PROSTATE-SPECIFIC ANTIGEN: THE BEST PROSTATIC TUMOR MARKER 0094-0143/97 $0.00 + .

20

PROSTATE-SPECIFIC ANTIGEN
AFTER ANATOMIC RADICAL
RETROPUBIC PROSTATECTOMY
Patterns of Recurrence and Cancer Control
Charles R. Pound, MD, Alan W. Partin, MD, PhD,
Jonathan I. Epstein, MD, and Patrick C. Walsh, MD

Prostate-specific antigen (PSA) is a valuable tu- dissection for clinically localized adenocarcinoma
mor marker in the follow-up evaluation of men of the prostate (Tl, T2, T3a) at our institution. All
who have been treated for prostate cancer. An procedures were performed by a single surgeon.
undetectable level of PSA serves as the gold stan- The average age was 59 & 6 years (range 34-76).
dard for tumor-free status in men after radical Of these men, 1623 (95%) were included in this
prostatectomy.'O,l2 In 1993, we reported our initial analysis. Table 1 illustrates the distribution of the
10-year experience with PSA values after anatomic clinical stages of these men.15 To evaluate the abil-
radical retropubic prostatectomy (RRP) for 955 ity of radical prostatectomy as single therapy to
men with localized disease.12 We now have ex- cure localized prostate cancer, 76 men (4.5%) were
panded this series to include 1623 men. In this excluded from the study group: 20 men (1.2%)
article we describe the actuarial likelihood of an were excluded based on inadequate postoperative
undetectable PSA level for this group of men who PSA data; 23 men (1.4%) underwent immediate
have undergone surgery at The Johns Hopkins adjuvant radiation therapy; 11 men (0.6%) were
Hospital since 1982. We examine the influence of treated with immediate adjuvant hormonal ther-
clinical and pathologic parameters, which alone apy; nine men (0.5%) were clinical stage DO/D2
and in combination are shown to have a significant
and were excluded for this reason; nine men (0.50/,)
effect on the actuarial rate of PSA recurrence after
received preoperative hormonal therapy; three
surgery. In addition, we show that anatomic RRP
with preservation of the neurovascular bundles men (0.2%) received preoperative radiation ther-
had no adverse effect on cancer control. A Gleason apy; and one patient (0.1%) died in the immediate
score of at least 8 or involvement of the seminal postoperative interval.
vesicles or lymph nodes is indicative of eventual Twenty of the 23 men excluded from the group
failure from distant metastases. We show that in because of immediate adjuvant radiation therapy
addition to these variables, the timing of the devel- received this therapy because of a positive surgical
opment of a detectable serum PSA also is im- margin. Pathologically, three of these men had
portant in predicting eventual local versus distant organ-confined disease, 12 men had capsular pene-
failure. tration only, five men had positive seminal vesicles
without involvement of the pelvic lymph nodes,
MATERIALS AND METHODS and three men had metastases to pelvic lymph
nodes. Eight of 23 men who received immediate
Patients adjuvant radiation therapy have undetectable PSA
Between May 1982 and March 1995, 1699 men levels 2 or more years after surgery without addi-
underwent anatomic RRP with pelvic lymph-node tional therapy. More than half (63%) of these men

From the Departments of Pathology (JIE) and Urology (CRP, AWP, JIE, PCW), James Buchanan Brady Urological
Institute, The Johns Hopkins University School of Medicine, Baltimore, Maryland

