,
B.S.
Cancer Center,
Between 1974 and 1983,472 patients with clinically-staged adenocarcinoma of the prostate treated by radiotherapy had baseline and follow-up prostatic acid phosphatase (SPAP) measurements by the enzymatic Roy method. The mean pretreatment SPAP was higher in Stage C (0.65 mIU/ml) than in combined Stages A2/B (0.43 mIU/ml), (p < 0.05). Likewise, the incidence of elevated SPAP (>0.8 mIU/ml) was also higher in Stage C (12%) than in Stages A2/B (3%), (p < 0.01). Only 3 of 113 patients in Stages A2/B had an elevated SPAP and all three remain disease-free. In Stage C elevated SPAP was an adverse prognostic factor, and patients with a normal SPAP fared worse if their value was in the upper half of normal (>0.4 mIU/ml) rather than in the lower half (10.4 mIU/ml). However, in Stage C, tumor grade was found to correlate with initial SPAP, so that the higher the grade, the higher was the mean SPAP and the greater was the incidence of elevated SPAP. When stratified for grade, the prognostic significance of low-normal versus high-normal SPAP in Stage C was lost. An elevated SPAP was, however, an independent adverse prognostic factor for patients with intermediate and high grade tumors. Following radiotherapy, mean SPAP values fell significantly within l-3 months. For patients with initially normal SPAP, this fall was of no prognostic significance. In 80% of the patients with baseline elevation of SPAP, the values normalized following treatment and the relapse rate in these patients was 51%, which was still higher than the relapse rate of patients with initially normal SPAP (33%) (p < 0.05) but was lower than the 89% relapse rate in patients whose postradiation SPAP did not normalize (p < 0.05). Pretreatment SPAP was of independent prognostic significance for only 6% of the study population and therefore has quite limited usefulness in the management of this disease. SPAP decreases following radiotherapy, but this is of prognostic significance only for the small group of patients with elevated pretreatment values. Carcinoma
of prostate,
Radiotherapy,
Prostatic
acid phosphatase.
in 11 of 15 patients with metastatic adenocarcinoma of the prostate (11). Shortly thereafter, investigators demonstrated a decrease in previously elevated acid phosphatase values following hormonal manipulation paralleling the clinical response of metastatic disease ( 14). Since that time, the measurement of serum activity of acid phosphatase has become well established as a useful monitor for response of metastatic prostate cancer to systemic therapy (18). With the development of assay methods more specific for the prostatic component of serum acid phosphatase, investigators have attempted to define more clearly the role of this marker in the management of early prostatic cancer. However. despite the long use of more specific enzymatic assays for
serum prostatic acid phosphatase (SPAP) and the more recent introduction of radioimmunoassay techniques, a clear role for SPAP in early prostate cancer has not been adequately defined. The significance of an elevated SPAP in patients with tumor clinically confined to the prostate and periprostatic tissues remains controversial. Surgical series report metastatic nodal disease at pelvic lymphadenectomy in 50 to 100% of patients with preoperative clinically localized disease and elevated SPAP (8. 10, 15. 26). This finding, along with an observed high relapse rate for clinicallystaged patients with apparently localized disease but elevated SPAP, has prompted some investigators to conclude that such patients are not curable and should be offered hormonal management (1. 26). Whitesel et ul. recommended a new metastatic staging category, Do, for such patients (26).
Reprint requests to: Gunar K. Zagars, M.D. ilcknclwled~ment-Supported in part by grant CA-06294 awarded by the National Cancer Institute. U.S. Department of
Health and Human Services, and the American Clinical Fellowship Award. Accepted for publication 2 I June IWO.
INTRODUCTION In 1938 Gutman of serum
and Gutman
reported
finding
elevation
acid phosphatase
1383
Cancer Society
1384
1. J. Radiation Oncology 0 Biology 0 Physics
With the advent of more specific assay techniques, the significance of high-normal SPAP versus low-normal SPAP has received attention. Several series have reported a worse prognosis for patients with a pretreatment SPAP in the upper half of the normal range compared to those with values in the lower half (2, 6, 19, 23). However, the relationship of such SPAP values to other prognostic variables such as grade and stage has not been adequately defined. Moreover. changes in SPAP following local-regional therapy have not been well-documented and the potential prognostic importance of early post-therapy SPAP has not been defined. Radiotherapy is the best established curative treatment modality for Stage C prostate cancer and is experiencing an increasing role in the management of Stage A2 and B disease. This retrospective review of 472 patients with Stages A2, B, and C prostatic cancer treated with external beam radiotherapy (EBRT) alone was performed to define better the role of SPAP in the radiotherapeutic management of this disease.
