Prostate Cancer and Prostatic Diseases (2007) 10, 60–65 & 2007 Nature Publishing Group All rights reserved 1365-7852/07 $30.00 www.nature.com/pcan ORIGINAL ARTICLE

Prostate-specific antigen adjusted for total prostatic tumor volume as a predictor for outcome after radical prostatectomy

N Hayashi1, M Urashima2, I Ikemoto1, H Kuruma1, Y Arai3, S Kuwao4, S Baba5 and S Egawa1 1Department of Urology, Jikei University School of Medicine, Minato-ku, Tokyo, Japan; 2Division of Clinical Research and Development, Jikei University School of Medicine, Minato-ku, Tokyo, Japan; 3Department of Urology, Tohoku University Graduate School of Medicine, Aoba-ku, Sendai, Japan; 4Department of Pathology, Higashi-Yamato Hospital, Higashi-Yamato, Tokyo, Japan and 5Department of Urology, Kitasato University School of Medicine, Sagamihara, Kanagawa, Japan

The aim of this study was to investigate the potential prognostic value of preoperative serum prostate-specific antigen levels adjusted for total tumor volume (PSA-TTV density) for outcome following radical prostatectomy for prostate cancer by retrospective review in 268 patients. Lower PSA-TTV density was not only associated with a significantly higher risk for biological failure (bF), systemic failure and cancer death but also an independent predictor for bF (hazard ratio, 6.3). Therefore, these data suggest that there are subsets of prostate cancer with lower PSA secretion levels, and this phenotype is associated with a higher risk of failure after surgery. Prostate Cancer and Prostatic Diseases (2007) 10, 60–65. doi:10.1038/sj.pcan.4500902; published online 3 October 2006

Keywords: prognosis; PSA; tumor volume; PSA tumor density

Introduction Patients and methods

Many clinicopathological variables have been shown to Patients correlate with the likelihood of biological failure (bF) The study population comprised 268 evaluable men after radical prostatectomy, with pathological stage 1 treated with radical retropubic prostatectomy and known to be most influential. However, it is still unclear bilateral pelvic lymph node dissection, without neoadju- which additional variable or combination of variables vant therapy of any type, at three medical school- will most accurately predict tumor recurrence. affiliated hospitals (Jikei University Hospital (n ¼ 69), The advantage of prostate-specific antigen (PSA) Kitasato University Hospital (n ¼ 137) and Kurashiki density over PSA for predicting adverse pathological Central Hospital (n ¼ 63)) between January 1988 and findings and bF after definitive therapy for prostate 2,3 November 2002. Several different PSA assays (Eiken cancer is controversial. Confounding factors include polyclonal radioimmunoassay (Eiken, Tokyo, Japan), the effects of benign prostatic hyperplasia (BPH) within Dainapack IMx PSA assay (Dinabot, Tokyo, Japan), the malignant gland: not only may the gland contain AxSYM PSA assay (Dinabot)) were used at Kitasato varying amounts of PSA-producing BPH tissue, but and Kurashiki during this period.6 All values at correlation between the volume of BPH and serum PSA 4 Jikei were measured by E test Tosoh IIPA (Tosoh Co., may be poor. On the other hand, a negative correlation Tokyo, Japan). For uniformity, Eiken PSA data between serum PSA and Gleason score adjusted for ¼  4 were inter-converted as follows: (IMx PSA) 1.39 tumor volume has been reported. Although PSA levels (Eiken PSA)À1.02.7 Results of other assays were not do not necessarily reflect tumor burden and pathological 4,5 inter-converted because they are considered virtually stage, low PSA values in the presence of a large tumor identical. may represent an aggressive phenotype irrespective of the amount of BPH. These considerations led us to investigate the signi- Staging and pathological examination ficance of PSA adjusted for total tumor volume as a All patients underwent pretreatment evaluation with predictor of advanced pathological findings and clinical digital rectal examination, chest radiograph, computer- outcome after radical prostatectomy. ized tomography or magnetic resonance imaging scan of the abdomen and pelvis, and bone scanning. Clinical Correspondence: Dr N Hayashi, Department of Urology, Jikei stage was determined according to the updated 2002 University School of Medicine, 3-25-8 Nishi-shimbashi, Minato-ku, version of the unified tumor node metastasis system.8 Tokyo 105-8461, Japan. E-mail: [email protected] Prostatic volume was determined at the time of transrectal Received 26 April 2006; revised and accepted 9 June 2006; published ultrasonography-guided biopsy with the use of following online 3 October 2006 formula: prostate length  width  height  p/6. PSA adjusted total tumor volume as a predictor N Hayashi et al