UROLOGIC CLINICS OF NORTH AMERICA

VOLUME 24 NUMBER 2 * MAY 1997 395

396 POUND et a1

Table 1. CLINICAL STAGE, PREOPERATIVE PSA Preoperative Evaluation and Pathologic

LEVEL, GLEASON SCORE, AND PATHOLOGIC Characteristics
STAGE IN MEN UNDERGOING ANATOMIC RADICAL
RETROPUBIC PROSTATECTOMY
Preoperative evaluation involved a digital rectal
No. of Percent of examination, a serum PSA assay measured in ng/
Variable Patients Total mL (Hybritech-Tandem, Hybritech, San Diego,
TNM
CA), a serum enzymatic prostatic acid phospha-
T1a 55 3 tase measurement (thymolphthalein monophos-
Tlb 109 7 phate assay; normal less than 0.8 IU/L), and a
T1c 343 21 nuclear bone scan. The PSA assay was performed
T2a 638 39 on serum that was collected before or at least 4
T2b 337 21 weeks after prostate biopsy or transurethral re-
T2c 97 6 section of the prostate. Pathologic diagnosis of
T3a 44 3 prostate cancer preoperatively was based on exam-
Total 1623 100 ination of prostatic tissue obtained through trans-
Serum PSA (ng/mL) rectally or digitally guided prostate biopsy, trans-
0-4 399 29 urethral resection of the prostate, or fine-needle
4.1-10 619 46 aspiration. Histologic scoring was performed us-
10.1-20 253 19 ing the Gleason scoring system. Pathologic evalua-
>20 83 6
1354 100
tion of the operative specimen was performed as
Total
previously described.'* Based on this evaluation,
Gleason score tumors were determined to be organ-confined dis-
2-4 61 4 ease, penetrating the prostatic capsule without ex-
5 308 19
35
tension to the seminal vesicles, involving the semi-
6 563
7 538 33 nal vesicles without nodal disease, or spread to
8-1 0 153 9 pelvic lymph nodes. A complete description and
Total 1623 100 justification of this pathologic staging system has
been p ~ b l i s h e d . ~ In
, ~ *our
' ~ previous report, we
Pathologic stage
Organ confined 658 41 subdivided the category of capsular penetration
Capsular penetration with 380 23 into focal capsular penetration (<2-3 glands be-
Gleason score less yond the prostatic capsule) and established capsu-
than 7 lar penetration (>2-3 glands beyond the prostatic
Capsular penetration with 373 23 capsule). Although Epstein et a15 have shown that
Gleason score greater there is a significant difference in recurrence be-
than or equal to 7 tween tumors with focal capsular penetration or
Involvement of the seminal 95 6 established capsular penetration, they also have
vesicles, negative lymph
nodes
determined that the best pathologic predictor of
Micrometastases to pelvic 115 7 disease progression was obtained using a multi-
lymph nodes variate model combining postoperative Gleason
Total 1621 100 score, surgical margin status, and the presence or
absence of and extent of capsular penetration. In
this work, we chose to group patients with capsu-
lar penetration into a single category that greatly
have been followed for at least 5 years postopera- simplifies interpretation of the graphic analyses.
tively: seven of 23 men have detectable PSA levels Table 1 describes the distribution of patients ac-
as the only indication of disease recurrence, but cording to the TNM clinical staging system, preop-
three received additional hormonal therapy, five erative PSA, postoperative Gleason score, and final
men have developed metastatic disease, and one pathologic stage.
man is dead from cancer.
None of the 11 patients excluded from our anal-
ysis because of immediate adjuvant hormonal
therapy showed organ-confined disease. Seven Patient Follow-up and Definitions of
men had microscopic metastatic spread to the pel- Progression and Potency
vic lymph nodes discovered in the final pathologic
specimen, and the four men with negative lymph Postoperative follow-up data were obtained
nodes had involvement of the seminal vesicles. through routine serum PSA assays and digital rec-
There were four men with positive surgical mar- tal examinations performed every 3 months for
gins. Only two of 11 men have an undetectable the first year, semiannually the second year, and
PSA after at least 2 years of follow-up, and two annually thereafter. Average length of follow-up
other men have a detectable PSA as their only was 5 f 3 years (range 1-13 years). One-hundred
evidence of disease recurrence. Seven men devel- ninety-three men (12%)were followed for at least
oped distant metastases, and six of these seven are 10 years postoperatively. More than half of these
dead of cancer. men (829/1623, 51%) have been followed for 5
 PROSTATE-SPECIFIC ANTIGEN AFTER ANATOMIC RADICAL RETROPUBIC PROSTATECTOMY 397

