- Email: [email protected]
Prostate Specific Antigen Detected Prostate Cancer (Clinical Stage T1C): An Interim Analysis
0022-5347/96/1553-0821$03.00/0 THEJUURNAL OF UROLOGY Copyright D 1996 by AMERICAN UROL~CICAL ASSOCIATION, INC
Vol. 155,821-826, March 1996 Printed in U.S.A.
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE TlC): AN INTERIM ANALYSIS SETH E. LERNER,* THOMAS M. SEAY, MICHAEL L. BLUTE, ERIK J. BERGSTRALH, DAVID BARRETT AND HORST Z1NCKE-t From the Department of Urology and Section of Biostatistics, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
ABSTRACT
Purpose: The majority of impalpable prostate cancers (stage Tlc) are biologically significant. We report the interim results in 257 patients with stage T l c prostate cancer treated with radical prostatectomy. Materials and Methods: Prostate specific antigen progression-free survival was assessed by the Kaplan-Meier method. Multivariate analyses were performed to determine which clinical and pathological variables independently correlated with progression. Comparisons among the various clinical substages ( T l a to T2b/c) were calculated. Results: Of the patients with stage T l c cancers 51% had stage pT2c or less and 91% had clinically significant tumors on the basis of pathological grade, deoxyribonucleic acid ploidy and tumor volume. High preoperative prostate specific antigen, poorly differentiated tumors and nondiploid status were strong independent predictors of progression. The 5-year survival rate free of progression was 84%. Patients with clinical stage T l c cancers had a significant progression-free survival advantage compared to those with clinical stage T2bk disease (p = 0.0005). Conclusions: Impalpable tumors should not be regarded a s insignificant or innocuous on the basis of pathological analysis. Disease-free survival in the stage T l c group was similar to that in the clinical stages T l a to T2a group but significantly better than that in the T2blc group. KEYWORDS:prostatic neoplasms; prostate-specific antigen; antigens, neoplasm; prostatectomy To our knowledge serum prostate specific antigen (PSA) is the most important and clinically significant tumor marker available to date. Since 1987, PSA has been commonly used to monitor response to cancer therapy and as a screening modality for the detection of potentially curable disease. As a result of its widespread use, adenocarcinoma of the prostate is presently the most commonly diagnosed visceral malignancy in men in the United States.’ Of greater importance is the fact that prostate cancer is second only to lung cancer as a cause of cancer death,’ and more than 40,000 men will die of the disease in 1995. Endorsements for PSA screening have been issued by the American Cancer Society2 and American Urological Association. PSA detects a significant number of prostate cancers missed by routine digital rectal examination.= Serum PSA elevation correlates with tumor volume, grade and pathological stage.68 According to previous studies PSA is increased by 2.2 ng./ml. (corrected from Yang to Hybritech assays) for each gram of malignant tissue present.6 Therefore, PSA testing is not likely to detect indolent malignancies because such tumors are not large enough to cause elevations in serum PSA. Two recent studies demonstrated that the majority of impalpable stage Tlc prostate cancers are biologically significant on the basis of tumor volume, grade and pathological stage, and only a few percent are similar to innocuous or latent autopsy cancers.g.10 Oesterling et a1 reported that stage T l c prostate cancers have pathological characteristics that essentially parallel those of palpably detected tumor^.^ We report the interim results of 257 patients with stage Tlc Prostate cancer treated with radical prostatectomy.
MATERIALS AND METHODS
Between 1987 and 1991, 2,203 patients 44 to 83 years old (median age 67) with clinically localized (stage T2c or less) adenocarcinoma of the prostate underwent bilateral pelvic lymphadenectomy and radical retropubic prostatectomy. Of these patients 257 (12%)had impalpable, PSA detected prostate cancer (clinical stage Tlc). The prostate gland was considered benign (no palpable evidence of asymmetry, nodularity, irregularity or induration) on the basis of digital rectal examination by usually a t least 2 examiners. Prostate cancer was diagnosed on the basis of tissue obtained usually by sextant transrectal needle biopsy. Preoperative PSA values were determined using the Hybritech Tandem-R PSA assay (reference range 0.0 to 4.0 ng./ml.)” The prostate gland was imaged with real-time transrectal ultrasonography as described previo~sly.~Our patients did not represent a screened population. In many patients serum PSA concentrations were measured as part of routine physical examinations. In addition, many patients had been referred to our institution with an established diagnosis of prostate cancer. All biopsy specimens obtained by local physicians were reviewed by pathologists at our clinic. The histological grades of the biopsy and radical prostatectomy specimens were determined using the Mayo grading system.12 The revised TNM staging system was used for clinical and pathological staging.13 All lymph nodes removed during the modified bilateral pelvic lymph node dissection were examined immediately by frozen section and later by formal paraffin section. The radical prostatectomy specimens were examined by the pathologist immediately after resection using gross inspection as well as multiple frozen secAccepted for publication September 29,1995. * Current address: Department of Urology, Albert Einstein Col- tions.14 All surgical margins, including the prostatic base, lege of Medicine, Montefiore Medical Center, 111 East 210th St., apex, urethra, bladder neck, capsule, penprostatic soft tissue Bronx, New York 10467. t Requests for re M t s : Department of Urology, Mayo Clinic, 200 and seminal vesicles, were evaluated. Multiple sections from the prostate were examined immediately by frozen section at First St., S.W.,h i e s t e r , MiMeSOta 55905. 821
822
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE T1C)
tectomy for clinically localized prostate cancer between 1987 and 1991 are summarized in table 1. Median preoperative PSA for patients with clinical stage Tlc disease was 10.0 ng./ml. (range 0.6 to 201). Of the 14 patients with a preoperative serum PSA of 4.0 ngJml. or less 8 had an increase of greater than 0.75 ng./ml. during the 12 months before diagnosis and 6 had abnormal findingson transrectal ultrasound. Overall, 12% of the patients who underwent radical prostatectomy for clinically localized disease during this interval had clinical stage Tlc cancer. However, the frequency of surgery for PSA detected tumors has increased steadily since 1987 (table 2). Of the patients who underwent radical retropubic prostatectomy in 1987 for clinically localized prostate cancer only 2% had stage Tlc disease, whereas 18% who underwent surgery in 1991 had PSA detected tumors. Pathological staging revealed that the incidence of organconfined prostate cancer (stage pT2c or less) ranged from 36% in patients with clinical stage TZb/c disease to 87% in those with clinical stage Tla prostate cancer (table 1).Of patients with clinical stage Tlc disease 51,36,9 and 4% had pathological stage pT2c or less, pT3a/b, pT3c and N+, respectively. A total of 36 patients (14%)had focally positive apical margins as the only evidence of extraprostatic disease and they were included in the pathological stage pT3a/b group. The relationship between preoperative PSA and pathological stage in patients with stage Tlc prostate cancer is shown in table 3. The incidence of pathologically organ-confined disease decreased with increasing preoperative serum PSA values. Of the patients with a normal preoperative PSA (4.0 ng./ml. or less) 79% had organ-confined disease on final pathological analysis, compared to only 33% and none of those with preoperative values of 20.1 to 50 ng./ml. and greater than 50 ng./ml., respectively. No patient with PSA of 10.0 ng./rnl. or less had disease spread to regional lymphatics. Table 1 includes data pertaining to estimated tumor volume in patients with clinically localized prostate cancer. Median tumor volume in patients with clinical stage Tlc disRESULTS ease was 4.0 cc (range 0.03 to 56). Of the stage Tlc cancers The characteristics of the 2,203 patients undergoing bilat- 89% had a total volume of at least 0.5 cc. Therefore, only 24 eral pelvic lymphadenectomy and radical retropubic prosta- of these tumors (11%)could be considered insignificant on
operation and later by routine fixed paraffin sections. Specimens with positive surgical margins but without other evidence of extraprostatic disease were classified within the stage pT3 subgroup. Total tumor volume was estimated from the predominant tumor foci in the gland, using width, length and height measurements, and summing the individual volumes. Tumor deoxyribonucleic acid (DNA) histograms, c ~ e ated by previously described techniques.15 were classified as diploid, tetraploid and aneuploid patterns. The patients were followed every 3 to 4 months for 2 years postoperatively, every 6 months for the next 3 years and then annually. A digital rectal examination, radionuclide bone scan, plain radiograph when indicated, chest x-ray and serum PSA test were performed at each followup visit. Bone scam were performed less often since the advent of PSA testing but were obtained at least once per year. In patients not returning to our institution the PSA concentrations were determined at our clinic via a mailed blood specimen, or the patients were contacted annually with additional medical information obtained from local physicians as indicated. Palpable local recurrence was confirmed by transrectal needle biopsy. Systemic progression was diagnosed on the basis of a positivebone scan andtor plain films.The common end point of progression was defined as either an elevated postoperative PSA level (more than 0.2 ngJml. on at least 2 occasions), local recurrence or distant metastases. Progression-free survival was assessed by the KaplanMeier method.16 To account for more than 1prognostic variable simultaneously, the clinical and pathological parameters were evaluated using a Cox proportional hazards model to determine which variables were independently correlated with progre~sion.~~ During the same period 30,97.449 and 1,370 patients with clinical stages Tla, Tlb, T2a and T2Wc disease, respectively, underwent radical prostatectomy. Comparisons among the various clinical substages with respect to the end point of progression-free survival were calculated.
