Salvage Radical Prostatectomy: Outcome Measured by Serum Prostate Specific Antigen Levels

0022-5347/95/153 1-0104$03.00/0

Vol. 153, 104-110,January 1995 Printed in U S A .

W E JOURNAL OF U R O U X ' Y

Copyright 0 1995 by

AMERICAN UROUXICAL ASSOCIATION, INC

SALVAGE RADICAL PROSTATECTOMY: OUTCOME MEASURED BY SERUM PROSTATE SPECIFIC ANTIGEN LEVELS EAMONN ROGERS, MAKOTO OHORI, VAHAN S. KASSABIAN, THOMAS M. WHEELER AND PETER T. SCARDINO From the Scott Department of Urology and Department of Pathology, Baylor College of Medicine, and Urology and Pathology Services, The Methodist Hospital, Houston, Texas

ABSTRACT

We reviewed our experience with salvage radical prostatectomy for locally recurrent cancer in 40 patients to assess the current complication rate and the results using prostate specific antigen

(PSA) as an indicator of treatment outcome and to identify better criteria for the selection of appropriate candidates for this operation. Most recurrent cancers were detected by digital rectal examination (26 patients) or increasing serum PSA levels (10). The operation was technically challenging, with 6 rectal injuries (15%), 2 requiring temporary colostomy. Serious technical complications were more common (31%)among the 29 patients who underwent pelvic lymphadenectomy a t the time of initial radiotherapy than among the 11 treated with external irradiation alone (9%). Urinary incontinence persisted in 18 of 31 evaluable patients (58%) and was successfully corrected with a n artificial urinary sphincter in 9. A total of 21 patients (54%)had pathologically advanced disease (seminal vesicle invasion and/or lymph node metastases). Preoperative PSA levels but not clinical stage or biopsy grade correlated with pathological stage (p <0.03). If the PSA was less than 10 ng./ml. only 15% of the patients had a n advanced pathological stage, compared to 86% if the PSA was 10 or more. After 2 to 97 months (mean 39) 2 patients died of metastatic prostatic cancer, 5 had distant metastases and none had symptomatic local recurrence. At 5 years the actuarial nonprogression rate measured by PSA was 55 % 20%. The only pretreatment factor predictive of progression was the serum PSA level. If the PSA was less than 10 ng./ml. the actuarial rate of progression was significantly lower than if the PSA was greater than 10 (p <0.05). The best results were in the subset of 18 patients with cancer confined to the prostate or immediate periprostatic tissue: 82% had no progression a t 5 years. Within each of these pathological stages the results of salvage prostatectomy were similar to those for standard radical prostatectomy in patients with no prior irradiation. Although technically challenging, salvage prostatectomy provides excellent control of radio-recurrent cancer confined to the prostate or immediate periprostatic tissue and is best performed before the preoperative PSA level increases to greater than 10 to 20 ng./ml. KEY WORDS:prostatic neoplasms; antigens, neoplasm; prostatectomy; radiotherapy

Local recurrence or persistence of cancer remains a major PSA as an indicator of treatment outcome and to identify problem after definitive radiotherapy for prostate cancer.' better criteria for the selection of appropriate candidates for Clinical local recurrence has been associated with an in- this operation. creased risk of distant metastases and death.'.2 By the time local recurrence is detected clinically, however, the tumor is MATERIALS AND METHODS usually extensive and incurable. If salvage therapy is to be Patient population. Otherwise healthy patients with a life successful, persistent cancer must be detected before clinical symptoms develop. One way to detect such cancer is with expectancy of 10 years or longer and with biopsy proved routine needle biopsy of the prostate 1 to 2 years after radio- cancer at least 12 months after definitive radiotherapy were therapy. Such biopsies show persistent malignant cells in 32 considered potential candidates for salvage radical prostatecto 93% of the and patients with a positive biopsy tomy. All patients had a clinically localized tumor (stage TI result have a significantly greater chance of clinical local to 3N0 or X,MO). If a pelvic lymph node dissection was perr e c ~ r r e n c e , ~although . some radiotherapists have disputed formed at the initial treatment, the nodes must have been free of metastases. Only 1 patient had received hormonal this concl~sion.~* Serum prostate specific antigen (PSA) levels have now therapy that could mask the extent of the tumor or the been shown to be a highly accurate marker for monitoring presence of distant metastases. Physical examination, transthe response to r a d i ~ t h e r a p y . ~A -'~ reasonable definition of rectal ultrasonography, serum PSA levels (available in 27 local recurrence after radiotherapy is a positive needle biopsy patients), enzymatic prostatic acid phosphatase levels and result in conjunction with an increasing serum PSA level in bone scan results were consistent with localized disease at a patient with no evidence of metastases.". l 2 Now that the recurrence. PSA levels were measured by the Hybritech Tanmeans are available to detect local treatment failure early, dem-R assay with an upper level of normal of 4.0 ng./ml. The clinicians are faced with the dilemma of managing this prob- preoperative PSA level used for this analysis was the level lem. We reviewed our experience with salvage radical pros- measured closest t o the time of the salvage prostatectomy tatectomy for locally recurrent cancer in 40 patients to assess (median 32 days, range 1to 81). In each case at least 3 weeks the current complications and results of this procedure, using had elapsed since the biopsy. Patients were further evaluated with cystoscopy, examinaAccepted for publication June 10, 1994. tion while under anesthesia, proctoscopy and an imaging 104