years or longer. Table 2 illustrates the length in detectable PSA (>0.2 ng/mL) was the only evi-
years of follow-up available for all 1623 patients. dence of recurrence in 129 men (7.9%). Forty-one
Disease recurrence is defined in three ways. A men (2.5%) have recurred locally without evidence
biochemical recurrence is a serum PSA level which of distant spread, and 88 men (5.4%) have distant
is detectable postoperatively (>0.2 ng/mL). A lo- metastases with or without evidence of local recur-
cal recurrence is defined as palpable induration at rence. Nine men (0.6%) who received radiation for
the operative site in the presence of an elevated an isolated elevation of PSA had a subsequent
serum PSA with or without a positive biopsy. A return of PSA to an undetectable level. Although
positive biopsy without palpable induration also their PSA continues to be undetectable, they are
is evidence of local recurrence. Distal recurrence is included in calculations of overall PSA progression
shown by a positive bone scan or radiographic as failures but are not included in determinations
evidence of retroperitoneal adenopathy. By our of type-specific recurrence in that their site of re-
exclusion criteria, no patient included in this anal- currence (local versus distant) has not been deter-
ysis received radiation or hormonal therapy post- mined. Kaplan-Meier statistical analysis (Fig. 1)
operatively before demonstration of disease recur- showed an overall actuarial 5- and 10-year recur-
rence as evidenced by a [Link] level or rence-free likelihood for any type of recurrence of
clinical evidence of local or distant failure. There- 80% and 68% (confidence intervals 78%-83% and
fore, adjuvant therapy had no impact on time to 63%-72%, respectively). The actuarial likelihoods
progression. of having an isolated elevation of serum PSA only
To determine if the anatomic approach with at 5 and 10 years after anatomic RRP were 10%
preservation of sexual function had an effect on and l8%, respectively. Overall, the actuarial 5- and
cancer control, we reviewed the postoperative po- 10-year PSA recurrence-free rates for men who
tency status of our patient group in relationship developed local recurrence without distant metas-
to likelihood of disease progression. only men tases were 96% and 92%. For men who developed
known to be potent before radical prostatectomy distant metastases with or without local recur-
were included in this analysis. Information regard- rence, the overall actuarial 5- and 10-year PSA
ing a patient’s preoperative as well as postopera- recurrence-free rates were 93% and 91%. The 5-
tive potency status was available on 1534 of 1623 and 10-year actuarial likelihood of remaining free
patients (95%). Potency following surgery is an of a PSA, local without distant, or distant with or
indication of surgical preservation of at least one without local recurrence is included in Figure 1.
neurovascular bundle. Men were considered po- Although it has been reported by other authors;
tent after radical prostatectomy if they had erec- no patient in our analysis was determined to have
tions that were sufficient for intercourse. Men developed a local recurrence or distant metastases
without erections, as well as those with partial with an undetectable PSA at the time of progres-
erections not sufficient for penetration were con- sion.
sidered impotent.

RESULTS AND DISCUSSION Clinical Stage and Progression

Overall Progression Figure 2 illustrates the Kaplan-Meier analysis of
Of one thousand six hundred twenty-three men, the likelihood of having an undetectable PSA
267 (17%) have shown a recurrence after RRP. A based on the TNM clinical stage, preoperative se-
rum PSA, and postoperative Gleason score. No
patient with Tla disease had a detectable PSA
Table 2. YEARS OF FOLLOW-UP AVAILABLE ON postoperatively. Patients with Tlc disease had an
1623 MEN AFTER ANATOMIC RADICAL 86% progression-free probability at 5 years. This
RETROPUBIC PROSTATECTOMY differs from our previous report, in which no Tlc
patient had evidence of progression at 7 years of
Years of Percent follow-up. This is undoubtedly because of the re-
Follow-up No. of Patients of Total
cent increase in the number of men treated with
1 1623 100 Tlc disease. Five- and 10-year actuarial progres-
2 1328 82 sion-free probabilities for the remainder of the clin-
3 1141 70 ical stages are provided in Table 3 (upper panel).
4 973 60 We had previously shown a significant difference
5 829 51 in the 10-year recurrence-free probability between
6 687 42 stages T2a and T2b; the present analysis suggests
7 535 33
8 41 1 25 that stage T2a and T2b tumors behave similarly
9 293 18 with extended follow-up.
10 193 12
11 103 6
12 42 3 Preoperative PSA and Progression
13 . 11 1
Totals 1623 100 Of the 1623 men, 1354 (83%) had preoperative
PSA values available. These patients were divided
398 POUND et a1

(I)
[Link] -

L Distant
Local

2 0.75- A - Only
e!
L
3
0
(I)
All Recurrence
cc
5 0.50-
-0
0
0
r
-
5
(I)
0.25-

0- I I I 1 I .I I I
0 2 4 6 8 10 12 14
Years Postoperative

Figure 1. Kaplan-Meier actuarial progression-free likelihood for overall

progression, isolated PSA elevation only, local recurrence, and distant
progression f local progression. One thousand six hundred twenty three
men were included in the analysis.