TABLE1. ChamcteTiStics of 2,203 patients with clinically localized prostate cancer treated with radical retropubic prostatectomy between 1987 and 1991 Clinical Stage
~~
~~
~
Median age at radical ret-
~
65.5
~~
(51-73)
~
~
~~
(39-82)
~~~
65.0 (48-75)
67.0 (38-81)
67.0
67.0 (4443)
4.5 (0.1-108) 49.4 23.4 14.3 13.0 20
6.5(0.4-260) 28.7 41.1 20.8 9.1 41
9.2 (0.2-321) 18 35.8 24.1 22.1 86
10.0(0.&201) 7 43.2 28.8 21.0 4
88.2 11.8 0.0 13
76.8 16.1 8.1 35
74.2 22.6 3.2 73
63.5 27.9 8.6 116
78.2 15.5 6.3 18
80.8 19.2
74.4 25.6
69.9 30.1
55.6 44.4
72.8 27.2
86.7 10.0 3.3
69.0 28.9 2.1
56.6 40.5 2.9
35.7 52.8 11.5
51.0 45.1 3.9
~
0.040
ropubic pmstateetomy (range)
h p . PSA (ngJml.f: Median (ranee) % 4.0 or lea8 % 4.1-10.0 % 10.1-20.0 % Greater than 20.0 No. unknown
DNA ploidy: % Diploid % Tetraploid % Aneuploid No. unknown % Pathological grade: LOW
High 8 TNM pfIthological.tagB: pTZc or l e d P n
N+
<0.001 3.0 (0.1-31.3) 56.0 36.0 4.0 4.0
5
<0.001
<0.001 <0.001
<0.001 Tumor vol.: Median cc (range) 3.4 (0.66-28) 4.2 (0.12-72) 3.0 (0.02-98) 6.1(0.01-143) 4.0(0.03-56) No. unknown 13 55 56 152 33 From Kruskal-Wallis or chi-square teat of the null hypothesis that all clinical stages have the same distribution. t Patients with positive surgical margins but without other evidence of extraprostatic disease were considered to have greater than stage pT2c disease.
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE TIC) TABLE 2. Distribution Yr.
No. Pts.
1987 1988 1989 1990 1991
216 355 385 494 753
of
clinical stage by year of surgery Clinical Stage (%)
Tla
Tlb
T2a
T2b
TIC
1 2 1 1
5
24 23 19 20 19
67 62 65 62 59
2 6 8 13 18
1
8 6 4 2
TABLE3. Relationship of preoperative serum PSA and pathological stage in patients with clinical stage T l c prostate cancer % Pathological Stage
PSA (ng./ml.)
No. Pts.
(255 pta.')
N+
pT3a-c
pT2c or Less
4.0 or less 14 12 28 4.1-10.0 60 40 112 10.1-20.0 41 49 75 20.lbO.O 49 33 53 50.1+ 5 0 100 * Two patients were missing preoperative PSA values.
0 0 4 14 -
the basis of volumetric analysis. Of these smaller lesions 3 (12.5%) were either poorly differentiated or nondiploid. Therefore, only 9% of PSA detected or impalpable cancers should be considered insignificant. Among the 24 patients with clinical stage Tlc prostate cancers less than 0.5 cc in total volume 20 (83%) were younger than 70 years and 12 (50%) were younger than 65 years. Survival and progression. Of the 257 patients 251 were alive at 1.3 to 7.4 years of followup (mean 3.5, median 3.2). Overall, 6 patients died, only 1of prostate cancer. Due to the few events (death) no analyses were done with respect to crude and cancer-specific survivals. There were 150 and 34 patients alive at 3 and 5 years of followup, respectively. Projected 3 and 5-year survival rates free of the combined end point of local, systemic and/or PSA progression (more than 0.2 ng./ml.) were 88%and 841,respectively (fig. 1). As expected, pathological stage impacted on progression-free survival. Patients with stage pT2c or less disease had significant progression-free survival advantages compared to those with extraprostatic or nodal tumor (fig. 2). Of the patients with stage pT2c or less cancer 96% were free of PSA progres-
sion at 5 years of followup. In contrast, disease-free survival rates decreased to 74% and 49%, respectively, in patients with stages pT3a,b and pT3c cancer during the same interval. A total of 50 patients received either adjuvant external beam radiotherapy or androgen ablation and 36 received adjuvant hormonal therapy (including all 10 with nodal involvement and 26 with extraprostatic disease). These patients were excluded from further analyses since adjuvant therapy may delay PSA progression, potentially masking the impact of advanced disease on clinical progression (figs. 3 and 4). As reviewed previously, pathological stage increased with preoperative PSA. Therefore, it was not surprising that preoperative PSA was inversely related to progression-free survival (fig. 3). None of our 14 patients with preoperative PSA values of 4.0 ng./ml. or less had evidence of progression. Progression-free survival rates at 5 years in patients with PSA values of 4.1 to 10.0 ngJml. were 87% compared to 75% for those with values of greater than 20.1 ngJml. Multivariate analysis using the Cox proportional hazards model for all 1,679patients treated with only radical pmstatectomy for clinically localized prostate cancer identified (inorder of significance) increased preoperative F'SA, high pathological grade, aneuploidy and tetraploidy as strong independent predictors of progression. Comparisons were made with respect to progression-free survival between patients with clinical stage Tlc cancers and the other clinical substages (Tla to T2Wc). The 5-year progression-fixx survival rate in the stage Tlc group was 84% whereas estimated survival rates for patients with stages Tla, Tlb, T2a and T2Wc cancers were 82%, 78%. 75% and 69%, respectively (fig. 4). When controlling for preoperative PSA, pathological grade and DNA ploidy using the Cox model, patients with clinical stage Tlc cancers had a s t a t i s t i d y significant progression-fi.ee survival advantage compared to those with clinical stage T2Wc disease (p = 0.0005). The differences between patients with stages Tlc versus T2a diaease a p proached statistical significance (p = 0.06). No significant difference was observed between patients with clinical stage Tlc versus Tla/b cancer (p = 0.26). DISCUSSION
The incidence of prostate cancer has increased dramatically during the last 8 years as a result of increased patient
"1 O ! 0
823
5yr survival
(W 84k2.9
1
1
I
2
I
3
I
i
4
5
Years after surgery ssion-free (locaVs stemidPSA greater than 0.2 ngJml.) survival estimate for 257 patients with clinical stage FIG. 1. Kaplan-Meier pro Tlc pmstate cancer aRer r a a prostatectomy 6987 to 1991)with or without aGuvant treatment. Numbers above lines represent number of patients a t nak.