105

SALVAGE RADICAL PROSTATECTOMY

study of the upper urinary tracts. Other treatment options, including observation, hormonal therapy and radical cystoprostatectomy, were discussed with each patient, who was fully informed of the risks associated with a salvage operation, including urinary incontinence, impotence, rectal injury and colostomy. From 1984 to 1992, 44 patients were considered for surgical excision of radio-recurrent prostate cancer. Two patients were unsuitable for radical prostatectomy because of radiation cystitis or locally extensive cancer. They were treated with cystoprostatectomy and urinary diversion. In 2 others lymph node metastases were detected by pelvic lymphadenectomy and the planned salvage prostatectomy was not performed. The remaining 40 patients underwent salvage radical prostatectomy with (11) or without (29) a simultaneous pelvic lymph node dissection performed in all patients who had not previously undergone lymphadenectomy. Patient and treatment characteristics are listed in table l. These patients were heavily pretreated (table 2): 29 (73%) underwent prior pelvic lymph node dissection and 26 (63%) had radioactive seed implantation as well. Seed implantation requires mobilization of the prostate with division of the endopelvic fascia and often induces periprostatic bleeding. Local recurrence of cancer was detected by digital rectal examination in 26 patients. Many patients were enrolled in this series before serum PSA levels were routinely measured at our institution but 10 tumors were detected because of an elevated PSA level. Two tumors were detected incidentally in a transurethral resection specimen during treatment of symptomatic bladder outlet obstruction. Two lesions were detected by routine needle biopsy of the prostate (PSA levels were not a ~ a i l a b l e ) . ~ The disease was staged by the tumor, nodes and metastasis classification.14 The stage at the time of recurrence was T l b in 2 patients (detected by transurethral resection of the prostate), T2a or T2b in 15 (palpable nodule in 1 lobe), T2c in 19 (both lobes but confined to the prostate) and T3 in 4. The grade of the biopsy specimen used to detect recurrent tumor was documented in 36 patients. The other 4 cases were recorded as adenocarcinoma of the prostate with no grade. The majority of the tumors (80%)were moderately differentiated, while 2 were well and 5 were poorly differentiated. Of 34 neoplasms graded by the Gleason system 17 (50%) contained a primary or secondary component of poorly differentiated cancer (Gleason grade 4 or 5). Surgical preparation and technique. A thorough mechanical and antibiotic bowel preparation was done to allow primary repair if a rectal injury occurred. In 12 patients the procedure was initiated through a perineal incision to separate the rectum from the prostate to the level of the seminal vesicles, then continued through a midline suprapubic incision. In the remaining 28 patients a standard retropubic approach was used.15 The operation was more difficult in patients in whom a n earlier pelvic lymphadenectomy had obliterated the normal anatomical planes and caused dense adhesions between the bladder and iliac vessels. In the 11 patients treated with external beam therapy alone the tissue planes were surprisingly normal, and mobilization of the bladder and prostate could be performed without difficulty. A combination of a previous lymphadenectomy and radioactive seed implantation resulted in marked fibrosis. The dense fibrosis between the prostate and rectum made this area of

TABLE2. Treatment hefore salvage Drostatectomv Pelvic lymphadenectomy: With radioactive seed implantation: '"iodine seed implants '"'iridium seed implants ""gold seed implants With external beam therapy External beam therapy alone Total

29 (73) 3 2 21 3

11(27) 40

dissection particularly hazardous. In some patients a retrograde approach worked we1116 but in 10 we found an antegrade or lateral approach to be more effective.17 The major advantage of the combined abdominoperineal incision was the avoidance of the difficult dissection between the prostate and rectum from above the tumor. Pathological examination. The whole-mount stepped sections of the prostatectomy specimen were prepared as previously described." Extracapsular extension referred to penetration through the capsule into the periprostatic soR tissue. Surgical margins were considered positive when tumor was found at the inked margins of the specimen. A mapping diagram was prepared by tracing directly from the glass slides the margins of the surgical specimen, prostatic capsule and exact area of carcinoma. Tumor volume was measured from these maps using a planimeter. Patient followup. Patients were evaluated every 3 months for 1 year, every 6 months until year 5 and then annually with a history and physical examination (including a digital rectal examination). Mean followup was 39 months (range 2 to 97). A serum prostatic acid phosphatase level was measured at each visit and a bone scan was repeated each year. Beginning in 1987 serum PSA levels were used to monitor these patients. Since then, bone scans have not been performed if the patient was asymptomatic and the PSA level undetectable. Postoperative PSA levels were available for 37 patients; 2 died before the serum PSA levels were measured, and 1 received hormonal therapy before salvage prostatectomy and was censored from the recurrence analyses. A persistent elevation of PSA to greater than 0.4 ng./ml. was considered disease recurrence. When the PSA level increased, a detailed evaluation, including a bone scan, abdominal and pelvic computerized tomography, and other studies as indicated, was performed to identify the site of recurrent cancer. Clinical recurrence was defined as palpable or biopsy proved local recurrence of cancer in the area of the radical prostatectomy or distant metastases proved by a bone scan or other imaging study in association with an increasing serum PSA level. Methods of analysis. The chi-square test for trends was used to correlate clinical T stage and pathological stage. Regression analysis was used to correlate PSA levels with tumor volume. The Kruskal-Wallis test was used to compare preoperative Gleason score, tumor markers and tumor volume with the clinical and pathological stage of the tumors. The overall survival rate, cancer specific survival rate, clinical nonprogression rate and PSA nonprogression rates were calculated by Kaplan-Meier life table analysis, and the results reported as the mean ? 2 standard errors to give the 95% confidence intervals. RESULTS