into four groups based upon this measurement: Gleason score 8 to 10 (46% versus 23%, P <
PSA less than 4 ng/mL, PSA 4.1 to 10 ng/mL, 0.00001). At 10 years, the likelihood of men with
PSA 10.1 to 20 ng/mL, and PSA greater than 20 Gleason 8 to 10 disease having a PSA recurrence
ng/mL. Table 1 and Figure 2 show actuarial pro- is greater than 80%.
gression-freeprobability based on the preoperative
PSA value. There was a statistically significant dif-
ference between recurrence-free likelihoods of men
in all PSA groups. Unlike our previous analysis, Pathologic Stage, Gleason Score,
the probability of remaining free of recurrence at and Surgical Margins in Relation
10 years was statistically significant ( P = 0.0003) to Progression
between men with preoperative PSA level of 10.1
to 20 ng/mL and those with a level greater than
20 ng/mL. The best stratification of actuarial progression-
free probabilities at 5 and 10 years is obtained
by grouping patients based on a combination of
pathologic stage, Gleason score, and surgical mar-
Gleason Score and Progression gin status, as was previously demonstrated by Ep-
stein et a1.5 Kaplan-Meier progression-free analyses
for men grouped in accordance with these vari-
In the analysis of Gleason score and its relation- ables are illustrated in Figure 3. In men with cap-
ship to disease recurrence, Gleason scores 2 to 4 sular penetration with Gleason scores 2 to 6, the
were considered as a single group, as were scores status of the surgical margin had a significant ( P
8 to 10. Kaplan-Meier analysis of Gleason score < 0.01) effect on outcome at 10 years, but the
with regard to progression-free probability is outcome in both groups was good (likelihood of
shown in Figure 2 as well as in Table 3. The an undetectable PSA at 10 years was 89% with a
recurrence-free probability of men with Gleason negative margin and 72% for men with a positive
score 7 was previously shown not to be signifi- margin). In men with capsular penetration but
cantly different from that of men with higher higher-grade tumors (CP+, Gleason score 2 7),
grade disease (Gleason score Although not the presence of a positive surgical margin had a
shown in our previous report, other reports have significant effect ( P < 0.00001) on actuarial pro-
shown a significant difference in progression be- gression-free likelihood (see Fig. 3). Also notable
tween tumors with Gleason score 7 and those of is the fact that men with capsular penetration,
higher grade (Gleason score 8-10).3,5 In the present high-grade disease, and a positive surgical margin
analysis, the likelihood of an having an undetect- had recurrence at a rate similar ( P = 0.6) to tumors
able PSA at 10 years was statistically different in men with involvement of the seminal vesicles
between men with Gleason score 7 and those with (see Fig. 3).
PROSTATE-SPECIFIC ANTIGEN AFTER ANATOMIC RADICAL RETROPUBIC PROSTATECTOMY 399

[Link] -
a
0)
a
a,
E 0.75-
m
c
0
a,
T2a
c
a, T2b
U
3 0.50-
c
0

0- '
I I I I I I I I
0 2 4 6 8 10 12 14
A Years Postoperative
[Link] -
c 4.0
a
0)
a 4.1-1 0.0
0.75-
m
-
c

V
0
0)
a,
K 0.50-
3 10.1-20.0
c
0
U
0 z 20.0
2
- 0.25-
a,
Y
1

0-
I I I I I I I I
0 2 4 6 8 10 12 14
B Years Postoperative

[Link] 2-4
a 5
0)
a
2 0.75
3 6
a
m
c
0
a,
c
a,
U
c 0.50
3
c
0

8-10

0
I I I I I I I I
0 2 4 6 8 10 12 14
C Years Postoperative

Figure 2. Kaplan-Meier actuarial likelihood of PSA recurrence by clinical

stage (A), preoperative serum PSA levels (ng/mL) (B), and Gleason
score (C).
400 POUNDetal

Table 3. CLINICAL STAGE, PREOPERATIVE PSA, respect to PSA recurrence.', 11, 16, l9 Table 4 sum-
AND GLEASON SCORE IN RELATION TO marizes the demographics and clinical and patho-
ACTUARIAL RECURRENCE-FREE RATE AT 5 AND 10 logic stages in these series. The Washington Uni-
YEARS AFTER ANATOMIC RADICAL RETROPUBIC versity series showed actuarial PSA progression
PROSTATECTOMY rates of 78% and 65% at 5 and 10 years respectively
after radical prostatectomy in a cohort of nearly
Actuarial Percentage (95% CI)
1000 men.' The Baylor group showed a 76% and
Variable 5Y 10 Y 73% 5- and 10-year PSA progression-free rate after
radical prostatectomy in more than 500 men."
TNM These data are similar to our previously reported
T1a 100 100 5- and 10-year PSA recurrence rates and to those of
Tlb 89 (81-94) 89 (81-94) the present series. Two other radical prostatectomy
T1c 86 (7593) 86 (75-93)
series do not confirm our PSA progression results.
T2a 85 (81-88) 68 (61-75)
T2b 69 (63-75) 57 (48-65) The UCLA group showed PSA progression-free
T2c 63 (51-73) 53 (38-66) rates of 69% and 47% at 5 and 10 years in a group
T3a 61 (42-75) 52* (28-71) of more than 600 men.16The Mayo Clinic series of
Serum PSA (ng/mL) more than 3000 men treated with radical prostatec-
0-4 94 (91-96) 87 (74-94) tomy showed only a 52% PSA progression-free
4.1-10 82 (76-86) 75 (67-81) rate at 10 years after s~rgery.'~
10.1-20 72 (63-79) 30 (5-62) The distribution of clinical stage T1 versus T2
>20 54 (39-66) 28 (9-51) patients in the present series differs from that of
Gleason score our earlier report. Whereas our 1993 report con-
2-4 100 94 (62-99) sisted of 17% T1 and 83% T2 patients, the present
5 97 (94-99) 91 (84-95) study is more heavily weighted to T1 patients
6 92 (88-95) 78 (66-86)
7 66 (61-71) 46 (34-56)
(31%; Tlc-21%), with only 66% T2 patients (see
8-1 0 41 (32-50) 23 (13-35) Table 1). This observation can be explained by
the recent increase in patients presenting with Tlc
'&year data. disease because of the increased use of PSA for
early detection of localized prostate cancer. The
present study also includes 3% T3 patients, which
were not included previously. These factors, in
Progression: Comparison with Other conjunction with the fact that additional follow-up
Reported Series has shown that prostate cancer does recur after 5
years, explains the slight decrease in the 5- and 10-
Several radical prostatectomy series have re- year progression-free likelihoods in the present se-
ported on the follow-up of men after surgery with ries.

