Predicting the Probability of Deferred Radical Treatment for Localised Prostate Cancer Managed by Active Surveillance

Transcription

1 european urology 54 (2008) available at journal homepage: Prostate Cancer Predicting the Probability of Deferred Radical Treatment for Localised Prostate Cancer Managed by Active Surveillance Nicholas J. van As a, Andrew R. Norman b, Karen Thomas b, Vincent S. Khoo a, Alan Thompson c, Robert A. Huddart a, Alan Horwich a, David P. Dearnaley a, Christopher C. Parker a, * a Academic Urology Unit, Institute of Cancer Research, Royal Marsden Hospital, Surrey, United Kingdom b Medical Statistics, Royal Marsden Hospital, Surrey, United Kingdom c Urology, Royal Marsden Hospital, Surrey, United Kingdom Article info Article history: Accepted February 28, 2008 Published online ahead of print on March 7, 2008 Keywords: Prostatic neoplasms Prostate-specific antigen Active surveillance Prognostic factors Abstract Objectives: Outcome data from a prospective study of active surveillance of localised prostate cancer were analysed to identify factors, present at the time of diagnosis, that predict subsequent radical treatment. Methods: Eligible patients had clinical stage T1 T2a, N0 Nx, M0 Mx adenocarcinoma of the prostate with serum PSA < 15 ng/ml, Gleason score 7, primary Gleason grade 3, and % positive biopsy cores (pbc) 50%. Monitoring included serial PSA measurement and repeat prostate biopsies. Radical treatment was initiated in the event of biochemical progression (PSA velocity > 1 ng/ml/yr) or histological progression (primary Gleason grade 4, or %pbc > 50%). Multivariate Cox regression analysis of baseline variables was performed with respect to time to radical treatment. Results: The 326 men recruited from 2002 to 2006 have been followed for a median of 22 mo. Median age was 67 yr, and median initial PSA (ipsa) 6.4 ng/ml. Sixty-five patients (20%) had deferred radical treatment, 16 (5%) changed to watchful waiting because of increasing comorbidity, 7 (2%) died of other causes, and 238 (73%) remain on surveillance. On multivariate Cox regression analysis, the free/total PSA ratio ( p < 0.001) and clinical T stage ( p = 0.006) were independent determinants of time to radical treatment. Conclusions: In addition to established prognostic factors, the free/total PSA ratio may predict time to radical treatment in patients with untreated, localised prostate cancer managed by active surveillance. This possibility warrants further study. # 2008 European Association of Urology. Published by Elsevier B.V. All rights reserved. * Corresponding author. Academic Urology Unit, Royal Marsden Hospital, Downs Road, Sutton, Surrey, SM2 5PT United Kingdom. Tel ; Fax: address: chris.parker@icr.ac.uk (C.C. Parker) /$ see back matter # 2008 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi: /j.eururo