824
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE T1C)
L--
60
1
s 40
Pt
Pathologic stage
----
20
--
spT2c pTW pT3c
3-yr survival
24
5 17 10
10
1
95.9 f 1.8 85.4 f 3.8 48.5 f 13.7 90.0f 9.5
92
-.- N +
O !
0
Events (no.)
(no.) 131
5-yr survival (04
( 0 4
95.9 f 1.a 73.6 f 6.6 40.5f 13.7 90.0 f 9.5
1
I
I
I
1
1
2
3
4
5
Years after surgery P = o.ooo1
FIG.2. Kaplan-Meier progression-free (local/systemic/PSA greater than 0.2 ng./ml.) survival estimates for 257 patients with clinical stage Tlc prostate cancer treated with radical rostatectomy with or without adjuvant therapy, according to pathological stage. Numbers above lines represent number of patients a t risR.
60
s 40 PrecpPSA (nglrnL)
.---4.1-10.0 --.5 4.0
20
10.1-20.0 20.1 +
O f
0
PI (no.) 14
Events (no.)
3-yr survival
0
100 58 37
8
100 93.8 f 2.5 06.7 f 4.7 75.1 f 7.2
5 y r survival
(%I
9
9
(O/O)
100 87.4 f 5.1 79.5 5 6.5 75.1 f 7.2
I
I
I
I
1
1
2
3
4
5
Years after surgery P = 0.022
FIG. 3. Kaplan-Meier progression-free (IocaYsystemidPSA greater than 0.2 ng./ml.) survival estimates for 208 patients with clinical stage T l c prostate cancer treated with radical prostatectomy without adjuvant therapy, according to preoperative serum PSA (ng./ml.). Numbers above lines represent number of patients at risk.
awareness, the routine use of serum PSA, and the ease and safety of transrectal ultrasound guided needle biopsies of the prostate.1,*5 While the absolute number of radical prostatectomies performed has increased (at least temporarily), the percentage done in patients with clinically palpable disease has been stable. In contrast, the percentage of patients treated for stage Tlc disease has increased by a factor of 9 during the same interval.
A prior study from our institution identified that stage Tlc prostate cancers are pathologically significant and indeed similar to digitally detected malignancies.9 Of our tumors only 11%could be considered insignificant on the basis of tumor volume criteria (less than 0.5 cc). Among 24 of these smaller cancers, however, 3 were either high grade andor had nondiploid chromosomal patterns. Therefore, the incidence of so-called clinically insignificant or irrelevant cancers
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE TIC)
825
100
80
60
;s? 40
-----
20
-I-
.-..-
0 0
Clinical stage T1a Tlb T2a Talc Tlc
F't
Events
3-yr survival
(no.)
(no.)
(%I
(%I
27 81
4 16 84
92.6f 5.0 83.6 f 4.2 83.0*1.9 74.6f 1.4 89.0 f 2.2
80.2 f 9.2 80.1 f 4.7 74.3 2.8 g5.2f1.9 84.2 f 3.2
383 977 211
295 26
5 y r survival
*
I
I
I
I
1
1
2
3
4
5
Years after surgery P =0 . m 1 FIG. 4. Kaplan-Meier rogression-free (1ocaVsystemiclPSA greater than 0.2 ngJml.) survival estimates for 1,679 patients after radical prostatectomy without aguvant therapy for stage T2Wc or less prostate cancer accurding to clinical stage. Numbere above lines represent number of patients a t risk.
decreases further to 9%. Finally, small diploid prostate cancers detected in young, otherwise healthy patients are not static but will eventually evolve into poorly differentiated malignancies with abnormal DNA content (50% of the patients with low volume tumors were younger than 65 years).18 In a recent study from our institution patients with pathologically organ-confined (stage pT2c or less) diploid prostate cancers experienced an 86% 5-year survival rate free of PSA andor clinical progression, whereas only 60% with pathologically confined aneuploid tumors were free of disease during the same interval.19 Therefore, once a prostate cancer evolves into an aneuploid tumor definitive therapy is significantly less effective. Among our referred patients carcinoma was confined to the prostate gland in only 51%. As a result, one may argue that rather than detecting clinically unimportant cancers with the use of PSA, a significant number of the tumors detected by this means may, indeed, not be curable with radical prostatectomy alone. However, it should be noted that a prior study from our institution demonstrated that limited extracapsular or residual microscopic disease did not significantly impact on cancer-specific or disease-fkee As expected, patients with stage pTzc or less prostate cancer had a significant survival advantage free of progression compared to those with extraprostatic or nodal disease (fig. 2). The PSA and clinical progression-free survival advantage observed in the patients with lymph node involvement compared to those with stage pT3a to c disease may be attributable to the few patients (10) and events (1)in this substage. Furthermore, all patients with positive nodes received adjuvant hormonal therapy, which w i l l impact on PSA progression. Since postoperative PSA elevation (more than 0.2 ng./ml.) was used as the progression end point in OUT study, further analyses excluded patients treated adjuvantly. As observed previously, preoperative serum PSA varied inversely with pathological ~tage.21-~~ The median preoperative PSA concentration was highest in the stage Tlc subgroup (table I), which may be attributable to the slightly larger volume of clinical stage Tlc compared to stage T2a
cancers (mean 4.0 cc versus 3.0 cc, respectively). However, selection bias associated with the diagnosis of impalpable tumors is perhaps the most plausible explanation. Elevated Berum PSA prompted the subsequent biopsy and diagnosis of prostate cancer in patients with stage Tlc disease rather than a palpable abnormality on digital rectal examination regardless of the PSA concentration. Of the patients with PSA values of 10.0ngJml. or less 60% had disease confined to the prostate. Multivariate analyses performed on clinical and morphological tumor characteristics in patients with clinically localized prostate cancer revealed that high preoperative serum PSA,high pathological grade and nondiploid chromosomal pattern had the greatest influence on progressionfree survival. When controlling for these variables, comparisons among the various clinical substages demonstrated that patients with clinical stage Tlc prostate cancers had progression-free survivals that were comparable to those with clinical stages Tla, T l b and T2a tumors but significantly superior to those with clinical stage T2Wc disease. This outcome is not surprising in light of the similar pathological characteristics observed among the different clinical stages (table 1). This interim analysis has demonstrated that clinical stage Tlc cancers have a better prognosis than clinical stage T2Wc tumors. However, progression-free survivals for patients with stage Tlc disease were similar to those of patients with small palpable nodules (stage T2a) and with cancers detected at transurethral prostatic resection (stages T l a to Tlb). Disappointingly, only 51% of the cancers were pathologically confined to the prostate at radical prostatectomy. Therefore, many clinical stage Tlc tumors are locally advanced or progressed systemically before becoming detectable on digital rectal examination. PSA based screening increases prostate cancer detection by 75% compared to screening by digital rectal examination alone.24 More importantly, 70% of the prostate cancers detected with PSA based screening are pathologically organ-confined in contrast to 51% of our referred patient population.26