1.00-
a
(I)
a
Q)
0.75-
m
4-
u
Q)
c
0
U
I 73- CP+, 2 7, SM-

-
K
3
0
0.50-

LN+ . sv+
I I I I I I I I
0 2 4 6 8 10 12 14
Years Postoperative

Figure 3. Kaplan-Meier actuarial likelihood of PSA recurrence by a combina-

tion of pathologic stage, Gleason score, and surgical margin status (N =
1623). See section of text on methods for complete definitions of pathologic
stages. OC = organ-confined; CP+ < 7, SM- = capsular penetration,
Gleason score less than 7, and negative surgical margins; CP + , 2 7, SM +
= capsular penetration, Gleason score greater than or equal to 7, and positive
surgical margins; SV+ = positive seminal vesicles with negative lymph
nodes; LN+ = positive lymph nodes.
 PROSTATE-SPECIFIC ANTIGEN AFTER ANATOMIC RADICAL RETROPUBIC PROSTATECTOMY 401

Table 4. COMPARISON OF SEVERAL RETROPUBIC PROSTATECTOMY SERIES

5-year PSA 10-year PSA
Months Progression- Progression-
Primary Mean Age Follow-up, Percent Percent free free
Institution Author No. (Y) Mean (Range) T1 T2 Likelihood Likelihood

Johns Hopkins Pattin12 894 59?6 53 (12-120) 17 83 87 77

Washington University Catalona’ 925 63c7 28 (0-123) 21 79 78 65
Baylor University Ohori” 500 63 36 (1-110) 22 78t 76 73
UCLA Trapassoq6 601 65 34’ (12-237) 27 73 69 47
The Mayo Clinic Zin~ke’~ 3170 65c4 60 7 93 70 52
Present study Pound 1623 59?6 60.4(12-156) 31 66$ 80 68

‘Median.
t 4 % stage T3.
$3% stage T3a.
Modified from Partin AW, Walsh PC: Surgical management of localized prostate cancer. In Ragharan D, Scher H. Leibel S, et al (eds): Principles and
Practice of Genitourinary Oncology. Philadelphia, Lippincott-Raven, 1997; with permission.

Progression and Postoperative Potency tastases with or without local recurrence (N = 35).
Variables that best distinguished local recurrence
In an effort to determine if the technique of from distant metastases were the timing of PSA
anatomic RRP with preservation of neurovascular recurrence, Gleason score, and pathologic stage. In
bundles influences cancer control, we compared general, those men who showed PSA recurrence
actuarial recurrence-free probabilities according to less than 2 years following surgery had tumors
pathologic stage and margin status between men with a Gleason score greater than 7 or had positive
who were potent and impotent postoperatively. seminal vesicles or positive lymph nodes at the
These two groups are similar with regard to age time of surgery and were more likely to demon-
distribution, Gleason score, and pathologic stage. strate distant metastases.
Within all pathologic stages, there was no differ- Nine of the men included in that report were
ence in actuarial recurrence-free probability be- excluded from the present series based on our
tween potent and impotent men after anatomic previously mentioned exclusion criteria. Since that
RRP (Fig. 4). In addition, there was no difference initial report, 87 additional patients have failed
in actuarial PSA recurrence rates between potent clinically with local recurrence (N = 29) or distant
and impotent men postoperatively if grouped ac- metastases (N = 58). We now have a combined
cording to low-grade (Gleason score 2-6) or high- total of 129 men (local, N = 41, and distant, N =
grade (Gleason score 7-10) disease. One would 58) who have had recurrence. We tested the ability
expect differences in cancer control secondary to of the previously determined predictive variables
preservation of a neurovascular bundle to be most to distinguish local recurrence from distant metas-
evident among men with capsular penetration and tases in this combined group of patients. Table 5
positive surgical margins. Yet, the actuarial PSA shows the percentage of men with local recurrence
recurrence rates for these two groups are identical versus distant metastases for the various clinical
(see Fig. 4). and pathologic variables. In summary, the vari-
ables that best distinguished between local recur-
rence and distant metastases were the timing of
PSA recurrence, a Gleason score greater than 7,
Timing of PSA Recurrence, Gleason and positive seminal vesicles or pelvic lymph
Score, and Pathologic Stage to nodes. We recommend that all three variables be
Distinguish Local Recurrence from evaluated in attempting to distinguish between
Distant Metastases With or Without eventual local recurrence or distant metastases as-
Local Recurrence sociated with a detectable postoperative PSA. The
most important predictive variable, however, is
We previously showed that the timing of PSA the timing of PSA recurrence, and this should be
recurrence, pathologic stage, and the Gleason score weighted accordingly.
provided useful information for distinguishing lo-
cal recurrence from distant metastases with or
without local rec~rrence.’~This information can be
used to aid in the decision of the timing and type Survival
of adjuvant therapy following radical prostatec-
tomy for men with an isolated elevation of PSA. Overall actuarial cause-specific survival rates at
In our previous study, 51 men were followed for 5 and 10 years in this group of 1623 men after
an isolated elevation in PSA until clinical evidence anatomic RRP were 99% and 93%, respectively.
of either local recurrence (N = 16 as defined in There was no difference in cause-specific survival
the methods section of this article) or distant me- among men when grouped by TNM stage or pre-
402 POUND et a1