TABLE1. Patient age, radiation dose, interval to recurrence and duration of followup Mean (range)

Pt.age (yrs.) Radiation dose (cGy.) Mos. to recurrence Mos. to prostatectomy

61.5 7,194 54.5 58.9

(5573) (4,627-8.800) (14-143) (14-144)

Complications. The operative time, blood loss and length of hospital stay are summarized in table 3. Mean estimated blood loss and transfusion requirements, although modest, were greater than for a standard radical prostatectomy at our institution." The mean operating time and average hospital stay were significantly longer. There were no operative deaths. Surgical complications

106

SALVAGE RADICAL PROSTATECTOMY TABLA?3. Perioperatiue statistics Mean (range)

Operative time ihrs.) Estimated blood loss (cc) Transfusions required (units) Hospitalization (days)

4.4 910 1.44 9.6

(2.S7.0) i350-2,200)

(0-8) (6-16)

directly related to this procedure are listed in table 4. A patient with a ureterovesical stricture requiring reimplantation 3 months postoperatively died of a presumed pulmonary embolus 1 week later. There were 6 rectal injuries. Two patients had vesicorectal fistulas that required secondary closure. Four small rectotomies were closed primarily in multiple layers without complications and in none of these cases was an omental flap interposed. One patient whose bladder had been heavily irradiated suffered ischemic necrosis of the bladder neck, which led to a spontaneous vesicoperineal fistula 9 months after the initial operation. This fistula was successfully treated by cystectomy and urinary diversion. Nine of the 29 patients who underwent bilateral pelvic lymph node dissection when initially treated with radiotherapy (table 2) had a major operative complication (5 rectal injuries, 2 postoperative hemorrhages, 1 ureteral transection and 1 ureteral reimplantation), compared to 1 (rectal injury) of the 11 irradiated patients who had not undergone lymph node dissection (chi-square 2.04, p <0.50). There was no difference in the incidence of rectal injuries whether an abdominoperineal or a retropubic approach was used. A total of 12 patients underwent transurethral resection or open prostatectomy before the salvage prostatectomy. Seven of these patients (58%) had a stricture at the vesicourethral anastomosis, including both patients with a documented bladder neck contracture following transurethral resection of the prostate. In contrast, only 4 of the 28 patients (14%)who had not previously undergone transurethral resection of the prostate had a stricture (chi-square 8.2, p <0.01). The strictures were easily treated with optical urethrotomy and/or bladder neck resection, and all resolved. Additionally, 5 patients required l or 2 urethral dilations but had no endoscopic evidence of an anastomotic stricture. Continence was defined as being completely dry or wearing no more than 1 small pad per day. Urinary continence was achieved in 13 of 31 evaluable patients (42%) who were continent before the operation, have been followed more than 1 year and have not undergone secondary cystectomy. The median interval to continence was 10 months. A total of 18 patients remained incontinent (2 or more pads per day) 1 year postoperatively; 9 regained continence after the insertion of an artificial urinary sphincter, which has been well tolerated with no instances of erosion or infection. The rate of incontinence was not significantly increased in patients who underwent bilateral pelvic lymph node dissection at the time

TABLE4. Intraoperative and postoperative complications in 19 of 40 patients (47.5%) treated with salvage radical prostatectomy No. (70) Type Anastomotic stricture (optical urethrotomy) 11 (27.5) Rectal injury: 6 (15) Closed without complication 4 Requiring further surgery (colostomy. vesicorectal fistula 2 repair) Reoperation for hemorrhage 2 (5) Ureteral transection (immediate repair) 1 (2.5) Ureterovesical junction stricture (delayed reimplantation) 1 (2.5) Vesicoperineal fistula (delayed cystectomy, urinary diversion) 1 (2.5) Septic shock 1 (2.5) Thrombophlebitis 1 (2.5) Wound infection 1 (2.5) Total 25