2 1298 european urology 54 (2008) Introduction Treatment options for localised prostate cancer are many and varied, ranging from immediate radical surgery to observation alone. On the one hand, radical prostatectomy has been shown in a goodquality randomised controlled trial to have an overall survival advantage compared with watchful waiting [1]. On the other hand, prostate cancer can often behave in an indolent fashion even without treatment, with no effect either on health or longevity [2,3]. In such cases, radical treatment, with its risks of incontinence and impotence, could be worse than the disease. So, the challenge of managing localised prostate cancer is to distinguish patients who have clinically relevant cancers and may benefit from radical treatment from the remainder who do not need any intervention. A conventional approach is to classify cases into risk groups in terms of serum prostate-specific antigen (PSA) level, biopsy Gleason score, and clinical T stage [4 6]. These risk groups have been shown to predict the probability of biochemical recurrence after radical treatment, and are used as a guide to treatment decision making. In particular, patients with high-risk localised prostate cancer are considered good candidates for immediate radical treatment rather than observation [7,8]. However, patients with low-risk and intermediate-risk localised disease are faced with the difficult decision of whether to have treatment, which requires weighing the potential survival benefit against the known morbidity. There is a major unmet need for markers of prostate cancer behavior within the low-and intermediaterisk groups that could be used to inform the decision whether or not to undergo radical treatment. Active surveillance, first described by Choo et al [9,10] in 2001 is an approach to the management of localised prostate cancer that aims to avoid overtreatment of men with indolent cancers, while still providing treatment with radical intent within a window of curability for those who need it [7]. In contrast to traditional watchful waiting a policy of observation with the use of palliative treatment for symptomatic progression active surveillance consists of close monitoring with radical treatment in the event of biochemical or histological progression. However, active surveillance has two important limitations: First, approximately 40% of men who embark on active surveillance will subsequently proceed to deferred radical treatment, and it is currently unknown whether this delay will adversely affect treatment efficacy. Second, living with untreated prostate cancer and not knowing whether treatment will be required may impose a psychological burden. For both these reasons, it would be ideal to have markers of prostate cancer progression available at the time of diagnosis that would predict the need for treatment. Patients at high risk of disease progression could then receive immediate, rather than deferred, radical treatment, whereas those truly at low risk of disease progression could be observed with greater confidence. A prospective cohort study of active surveillance was initiated at the Royal Marsden Hospital in We have analysed early outcome data from this study with the primary objective of evaluating baseline clinical variables with respect to freedom from subsequent treatment so as to identify factors that would inform the decision of whether to undergo immediate radical treatment at the time of diagnosis. Secondary objectives were to describe the feasibility and early clinical outcomes of active surveillance. 2. Methods Eligible patients had clinical stage T1 2a, N0 x, M0 x (TNM 2002) histologically proven adenocarcinoma of the prostate, with a serum PSA level less than 15 ng/ml, a Gleason score 7 with primary Gleason grade 3, and cancer present in 50% of the total number of biopsy cores. There is no established prostate cancer screening programme in the United Kingdom, so these patients represent a combination of cases presenting as a result of lower urinary tract symptoms and those who had elected to have PSA monitoring. Patients were diagnosed in a number of centres with a range of different biopsy techniques, reflecting standard UK clinical practice. Patients were aged yr and were fit for radical treatment, but had chosen active surveillance as their initial management. All patients gave their informed consent to the study, which was approved by the local research ethics committee. Baseline staging investigations (bone scan and magnetic resonance imaging [MRI] scan of the pelvis) were not mandatory. The active surveillance protocol (Table 1) consisted of monthly serum PSA levels in the first year, every 3 mo in the second year, and every 6 mo thereafter. The Abbott Architect Table 1 Active surveillance protocol Eligibility Clinical T stage T1/T2a Gleason score 7 (3+4) PSA 15 ng/ml 50% of biopsy cores positive Monitoring Serum PSA monthly in year 1, every 3 mo in year 2, then every 6 mo DRE every 3 mo for 2 yr, then every 6 mo Repeat biopsy at 18 mo 2 yr Indications for radical treatment PSA velocity >1 ng/ml/yr Gleason score 4+3, or > 50% cores involved PSA, prostate-specific antigen; DRE, digital rectal examination.