[Link] -
I Potent
a
0 Impotent
a

-
-
a, 0.75-
n
m
0
a
c
a
? 0.50-
3
c
0
U
0
2- 0.25-
f
0-
I I I I I I I
0 2 4 6 1.0 12 14
A Years Postoperative
[Link] -

3
a
2 0.75- J o Potent
t ent
m
-
c
0
a,
a
U
c 0.50-
3
c
0
U
0
0
5 0.25-
a
Y

I I I I I I I I
0 2 4 6 8 10 12 14
B Years Postoperative
1.00-
a
v)
a
2 0.75 -
> i Potent

-
n
m
0
a,
c
Impotent

a
c
U 0.50 -
3
c
0
U
0
- 0.25 -
2
;
0- I
I I I I I I I I
0 2 4 6 8 10 12 14
C Years Postoperative

Figure 4. Kaplan-Meier actuarial likelihood of PSA recurrence by various

pathologic stages for men who were preoperatively potent (N = 1573).
See section of text on methods for complete definitions of pathologic
stages. PSA recurrence data are presented based on postoperative
potency status. Table 1 details the likelihood of PSA recurrence for all
men in the study (N = 1623) by the various pathologic stages.
Illustration continued on opposite page
PROSTATE-SPECIFIC ANTIGEN AFTER ANATOMIC RADICAL RETROPUBIC PROSTATECTOMY 403

1 .oo

a
0
a I I

-
-
a, 0.75
a
m
a,
c
0

a,
U
c 0.50
3
c
0
U
0
2
- 0.25
;
0 I
I I I I I I I I
0 .2 4 6 8 10 12 14
D Years Postoperative

a
6
0.75 - -
c
m
c
0
a, -
0
0.50 - 1
3
c
0
-0 Potent
0
2 0.25 -
-
._ Impotent
(I)
z
1

0-

[Link]
a
2
2m 0.75
4.-
0
c
a,
a,
U
5 0.50
c
0
-0
0
0
g 0.25
f
0 I
I I I I I I I I
0 2 4 6 8 10 12 14
F Years Postoperative

Figure 4 (Continued). Organ-confined (A), capsular penetration (B), cap-

sular penetrationwith negative surgical margins (C), capsular penetration
with positive surgical margins (D), seminal vesicle involvement (€), and
lyrnpn noae invoivement (f).
404 POUND et a1

Table 5. COMPARISON OF GLEASON SCORE, PATHOLOGIC STAGE, AND TIMING OF PSA RECURRENCE BY
SITE OF RECURRENCE AFTER ANATOMIC RADICAL PROSTATECTOMY
Local Distant Metastases With or
Variable Recurrence Without Local Recurrence
Number 41 (34%) 88 (66Yo)
Gleason score
2-4 0% 0%
5-6 55% 45%
7 39% 61Yo
8-1 0 11% 89%
Pathologic stage
Organ-confined 40% 60%
Capsular penetration with negative 54% 46%
surgical margins
Capsular penetration with positive 48% 52%
surgical margins
Involvement of seminal vesicles 16% 84%
Micrometastases to pelvic lymph nodes 7% 93%
Timing of PSA recurrence
In year 1 7% 93%
Within years 1-2 10% 90%
After year 2 61 % 39%
After year 3 74% 26%

