Prostate Cancer and Prostatic Diseases (2000) 3, 269±274
ß 2000 Macmillan Publishers Ltd All rights reserved 1365±7852/00 $15.00
www.nature.com/pcan
Use of pretreatment prostatespeci®c antigen doubling time to
predict outcome after radical
prostatectomy
S Egawa1*, Y Arai2, K Tobisu3, S Kuwao4, T Kamoto5, Y Kakehi5 & S Baba1
1
Department of Urology, Kitasato University School of Medicine, Sagamihara, Kanagawa,
Japan; 2Department of Urology, Kurashiki Central Hospital, Kurashiki, Okayama, Japan;
3Department of Urology, National Cancer Center Hospital, Chuo-ku, Tokyo, Japan;
4
Department of Pathology, Kitasato University School of Medicine, Sagamihara, Kanagawa,
Japan; and 5Department of Urology, Faculty of Medicine, Kyoto University, Shogoin,
Sakyo, Kyoto, Japan
The objective of this study was to better understand the implications of the rate
of prostate-speci®c antigen (PSA) changes in prostate carcinoma. We retrospectively calculated PSA doubling times prior to surgery in 62 patients with prostate
carcinoma. The calculated values were compared with ®nal pathologic ®ndings
and with rates of PSA failure after surgery. PSA values increased during the
period of observation in 82.3% of the patients, whereas 17.7% had levels that
remained stable. The median calculated PSA doubling time in those with
increasing levels was 25.8 months, with doubling times 24 months observed in
37.1% of the patients. Stage pT3 disease was more common in patients with PSA
doubling times of 36 months than in those with doubling times > 36 months
(P ˆ 0.02). Biochemical failure was more common in patients with rapid PSA
doubling times (P < 0.01). The calculated PSA doubling time prior to radical
surgery is signi®cantly associated with the ®nal pathologic ®ndings. Early PSA
failure is more common in patients with rapid PSA doubling times prior to
radical surgery. Prostate Cancer and Prostatic Diseases (2000) 3, 269±274.
Keywords: prostate cancer; radical prostatectomy; prostate speci®c antigen
doubling time; biochemical failure
Introduction
Effective methods are needed to distinguish clinically
signi®cant prostate cancers requiring intervention from
those that might possibly be better treated by expectant
management. It has been suggested that calculated
prostate speci®c antigen (PSA) doubling time predicts
cancer aggressiveness.1 ± 5 In contrast, several investigators failed to ®nd any signi®cant correlation between
PSA doubling time and speci®c clinical parameters in
patients followed expectantly.6 ± 8 Whether PSA doubling
time prior to therapy is of value for predicting
the prognosis of individual patients thus remains
controversial.
To better understand the relationship between PSA
doubling time and cancer aggressiveness, we conducted
a retrospective study to determine whether pre-radical
prostatectomy PSA doubling time predicts pathological
stage at radical prostatectomy or predicts PSA failure
after surgery.
Materials and methods
*Correspondence: S Egawa, Department of Urology, Kitasato
University School of Medicine, 1-15-1 Kitasato, Sagamihara,
Kanagawa 228-8555, Japan.
E-mail: [email protected]
Received 8 June 2000; revised 26 July 2000; accepted 14 August 2000
We reviewed the medical records of 61 patients who
underwent radical retropubic prostatectomy at four urologic centers (Kitasato University Hospital, 36 patients;
Kurashiki Central Hospital, 14 patients; National Cancer
Center Hospital, seven patients; Kyoto University
PSA doubling time in prostate cancer
S Egawa et al
270
Hospital, four patients) between November 1991 and
June 1999 and who had serial PSA measurements preoperatively. One additional patient (at Kurashiki), whose
surgery was aborted owing to intraoperative discovery
of gross tumor involvement outside the gland, was also
included for analysis. This case, in which tumor invasion
was con®rmed by biopsy, was treated and analyzed as
having positive extracapsular extension, but excluded
from other assessments such as seminal vesicle involvement, nodal involvement and margin status. His
PSA never became undetectable postoperatively and
androgen deprivation therapy was initiated.