826
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE T1C)
14. Zincke, H., Blute, M. L., Fallen, M.J. and Farrow, G. M.: Radical prostatectomy for stage A adenocarcinoma of the prostate: staging errors and their implications for treatment recommendations and disease outcome. J. Urol., 146: 1053. 1991. 15.Winkler, H. Z., Rainwater, L. M., Myers, R. P., Farrow, G. M., Therneau, T. M., Zincke, H. and Leiber, M. M.: Stage D1 prostatic adenocarcinoma: significance of nuclear DNA ploidy patterns studied by flow cytometry. Mayo Clin. Proc., 63: 103, 1988. 16. Kaplan, E. L. and Meier, P.: Nonparametric estimation from incomplete observations. J. Amer. Stat. ASS.,53: 457, 1958. 17. Cox, D. R.: Regression models and life-tables. J. Roy. Stat. Soc., REFERENCES 34: 187,1972. 1. Boring, C. C., Squires, T. S., Tong, T. and Montgomery, S.: 18. Adolfsson, J . and Tribukait, B.: Evaluation of tumor progression Cancer statistics, 1994.CA, 44:7, 1994. by repeated fine needle biopsies in prostate adenocarcinoma: 2. Mettlin, C., Jones, G., Averette, H., Gusberg, S. B. and Murphy, modal deoxyribonucleic acid value and cytological differentiaG. P.: Defining and updating the American Cancer Society tion. J. Urol., 144. 1408,1990. guidelines for cancer-related checkup: prostate and endome- 19. Lerner, S. E., Blute, M. L., Eickholt, J. T. and Bergstralh, trial cancers. CA, 43:42, 1993. E. J.: Analysis of risk factors for progression in patients 3. Catalona, W. J., Smith, D. S., Ratliff, T. L., Dodds, K M., Coplen, with pathologically confined prostate cancers after radical D. E., Yuan, J. J . J., Petros, J. A. and Andriole, G. E.: Measretropubic prostatectomy. J . Urol., part 2, 153: 390A, aburement of prostate-specific antigen in serum as a screening stract 648, 1995. test for prostate cancer. New Engl. J . Med., 324: 1156,1991. 20. Cheng, W. S.,Frydenberg, M., Bergstralh, E. J., Larson-Keller, 4. Brawer, M. R, Chetner, M. P., Beatie, J., Buchner, D. M., J . J. and Zincke, H.: Radical prostatectomy for pathologic Vessella, R. L. and Lange, P. H.: Screening for prostatic carstage C prostate cancer: influence of pathologic variables and cinoma with prostate specific antigen. J . Urol., part 2, 147: adjuvant treatment on disease outcome. Urology, 42 283, 841, 1992. 1993. 5. Guthman, D. A,, Wilson, T. M., Blute, M. L., Bergstralh, E. J., 21. Ruckle, H., Oesterling, J. E. and Zincke, H.: Relationship of high Zincke, H. and Oesterling, J. E.: Biopsy-proved prostate cancer pre-prostatectomy prostate-specific antigen tPSAI levels to in 100 consecutive men with benign digital rectal examination postoperative PSA levels and outcome in prostate cancer (PC). and elevated serum prostate-specific antigen level. Prevalence J. Urol., part 2,151: 340A, abstract 450, 1994. and pathologic characteristics. Urology, 42: 150,1993. 6. Stamey, T. A,, Kabalin, J . N., McNeal, J . E., Johnstone, I. M., 22. Kleer, E., Larson-Keller, J. J., Zincke, H. and Oesterling, J. E.: Ability of preoperative serum prostate-specific antigen value Freiha, F., Redwine, E. A. and Yang, N.: Prostate specific to predict pathologic stage and DNA ploidy. Influence of clinantigen in the diagnosis and treatment of adenocarcinoma of ical stage and tumor grade. Urology, 41: 207, 1993. the prostate. 11. Radical prostatectomy treated patients. J. 23. Hudson, M. A., Bahnson, R. R. and Catalona, W. J.: Clinical use Urol., 141: 1076,1989. of prostate specific antigen in patients with prostate cancer. J . 7. Partin, A. W.,Carter, H. B., Chan, D. W., Epstein, J . I., Urol., 142. 1011,1989. Oesterling, J. E., Rock, R. C., Weber, J. P. and Walsh, P. C.: Prostate specific antigen in the staging of localized prostate 24. Catalona, W. J., Richie, J . P., Ahmann, F. R., Hudson, M. A., Scardino, P. T., Flanigan, R. C., deKernion, J . B., Ratliff, T. L., cancer: influence of tumor differentiation, tumor volume and Kavoussi, L. R., Dalkin, B. L., Waters, W. B., MacFarlane, benign hyperplasia. J. Urol., 143: 747, 1990. M. T. and Southwick, P. C.: Comparison of digital rectal ex8. Blackwell, K L., Bostwick, D. G., Myers, R. P., Zincke, H. and amination and serum prostate specific antigen in the early Oesterling, J . E.: Combining prostate specific antigen with detection of prostate cancer: results of a multicenter clinical cancer and gland volume to predict more reliably pathological trial of 6,630men. J . Urol., 151: 1283,1994. stage: the influence of prostate specific antigen cancer density. 25. Catalona, W. J., Smith, D. S., Ratliff, T. L. and Basler, J . W.: J. Urol., 151: 1565,1994. Detection of organ-confined prostate cancer is increased 9. Oesterling, J. E., Suman, V. J., Zincke, H. and Bostwick, D. G.: through prostate-specific antigen-based screening. J.A.M.A., PSA-detected (clinical stage Tlc or BO) prostate cancer. Patho270 948,1993. logically significant tumors. Urol. Clin. N. Amer., 20: 687, 26. Carter, H. B., Pearson, J . D., Metter, E. J., Brant, L. J., Chan, 1993. D. W., Andres, R., Fozard, J. L. and Walsh, P. C.: Longitudinal 10. Epstein, J. I., Walsh, P. C., Carmichael, M. and Brendler, C. B.: evaluation of prostate-specific antigen levels in men with and Pathologic and clinical findings to predict tumor extent of without prostate disease. J.A.M.A., 267: 2215,1992. nonpalpable (stage Tlc) prostate cancer. J.A.M.A., 271: 368, 27. Oesterling, J . E., Jacobsen, S. J., Chute, C. G., Guess, H. A., 1994. 11. Oesterling, J. E.: Prostate specific antigen: a critical assessment Girman, C. J., Panser, L. A. and Lieber, M. M.: Serum prosof the most useful tumor marker for adenocarcinoma of the tate-specific antigen in a community-based population of prostate. J. Urol., 145: 907, 1991. healthy men. Establishment of age-specific reference ranges. 12. Utz, D. C. and Farrow, G. M.: Pathologic differentiation and J.A.M.A., 270 860,1993. prognosis of prostatic carcinoma. J.A.M.A., 209: 1701, 1969. Significance of different molecular forms of serum PSA. 28. Lilja, H.: 13. Bostwick, D. G., Myers, R. P. and Oesterling, J. E.: Staging of The free, noncomplexed form of PSA versus that comdexed to prostate cancer. Sem. Surg. Oncol., 1 0 60,1994. alpha l-antichym&ypsin. Urol. Clin. N. Amer., 2 0 681,1993. CONCLUSIONS
PSA based screening programs, directed at patients with at least a 10-year life expectancy, may improve the ability to diagnose prostate cancer confined to the prostate gland and, thus, still curable.25 Furthermore, measures, such as PSA velocity,26 age-specific reference range27 and free, total and complex PSA values,2S may improve detection of pathologically confined but clinically relevant prostate cancers, thus potentially decreasing the prostate cancer death rate.