of initial radiotherapy and/or had a history of transurethral resection of the prostate or open prostatectomy. Pathological findings. The pathological findings in 39 patients (1 excluded because of orchiectomy) are summarized in table 5. In the 37 patients with negative lymph nodes the cancer was organ confined in 8 (22%),specimen confined in 15 (41%)and margin positive in 14 (37%).Overall, 15 of the patients (38%) had positive surgical margins (including 1 with lymph node metastases). Surgical margins were most often positive ( 7 patients) at the apex or adjacent anterior surface and at the bladder neck (4).Three patients had tumor in bladder neck tissue, which did not extend to the surgical margin. Bladder neck invasion was a poor prognostic sign whether or not the surgical margin was positive in this area. All 7 of these patients had seminal vesicle invasion as well, and 5 had positive margins elsewhere in the gland. All 4 patients with bladder neck invasion who have been followed for more than 1 year have an elevated PSA level. Most of the tumors (92%) in the radical prostatectomy specimens were moderately differentiated (Gleason score 5 to 7) and the remainder were poorly differentiated (score 8 to 10). A primary or secondary component of poorly differentiated cancer (Gleason grade 4 or 5) was present in 29 cases (74%). Correlation of clinical features with pathological stage. PSA levels before the operation (27 patients) correlated with total tumor volume (r = 0.6, fig. 1). Neither clinical stage (chi-square 4.2, p <0.3), biopsy Gleason score (chi-square 4.16, p c0.3) nor preoperative acid phosphatase levels (chisquare 3.38, p C0.3)correlated with the pathological extent of the tumor. In contrast, preoperative PSA levels and tumor volume correlated significantly with pathological stage (chisquare 11.97, p <0.03 and (chi-square 17.013, p <0.001, respectively, fig. 2). A PSA level of 10 or tumor volume of 3 ~ mseparated . ~ the patients into those with a low or high risk of having a pathologically advanced stage (seminal vesicle invasion or lymph node metastases). Of 13 patients with a PSA level of less than 10 ng./ml. only 2 (15%) had an advanced pathological stage compared to 12 of 14 (86%)with a PSA of 10 ng./ml. or more. Survival and recurrence. Table 6 shows the current status of the 39 uncensored patients. Four patients died: 2 of causes unrelated to the cancer with no evidence of prostate cancer at autopsy, and 2 of metastatic prostatic cancer at 42 and 68 months. The actuarial cancer specific survival rate was 95 2 8%(mean 5 2 standard errors) at 5 years and 87 2 18%at 8 years (fig. 3). No patient had symptomatic local recurrence. Distant metastases developed in 7 patients and 2 of these also had biopsy proved local recurrence. Actuarial freedom from local or distant recurrence was achieved in 83 ? 14%at 5 years and 67 2 24% at 8 years (fig. 3). When PSA levels were used to detect recurrence, the actuarial nonprogression rate at 5 and 8 years was 55 & 20% and 33 2 24%,respectively (fig. 4). The 5-year actuarial nonprogression rate was 100%for patients with organ confined cancer, 71 2 34%with extracapsular extension and 28 2 26% with seminal vesicle invasion (table 7). Of patients with cancer limited to the

TABLE5. Pathological stage in 39 patients (2 was excluded because of prior orchiectomvi Stage pTl-24onfined to prostate pT3a,&xtracapsular extension: Neg. surgical margins Pos. surgical margins pT3c-seminal vesicle invasion: Neg. surgical margins Pos. surgical margins pN+--pos. pelvic lymph nodes Total

No.

(%)

8 (20)

10 (26) 7 (18) 3 (8) 19 (49) 8 (21) 11 (28) 2 (5) 39 (100)

107

SALVAGE RADICAL PROSTATECTOMY

TABLE6. Current status o f 3 9 patients followed for 2 to 97 months (mean 39.3)

100

No. (%) 35 (90)

Alive: No recurrence (PSA 0.4 or less) Recurrence (PSA more than 0.4, pos. bone scan or pos. nodes) Dead: No recurrence Due to prostate Ca

10

PSA ng/ml

J

0

0

0.75-

.a 2 0.50-

I

I

I

1

4

8

12

16

I 20

FIG. 1. Correlation between PSA levels (logarithmic scale) before salvage prostatectomy and total tumor volume ( ~ m .of ~ )irradiated tumor calculated from whole-mount stepped sections of prostate in 22 patients with PSA levels available.

prostate or immediate periprostatic tissue 82 2 23%had an undetectable PSA level at 5 years compared to only 31 ? 24% of 21 with seminal vesicle invasion or positive nodes (fig. 5). The clinical stage and grade in the biopsy specimen were not predictive of disease recurrence. The only pretreatment factor that predicted outcome was the serum PSA level (fig. 6). Among 13 patients with a PSA level of less than 10 ng./ml. the actuarial nonprogression rate at 4 years was 50% compared to 29%for 14 patients with a PSA level of greater than 10 ng./ml. (p (0.05).