All patients had had two or more preoperative PSA
measurements covering an interval of at least 6 months.
No patients received any form of preoperative hormonal
ablation or chemotherapy.
PSA was measured at various laboratories using
various techniques. These include Eiken polyclonal
radioimmunoassay (Eiken, Tokyo, Japan), Dainapack
IMx PSA assay (Dinabot, Tokyo, Japan), AxSYM PSA
assay (Dinabot, Tokyo, Japan), Del®a PSA assay (Pharmacia and Upjohn Inc., Tokyo, Japan) and Lumipulse
PSA assay (Fujirebio Inc., Tokyo, Japan). Eiken PSA
assay was used to obtain initial PSA data in 12 cases.
For uniformity, Eiken PSA data were inter-converted as
follows: (IMx PSA) ˆ 1.39 6 (Eiken PSA) 7 1.02.9 Results
of other assays were not inter-converted, because they
are considered virtually the same.6,9 ± 11 After surgery,
patients were evaluated at 3 ± 6 month intervals with a
PSA measurement and digital rectal examination. Postoperative PSA values were considered elevated (PSA
failure) if values of 0.1 ng/ml or greater were obtained
on two consecutive visits 1 month apart for IMx, AxSYM
and Lumipulse PSA assays, and 0.2 ng/ml or greater for
the Del®a PSA assay.
Several different surgeons conducted radical retropubic prostatectomy at each institution. Histologic ®ndings were con®rmed by ultrasound-guided, six-sextant
biopsy or by transurethral resection. Tumors were
staged clinically according to the 1992 TNM classi®cation.12 All prostatectomy specimens but one were examined by the whole-organ step-section technique: the
specimens were ®xed in 10% formalin and sectioned at
4 or 5 mm intervals in a plane perpendicular to the long
axis of the gland, from the prostatic apex to the tip of the
seminal vesicles, followed by staining with hematoxylin
and eosin. One specimen was sectioned randomly.
All histology slides were examined and reviewed by a
single, central pathologist (SK). Tumor grade and depth
of capsular penetration were determined on the basis of
criteria speci®ed previously.13 Seminal vesicle involvement in the sections was searched for and assessed.
Tumor area was measured using a planimeter. Allowance was made for prostate section thickness by multiplying the measured area by the actual thickness of a cut
section. The volume of each tumor was determined in 60
patients by multiplying the measured volume by a
shrinkage factor of 1.33.
For each patient, we calculated a linear regression of
log PSA on time and computed the doubling time as log
2 divided by the slope. Prostatic volume was determined sonographically and calculated either as prostatic
length 6 width 6 height 6 (p/6) (Kitasato, Kyoto) or as
prostatic height 6 width2 6 (p/6) (Kurashiki). No
Prostate Cancer and Prostatic Diseases
records of prostatic volume were available for the
remaining institution.
The association of PSA doubling time with clinical
stage, PSA level at diagnosis, specimen Gleason score,
pathological ®ndings, PSA failure and risk-group assessment was studied. Risk-group strati®cation similar to
that of D'Amico et al 14 was adopted; the only difference
was the use of specimen rather than biopsy Gleason
score.
Fisher's exact probability test, the chi-square test, the
Mann ± Whitney U-test and Kruskal ± Wallis test were
used for comparison of variables with P < 0.05 considered signi®cant. Multivariate logistic regression analysis
was conducted to assess the capacity of parameters for
predicting pathological stage and biochemical failure
with other variables controlled. These parameters
included age, clinical stage, preoperative PSA, specimen
tumor grade, PSA doubling time and risk-group classes.
Results
Patient characteristics and outcome of surgery
Median age of the 62 patients at radical prostatectomy
was 67.0 y (range, 49 ± 78 y; mean, 65.7 y). The median
follow-up period after radical surgery was 23.0 months
(range, 1.0 ± 92.0 months; mean, 29.6 months). The diagnosis of prostate carcinoma was made by needle biopsy in
61 patients and at transurethral resection of the prostate
gland in one patient. In 12 patients, the diagnosis was
made at repeat biopsy following initially negative results.