The whole mount step-sectioning technique was used istics are listed in Table 1. Median age at surgery was 61 in these patients. All hematoxylin & eosin pathological 65.0 years (range, 49–80 years). Median preoperative slides of radical prostatectomy specimens were reviewed PSA was 9.0 ng/ml (range, 0.44–136 ng/ml). Median total by a single pathologist (SK) and tumor grade was tumor volume and prostatic volume were 2.7 cm3 assigned based on the Gleason grading system. Tumor (range, 0.01–44.0 cm3)and27.2cm3 (range, 8.4–107.7 cm3), volumes were calculated by the method described respectively. previously:9 the tumor area is measured using a Specimen tumor grade was classified as Gleason score planimeter and multiplied by 5 mm thickness and by a 2–6 in 106 patients (40%), 7 in 108 patients (40%) and shrinkage correction factor of 1.33. Tumor extent was 8–10 in 54 patients (20%). Approximately half (n ¼ 125, determined according to the General Rule for Clinical 47%) had pT3 disease. Extracapsular extension, seminal and Pathological Studies on Prostatic Cancer.10 Briefly, vesicle involvement, lymph node involvement and the prostatic capsule on each slide was carefully positive surgical margins were found in 113 (42%), 46 inspected microscopically for areas of complete cancer (17%), 20 (7%) and 99 (40%) of the patients, respectively. penetration through the capsule surface (extracapsular Twenty-seven (10%) of these were treated with adjuvant extension). The presence of seminal vesicle involvement radiation therapy without endocrine manipulation. After or nodal involvement and the status of the surgical bF, 33% of the patients received some type of salvage margins were also assessed. therapy – local irradiation in 60 patients, endocrine therapy in 33 patients. PSA-TTV density (ng/ml cm3) was stratified into three Follow-up and end points groups: Group I, o0.7 (n ¼ 16); Group II, 0.7–6.0 (n ¼ 161) After surgery, patients were evaluated at 3- to 6-month and Group III, 46.0 (n ¼ 91). Cutoff points were set intervals with a PSA measurement and digital rectal to maximize accuracy of bF prediction using a log- examination. Bone scans were performed annually. rank test. When compared to other groups, Group I Postoperative PSA values were considered elevated patients appeared to have significantly larger tumors (bF) if values of 0.1 ng/ml or greater were obtained on (Po0.0001). These tumors also appeared more aggres- two consecutive visits. The date of bF was defined as sive, with significantly higher Gleason scores and higher time of the first detectable PSA. If PSA never became incidences of pT3, extracapsular extension, seminal undetectable postoperatively, then bF was considered to vesicle involvement, lymph node involvement and have occurred at time zero. Systemic progression was positive surgical margins (Po0.05). No statistically defined clinically as the appearance of new lesions on an significant difference in prostatic volume (P40.05) was annual bone scan or appearance of soft-tissue metastasis found among Group I (27.3714.2 cm3), Group II indicated by biopsy. Causes of death were determined (28.5713.0 cm3) and Group III (26.3716.3 cm3). based on hospital records and/or death certificates.