Vol. 155,821-826, March 1996 Printed in U.S.A.
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE TlC): AN INTERIM ANALYSIS SETH E. LERNER,* THOMAS M. SEAY, MICHAEL L. BLUTE, ERIK J. BERGSTRALH, DAVID BARRETT AND HORST Z1NCKE-t From the Department of Urology and Section of Biostatistics, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
ABSTRACT
Purpose: The majority of impalpable prostate cancers (stage Tlc) are biologically significant. We report the interim results in 257 patients with stage T l c prostate cancer treated with radical prostatectomy. Materials and Methods: Prostate specific antigen progression-free survival was assessed by the Kaplan-Meier method. Multivariate analyses were performed to determine which clinical and pathological variables independently correlated with progression. Comparisons among the various clinical substages ( T l a to T2b/c) were calculated. Results: Of the patients with stage T l c cancers 51% had stage pT2c or less and 91% had clinically significant tumors on the basis of pathological grade, deoxyribonucleic acid ploidy and tumor volume. High preoperative prostate specific antigen, poorly differentiated tumors and nondiploid status were strong independent predictors of progression. The 5-year survival rate free of progression was 84%. Patients with clinical stage T l c cancers had a significant progression-free survival advantage compared to those with clinical stage T2bk disease (p = 0.0005). Conclusions: Impalpable tumors should not be regarded a s insignificant or innocuous on the basis of pathological analysis. Disease-free survival in the stage T l c group was similar to that in the clinical stages T l a to T2a group but significantly better than that in the T2blc group. KEYWORDS:prostatic neoplasms; prostate-specific antigen; antigens, neoplasm; prostatectomy To our knowledge serum prostate specific antigen (PSA) is the most important and clinically significant tumor marker available to date. Since 1987, PSA has been commonly used to monitor response to cancer therapy and as a screening modality for the detection of potentially curable disease. As a result of its widespread use, adenocarcinoma of the prostate is presently the most commonly diagnosed visceral malignancy in men in the United States.’ Of greater importance is the fact that prostate cancer is second only to lung cancer as a cause of cancer death,’ and more than 40,000 men will die of the disease in 1995. Endorsements for PSA screening have been issued by the American Cancer Society2 and American Urological Association. PSA detects a significant number of prostate cancers missed by routine digital rectal examination.= Serum PSA elevation correlates with tumor volume, grade and pathological stage.68 According to previous studies PSA is increased by 2.2 ng./ml. (corrected from Yang to Hybritech assays) for each gram of malignant tissue present.6 Therefore, PSA testing is not likely to detect indolent malignancies because such tumors are not large enough to cause elevations in serum PSA. Two recent studies demonstrated that the majority of impalpable stage Tlc prostate cancers are biologically significant on the basis of tumor volume, grade and pathological stage, and only a few percent are similar to innocuous or latent autopsy cancers.g.10 Oesterling et a1 reported that stage T l c prostate cancers have pathological characteristics that essentially parallel those of palpably detected tumor^.^ We report the interim results of 257 patients with stage Tlc Prostate cancer treated with radical prostatectomy.
MATERIALS AND METHODS
Between 1987 and 1991, 2,203 patients 44 to 83 years old (median age 67) with clinically localized (stage T2c or less) adenocarcinoma of the prostate underwent bilateral pelvic lymphadenectomy and radical retropubic prostatectomy. Of these patients 257 (12%)had impalpable, PSA detected prostate cancer (clinical stage Tlc). The prostate gland was considered benign (no palpable evidence of asymmetry, nodularity, irregularity or induration) on the basis of digital rectal examination by usually a t least 2 examiners. Prostate cancer was diagnosed on the basis of tissue obtained usually by sextant transrectal needle biopsy. Preoperative PSA values were determined using the Hybritech Tandem-R PSA assay (reference range 0.0 to 4.0 ng./ml.)” The prostate gland was imaged with real-time transrectal ultrasonography as described previo~sly.~Our patients did not represent a screened population. In many patients serum PSA concentrations were measured as part of routine physical examinations. In addition, many patients had been referred to our institution with an established diagnosis of prostate cancer. All biopsy specimens obtained by local physicians were reviewed by pathologists at our clinic. The histological grades of the biopsy and radical prostatectomy specimens were determined using the Mayo grading system.12 The revised TNM staging system was used for clinical and pathological staging.13 All lymph nodes removed during the modified bilateral pelvic lymph node dissection were examined immediately by frozen section and later by formal paraffin section. The radical prostatectomy specimens were examined by the pathologist immediately after resection using gross inspection as well as multiple frozen secAccepted for publication September 29,1995. * Current address: Department of Urology, Albert Einstein Col- tions.14 All surgical margins, including the prostatic base, lege of Medicine, Montefiore Medical Center, 111 East 210th St., apex, urethra, bladder neck, capsule, penprostatic soft tissue Bronx, New York 10467. t Requests for re M t s : Department of Urology, Mayo Clinic, 200 and seminal vesicles, were evaluated. Multiple sections from the prostate were examined immediately by frozen section at First St., S.W.,h i e s t e r , MiMeSOta 55905. 821
822
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE T1C)
tectomy for clinically localized prostate cancer between 1987 and 1991 are summarized in table 1. Median preoperative PSA for patients with clinical stage Tlc disease was 10.0 ng./ml. (range 0.6 to 201). Of the 14 patients with a preoperative serum PSA of 4.0 ngJml. or less 8 had an increase of greater than 0.75 ng./ml. during the 12 months before diagnosis and 6 had abnormal findingson transrectal ultrasound. Overall, 12% of the patients who underwent radical prostatectomy for clinically localized disease during this interval had clinical stage Tlc cancer. However, the frequency of surgery for PSA detected tumors has increased steadily since 1987 (table 2). Of the patients who underwent radical retropubic prostatectomy in 1987 for clinically localized prostate cancer only 2% had stage Tlc disease, whereas 18% who underwent surgery in 1991 had PSA detected tumors. Pathological staging revealed that the incidence of organconfined prostate cancer (stage pT2c or less) ranged from 36% in patients with clinical stage TZb/c disease to 87% in those with clinical stage Tla prostate cancer (table 1).Of patients with clinical stage Tlc disease 51,36,9 and 4% had pathological stage pT2c or less, pT3a/b, pT3c and N+, respectively. A total of 36 patients (14%)had focally positive apical margins as the only evidence of extraprostatic disease and they were included in the pathological stage pT3a/b group. The relationship between preoperative PSA and pathological stage in patients with stage Tlc prostate cancer is shown in table 3. The incidence of pathologically organ-confined disease decreased with increasing preoperative serum PSA values. Of the patients with a normal preoperative PSA (4.0 ng./ml. or less) 79% had organ-confined disease on final pathological analysis, compared to only 33% and none of those with preoperative values of 20.1 to 50 ng./ml. and greater than 50 ng./ml., respectively. No patient with PSA of 10.0 ng./rnl. or less had disease spread to regional lymphatics. Table 1 includes data pertaining to estimated tumor volume in patients with clinically localized prostate cancer. Median tumor volume in patients with clinical stage Tlc disRESULTS ease was 4.0 cc (range 0.03 to 56). Of the stage Tlc cancers The characteristics of the 2,203 patients undergoing bilat- 89% had a total volume of at least 0.5 cc. Therefore, only 24 eral pelvic lymphadenectomy and radical retropubic prosta- of these tumors (11%)could be considered insignificant on
operation and later by routine fixed paraffin sections. Specimens with positive surgical margins but without other evidence of extraprostatic disease were classified within the stage pT3 subgroup. Total tumor volume was estimated from the predominant tumor foci in the gland, using width, length and height measurements, and summing the individual volumes. Tumor deoxyribonucleic acid (DNA) histograms, c ~ e ated by previously described techniques.15 were classified as diploid, tetraploid and aneuploid patterns. The patients were followed every 3 to 4 months for 2 years postoperatively, every 6 months for the next 3 years and then annually. A digital rectal examination, radionuclide bone scan, plain radiograph when indicated, chest x-ray and serum PSA test were performed at each followup visit. Bone scam were performed less often since the advent of PSA testing but were obtained at least once per year. In patients not returning to our institution the PSA concentrations were determined at our clinic via a mailed blood specimen, or the patients were contacted annually with additional medical information obtained from local physicians as indicated. Palpable local recurrence was confirmed by transrectal needle biopsy. Systemic progression was diagnosed on the basis of a positivebone scan andtor plain films.The common end point of progression was defined as either an elevated postoperative PSA level (more than 0.2 ngJml. on at least 2 occasions), local recurrence or distant metastases. Progression-free survival was assessed by the KaplanMeier method.16 To account for more than 1prognostic variable simultaneously, the clinical and pathological parameters were evaluated using a Cox proportional hazards model to determine which variables were independently correlated with progre~sion.~~ During the same period 30,97.449 and 1,370 patients with clinical stages Tla, Tlb, T2a and T2Wc disease, respectively, underwent radical prostatectomy. Comparisons among the various clinical substages with respect to the end point of progression-free survival were calculated.