One patient was initially followed expectantly after diagnosis of prostate carcinoma, but later underwent radical
surgery. Median preoperative serum PSA was 9.0 ng/ml
(range, 1.8 ± 86.4 ng/ml; mean, 12.3 ng/ml). The median
number of PSA recordings per patient was 4.0 (range, 2 ±
15 recordings) over a median period of 17.0 months
(range, 6.0 ± 74.0 months) prior to radical prostatectomy.
Nine patients had only two available PSA values. Seven
patients (11.3%) had PSA levels > 20.0 ng/ml. The
median gland volume was 32.9 cm3 (range, 12.5 ±
64.4 cm3; mean, 33.5 cm3).
The 62 patients were classi®ed clinically as stage T1b
(one patient), T1c (33 patients), T2a (12 patients), T2b (12
patients), T2c (two patients), and T3a (two patients). Specimen tumor grade was classi®ed as well differentiated
(Gleason score 2 ± 4) and moderately differentiated (Gleason score 5 or 6) in ®ve (8.1%) and 22 (35.5%) patients,
respectively. Thirty-®ve patients had poorly differentiated
tumors, with Gleason scores of 7 in 31 patients (50.0%)
and of 8 ± 10 in four patients (6.5%). Total tumor volume
ranged from 0.1 to 17.8 cm3 (mean, 2.8 cm3; median,
1.6 cm3).
Twenty-two patients (35.5%) had locally advanced
disease on ®nal pathologic examination; extracapsular
extension was seen in 20 patients (32.3%), seminal
vesicle involvement in seven (11.3%), and microscopic
nodal involvement in one (1.6%). Thirteen patients
(21.0%) had positive surgical margins. To date, nine
patients have experienced PSA failure at a median of
11.0 months following surgery (range, 0.0 ± 24.0). No
clinical progression was reported.
PSA doubling time in prostate cancer
S Egawa et al
Calculated PSA doubling times and
clinicopathological parameters
Table 1 Preoperative prostate speci®c antigen doubling times for
Japanese patients who underwent radical retropubic prostatectomy
Plots of serial PSA levels during the observation period
are shown in Figure 1. PSA values increased during the
period of observation in 51 of the 62 patients (82.3%)
(Table 1). Twenty-three patients (37.1%) had PSA doubling times of 2 y, whereas 11 patients (17.7%) had
levels that remained stable. The proportions were almost
the same when analysis was limited to the 54 patients
with T1/T2 disease and an initial PSA of 20.0 ng/ml.
The median calculated PSA doubling time in patients
with increasing levels was 25.8 months (range, 3.5 ±
228.8 months; mean, 39.7 months).
The relationships of PSA doubling time to clinical
stage, PSA level at diagnosis, tumor grade, pathological
®ndings, risk group assessment, and PSA failure are
outlined in Table 2. Patients with T2/T3 cancers or
seminal vesicle involvement exhibited more rapid PSA
doubling times than did those without those characteristics (P < 0.05). Though not statistically signi®cant,
pathologically more advanced cancers tended to show
increasing rather than stable PSA levels compared to
con®ned disease (90.9% vs 67.5%, P ˆ 0.06). Those with
PSA failure were more likely to have shorter PSA
doubling times (15.1 months vs 30.5 months, P ˆ 0.01).
No substantial difference in doubling time or frequency
of stable PSA levels was noted among risk groups
PSA doubling time
(months)
0 ± 12.0
12.1 ± 24.0
24.1 ± 36.0
36.1 ± 48.0
48.1 ± 60.0
60.1 ± 72.0
72.1 ± 228.8
Stable
Total
T1 ± T2NOMO
PSA 20.0 ng/ml
no. (%)
7
12
11
5
1
1
7
10
54
(13.0)
(22.2)
(20.4)
(9.3)
(1.8)
(1.8)
(13.0)
(18.5)
(100.0)
271
T1 ± T3NOMO
no. (%)
9
14
13
5
1
1
8
11
62
(14.5)
(22.6)
(21.0)
(8.1)
(1.6)
(1.6)
(12.9)
(17.7)
(100.0)
PSA, prostate speci®c antigen.