Kaplan–Meier survival analysis in each end point Statistical analysis To date, 112 patients (42%) have experienced bF at a w2 test and simple regression analysis were used to median of 15 months (range, 0–128 months) following evaluate the relationship between comparisons of vari- surgery. Systemic failure was diagnosed in 23 patients ables, with Po0.05 as significant. Survival curves of the (9%) at a median of 38 months (range, 3–90 months). patients were compared using the Kaplan–Meier method Eighteen patients (7%) died of prostate cancer at a and analyzed by the log-rank test; the level of signifi- median of 61 months (range, 33–109 months) after bF. cance was again set at 5%. Cox proportional hazards Seven (3%) patients died of intercurrent causes at a models were used to assess hazard ratio with 95% median of 44 months (range, 2–71 months) following confidence interval (95%CI) under univariate or multi- surgery. variate analysis. Log-likelihood ratio test was used to Figures 1–3 depict Kaplan–Meier projection of survival examine whether addition of prostate-specific antigen curves for bF, systemic failure and death from prostate levels adjusted for total tumor volume (PSA-TTV) cancer progression for the three PSA-TTV density density to previously known prognostic factors im- groups. The log-rank test indicated significant differ- proved a model’s prediction ability. Parameters included ences between groups for all these end points, with age, PSA, tumor volume, specimen Gleason score, Group I patients appearing to have the most dire pathological category, extracapsular extension, seminal prognosis. vesicle involvement, nodal involvement, margin status Five-year bF-free survival was 27, 57 and 82% for and adjuvant therapy. End points were defined as the Group I, Group II and Group III, respectively advent of bF, systemic failure and death from prostate (Po0.0001). Similarly, systemic progression-free and cancer progression. The interval to each end point was cancer-specific survival of these patients at 5 years were calculated from the date of surgery. All statistical 49, 94 and 96%, and 73, 97 and 99%, respectively analyses were conducted using STATA8.0 (STATA (Po0.0001). Corporation, College Station, TX, USA).

Cox proportional hazards model for predicting bF Results In univariate analysis, lower PSA-TTV density, higher preoperative PSA, greater total tumor volume, higher Patient characteristics and PSA-TTV density Gleason score, pT3, extracapsular extension, seminal Median follow-up after surgery in 268 evaluable patients vesicle involvement, lymph node involvement and was 84 months (range, 2–152 months). Patient character- positive surgical margins all indicated a significantly

Prostate Cancer and Prostatic Diseases PSA adjusted total tumor volume as a predictor N Hayashi et al 62 Table 1 Clinicopathological characteristics, stratified by PSA-TTV density, of 268 men who underwent radical prostatectomy Variables Total N ¼ 268 Group I PSA-TTV Group II PSA-TTV Group III PSA-TTV Statistical No (%) density o0.7 ng/ml cm3 density 0.7–6.0 ng/ density 46.0 ng/ml cm3 significance N ¼ 16 No (%) ml cm3 N ¼ 161 No (%) N ¼ 91 No (%)

Age at surgery (years) 49–59 64 (22) 4 (25) 32 (20) 24 (26) NS 60–69 156 (55) 11 (69) 93 (58) 45 (50) X70 65 (23) 1 (6) 36 (22) 22 (24)

Preoperative PSA (ng/ml) o4.0 33 (12) 4 (25) 20 (12) 9 (10) NS 4.0po10.0 113 (42) 11 (69) 61 (38) 41 (45) 10.0po20.0 80 (30) 1 (6) 53 (33) 26 (29) 20.0p 42 (16) 0 (0) 27 (17) 15 (16) Median7s.e. 9.070.8 5.770.7 9.971.2 8.771.9 NS

Tumor volume (cm3) o1.1 67 (25) 0 (0) 10 (6) 57 (63) o0.001 1.1po2.7 67 (25) 0 (0) 43 (27) 24 (26) 2.7po6.0 65 (24) 3 (19) 54 (34) 8 (9) 6.0p 69 (26) 13 (81) 54 (33) 2 (2) Median7s.e. 2.770.4 471.8 370.5 170.2 o0.0001

Specimen Gleason score 2–6 106 (40) 1 (6) 64 (40) 41 (45) 0.002 7 108 (40) 6 (38) 65 (40) 37 (41) 8–10 54 (20) 9 (56) 32 (20) 13 (14)

Pathological category T2 142 (53) 1 (6) 76 (47) 66 (72) o0.001 T3 126 (47) 15 (94) 86 (53) 25 (28) Extracapsular extension 113 (42) 13 (81) 80 (50) 20 (22) o0.001 Seminal vesicle 46 (17) 12 (75) 28 (18) 6 (7) o0.001 involvement Lymph node involvement 20 (8) 5 (31) 13 (8) 2 (2) o0.001 Positive surgical margin 99 (37) 7 (44) 68 (43) 24 (27) 0.034 Adjuvant therapy 27 (10) 3 (19) 18 (11) 6 (7) NS

Abbreviations: NS, not significant; PSA-TTV, prostate-specific antigen levels adjusted for total tumor volume; s.e., standard error. Statistical significance, comparison between Groups I–III.