MATERIALS
AND
METHODS
This series includes all patients with clinical Stage A2, B, and C prostatic adenocarcinoma who were treated definitively with external beam radiotherapy at U.T. M. D. Anderson Cancer Center (UT MDACC) between 1974 and 1983, who had no hormonal manipulation prior to relapse, and who had pretreatment and follow-up SPAP measurements performed by the Roy et al. method (22). Four hundred seventy-two patients fulfilled all these criteria and are the subject of this report. Stage A2 disease was defined as multifocal or diffuse carcinoma not evident on rectal palpation of the prostate. Stage B lesions were palpable without palpatory evidence of extension beyond the prostate, whereas Stage C disease was defined as tumor extension beyond the prostatic capsule on digital examination. Stage C included disease extension to the bladder, to the seminal vesicles, or to the pelvic sidewall. Of the 472 patients, 46 were staged as having A2 disease, 67 as having Stage B, and 359 as having Stage C. The mean and median ages for the population were 65 and 66 years, respectively, with a range from 44 to 80 years of age. None of the patients had evidence of lymphatic or hematogenous metastases by staging evaluation prior to start of radiotherapy. Radionuclide bone scans were performed in 388 of the cases (82%). Bipedal lymphangiography was performed in 388 of the cases (82%), and no patient underwent staging lymphadenectomy. All patients had a histological confirmation of prostatic adenocarcinoma by pathologists at our institution. Histologic grading by the M. D. Anderson system, the MDA grade (4). was available for 390 patients (83%). All prostatic acid phosphatase measurements were determined at our institution by enzymatic assay using thymolphthalein monophosphate as a substrate according to
December 1990. Volume 19. Number 6
the method described by Roy et ul. (23). Normal range has been determined to include values up to 0.80 mIU/ml. External beam radiotherapy was delivered using high energy photon beams ( 18-25 MV) according to techniques described previously (27, 28). Following treatment, most patients were seen regularly at U.T. M.D.A.C.C. for follow-up evaluation. The remaining patients were followed by their referring urologists, and follow-up information was obtained from these physicians or from correspondence with the patient. In general, follow-up evaluations were performed every 3 months for 2 to 3 years, then every 6 months up to 5 years, and then annually thereafter. Routine follow-up evaluation consisted of a digital rectal exam, pelvic and lumbar spine radiograph, routine blood studies, and serum SPAP. Radionuclide bone scans were routinely performed annually or more often when clinically indicated. The duration of follow-up for the 338 patients alive at the time of analysis ranged from 7 to 148 months with a mean of 5.6 years and a median of 5.1 years. Only eight patients were followed less than 3 years because of loss of contact. Datu analNs rnethou’s Disease-free survival (DFS) curves were calculated using the actuarial Berkson and Gage method (3). These curves represent the probability that a patient is free of evidence of disease on the assumption he is alive at the time in question. This does not represent the probability that the patient is both alive and free from disease (27,28). Diseasefree survival was selected as an endpoint in this study, as overall survival is influenced heavily by intercurrent disease in this elderly population. A limited analysis using survival as an endpoint revealed no major departures from the conclusions derived from DFS analysis. An increasing SPAP or a change from a normal to an elevated value was not considered an adequate reason to score a patient as a treatment failure. Such patients were considered disease-free until there was radiographic, bone scan, or-histologic evidence of relapse. Local control was defined as a prostate gland persistently normal by palpation. even in the presence of obstructive symptoms if transurethral resection of the prostate (TURP) revealed no cancer. To allow for the slow regression of the prostatic lesion after EBRT, we adopted the following convention. All patients who achieved complete clinical regression of their primary tumor before beginning hormonal management for metastatic disease were scored as being in remission at the time EBRT began. Those patients who never achieved complete clinical regression were scored as local failures at the commencement of EBRT. Tests of significance between actuarial curves were done with the log-rank statistic (7), and tests of significance for differences between proportions were done with the Chisquare statistic or the one-sided Fisher exact test. Statistical
Prostatic acid phosphatase 0 J. C. CARLTON elal
comparisons of means were performed using T-tests by the pooled-variance estimate method. All follow-up times were calculated from the start of EBRT.