TABLE1. ChamcteTiStics of 2,203 patients with clinically localized prostate cancer treated with radical retropubic prostatectomy between 1987 and 1991 Clinical Stage
~~
~~
~
Median age at radical ret-
~
65.5
~~
(51-73)
~
~
~~
(39-82)
~~~
65.0 (48-75)
67.0 (38-81)
67.0
67.0 (4443)
4.5 (0.1-108) 49.4 23.4 14.3 13.0 20
6.5(0.4-260) 28.7 41.1 20.8 9.1 41
9.2 (0.2-321) 18 35.8 24.1 22.1 86
10.0(0.&201) 7 43.2 28.8 21.0 4
88.2 11.8 0.0 13
76.8 16.1 8.1 35
74.2 22.6 3.2 73
63.5 27.9 8.6 116
78.2 15.5 6.3 18
80.8 19.2
74.4 25.6
69.9 30.1
55.6 44.4
72.8 27.2
86.7 10.0 3.3
69.0 28.9 2.1
56.6 40.5 2.9
35.7 52.8 11.5
51.0 45.1 3.9
~
0.040
ropubic pmstateetomy (range)
h p . PSA (ngJml.f: Median (ranee) % 4.0 or lea8 % 4.1-10.0 % 10.1-20.0 % Greater than 20.0 No. unknown
DNA ploidy: % Diploid % Tetraploid % Aneuploid No. unknown % Pathological grade: LOW
High 8 TNM pfIthological.tagB: pTZc or l e d P n
N+
<0.001 3.0 (0.1-31.3) 56.0 36.0 4.0 4.0
5
<0.001
<0.001 <0.001
<0.001 Tumor vol.: Median cc (range) 3.4 (0.66-28) 4.2 (0.12-72) 3.0 (0.02-98) 6.1(0.01-143) 4.0(0.03-56) No. unknown 13 55 56 152 33 From Kruskal-Wallis or chi-square teat of the null hypothesis that all clinical stages have the same distribution. t Patients with positive surgical margins but without other evidence of extraprostatic disease were considered to have greater than stage pT2c disease.
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE TIC) TABLE 2. Distribution Yr.
No. Pts.
1987 1988 1989 1990 1991
216 355 385 494 753
of
clinical stage by year of surgery Clinical Stage (%)
Tla
Tlb
T2a
T2b
TIC
1 2 1 1
5
24 23 19 20 19
67 62 65 62 59
2 6 8 13 18
1
8 6 4 2
TABLE3. Relationship of preoperative serum PSA and pathological stage in patients with clinical stage T l c prostate cancer % Pathological Stage
PSA (ng./ml.)
No. Pts.
(255 pta.')
N+
pT3a-c
pT2c or Less
4.0 or less 14 12 28 4.1-10.0 60 40 112 10.1-20.0 41 49 75 20.lbO.O 49 33 53 50.1+ 5 0 100 * Two patients were missing preoperative PSA values.
0 0 4 14 -
the basis of volumetric analysis. Of these smaller lesions 3 (12.5%) were either poorly differentiated or nondiploid. Therefore, only 9% of PSA detected or impalpable cancers should be considered insignificant. Among the 24 patients with clinical stage Tlc prostate cancers less than 0.5 cc in total volume 20 (83%) were younger than 70 years and 12 (50%) were younger than 65 years. Survival and progression. Of the 257 patients 251 were alive at 1.3 to 7.4 years of followup (mean 3.5, median 3.2). Overall, 6 patients died, only 1of prostate cancer. Due to the few events (death) no analyses were done with respect to crude and cancer-specific survivals. There were 150 and 34 patients alive at 3 and 5 years of followup, respectively. Projected 3 and 5-year survival rates free of the combined end point of local, systemic and/or PSA progression (more than 0.2 ng./ml.) were 88%and 841,respectively (fig. 1). As expected, pathological stage impacted on progression-free survival. Patients with stage pT2c or less disease had significant progression-free survival advantages compared to those with extraprostatic or nodal tumor (fig. 2). Of the patients with stage pT2c or less cancer 96% were free of PSA progres-
sion at 5 years of followup. In contrast, disease-free survival rates decreased to 74% and 49%, respectively, in patients with stages pT3a,b and pT3c cancer during the same interval. A total of 50 patients received either adjuvant external beam radiotherapy or androgen ablation and 36 received adjuvant hormonal therapy (including all 10 with nodal involvement and 26 with extraprostatic disease). These patients were excluded from further analyses since adjuvant therapy may delay PSA progression, potentially masking the impact of advanced disease on clinical progression (figs. 3 and 4). As reviewed previously, pathological stage increased with preoperative PSA. Therefore, it was not surprising that preoperative PSA was inversely related to progression-free survival (fig. 3). None of our 14 patients with preoperative PSA values of 4.0 ng./ml. or less had evidence of progression. Progression-free survival rates at 5 years in patients with PSA values of 4.1 to 10.0 ngJml. were 87% compared to 75% for those with values of greater than 20.1 ngJml. Multivariate analysis using the Cox proportional hazards model for all 1,679patients treated with only radical pmstatectomy for clinically localized prostate cancer identified (inorder of significance) increased preoperative F'SA, high pathological grade, aneuploidy and tetraploidy as strong independent predictors of progression. Comparisons were made with respect to progression-free survival between patients with clinical stage Tlc cancers and the other clinical substages (Tla to T2Wc). The 5-year progression-fixx survival rate in the stage Tlc group was 84% whereas estimated survival rates for patients with stages Tla, Tlb, T2a and T2Wc cancers were 82%, 78%. 75% and 69%, respectively (fig. 4). When controlling for preoperative PSA, pathological grade and DNA ploidy using the Cox model, patients with clinical stage Tlc cancers had a s t a t i s t i d y significant progression-fi.ee survival advantage compared to those with clinical stage T2Wc disease (p = 0.0005). The differences between patients with stages Tlc versus T2a diaease a p proached statistical significance (p = 0.06). No significant difference was observed between patients with clinical stage Tlc versus Tla/b cancer (p = 0.26). DISCUSSION
The incidence of prostate cancer has increased dramatically during the last 8 years as a result of increased patient
"1 O ! 0
823
5yr survival
(W 84k2.9
1
1
I
2
I
3
I
i
4
5
Years after surgery ssion-free (locaVs stemidPSA greater than 0.2 ngJml.) survival estimate for 257 patients with clinical stage FIG. 1. Kaplan-Meier pro Tlc pmstate cancer aRer r a a prostatectomy 6987 to 1991)with or without aGuvant treatment. Numbers above lines represent number of patients a t nak.