1.00 1.00 NS

0.75 0.75 p = 0.0008 p < 0.0001

0.50 0.50

p = 0.0019 0.25 0.25

Log-rank test: p < 0.0001 free survival (%) Systemic failure Log-rank test: p < 0.0001

Biochemical failure free survival (%) Biochemical failure 0.00 0.00 0 50 100 150 0 50 100 150 Months after surgery (M) Months after surgery (M)

Group I 16 4 1 Group I 16 9 2 Group II 161 65 23 Group II 161 127 50 Group III 91 67 25 Group III 91 77 29 Group I: PSA-TTV density, < 0.7 ng/ml+cc (Group I) Group I: PSA-TTV density, < 0.7 ng/ml+cc (Group I) Group II: PSA-TTV density, 0.7-6.0 ng/ml+cc (Group II) Group II: PSA-TTV density, 0.7-6.0 ng/ml+cc (Group II) Group III: PSA-TTV density, > 6.0- ng/ml+cc (Group III) Group III: PSA-TTV density, > 6.0- ng/ml+cc (Group III) Figure 2 Kaplan–Meier systemic failure survival curves for the Figure 1 Kaplan–Meier bF survival curves for the three PSA-TTV 3 3 three PSA-TTV density groups. Group I, o0.7 ng/ml cm ; Group II, density groups. Group I, o0.7 ng/ml cm ; Group II, 0.7–6.0 ng/ 0.7–6.0 ng/ml cm3 and Group III, 46.0 ng/ml cm3. Statistical signi- ml cm3 and Group III, 46.0 ng/ml cm3. Statistical significance was ficance was estimated using the log-rank test. estimated using the log-rank test.

greater risk of bF (Table 2). Multivariate analysis Gleason score remained significant in multivariate demonstrated a relative bF risk of 6.3 (95%CI: 1.7–23.5) analysis, tumor volume, pathological category, extra- for Group 1 versus Group 3 (P ¼ 0.006) after adjusting for capsular extension, seminal vesicle involvement, lymph other parameters. Although preoperative PSA and node involvement and positive surgical margins lost

Prostate Cancer and Prostatic Diseases PSA adjusted total tumor volume as a predictor N Hayashi et al significance (Table 2). Age at surgery and adjuvant PSA-TTV density to a Cox proportional model using the 63 therapy also remained significant under multivariate variables age, PSA, tumor volume, specimen Gleason analysis. In the log-likelihood ratio test, addition of score, pathological category, extracapsular extension, seminal vesicle involvement, nodal involvement, margin

1.00 status and adjuvant therapy produced a significant NS improvement in the accuracy of bF prediction (P ¼ 0.035).

0.75 p < 0.0001 Cox proportional hazards model for predicting systemic failure and death from prostate cancer 0.50 Univariate analyses indicated Group I, larger tumor volume, higher Gleason score, pT3, extracapsular exten-

0.25 sion, seminal vesicle involvement, lymph node involve-

Cancer-specific survival (%) ment and positive surgical margins were significantly

Log-rank test: p < 0.0001 associated with increased risk of systemic failure and 0.00 cancer-specific death (Table 3). Preoperative PSA did not 0 50 100 150 show any significant association with these end points. Months after surgery (M) Multivariate analysis indicated lymph node involvement Group I 16 12 5 and positive surgical margins were significant for Group II 161 128 50 predicting systemic failure (data not shown). No vari- Group III 91 77 29 Group I: PSA-TTV density, < 0.7 ng/ml+cc (Group I) ables were significant in prediction of cancer death. Group II: PSA-TTV density, 0.7-6.0 ng/ml+cc (Group II) Group III: PSA-TTV density, > 6.0- ng/ml+cc (Group III) Figure 3 Kaplan–Meier survival curves for death from prostate Discussion cancer for the three PSA-TTV density groups. Group I, o0.7 ng/ ml cm3; Group II, 0.7-6.0 ng/ml cm3 and Group III, 46.0 ng/ml cm3. Although the relation of serum PSA values to prostate Statistical significance was estimated using the log-rank test. cancer was recognized 20 years ago, their relation to BPH

Table 2 Univariate and multivariate Cox regression analyses of the HR for biochemical failure after radical prostatectomy Variables Univariate analysis Multivariate analysis