(P > 0.05). There was no association of PSA doubling
time with PSA level at diagnosis, tumor grade, or
extracapsular extension or surgical margin status
(P > 0.05).
Calculated PSA doubling times and outcome of
surgery
The relationships of the pathological ®ndings and of
PSA failure to preoperative PSA doubling time are
shown in Table 3. Stage pT3 disease was more
common in patients with PSA doubling times of 36
months than in those with doubling times > 36 months
(47.2% vs 19.2%, P ˆ 0.02). Strati®cation by PSA doubling times of or > 24 months did not produce
statistical signi®cance. Similar ®ndings were obtained
for seminal vesicle involvement. More patients with PSA
doubling times 24 months had seminal vesicle involvement than did those with doubling times 24 months
(26.1% vs 5.1%, P ˆ 0.04). No signi®cant difference
between PSA doubling time classes was found for the
incidence of extracapsular extension or positive surgical
margins (P > 0.05). Total tumor volume and specimen
Gleason score also did not correlate with calculated PSA
doubling times (P > 0.05). Biochemical failure developed
only in patients with relatively rapid PSA doubling
times (25.0% vs 0.0% for those with doubling times
or > 36 months, P < 0.01). Nine patients with PSA
failure had PSA doubling times of 36 months. Multivariate logistic regression analysis indicated that neither
age, clinical stage, preoperative PSA, specimen tumor
grade, PSA doubling time or risk-group classes could
independently predict pathological stage or biochemical
failure (P > 0.05).
Discussion
Figure 1 Serial PSA changes prior to surgery in 63 patients with T1 ± T3
nonmetastatic prostate cancer. (a) Plots of actual PSA levels; (b) plots of
percent change in PSA levels.
The optimum treatment for clinically localized prostate
carcinoma has been controversial.15 This is largely due
to the paucity of data regarding the natural history of
this disease and the resulting inability to predict the
biological course of early stage prostate carcinoma.
The rate of change in PSA level has been suggested as
a useful way to assess the status of prostate carcinoma
Prostate Cancer and Prostatic Diseases
PSA doubling time in prostate cancer
S Egawa et al
272
Table 2 Correlation between clinicopathological parameters and calculated PSA doubling times in patients who underwent surgical intervention
PSADT (months)
Increasing
Parameter
Clinical stage
T1
T2
T3
P value
PSA at diagnosis (ng/ml)
± 4.0
4.1 ± 10.0
10.1 ± 20.0
20.1 ‡
P value
Specimen Gleason score
2±6
7 ± 10
P value
Pathological stage
pT2
pT3
P value
ECE
Negative
Positive
P value
SVI
Negative
Positive
P value
Surgical margin
Negative
Positive
P value
Risk-group
Low
Intermediate
High
P value
PSA failure
No
Yes
P value
n
No. (%)
Median s.e.