824
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE T1C)
L--
60
1
s 40
Pt
Pathologic stage
----
20
--
spT2c pTW pT3c
3-yr survival
24
5 17 10
10
1
95.9 f 1.8 85.4 f 3.8 48.5 f 13.7 90.0f 9.5
92
-.- N +
O !
0
Events (no.)
(no.) 131
5-yr survival (04
( 0 4
95.9 f 1.a 73.6 f 6.6 40.5f 13.7 90.0 f 9.5
1
I
I
I
1
1
2
3
4
5
Years after surgery P = o.ooo1
FIG.2. Kaplan-Meier progression-free (local/systemic/PSA greater than 0.2 ng./ml.) survival estimates for 257 patients with clinical stage Tlc prostate cancer treated with radical rostatectomy with or without adjuvant therapy, according to pathological stage. Numbers above lines represent number of patients a t risR.
60
s 40 PrecpPSA (nglrnL)
.---4.1-10.0 --.5 4.0
20
10.1-20.0 20.1 +
O f
0
PI (no.) 14
Events (no.)
3-yr survival
0
100 58 37
8
100 93.8 f 2.5 06.7 f 4.7 75.1 f 7.2
5 y r survival
(%I
9
9
(O/O)
100 87.4 f 5.1 79.5 5 6.5 75.1 f 7.2
I
I
I
I
1
1
2
3
4
5
Years after surgery P = 0.022
FIG. 3. Kaplan-Meier progression-free (IocaYsystemidPSA greater than 0.2 ng./ml.) survival estimates for 208 patients with clinical stage T l c prostate cancer treated with radical prostatectomy without adjuvant therapy, according to preoperative serum PSA (ng./ml.). Numbers above lines represent number of patients at risk.
awareness, the routine use of serum PSA, and the ease and safety of transrectal ultrasound guided needle biopsies of the prostate.1,*5 While the absolute number of radical prostatectomies performed has increased (at least temporarily), the percentage done in patients with clinically palpable disease has been stable. In contrast, the percentage of patients treated for stage Tlc disease has increased by a factor of 9 during the same interval.
A prior study from our institution identified that stage Tlc prostate cancers are pathologically significant and indeed similar to digitally detected malignancies.9 Of our tumors only 11%could be considered insignificant on the basis of tumor volume criteria (less than 0.5 cc). Among 24 of these smaller cancers, however, 3 were either high grade andor had nondiploid chromosomal patterns. Therefore, the incidence of so-called clinically insignificant or irrelevant cancers
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE TIC)
825
100
80
60
;s? 40
-----
20
-I-
.-..-
0 0
Clinical stage T1a Tlb T2a Talc Tlc
F't
Events
3-yr survival
(no.)
(no.)
(%I
(%I
27 81
4 16 84
92.6f 5.0 83.6 f 4.2 83.0*1.9 74.6f 1.4 89.0 f 2.2
80.2 f 9.2 80.1 f 4.7 74.3 2.8 g5.2f1.9 84.2 f 3.2
383 977 211
295 26
5 y r survival
*
I
I
I
I
1
1
2
3
4
5
Years after surgery P =0 . m 1 FIG. 4. Kaplan-Meier rogression-free (1ocaVsystemiclPSA greater than 0.2 ngJml.) survival estimates for 1,679 patients after radical prostatectomy without aguvant therapy for stage T2Wc or less prostate cancer accurding to clinical stage. Numbere above lines represent number of patients a t risk.
decreases further to 9%. Finally, small diploid prostate cancers detected in young, otherwise healthy patients are not static but will eventually evolve into poorly differentiated malignancies with abnormal DNA content (50% of the patients with low volume tumors were younger than 65 years).18 In a recent study from our institution patients with pathologically organ-confined (stage pT2c or less) diploid prostate cancers experienced an 86% 5-year survival rate free of PSA andor clinical progression, whereas only 60% with pathologically confined aneuploid tumors were free of disease during the same interval.19 Therefore, once a prostate cancer evolves into an aneuploid tumor definitive therapy is significantly less effective. Among our referred patients carcinoma was confined to the prostate gland in only 51%. As a result, one may argue that rather than detecting clinically unimportant cancers with the use of PSA, a significant number of the tumors detected by this means may, indeed, not be curable with radical prostatectomy alone. However, it should be noted that a prior study from our institution demonstrated that limited extracapsular or residual microscopic disease did not significantly impact on cancer-specific or disease-fkee As expected, patients with stage pTzc or less prostate cancer had a significant survival advantage free of progression compared to those with extraprostatic or nodal disease (fig. 2). The PSA and clinical progression-free survival advantage observed in the patients with lymph node involvement compared to those with stage pT3a to c disease may be attributable to the few patients (10) and events (1)in this substage. Furthermore, all patients with positive nodes received adjuvant hormonal therapy, which w i l l impact on PSA progression. Since postoperative PSA elevation (more than 0.2 ng./ml.) was used as the progression end point in OUT study, further analyses excluded patients treated adjuvantly. As observed previously, preoperative serum PSA varied inversely with pathological ~tage.21-~~ The median preoperative PSA concentration was highest in the stage Tlc subgroup (table I), which may be attributable to the slightly larger volume of clinical stage Tlc compared to stage T2a
cancers (mean 4.0 cc versus 3.0 cc, respectively). However, selection bias associated with the diagnosis of impalpable tumors is perhaps the most plausible explanation. Elevated Berum PSA prompted the subsequent biopsy and diagnosis of prostate cancer in patients with stage Tlc disease rather than a palpable abnormality on digital rectal examination regardless of the PSA concentration. Of the patients with PSA values of 10.0ngJml. or less 60% had disease confined to the prostate. Multivariate analyses performed on clinical and morphological tumor characteristics in patients with clinically localized prostate cancer revealed that high preoperative serum PSA,high pathological grade and nondiploid chromosomal pattern had the greatest influence on progressionfree survival. When controlling for these variables, comparisons among the various clinical substages demonstrated that patients with clinical stage Tlc prostate cancers had progression-free survivals that were comparable to those with clinical stages Tla, T l b and T2a tumors but significantly superior to those with clinical stage T2Wc disease. This outcome is not surprising in light of the similar pathological characteristics observed among the different clinical stages (table 1). This interim analysis has demonstrated that clinical stage Tlc cancers have a better prognosis than clinical stage T2Wc tumors. However, progression-free survivals for patients with stage Tlc disease were similar to those of patients with small palpable nodules (stage T2a) and with cancers detected at transurethral prostatic resection (stages T l a to Tlb). Disappointingly, only 51% of the cancers were pathologically confined to the prostate at radical prostatectomy. Therefore, many clinical stage Tlc tumors are locally advanced or progressed systemically before becoming detectable on digital rectal examination. PSA based screening increases prostate cancer detection by 75% compared to screening by digital rectal examination alone.24 More importantly, 70% of the prostate cancers detected with PSA based screening are pathologically organ-confined in contrast to 51% of our referred patient population.26