HR (95%CI) Statistical significance HR (95%CI) Statistical significance

PSA-TTV density (ng/ml cm3) Group I: o0.7 5.5 (2.7–11.2) o0.001 6.3 (1.7–23.5) 0.006 Group II: 0.7–6.0 2.2 (1.4–3.5) 0.001 2.0 (0.9–4.4) NS Group III: 46.0 Reference Reference

Age at prostatectomy (years) 49–59 1.0 (0.6–1.9) NS 2.0 (0.9–4.0) NS 60–69 1.2 (0.7–2.0) NS 1.7 (1.0–3.0) 0.049 X70 Reference Reference

Preoperative PSA (ng/ml) o4.0 0.2 (0.1–0.4) o0.001 0.1 (0.0–0.5) 0.001 4.0po10.0 0.2 (0.1–0.4) o0.001 0.2 (0.1–0.4) o0.001 10.0po20.0 0.5 (0.3–0.8) 0.007 0.5 (0.3–0.9) 0.049 20.0p Reference Reference

Tumor volume (cm3) o1.1 Reference Reference 1.1po2.7 2.0 (1.0–3.9) 0.047 0.9 (0.4–2.0) NS 2.7po6.0 2.9 (1.5–5.7) 0.002 0.8 (0.3–2.3) NS 6.0p 7.0 (3.8–13.1) o0.001 0.8 (0.3–2.8) NS

Specimen Gleason score 2–6 Reference Reference 7 1.8 (1.1–2.8) 0.017 1.3 (0.7–2.1) NS 8–10 5.1 (3.1–8.4) o0.001 2.9 (1.6–5.5) 0.001

Pathological category T2 Reference Reference T3 3.7 (2.4–5.5) o0.001 1.1 (0.4–3.3) NS Extracapsular extension 3.6 (2.4–5.4) o0.001 1.5 (0.5–4.1) NS Seminal vesicle involvement 3.4 (2.2–5.2) o0.001 1.1 (0.6–2.0) NS Lymph node involvement 3.1 (1.7–5.5) 0.007 1.3 (0.6–3.2) NS Positive surgical margin 2.9 (2.0–4.3) o0.001 1.4 (0.8–2.3) NS Adjuvant therapy 1.8 (0.9–3.5) NS 0.3 (0.1–0.8) 0.013

Abbreviations: CI, confidence interval; HR, hazard ratio; NS, not significant; PSA-TTV, prostate-specific antigen levels adjusted for total tumor volume.

Prostate Cancer and Prostatic Diseases PSA adjusted total tumor volume as a predictor N Hayashi et al 64 Table 3 Univariate Cox regression analyses of the HR for systemic failure and death from prostate cancer after radical prostatectomy Variables Systemic failure Cancer death

HR (95%CI) Statistical significance HR (95%CI) Statistical significance

PSA-TTV density (ng/ml cm3) Group I: o0.7 27.2 (7.5–99.1) o0.001 23.0 (4.8–111) o0.001 Group II: 0.7–6.0 2.0 (0.5–7.1) NS 2.7 (0.6–12.6) NS Group III: 46.0 Reference Reference

Age at surgery (years) 49–59 3.0 (0.6–15.1) NS 1.1 (0.2–5.3) NS 60–69 3.0 (0.7–13.2) NS 1.6 (0.4–5.5) NS X70 Reference Reference

Preoperative PSA (ng/ml) o4.0 0.4 (0.1–1.7) NS 0.2 (0.0–1.6) NS 4.0po10.0 0.6 (0.2–1.5) NS 0.4 (0.2–1.2) NS 10.0po20.0 0.3 (0.1–1.0) NS 0.2 (0.0–1.1) NS 20.0p Reference Reference

Tumor volume (cm3) o1.1 Reference — 1.1po2.7 2.0 (0.2–22.0) NS 0.4 (0.0–4.1) NS 2.7po6.0 2.3 (0.2–25.4) NS Reference 6.0p 21.4 (2.9–160) 0.003 6.2 (1.4–27.3) 0.016

Specimen Gleason score 2–6 Reference Reference 7 10.6 (1.4–83.1) 0.024 7.8 (0.9–63.5) NS 8–10 30.5 (4.0–235) 0.001 25.3 (3.2–198) 0.002