34
26
2
28 (82.4)
21 (80.8)
2 (100.0)
33.7 9.7
18.9 7.7
7.6 2.0
0.003 (T1 vs T2/3)
6
27
22
7
4 (66.7)
24 (88.9)
17 (77.3)
6 (85.7)
27
35
Range
DT < 24 months (%)
Stable no. (%)
P-value
5.4 ± 228.8
3.5 ± 162.1
5.5 ± 9.6
8 (23.5)
13 (50.0)
2 (100.0)
6 (17.6)
5 (19.2)
0 (0.0)
NS
89.2 31.8
32.3 9.3
21.7 4.9
22.7 28.6
NS
7.9 ± 162.1
3.5 ± 228.8
9.6 ± 86.1
5.6 ± 188.8
1 (16.7)
9 (33.3)
10 (45.5)
3 (42.9)
2 (33.3)
3 (11.1)
5 (22.7)
1 (14.3)
NS
21 (77.8)
30 (85.7)
32.9 13.9
23.6 3.5
NS
3.5 ± 228.8
5.6 ± 86.1
7 (25.9)
16 (45.7)
6 (22.2)
5 (14.3)
NS
40
22
27 (67.5)
20 (90.9)
35.4 10.6
20.2 4.3
NS
10.1 ± 228.8
5.6 ± 86.1
9 (22.5)
10 (45.5)
13 (32.5)
2 (9.1)
NS
42
20
33 (78.6)
18 (90.0)
32.3 9.3
24.2 4.6
NS
3.5 ± 228.8
5.6 ± 86.1
15 (35.7)
8 (40.0)
9 (21.4)
2 (10.0)
NS
53
8
42 (79.2)
8 (100.0)
30.5 7.5
15.3 3.7
0.02
3.5 ± 228.8
5.6 ± 37.1
16 (30.2)
6 (75.0)
11 (20.8)
0 (0.0)
NS
48
13
38 (79.2)
12 (92.3)
27.4 7.4
25.0 13.8
NS
3.5 ± 228.8
15.1 ± 188.8
18 (37.5)
4 (30.8)
10 (20.8)
1 (7.7)
NS
17
34
11
14 (82.4)
28 (82.4)
9 (81.8)
37.0 17.3
23.9 19.5
19.6 19.3
NS
4.7 ± 228.8
3.5 ± 86.1
5.6 ± 188.8
4 (23.5)
14 (41.2)
5 (45.5)
3 (17.6)
6 (17.6)
2 (18.2)
NS
53
9
42 (79.2)
9 (100.0)
30.5 7.4
15.1 3.2
0.01
4.7 ± 228.8
3.5 ± 34.5
16 (30.2)
7 (77.8)
11 (20.8)
0 (0.0)
NS
PSA, prostate-speci®c antigen; PSADT, PSA doubling time; s.e. standard error;
ECE, extracapsular extension; SVI, seminal vesicle involvement; NS, not signi®cant.
followed expectantly.1 ± 8,16 It has previously been shown
that the median PSA doubling time of nonmetastatic
prostate carcinoma is approximately 36 months (23 ± 56
months).1 ± 8,16 However, only between 68% and 86% of
such patients show increasing levels; no increase is seen
in up to 30%. On the other hand, a short PSA doubling
time (< 24 months) is observed in 7 ± 47% of patients. It
has been speculated that tumors with relatively stable or
declining PSA levels may be in their linear phase of
growth and may safely be followed for years without
treatment.6,8,17, Conversely, tumors with rapidly increasing PSA levels may be in the exponential growth phase
and thus signi®cant.
The median calculated PSA doubling time in patients
with increasing levels was 25.8 months in this study.
Although this value is slightly lower than the one we
previously calculated,6 it has been claimed that tumor
doubling time may be underestimated when the calculaProstate Cancer and Prostatic Diseases
tion is based on serial PSA changes.6,18 Nevertheless,
there appear to be subpopulations of tumors with PSA
levels that are, respectively, increasing and stable for a
period of time prior to radical surgery.
Several investigators have suggested use of this parameter to predict treatment outcome prior to de®nitive
therapy.1,4,5 D'Amico and Hanks found that, following
radiation therapy for prostate cancer, PSA doubling time
was linearly correlated with the interval to clinical
relapse after PSA failure.5 They recommended observation rather than treatment for patients with PSA doubling times of 18 months, because the disease may not
become clinically apparent during the patient's life time.
Hanks et al considered PSA doubling time a predictor of
biochemical disease-free survival following radiation
therapy.1 In their view, patients with pretreatment PSA
doubling times < 12 months have aggressive disease
and should be candidates for multimodal therapy. On
PSA doubling time in prostate cancer
S Egawa et al
Table 3 Relationship of pathological ®ndings and of PSA failure following surgery to preoperative PSA doubling time
3
PSA doubling Patients, Total TV (cm ), Speciman Gleason pT3, no.
time (months) no.(%)
median s.e. score, median s.e.