826
PROSTATE SPECIFIC ANTIGEN DETECTED PROSTATE CANCER (CLINICAL STAGE T1C)
14. Zincke, H., Blute, M. L., Fallen, M.J. and Farrow, G. M.: Radical prostatectomy for stage A adenocarcinoma of the prostate: staging errors and their implications for treatment recommendations and disease outcome. J. Urol., 146: 1053. 1991. 15.Winkler, H. Z., Rainwater, L. M., Myers, R. P., Farrow, G. M., Therneau, T. M., Zincke, H. and Leiber, M. M.: Stage D1 prostatic adenocarcinoma: significance of nuclear DNA ploidy patterns studied by flow cytometry. Mayo Clin. Proc., 63: 103, 1988. 16. Kaplan, E. L. and Meier, P.: Nonparametric estimation from incomplete observations. J. Amer. Stat. ASS.,53: 457, 1958. 17. Cox, D. R.: Regression models and life-tables. J. Roy. Stat. Soc., REFERENCES 34: 187,1972. 1. Boring, C. C., Squires, T. S., Tong, T. and Montgomery, S.: 18. Adolfsson, J . and Tribukait, B.: Evaluation of tumor progression Cancer statistics, 1994.CA, 44:7, 1994. by repeated fine needle biopsies in prostate adenocarcinoma: 2. Mettlin, C., Jones, G., Averette, H., Gusberg, S. B. and Murphy, modal deoxyribonucleic acid value and cytological differentiaG. P.: Defining and updating the American Cancer Society tion. J. Urol., 144. 1408,1990. guidelines for cancer-related checkup: prostate and endome- 19. Lerner, S. E., Blute, M. L., Eickholt, J. T. and Bergstralh, trial cancers. CA, 43:42, 1993. E. J.: Analysis of risk factors for progression in patients 3. Catalona, W. J., Smith, D. S., Ratliff, T. L., Dodds, K M., Coplen, with pathologically confined prostate cancers after radical D. E., Yuan, J. J . J., Petros, J. A. and Andriole, G. E.: Measretropubic prostatectomy. J . Urol., part 2, 153: 390A, aburement of prostate-specific antigen in serum as a screening stract 648, 1995. test for prostate cancer. New Engl. J . Med., 324: 1156,1991. 20. Cheng, W. S.,Frydenberg, M., Bergstralh, E. J., Larson-Keller, 4. Brawer, M. R, Chetner, M. P., Beatie, J., Buchner, D. M., J . J. and Zincke, H.: Radical prostatectomy for pathologic Vessella, R. L. and Lange, P. H.: Screening for prostatic carstage C prostate cancer: influence of pathologic variables and cinoma with prostate specific antigen. J . Urol., part 2, 147: adjuvant treatment on disease outcome. Urology, 42 283, 841, 1992. 1993. 5. Guthman, D. A,, Wilson, T. M., Blute, M. L., Bergstralh, E. J., 21. Ruckle, H., Oesterling, J. E. and Zincke, H.: Relationship of high Zincke, H. and Oesterling, J. E.: Biopsy-proved prostate cancer pre-prostatectomy prostate-specific antigen tPSAI levels to in 100 consecutive men with benign digital rectal examination postoperative PSA levels and outcome in prostate cancer (PC). and elevated serum prostate-specific antigen level. Prevalence J. Urol., part 2,151: 340A, abstract 450, 1994. and pathologic characteristics. Urology, 42: 150,1993. 6. Stamey, T. A,, Kabalin, J . N., McNeal, J . E., Johnstone, I. M., 22. Kleer, E., Larson-Keller, J. J., Zincke, H. and Oesterling, J. E.: Ability of preoperative serum prostate-specific antigen value Freiha, F., Redwine, E. A. and Yang, N.: Prostate specific to predict pathologic stage and DNA ploidy. Influence of clinantigen in the diagnosis and treatment of adenocarcinoma of ical stage and tumor grade. Urology, 41: 207, 1993. the prostate. 11. Radical prostatectomy treated patients. J. 23. Hudson, M. A., Bahnson, R. R. and Catalona, W. J.: Clinical use Urol., 141: 1076,1989. of prostate specific antigen in patients with prostate cancer. J . 7. Partin, A. W.,Carter, H. B., Chan, D. W., Epstein, J . I., Urol., 142. 1011,1989. Oesterling, J. E., Rock, R. C., Weber, J. P. and Walsh, P. C.: Prostate specific antigen in the staging of localized prostate 24. Catalona, W. J., Richie, J . P., Ahmann, F. R., Hudson, M. A., Scardino, P. T., Flanigan, R. C., deKernion, J . B., Ratliff, T. L., cancer: influence of tumor differentiation, tumor volume and Kavoussi, L. R., Dalkin, B. L., Waters, W. B., MacFarlane, benign hyperplasia. J. Urol., 143: 747, 1990. M. T. and Southwick, P. C.: Comparison of digital rectal ex8. Blackwell, K L., Bostwick, D. G., Myers, R. P., Zincke, H. and amination and serum prostate specific antigen in the early Oesterling, J . E.: Combining prostate specific antigen with detection of prostate cancer: results of a multicenter clinical cancer and gland volume to predict more reliably pathological trial of 6,630men. J . Urol., 151: 1283,1994. stage: the influence of prostate specific antigen cancer density. 25. Catalona, W. J., Smith, D. S., Ratliff, T. L. and Basler, J . W.: J. Urol., 151: 1565,1994. Detection of organ-confined prostate cancer is increased 9. Oesterling, J. E., Suman, V. J., Zincke, H. and Bostwick, D. G.: through prostate-specific antigen-based screening. J.A.M.A., PSA-detected (clinical stage Tlc or BO) prostate cancer. Patho270 948,1993. logically significant tumors. Urol. Clin. N. Amer., 20: 687, 26. Carter, H. B., Pearson, J . D., Metter, E. J., Brant, L. J., Chan, 1993. D. W., Andres, R., Fozard, J. L. and Walsh, P. C.: Longitudinal 10. Epstein, J. I., Walsh, P. C., Carmichael, M. and Brendler, C. B.: evaluation of prostate-specific antigen levels in men with and Pathologic and clinical findings to predict tumor extent of without prostate disease. J.A.M.A., 267: 2215,1992. nonpalpable (stage Tlc) prostate cancer. J.A.M.A., 271: 368, 27. Oesterling, J . E., Jacobsen, S. J., Chute, C. G., Guess, H. A., 1994. 11. Oesterling, J. E.: Prostate specific antigen: a critical assessment Girman, C. J., Panser, L. A. and Lieber, M. M.: Serum prosof the most useful tumor marker for adenocarcinoma of the tate-specific antigen in a community-based population of prostate. J. Urol., 145: 907, 1991. healthy men. Establishment of age-specific reference ranges. 12. Utz, D. C. and Farrow, G. M.: Pathologic differentiation and J.A.M.A., 270 860,1993. prognosis of prostatic carcinoma. J.A.M.A., 209: 1701, 1969. Significance of different molecular forms of serum PSA. 28. Lilja, H.: 13. Bostwick, D. G., Myers, R. P. and Oesterling, J. E.: Staging of The free, noncomplexed form of PSA versus that comdexed to prostate cancer. Sem. Surg. Oncol., 1 0 60,1994. alpha l-antichym&ypsin. Urol. Clin. N. Amer., 2 0 681,1993. CONCLUSIONS
PSA based screening programs, directed at patients with at least a 10-year life expectancy, may improve the ability to diagnose prostate cancer confined to the prostate gland and, thus, still curable.25 Furthermore, measures, such as PSA velocity,26 age-specific reference range27 and free, total and complex PSA values,2S may improve detection of pathologically confined but clinically relevant prostate cancers, thus potentially decreasing the prostate cancer death rate.