Pathological category T2 Reference — T3 27.6 (3.7–205) 0.001 — — Extracapsular extension 16.1 (3.8–68.6) o0.001 — o0.001 Seminal vesicle involvement 12.5 (5.2–30.5) o0.001 13.3 (4.7–37.3) o0.001 Lymph node involvement 13.6 (5.9–31.1) o0.001 9.9 (3.9–25.5) o0.001 Positive surgical margin 3.8 (1.6–9.0) 0.002 4.1 (1.5–10.9) 0.005 Adjuvant therapy 3.1 (1.1–8.3) 0.026 2.8 (0.9–8.4) NS

Abbreviations: CI, confidence interval; HR, hazard ratio; NS, not significant; PSA-TTV, prostate-specific antigen levels adjusted for total tumor volume.

became apparent only in the post-PSA screening era.5 pathologically advanced cancer with preoperative PSA The heterogeneous nature of prostate cancer, together levels less than 10 ng/ml. Although they represented with biological variables such as tumor vascularity and only about 5% of our patient population, they accounted multifocality, may also influence serum PSA levels. for approximately one-fifth of high grade (Gleason score Furthermore, it has been shown that poorly differen- 8–10) and large volume (X6.0 cm3) tumors. This patient tiated cancers tend to produce less PSA per volume of subset may be different from those in the era before PSA tumor tissue.4 Attempts to accurately predict pathologi- testing, who tended to have poorly differentiated, high cal characteristics of prostate cancer solely as a function volume tumors and advanced disease and may be of serum PSA below 10 ng/ml have thus had little candidates for more interventional therapeutic approach. success to date. To address the many clinicopathological variables that We hypothesized that the combination of a large tumor have been shown to correlate with the likelihood of bF and low serum PSA may reflect an aggressive phenotype after radical prostatectomy,1,6 Kattan et al.11 developed a irrespective of the amount of BPH; consequently, PSA- postoperative nomogram for disease recurrence. Vari- TTV density may have prognostic value after radical ables in this nomogram included preoperative PSA, prostatectomy for prostate cancer. Increasing dediffer- specimen Gleason sum, prostatic capsular invasion, entiation of prostate cancer with increasing tumor surgical margin status and seminal vesicle and lymph volume may result in diminution of PSA production by node involvement. Other investigators, however, have prostate cancer cells, thus perturbing the otherwise emphasized the need to include additional factors, expected relationship between serum PSA and tumor possibly not yet analyzed or determined.12 In the log- volume. Our patients with the lowest PSA-TTV density, likelihood ratio test, adding PSA-TTV density to the Cox o0.7 ng/ml cm3 (Group I) had a risk of bF 6.3 times that proportional model showed significant (P ¼ 0.035) im- of patients with values of 6.0 ng/ml cm3 or greater provement in bF prediction accuracy. A postoperative (Group III). This low PSA-TTV patient subset was also nomogram incorporating PSA-TTV density may allow at significantly higher risk for systemic progression and more accurate prediction of treatment outcome. cancer death in univariate analysis (Tables 2 and 3; The advantage of PSA density over PSA for predicting Figures 1–3). Most (490%) patients in this subset had adverse pathological findings and bF after definitive