(%)
0 ± 12.0
12.1 ± 24.0
24.1 ± 36.0
36.1 ± 48.0
48.1 ± 60.0
60.1 ± 72.0
72.1 ± 228.8
Stable
Total
9
14
13
5
1
1
8
11
62
(14.5)
(22.6)
(21.0)
(8.1)
(1.6)
(1.6)
(12.9)
(17.7)
(100.0)
0.8 1.9
1.5 0.6
2.9 0.8
1.5 1.3
2.1
2.2
0.6 0.9
1.4 0.7
1.6 0.4
6.0 0.4
7.0 0.3
7.0 0.3
7.0 0.5
7
5
5.0 0.5
6.0 0.4
6.0 0.4
4
6
7
1
1
0
1
2
22
273
ECE, no. SVI, no. Positive node, Positive margin, PSA failure,
(%)
(%)
no. (%)
no. (%)
no. (%)
(44.4)
3
(42.9)
5
(53.8)
7
(20.0)
1
(100.0) 1
(0.0)
0
(12.5)
1
(18.2)
2
(35.5) 20
(33.3)
(35.7)
(53.8)
(20.0)
(100.0)
(0.0)
(12.5)
(18.2)
(32.3)
2
4
1
1
0
0
0
0
8
(22.2)
(28.6)
(7.7)
(20.0)
(0.0)
(0.0)
(0.0)
(0.0)
(12.9)
0
1
0
0
0
0
0
0
1
(0.0)
(7.1)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(1.6)
0
4
4
2
0
0
1
1
12
(0.0)
(28.6)
(30.8)
(40.0)
(0.0)
(0.0)
(12.5)
(9.1)
(19.4)
3
4
2
0
0
0
0
0
9
(33.3)
(28.6)
(15.4)
(0.0)
(0.0)
(0.0)
(0.0)
(0.0)
(14.5)
PSA, prostate-speci®c antigen; ECE, extracapsular extension; SVI, seminal vesicle involvement.
the other hand, Goluboff et al found that, in 56 patients
with three or more PSA values prior to radical prostatectomy, calculated PSA doubling time did not correlate
with PSA failure, ®nal PSA, or Gleason score, but
paralleled the pathologic stage.4 However, the median
follow-up after radical surgery was only 17.3 months in
this study. With longer follow-up, the association of PSA
doubling time with pathologic stage would presumably
be translated into a correlation with PSA failure.
The median follow-up period after radical prostatectomy in our study was slightly longer, 23.0 months,
which may in part explain why we found an association
between calculated PSA doubling times and PSA failure.
It is of interest that all nine patients with PSA failure had
preoperative PSA doubling times 36 months. Whether
preoperative PSA doubling times can be used to drive
clinical decision-making, especially in individual
patients, still remains uncertain. Nevertheless, it appears
that early PSA failure is more common in those with
faster preoperative PSA doubling times.
Compared to pathologically more advanced cancers,
organ-con®ned disease tended to show stable rather
than increasing PSA levels (P ˆ 0.06). Stage pT3 disease
was more common in patients with fast PSA doubling
times ( 36) months than in those with doubling times
> 36 months (P ˆ 0.02), although the group with PSA
doubling times > 36 months included ®ve of the 23
patients with pT3 disease (Table 3). This broad spectrum
of distribution may in part account for the variable
natural history of progression following radical prostatectomy.19 Those with PSA failure were more likely to
have shorter PSA doubling times (15.1 months vs 30.5
months, P ˆ 0.01). There was no association of PSA
doubling time with PSA level at diagnosis, tumor
grade, or risk-group assessment.
Our study is limited by its retrospective nature, by
short follow-up time and by the small number of
patients who had radical surgery. This may account for
the inability of multivariate logistic regression analysis
to detect any independent parameter for predicting
pathological stage and biochemical failure. The variety
of laboratories performing the PSA assays may be
problematic as well. Use of the ®rst and last available
PSA values to obtain PSA doubling time in some
patients may also be a problem, but the values calculated by this method have been shown to correlate
highly with those calculated by regression analysis.16
A multi-institutional, prospective study is now underway. Further investigation is warranted.