operative PSA level. With advancing histologic and 15 years respectively. These progression statis-
grade and pathologic stage, however, there was a tics, although skewed by favorable selection bias
significant difference in the cause-specific survival. within the individual studies, show the best data
Table 6 shows 5- and 10-year actuarial metastases- to date documenting the natural history of un-
free rates based on Gleason score and pathologic treated clinical stage Tlc and T2 prostate cancer.
stage. Note that in this table, we have grouped Eighty-six percent of men with clinically localized
Gleason scores into three categories-I (24), I1 prostate cancer in our present series presented
(5-7), and I11 (8-l0)-for the purpose of compari- with grade I1 disease (Gleason score 5-7), and of
son with previously reported data.z Chodak et alz this group, 87% were at clinical stage Tlc or T2.
summarized the results of six nonrandomized This group of men (grade 11) showed only a 16%
studies totaling more than 800 men who were actuarial rate of development of metastatic disease
treated conservatively with observation and de- (positive bone scan) 10 years after anatomic RRP.
layed hormonal therapy for clinically localized We believe that the higher failure rate for men
prostate cancer. Actuarial survival statistics
showed that grade was the most significant vari-
able in predicting survival, with 81% of men with Table 6. ACTUARIAL 5- A N D 10-YEAR METASTASIS-
grade I (Gleason score 2 4 ) , 58% of men with FREE SURVIVAL IN RELATION TO GLEASON SCORE
grade I1 (Gleason score 5-7), and only 26% of AND PATHOLOGIC STAGE
men with grade I11 (Gleason score 8-10) tumors
showing metastases-free survival at 10 years. The Actuarial Metastasis-
free Survival f%)
series analyzed and the summary presented by
Chodak et al2 have been criticized severely for Variable 5Y 10 Y
selection bias.17 The study Chodak et a12 included
by Whitmore et all8 consisted of 65 hand-picked Grade (Gleason score)
men who were selected for conservative manage- I(2-4) 100 100
II (5-7) 95 84
ment during a 40-year period. Johansson et a16 111 (8-10) 76 38
excluded 83 of 306 patients for their hand-picked Pathologic stage
series. Moskowitz et als excluded 38% of men and Organ-confined 99 99
Jones7reported on only 233 of 312 men who un- Capsular penetration with 99 99
derwent observation for clinically localized pros- Gleason score less than 7
tate cancer. As further evidence of selection bias Capsular penetration with 98 91
in this meta-analysis, nearly 60% of the men in the Gleason score greater
Chodak analysis had grade I tumors, compared than or equal to 7
with only 4% in the present series. Involvement of seminal 92 77
vesicles
Chodak and associatesZshowed that 42% and Micrometastases to pelvic 96 68
70% of the men with grade I1 disease, followed lymph nodes
conservatively, developed metastatic disease at 10
 PROSTATE-SPECIFIC ANTIGEN AFTER ANATOMIC RADICAL RETROPUBIC PROSTATECTOMY 405

treated with conservative therapy in comparison erwise healthy and have a greater than 10-year life
to those with surgical treatment justifies radical expectancy.
prostatectomy as the treatment of choice for men
with clinically localized (Gleason grades 2-7, TNM
T1-T2) prostate cancer who are otherwise healthy ACKNOWLEDGMENT
and have a greater than 10-year life expectancy.
We would like to express our sincere appreciation to
Bonnie Baxley for her assistance in the preparation of the
tables and references in this manuscript.
SUMMARY