In this series of 40 patients treated with radical prostatectomy for locally recurrent prostate cancer after definitive radiotherapy only 2 died of cancer and 7 have clinical recur-

Ca deaths 2 Clinical recurrence 7 Total patients 39

Baylor, 1994

0-

d

i

h

A

4

L

Time (years)

i

b

S

Q

FIG. 3. Actuarial cancer specific (Cu spec) survival and clinical recurrence (local or distant) after salvage prostatectomy (mean % 2 standard errors).

rence. The series is not sufficiently mature (mean followup 39 months) for these end points to assess the outcome adequately. ARer radical prostatectomy PSA is a more sensitive indicator of progression. In our series of more than 500 standard radical prostatectomies no patient had clinical recurrence without an elevated PSA level." If no adjuvant hormonal therapy is administered, PSA levels usually become elevated within 3 years of radical prostatectomy in patients with recurrence.21 Consequently, the rate of progression, measured as the time until the PSA level first becomes detectable, can serve as a reasonable intermediate end point to assess the results of salvage radical prostatectomy in our patient population.

B

I

10

2

a 0.25-

Total Tumor Volume (cm3)

A

2 (5)

Free recurrence of clinical

h

0

4 (10) 2 (5)

Ca spec survival

c

r = 0.6; p c 0.01.

16 (43)

w

1.oo-

1

0.1

19 (47)

i

Totat Tumor Volume (cm3)

I

0.1

II

PSA

0

0

nglml

..O

n=27

n=22 pT1-2 pT3a.b (Confined) (ECE)

pT3c (SVI)

Pathologic Stage

"

nr

Y. 1

pT1-2 pT3a.b p T k (Confined) (ECE) (SVO

Pathologic Stage.

FIG. 2. A, correlation of total tumor volume (logarithmic scale) and pathological stage in 22 patients with tumor volumes and PSA levels available. B, correlation of preoperative PSA levels and pathological stage in 27 patients. ECE, extracapsular extension. SIT, seminal vesicle invasion. N+,positive pelvic lymph nodes.

108

SALVAGE RADICAL PROSTATECTOMY

0.75-

Confined (n=8) or ECE (n=10) 0.50-

SVI (n=19) or LN+ (n=2)

0.25

0.251

a

Logrank test

I....__....______.._

(n=39) 0

---- 95% C o n f i i Limita I

I

0

I

1

I

2

3

,

4

Baylor. 1994 I

5

Time (years)

I

6

I

7

I

8

IF'Ic. 4. Actuarial nonprogression rates based on postoperative PSA levels after salvage prostatectomy. Progression was defined as PSA level of 0.4 ng./ml. or more and increasing.

TABLE7. Comparison of 5-year actuarial nonprogression rates (undetectable serum PSA level) in 39 patients treated with radical prostatectomy after radiotherapy (salvage) and 500 treated with radical Drostatectomv with no Drior theraDv (standard) Mean Nonprogression Rates (95% confidence intervals) Salvage No.(%) Standard NO.(%) Totals Preop. PSA Less than 10 10 or more Pathological stage: pT1-2 (confined) pT3a.b (extracapsular extension) pT3c (seminal vesicle invasion) or pN+ Pas. surgical margins

39

( 5 5 2 20)

500

(7625)

13 14

(50 2 50) ( 2 9 2 26)

300

140

(82 2 8) (64 2 12)

8

10

(100) (71 2 34)

226 157

(94 2 4) (75 2 10)

21

(28224)

94

(442 12)

15

(46 2 32)

83

(61 2 16)

As shown in figure 4,55% of our patients had no evidence of progression at 5 years and 33% at 8 years after salvage prostatectomy. The best results were in the subset of 18 patients (46%)with cancer confined to the prostate (stage pT1 to 2) or immediate periprostatic tissue (stage pT3a,b): 82% of these patients had no progression at 5 years. In fact, when the outcome was determined for each pathological stage, the results of salvage prostatectomy were essentially the same as for standard radical prostatectomy in patients with no prior irradiation (table 7).20-21The major difference in the overall results (55%versus 76% free of progression at 5 years) reflects the more advanced pathological stage of the salvage prostatectomy patients. Only 15to 20% of previously untreated patients in radical prostatectomy series have seminal vesicle invasion (stage pT3c) or positive lymph nodes (stage pN+) compared to 54% of those in our salvage prostatectomy series (table 5)z0*21 Positive surgical margins, which suggest that the tumor has not been completely eradicated, were present in 38% of salvage prostatectomy patients compared to 17% undergoing standard prostatectomy, and 46% of these patients remain free of progression a t 5 years (table 7). Thus, salvage prostatectomy provides excellent control of cancer confined to the prostate or immediate periprostatic tissue but i t is particularly difficult to determine the pathological stage of an irradiated tumor. In our series neither clinical stage, Gleason grade in the biopsy specimen nor serum acid phosphatase levels (within the range of normal) accurately predicted pathological stage. Preoperative serum PSA levels, however, did discriminate between early (patho-

pc.005

0-

Baylor, 1994

I

I

I

I

0

1

2

3

I

I

I

I

4

5

6

7

Time (years)

4

FIG. 5. Actuarial nonprogression rates by pathological stage. ECE, extracapsular extension. SVZ,seminal vesicle invasion. L N + , positive pelvic lymph nodes.