RESULTS A histogram displaying the frequency distribution of pretreatment SPAP values is shown in Figure I. For values within the normal range, the distribution is approximately Gaussian. Four hundred twenty-five patients had pretreatment SPAP values within the normal range (O-0.80 mIU/ml), and 47 patients had elevated values. In the 47 patients with elevated SPAP. the values ranged from 0.82 to 27.32 mIU/ml, and the median elevated SPAP value was 1.33 mIU/ml. The incidence of elevated SPAP by stage was as follows: Stage A2. 0 of 46; Stage B. 3 of 67 (4%); Stage C 44 of 359 ( 12%). The incidence of elevated SPAP in Stage C disease was significantly greater than the incidence of elevated SPAP in Stages A2 and B combined (p = ~0.0 1). In keeping with this finding, the mean SPAP value in all patients with Stage C disease (0.65 mIU/ml) was significantly higher than the mean value of (0.43 mIU/ ml) in patients with Stages A2 or B (p = ~0.05). Pretreatment SPAP was of no prognostic significance in patients with Stages A2 or B disease since the three patients with elevated values ( 1.09 to 3.83 mIU/ml) have remained continuously disease-free. The 5-year DFS rate for patients with Stages A2 or B was 9 1% (89% for A2, 93% for B). For Stage C disease, however, pretreatment SPAP was a significant prognostic factor. Figure 2 shows that the 5year DFS rate was 59% for those patients with normal SPAP value and 38%) for those with an elevated value (p = ~0.02). Of the 44 patients in stage C with elevated SPAP, 38 (86%) were treated with fields that encompassed the pelvic lymph nodes. whereas 25 1 of 3 15 patients (80%) with normal SPAP received elective pelvic node treatment. Thus, the extent of the treatment fields was approximately in the two patient groups. Moreover. even comparable
__
SPAP Elevated
1
20
No
_.......
yes
Oc----1 0
1
Time
Fig. 2. Disease-free disease: pretreatment I 0.02).
48
24
12
80
in Monk
survival curves for patients with Stage C SPAP in normal range versus elevated (p
for patients whose SPAP was normal, we found that values in the upper half of the normal range (0.80 2 SPAP > 0.4) were associated with a significantly poorer 5 year DFS rate (53%) than values in the lower half of the normal range (DFS 65%) (p = ~0.01) (Fig. 3). No such correlation was found for Stages A2 and B. Within the Stage C group. initial SPAP was found to correlate with tumor grade so that the higher the grade, the higher the SPAP value and the greater was the incidence of elevated SPAP values (see Fig. 4). For M. D. Anderson grade 4 lesions, 8/33 (24%) had elevated SPAP values as compared with only 6/89 (7%) elevated SPAP in patients with M. D. Anderson grade 1 (p = 0.017). Similarly. an SPAP I 0.4 was found in 55189 (62%) patients with grade 1 lesions, whereas only lo/33 (30%) patients with grade 4 lesions had such low SPAP values (p = 13
(472
Oi
0
I
40
12
T21e
80
in MonZs
Fig. 3. Disease-free survival curves of patients with Stage C disease and normal pretreatment SPAP: SPAP in lower half of normal range versus SPAP in upper half of normal range (p 4 0.01).
1386
I. J. Radiation Oncology 0 Biology 0 Physics SPAP Value 0 -
MDA 1
0.4
MDA 2-3
.41 -
0.8
> 0.8
MDA 4
Grade
Fig. 4. Correlation between pretreatment SPAP and M. D. Anderson grade for patients with Stage C disease (p < 0.05).
SPAP category, prognosis worsened as grade increased. However, for low grade lesions (MDA grade l), SPAP was of no prognostic significance. For intermediate (MDA grades 2 and 3 combined) and high-grade lesions (MDA grade 4), SPAP values within the normal range had no prognostic significance. For intermediate and high-grade lesions, elevated SPAP, however, appears to predict independently for a worse outcome. For grades 2 and 3, the difference in DFS at 5 years was statistically significant (p =