In a series of 1623 men with a follow-up of 5 -t 3

years (range 1-13) after anatomic RRP for clinically References
localized prostate cancer, 17% (276/1623) have
shown recurrence. A detectable PSA was the only 1. Catalona WJ, Smith DJ: Five-year tumor recurrence
evidence of recurrence in 7.9%, whereas 2.5% have rates after anatomic radical retropubic prostatectomy
recurred locally and 5.4% have developed distant for prostate cancer. J Urol 152:1837-1842, 1994
metastases. The overall actuarial progression-free 2. Chodak GW, Thisted RA, Gerber GS, et al: Results
rate for these men at 10 years was 68%. Actuarial of conservative management of clinically localized
rates at 10 years were 18% for development of an prostate cancer. N Engl J Med 330:242-248, 1994
isolated PSA recurrence, 8% for local recurrence, 3. Epstein JI, Pizov G, Walsh PC: Correlation of patho-
and 9% for distant recurrence. logic findings with progression following radical ret-
ropubic prostatectomy. Cancer 71:3582-3593, 1993
The actuarial likelihood of a postoperative recur- 4. Epstein JI, Carmichael MJ, Pizov G, et al: Influence
rence increased with increasing clinical stage, of capsular penetration on progression following rad-
Gleason score, preoperative PSA level, and patho- ical prostatectomy: A study of 196 cases with long-
logic stage. Although not shown in our previous term follow-up. J Urol 150:135-141, 1993
report, the actuarial rate of recurrence of tumors 5. Epstein JI, Partin AW, Sauvageot J, et a1 Prediction
with a Gleason score of 7 was statistically different of progression following radical prostatectomy: A
from that of tumors of higher Gleason score (8-10). multivariate analysis of 721 men with long-term fol-
As well, men with preoperative PSA levels of 10.1 low-up. Am J Surg Pathol20:286-292, 1996
to 20 ng/mL experienced recurrence at a signifi- 6. Johansson JE, Adami HO, Andersson SO, et al: High
10-year survival rate in patients with early, untreated
cantly lower rate than did men with preoperative prostatic cancer. JAMA 2672191-2196, 1992
PSA levels greater than 20 ng/mL. 7. Jones G W Prospective, conservative management of
By using a combination of Gleason score, patho- localized prostate cancer. Cancer 70 (suppl): 307-
logic stage, and surgical margin status, we demon- 310, 1992
strated that the presence of a positive surgical 8. Moskowitz B, Nitecki S, Richter LD, et al: Cancer of
margin did not dramatically affect recurrence in the prostate: Is there a need for aggressive treatment?
tumors of Gleason scores 2 to 6 with capsular Urol Int 4249, 1987
penetration. Surgical margin status was important 9. Oefelein MG, Smith N, Carter M, et al: The incidence
in high-grade tumors with capsular penetration. of prostate cancer progression with undetectable se-
rum prostate specific antigen in a series of 394 radical
In fact, tumors with capsular penetration, Gleason prostatectomies. J Urol 154:2128-2131, 1995
score of at least 7, and a positive surgical margin 10. Oesterling J: Prostate-specific antigen: A critical as-
behaved similarly to tumors with invasion of the sessment of the most useful tumor marker for adeno-
seminal vesicles. carcinoma of the prostate. J Urol 145:907, 1991
Preservation of potency did not adversely in- 11. Ohori M, Goad JR, Wheeler TM, et al: Can radical
fluence cancer control. The Gleason score, presence prostatectomy alter the progression of poorly differ-
or absence of seminal vesicle or lymph node entiated prostate cancer? J Urol 154:1818-1824, 1995
involvement, and the timing of PSA recurrence are 12. Partin A, Pound C, Clemons J, et al: Serum PSA after
all important variables in predicting eventual local anatomic radical prostatectomy: The Johns Hopkins
Experience after 10 years. Urol Clin North Am
versus distant failure associated with an isolated 20:713, 1993
rise in serum PSA. 13. Partin AW, Carter HB, Chan DW, et al: Prostate
Overall actuarial cause-specific survival at 5 and specific antigen in the staging of localized prostate
10 years was 99% and 93%. Although there was cancer: Influence of tumor differentiation, tumor vol-
no difference in survival among men grouped by ume and benign hyperplasia. J Urol 143:747-752,
TNM stage or preoperative PSA, advancing histo- 1990
logic grade and pathologic stage did have an effect 14. Partin AW, Pearson JD, Landis PK, et al: Evaluation
on actuarial cause-specific survival. of serum prostate-specific antigen velocity after radi-
Men undergoing RRP for clinically localized cal prostatectomy to distinguish local recurrence
from distant metastases. Urology 43:649-659, 1994
prostate cancer showed a 16% actuarial rate of 15. Schroder FH, Hermanek P, Denis L, et al: The
development of metastatic disease at 10 years. This TNM classification of prostate cancer. Prostate
is considerably better than conservative therapy ~ ( s u P P ~129-1
) : 38, 1992
and justifies RRP as the treatment of choice for 16. Trapasso JG, DeKernion JB, Smith RB, et al: The
men with clinically localized disease who are oth- incidence and significance of detectable levels of se-
406 POUND et a1

rum PSA after radical prostatectomy. J Urol Expectant management of localized prostatic cancer.
152:1821-1825, 1994 Cancer 671091-1096, 1991
17. Walsh PC: The natural history of localized prostate 19. Zincke H, Oesterling JE, Blute ML, et a1 Long-term
cancer: A guide to therapy. In Cambell’s Urology, ed (15 years) results after radical prostatectomy for clini-
6, update 13. Philadelphia, WB Saunders, 1995, p 4 cally localized (stage T2c or lower) prostate cancer. J
18. Whitmore WF Jr, Warner JA, Thompson IM Jr, et al: Urol 152:1850-1857, 1994

Address reprint requests to

Charles R. Pound, MD
James Buchanan Brady Urological Institute
Department of Urology
The Johns Hopkins Hospital
600 North Wolfe Street
Baltimore, MD 21287-2101