Acknowledgements
Thanks are due to WA Thomasson PhD, for expert
editorial assistance. This work was supported in part
by a grant from the Ministry of Health and Welfare of
Japan (11-10).
References
1 Hanks GER et al. Pretreatment prostate-speci®c antigen doubling
times: clinical utility of this predictor of prostate cancer behavior.
Int J Radiat Oncol Biol Phys 1996; 34: 549 ± 553.
2 Lee WR, Hanks GE, Corn BW, Schultheiss TE. Observations of
pretreatment prostate-speci®c antigen doubling time in 107
patients referred for de®nitive radiotherapy. Int J Radiat Oncol
Biol Phys 1995; 31: 21 ± 24.
3 Schmid H-P, McNeal JE, Stamey TA. Observations on the doubling time of prostate cancer. The use of serial prostate-speci®c
antigen in patients with untreated disease as a measure of
increasing cancer volume. Cancer 1993; 71: 2031 ± 2040.
4 Goluboff ET et al. Pretreatment prostate speci®c antigen doubling
times: use in patients before radical prostatectomy. J Urol 1997;
158: 1876 ± 1879.
5 D'Amico AV, Hanks GE. Linear regressive analysis using prostate-speci®c antigen doubling time for predicting tumor biology
and clinical outcome in prostate cancer. Cancer 1993; 72: 2638 ±
2643.
6 Egawa S et al. Observations of prostate speci®c antigen doubling
time in Japanese patients with nonmetastatic prostate carcinoma.
Cancer 1999; 86: 463 ± 469.
7 Gerber GS et al. Evaluation of changes in prostate speci®c antigen
in clinically localized prostate cancer managed without initial
therapy. J Urol 1998; 159: 1243 ± 1246.
8 Nam RK et al. Prostate speci®c antigen velocity as a measure of
the natural history of prostate cancer: de®ning a `rapid riser'
subset. Br J Urol 1998; 81: 100 ± 104.
9 Machida T et al. Clinical evaluation of a new kit (IMx PA
Dainapack) for detection of serum prostate speci®c antigen.
Acta Urol Japonica 1993; 39: 977 ± 984.
10 Kuriyama M et al. Comparison of various assay systems for
prostate-speci®c antigen standardization. Jpn J Clin Oncol 1992;
22: 393 ± 399.
11 Nodate K, Tsukada T. Evaluation of chemiluminescent enzyme
immunoassay `Lumipulse PSA'. Jpn J Med Pharm Sci 1996; 36:
1395 ± 1402.
12 International Union against Cancer. TNM Atlas, 3rd edn, 2nd
revision. Springer: New York, 1992, pp 241 ± 250.
13 Greene DR et al. A comparison of the morphological features of
cancer arising in the transition zone and in the peripheral zone of
the prostate. J Urol 1991; 146: 1069 ± 1076.
Prostate Cancer and Prostatic Diseases
PSA doubling time in prostate cancer
S Egawa et al
274
14 D'Amico AV et al. Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation
therapy for clinically localized prostate cancer. JAMA 1998; 280:
969 ± 974.
15 Abbas F, Scardino PT. The natural history of clinical prostate
carcinoma. Cancer 1997; 80: 827 ± 833.
16 McLaren DB, McKenzie M, Duncan G, Pickles T. Watchful waiting or watchful progression? Prostate speci®c antigen doubling
times and clinical behavior in patients with early untreated
prostate carcinoma. Cancer 1998; 82: 342 ± 348.
Prostate Cancer and Prostatic Diseases
17 Carter HB et al. Estimation of prostatic growth using serial
prostate-speci®c antigen measurements in men with and without
prostate disease. Cancer Res 1992; 52: 3323 ± 3328.
18 Berges RR et al. Implication of cell kinetic changes during the
progression of human prostatic cancer. Clin Cancer Res 1995; 1:
473 ± 480.
19 Pound CR et al. Natural history of progression after PSA elevation following radical prostatectomy. JAMA 1999; 281: 1591 ±
1597.