logically confined or with extracapsular extension only) and advanced stages (fig. 2, B ) . Serum PSA levels correlated well with the volume of these irradiated tumors (fig. 2, A). The correlation was similar to that reported in nonirradiated disease." Since tumor volume correlates with pathological stage in irradiated (fig. 2, A) as well as nonirradiated can~ e r serum ? ~ PSA levels provide the clinician with a reasonably accurate way to predict pathological stage in patients who may be candidates for salvage radical prostatectomy. Nearly 85%of our patients with a PSA level of less than 10 ng./ml. had an early stage tumor. Definitive irradiation therapy for clinically localized prostate cancer has a profound effect on serum PSA levels, which begin to decrease within 3 months and reach a nadir 3 to 18 months later.'-13 Within 5 years, however, PSA levels begin to increase consistently in more than half of the patients. A needle biopsy will document recurrent cancer in more than 80% of these patients5*11, 24 In the absence of evidence of metastases, an increasing serum PSA level combined with a positive biopsy result documents local recurrence of the cancer. Since there is no established curative therapy for radiorecurrent cancer, however, there has been little motivation for the early detection of recurrence in asymptomatic men. Our results suggest that many radio-recurrent tumors can be controlled if they are treated early and that serum PSA levels are the best indicator of the pathological stage of the cancer. Patients who are potential candidates for salvage prostatectomy should be treated before the serum PSA level increases above 10 to 20 ng./ml. Two of our patients had a PSA level between 10 and 20 ng./ml., and 1 of them had advanced cancer. Eleven of the 12 patients (92%)with a PSA level of 20 ng./ml. or greater had advanced cancer. A number of treatment options are available for patients with local recurrence after radiotherapy. Most are managed expectantly but the median interval to symptomatic local progression with expectant management or watchful waiting is 6 to 7 years in nonirradiated patientsz5, 26 and may be even shorter when irradiation fails.27 Deferring therapy until occurrence of symptomatic progression will, therefore, be feasible only temporarily. Androgen deprivation will provide effective tumor control and delay local morbidity but androgen insensitivity and late recurrence are inevitable in the long term.6 The experience with cryotherapy for these tumors is limited. Its safety and efficacy are not clear, and long-term cancer control has not been documented.28 Further irradiation therapy invites substantial risks and there is little evidence that it will alter the course of a radio-recurrent cancer.29*30 Pelvic exenteration is usually reserved for those symptomatic advanced tumors that have invaded the bladder neck or rectum and that are refractory to hormonal thera-

109

SALVAGE RADICAL PROSTATECTOMY

PSA<10 (n=13) 1

PSA>10 ( ~ 1 4 )

Logrank test

pc.05

0-

d

i

Baylor. 1994

1

Time (years)

4

k

6

FIG.6. Comparison of nonprogression rates (PSA 0.4 ng./ml. or less) for patients with preoperative PSA levels of greater or less than 10 ng./ml.

py.31 Although local control can be achieved, few such patients are cured.32.33 Whether radical cystoprostatectomy can cure some patients with locally extensive cancer remains unclear. All 7 of our patients with bladder neck invasion also had seminal vesicle invasion, suggesting that most tumors would have recurred regardless of the extent of surgical resection. While salvage prostatectomy provides excellent local control and a high probability of long-term eradication of radiorecurrent cancer detected at an early pathological stage, the operation is technically challenging. Operative time is longer than for standard radical prostatectomy because of the fibrosis associated with radiotherapy and the need for sharp dissection of the prostate from the rectum. The procedure is associated with a high incidence of rectal injuries, incontinence and anastomotic stricture^.^^. 32, 34, 35 Our complication rate may have been higher than in other series because most of our patients underwent pelvic lymph node dissection and mobilization of the prostate for radioactive seed implantation as well as external beam irradiation, resulting in severe pelvic fibrosis. Serious intraoperative complications, especially rectal injuries, were more common (31%)in patients who had previously undergone pelvic surgery plus radiotherapy than in patients who had received radiotherapy alone (9%).We were not able to decrease the risk of rectal injury by initiating the operation through a perineal incision. However, with adequate mechanical and antibiotic preparation of the bowel before the operation, most rectal injuries were closed primarily without incident. One of the 2 patients in our series who required colostomy also required resection of the entire anterior rectal wall to remove the cancer completely, while in the other a vesicorectal fistula developed on postoperative day 8. Anastomotic strictures were common in our series but most occurred before we began to evert the mucosa over the reconstructed bladder neck.I6 Nevertheless, these patients are at increased risk for a stricture because of the decreased blood supply to the bladder neck and the fibrosis induced by the radiation, which compromise anastomotic healing. A simple prostatectomy (transurethral resection of the prostate or open) was the most important factor predisposing patients to an anastomotic stricture. Most strictures resolved after a single optical urethrotomy. Of the patients followed for more than 1 year 56% were incontinent, a higher rate than we previously reported.15,36 Incontinence is significantly more common after salvage than after standard radical p r o s t a t e c t ~ m y32. , ~34. ~ ~36 perhaps because radiation-induced fibrosis combined with some degree of paresis to the external sphincter exacerbates the effects of the operation itself on the external sphincter mech-