Prostate Cancer and Prostatic Diseases PSA adjusted total tumor volume as a predictor N Hayashi et al 65 therapy for prostate cancer has been reported as margin- predicting outcome for clinically localized prostate cancer. al to nonexistent.2,3 We adjusted PSA levels using tumor Urology 1997; 50: 73–78. volume data measured from the surgical specimen rather 3 Freedland SJ, Kane CJ, Presti Jr JC, Terris MK, Amling CL, Dorey than total prostatic volume determined by imaging F et al. Comparison of preoperative prostate specific antigen studies. One of the disadvantages of using values based density and prostate specific antigen for predicting recurrence on tumor volume in a daily practice is the labor-intensive after radical prostatectomy: results from the search data base. JUrol2003; 169: 969–973. nature of the measurement procedure. Whether values of 4 Partin AW, Carter HB, Chan DW, Epstein JI, Oesterling JE, Rock PSA-TTV calculated using a simpler technique such as RC et al. Prostate specific antigen in the staging of localized tumor volume estimates may provide adequate prog- prostate cancer: influence of tumor differentiation, tumor 13 nostic accuracy remains to be seen. volume and benign hyperplasia. JUrol1990; 143: 747–752. Our study is limited owing to the small number of 5 Stamey TA, Caldwell M, McNeal JE, Nolley R, Hemenez M, patients and its retrospective nature. Numbers of events Downs J. The prostate specific antigen era in the United States is were too small to examine the independent impact of over for prostate cancer: what happened in the last 20 years? PSA-TTV density on systemic failure and cancer death. JUrol2004; 172: 1297–1301. Volumes of the malignant and benign portions of the 6 Egawa S, Matsui Y, Matsumoto K, Suyama K, Arai Y, Kuwao S et al. Impact of biochemical failure on long-term clinical outcome gland could not be discriminated to accurately assess the 14 after radical prostatectomy for prostate cancer in Japan. Prostate importance of PSA-TTV density. Yang et al. demon- Cancer Prostatic Dis 2004; 7: 152–157. strated higher tissue–PSA values in tissue samples from 7 Machida T, Ohishi Y, Wada T, Akimoto S, Shimazaki J, Oda H BPH than in samples from prostate cancer. Although this et al. Clinical evaluation of a new kit (IMX PA Dainapack) for finding may strengthen the impact of PSA-TTV density detection of serum prostate specific antigen. Hinyokika Kiyo 1993; in this study, measurement and calculation based on the 39: 977–984. more cancer-specific isoform of PSA may be more 8 UICC International Union. In: Sobin LH, Wittekind CH (eds). informative.15 Data on serum levels of testosterone in UICC International Union Against Cancer, TNM classification of this elderly patient population were not available,16 so malignant tumors, 6th edn. John Wiley & Sons, Inc: New York, association of free and total testosterone with PSA-TTV 2002, pp 184–187. 9 Egawa S, Takashima R, Matsumoto K, Mizoguchi H, Kuwao S, density and tumor characterization could not be Baba S. Infrequent involvement of the anterior base in low-risk assessed. More detailed data from a greater number of patients with clinically localized prostate cancer and its possible patients are needed. Further work is warranted. significance in definitive radiation therapy. Jpn J Clin Oncol 2000; 30: 126–130. 10 Japanese Urological Association. The Japanese Pathological Society. General Rule for Clinical and Pathological Studies on Prostatic Cancer, Conclusions 3rd edn. 2001. 11 Kattan MW, Wheeler TM, Scardino PT. Postoperative nomogram Patients with low PSA-TTV density levels may represent for disease recurrence after radical prostatectomy for prostate a subset with dire prognosis after radical prostatectomy cancer. J Clin Oncol 1999; 17: 1499–1507. and thus may be candidates for a more aggressive 12 D’Amico AV, Whittington R, Malkowicz SB, Fondurulia J, Chen therapeutic approach. MH, Kaplan I et al. Pretreatment nomogram for prostate-specific antigen recurrence after radical prostatectomy or external-beam radiation therapy for clinically localized prostate cancer. J Clin Oncol 1999; 17: 168–172. Acknowledgements 13 Noguchi M, Stamey T A, McNeal JE, Yemoto CE. Assessment of We thank WA Thomasson, PhD, for expert editorial morphometric measurements of prostate carcinoma volume. Cancer 2000; 89: 1056–1064. assistance. 14 Yang Y, Chisholm GD, Habib FK. The distribution of PSA, cathepsin-D, and pS2 in BPH and cancer of the prostate. Prostate 1992; 21: 201–208. References 15 Catalona WJ, Bartsch G, Rittenhouse HG, Evans CL, Linton HJ, Horninger W et al. Serum pro-prostate specific antigen prefer- 1 Ohori M, Wheeler TM, Dunn JK, Stamey TA, Scardino PT. The entially detects aggressive prostate cancers in men with 2 to pathological features and prognosis of prostate cancer detectable 4 ng/ml prostate specific antigen. JUrol2004; 171: 2239–2244. with current diagnostic tests. JUrol1994; 152: 1714–1720. 16 Hoffman MA, DeWolf WC, Morgentaler A. Is low serum free 2 Ingenito AC, Ennis RD, Hsu IC, Begg MD, Benson MC, Schiff testosterone a marker for high grade prostate cancer? JUrol2000; PB. Re-examining the role of prostate-specific antigen density in 163: 824–827.

Prostate Cancer and Prostatic Diseases