anism. The return to continence has taken up to 10 months in some patients, and so we prefer to delay insertion of an artificial urinary sphincter for at least 1 year. A successful outcome can be anticipated in most patients requiring an artificial sphincter. When a prosthesis was used, it was placed around the bulbous urethra, outside of the field of radiation. Nine prostheses have been inserted in 18 incontinent patients without complication. Throughout the last decade there have been sporadic reports of radical prostatectomy for radio-recurrent cancer. Most series are small and many patients received androgen deprivation therapy preoperatively. To our knowledge our series of 40 patients is the largest reported to date, with morphological and survival data on 39 patients who had not received hormonal therapy. When the tumor was confined to the prostate or immediate periprostatic tissue, the likelihood of cancer control after salvage radical prostatectomy was comparable to the results of radical prostatectomy in nonirradiated patients. Only 46% of our patients had early stage cancer. However, when our patients with a locally recurrent prostate cancer were considered for salvage prostatectomy before the PSA level was greater than 10 ng./ml. 85% had early stage disease. We believe that salvage prostatectomy should be considered only in men with a biopsy proved tumor 1year or longer after the completion of irradiation therapy, who initially had a low stage (T1 to 2) cancer with no evidence of nodal metastases if a pelvic node dissection was performed and who have received no hormonal therapy that would mask the presence of metastases. These patients should be in good general health with a life expectancy of greater than 10 years and have no evidence of persistent radiation proctitis or cystitis. Patients who may be candidates for salvage radical prostatectomy after irradiation therapy should be monitored regularly with serum PSA levels in addition to digital rectal examinations. Transrectal ultrasonography and ultrasound guided needle biopsies should be performed if the PSA level begins to increase. If the patient elects salvage prostatectomy, it should be performed before the serum PSA increases to greater than 10 to 20 ng./ml. REFERENCES

1. Kaplan, I. D.,Prestidge, B. R., Bagshaw, M. A. and Cox, R. S.: The importance of local control in the treatment of prostatic cancer. J. Urol., 147: 917, 1992. 2. Fuks, Z.,Leibel, S. A,, Wallner, K. E., Begg, C. B., Fair, W. R., Anderson, L. L., Hilaris, B. S. and Whitmore, W. F.: The effect of local control on metastatic dissemination in carcinoma of the prostate: long term results in patients treated with ''1 implantation. Int. J. Rad. Oncol. Biol. Phys., 21: 537, 1991. 3. Scardino, P. T. and Wheeler, T. M.: Local control of prostate cancer with radiotherapy: frequency and prognostic significance of positive results of postirradiation prostate biopsy. NCI Monogr., 7: 95,1988. 4. Lee, F., Torp-Pedersen,S., Meiselman, L., Siders, D. B., Littrup, P., Dorr, R. P. and Pauli, F. J.: Transrectal ultrasound in the diagnosis and staging of local disease after I125 seed implantation for prostate cancer. Int. J. Rad. Oncol. Biol. Phys., 1 5 1453, 1988. 5. 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., 142: 326, 1989. 6. Kuban, D.A,, el-Mahdi, A. M. and Schellhammer, P. F.: Prognostic significance of post-irradiationbiopsies. Oncology, 7: 29, 1993. 7. Cox, J. D. and Stoffel, T. J.: The significance of needle biopsy after irradiation for stage C adenocarcinoma of the prostate. Cancer, 4 0 156, 1977. 8. Hanks, G.E.:External beam radiation treatment for prostate cancer: still the gold standard. Oncology, 6 79, 1992. 9. Stamey, T. A., Kabalin, J. N. and Ferrari, M.: Prostate s&c

110

SALVAGE RADICAL PROSTATECTOMY

antigen in the diagnosis and treatment of adenocarcinoma of the prostate. 111. Radiation treated patients. J. Urol., 141: 1084, 1989. 10. Kaplan, I. D., Cox, R. S. and Bagshaw, M. A.: Prostate specific antigen after external beam radiotherapy for prostate cancer: followup. J. Urol., 149 519, 1993. 11. Goad, J. R., Chang, S. J., Ohori, M. and Scardino, P. T.: PSA after definitive radiotherapy for clinically localized prostate cancer. Urol. Clin. N. h e r . , 2 0 727, 1993. 12. Zagars, G. K. and von Eschenbach, A. C.: Prostate-specific antigen: an important marker for prostate cancer treated by external beam radiation. Cancer, 7 2 538, 1993. 13. Schellhammer, P. F., El-Mahdi, A. M., Wright, G. L., Jr., Kolm, P. and Ragle, R.: Prostate-specific antigen to determine progression-free survival after radiation therapy for localized carcinoma of the prostate. Urology, 42: 13, 1993. 14. Schroder, F. H., Hermanek, P., Denis, L., Fair, W. R., Gospodarowicz, M. K. and Pavone-Macaluso, M.: The TNM classification of prostate cancer. Prostate, suppl., 4 129,1992. 15. Goldstone, L. M., Scardino, P. T., Cantini, M. and Egawa, S.: Salvage radical prostatectomy. In: Problems in Urology: Controversies in Prostate Cancer Management. Edited by D. F. Paulson and R. D. Williams. Philadelphia: J. B. Lippincott Co., 1990. 16. Walsh, P. C.: Radical retropubic prostatectomy: surgical anatomy. In: Campbell’s Urology, 6th ed. Edited by P. C. Walsh, A. B. Retik, T. A. Stamey and E. D. Vaughan, Jr.: Philadelphia: W. B. Saunders Co., vol. 3, chapt. 78, p. 2865, 1992. 17. Campbell, E. W.: Total prostatectomy with preliminary ligation of the vascular pedicles. J. Urol., 81: 464, 1959. 18. Wheeler, T. M.: Anatomic considerations in carcinoma of the prostate. Urol. Clin. N. Amer., 1 6 623, 1989. 19. Fitzgerald, R B., Goad, J . R., Collini, P., Kassabian, V. and Scardino, P. T.: Declining intraoperative blood loss during radical prostatectomy: is autologous blood donation necessary? J. Urol., part 2, 149 378A, abstract 661, 1993. 20. Goad, J . R., Kassabian, V. S., Weaver, R. L. and Scardino, P. T.: PSA as a measure of recurrent prostate cancer after radical prostatectomy. J . Urol., part 2, 149 447A. abstract 939, 1993. 21. Partin, A. W., Pound, C. R., Clemens, J . Q., Epstein, J . I. and Walsh, P. C.: Serum PSA after anatomic radical prostatectomy. The Johns Hopkins experience after 10 years. Urol. Clin. N. Amer., 20.713, 1993. 22. 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. 11. Radical prostatectomy treated patients. J . Urol., 141: 1076, 1989. 23. Stamey, T. A,, Freiha, F. S., McNeal, J . E., Redwine, E. A,,

Whittemore, A. S. and Schmid, H. P.: Localized prostate cancer. Relationship of tumor volume to clinical significance for treatment of prostate cancer. Cancer, suppl. 3, 71: 933, 1993. 24. Dugan, T. C., Shipley, W. U., Young, R. H., Verhey, L. J., Althausen, A. F., Heney, N. M., McManus, P. L. and Abraham, E. H.: Biopsy after external beam radiation therapy for adenocarcinoma of the prostate: correlation with original histological grade and current prostate specific antigen levels. J. Urol., 146 1313, 1991. 25. Whitmore, W. F., Jr., Warner, J . A. and Thompson, I. M., Jr.: Expectant management of localized prostatic cancer. Cancer, 67: 1091, 1991. 26. Adolfsson, J., Carstensen, J. and Ldwhagen, T.: Deferred treatment in clinically localised prostatic carcinoma. Brit. J . Urol., 6 9 183, 1992. 27. Holzman, M., Carlton, C. E., J r . and Scardino, P. T.: The frequency and morbidity of local tumor recurrence after definitive radiotherapy for stage C prostate cancer. J . Urol., 146 1578, 1991. 28. Cohen, J . R, Onik, G. and Miller, R.: Cryosurgical ablation of the prostate in patients who have failed radiation therapy as primary treatment for adenocarcinoma of the prostate. J . Urol., part 2,149256A. abstract 170, 1993. 29. Goffinet, D. R., Martinez, A., Freiha, F., Pooler, D. M., Pistenma, D. A., Cumes, D. and Bagshaw, M. A,: 125Iodine prostate implants for recurrent carcinomas after external beam irradiation: preliminary results. Cancer, 4 5 2717, 1980. 30. Cumes, D. M., Goffinet, D. R., Martinez, A. and Stamey, T. A,: Complications of lZ5iodineimplantation and pelvic lymphadenectomy for prostatic cancer with special reference to patients who had failed external beam therapy as their initial mode of therapy. J . Urol., 126 620, 1981. 31. Aherling, T. E., Lieskovsky, G. and Skinner, D. G.: Salvage surgery plus androgen deprivation for radioresistant prostatic adenocarcinoma. J . Urol., 147: 900, 1992. 32. Mador, D. R., Huben, R. P., Wajsman, Z. and Pontes, J. E.: Salvage surgery following radical radiotherapy for adenocarcinoma of the prostate. J . Urol., 133 58, 1985. 33. Moul, J . W. and Paulson, D. F.: The role of radical surgery in the management of radiation recurrent and large volume prostate cancer. Cancer, 6 8 1265, 1991. 34. Link, P. and Freiha, F. S.: Radical prostatectomy aRer definitive radiation therapy for prostate cancer. Urology, 37: 189, 1991. 35. Zincke, H.: Radical prostatectomy and exenterative procedures for local failure after radiotherapy with curative intent: comparison of outcomes. J . Urol., 147:894, 1992. 36. Neerhut, G. J., Wheeler, T., Cantini, M. and Scardino, P. T.: Salvage radical prostatectomy for radiorecurrent adenocarcinoma of the prostate. J . Urol., 140: 544, 1988.