3 european urology 54 (2008) assay was used. Biochemical disease progression was defined as a PSA velocity greater than 1 ng/ml/yr, which was based on a minimum of four values observed over a minimum of 6 mo. PSA density was calculated by dividing PSA by prostate volume, as calculated with the prolate ellipsoid formula. Measurements were taken in three linear dimensions, leftright (LR), anterior-posterior (AP), and base-apex (BA) with the use of either MRI or transrectal ultrasound (TRUS). The three measurements were multiplied by pi/6 (0.52 LR AP BA) [11]. A TRUS-guided octant prostate biopsy was performed after mo on surveillance and every 2 yr thereafter. Both the initial diagnostic and the repeat prostate biopsy specimens were centrally reviewed by a specialist uropathologist. Histological disease progression on repeat biopsy was defined as primary Gleason grade 4 or the presence of cancer in >50% of the total number of cores. Radical treatment was recommended for those with biochemical or histological disease progression. Biochemical failure after radical treatment was defined as a PSA > 0.2 ng/ml after radical prostatectomy, or by the Phoenix criteria (nadir + 2) after radical radiotherapy Statistical methods Baseline clinical variables (initial PSA level, Gleason score, clinical T stage, free/total PSA ratio, PSA density, % positive cores, number of positive scores, prostate volume, and maximum involvement of any core) were analysed with respect to time to radical treatment. All variables were considered as continuous, except where stated. Time to radical treatment was examined with the use of survival type analysis including Cox regression for uni- and multivariate analysis. A forward stepwise model selection process was used for multivariate Cox regression. The time to radical treatment was measured from the date of study enrollment to the date of radical treatment. Patients that have not yet received radical treatment were censored on their last followup date. P values less than 0.05 were considered statistically significant. Patients were subdivided according to their initial PSA and free/total PSA ratio into three groups. Group 1 had both a PSA less than the median value (6.4 ng/ml) and a free/ total PSA ratio greater than or equal to the median (18%); group 2 had one, but not both, of these factors, and group 3 had both a PSA 6.4 ng/ml and a free/total PSA ratio < 18%. Kaplan Meier curves were constructed for each of the above groups including the 95% confidence intervals (95%CI) at yearly time points. The distribution of PSA and free/total PSA ratios was illustrated with scatter plots to demonstrate their utility in discriminating between patients who did or did not proceed to radical treatment. For the group of patients with at least 24-mo follow-up, receiver operator characteristic (ROC) curves were plotted to show the value of free/total PSA ratio as a predictor of radical treatment within 2 yr from start of active surveillance. 3. Results Three hundred twenty-six men were recruited between November 2002 and September The median age was 67 yr, median initial PSA was 6.4 ng/ ml, median percentage of biopsy cores involved 17%, and the median total number of cores taken at diagnosis was 8. According to the D Amico risk group criteria [4], 238 (73%) had low-risk, and 88 (27%) intermediate-risk disease. Patient characteristics are shown in Table 2. At a median follow-up of 22 mo (range, 1 56 mo), 238 patients (73%) remain on surveillance, 65 (20%) have had radical treatment, 16 (5%) have switched to watchful waiting because of increasing comorbidity, and 7 (2%) have died of other causes. Of the 65 patients who have had radical treatment, median time to treatment was 15 mo (range, 1 40 mo). Thirty-two had histological progression on repeat biopsy, 42 biochemical progression, and 5 elected to proceed with radical treatment without having reached the threshold for either biochemical or histological progression. Eighteen patients (6%) were Table 2 Patient characteristics All pts n = 326 No radical treatment n = 261 Radical treatment n = 65 Median age (yr) 67 (50 79) 67 (50 79) 66.5 (51 77) Gleason score (88%) 223 (85%) 50 (77%) (12%) 38 (15%) 15 (23%) Clinical T stage T1 281 (86%) 232 (89%) 49 (75%) T2a 45 (14%) 29 (11%) 16 (25%) Median % +ve cores (range) 17% (4 50%) 17% (4 50%) 25% (12 50%) Median initial PSA (range) 6.4 ng/ml ( ) 5.9 ng/ml ( ) 7.9 ng/ml (1 14.8) Median % F/T PSA (range) 18% (4 68%) 19% (4 68%) 10% (4 27%) Median prostate volume (range) 42 ml ( ) 53 ml ( ) 35 ml ( ) Median PSA density (range) 0.14 ng/ml/ml ( ) 0.12 ng/ml/ml ( ) 0.2 ng/ml/ml ( ) PSA velocity (range) 0.6 ng/ml/yr ( 7.6 to 11.2) 0.3 ng/ml/yr ( 7.6 to 11.2) 1.4 ng/ml/yr ( 2.2 to 9.5) PSA, prostate-specific antigen; % +ve cores, percentage of positive cores; % F/T PSA, percentage of free to total PSA.

4 1300 european urology 54 (2008) treated with radical prostatectomy, 45 (14%) with external-beam radiotherapy, and 2 (0.6%) with brachytherapy. There have been no prostate cancer deaths, and no patient has developed metastatic disease. At present, 61 of 65 patients (95%) remain free from biochemical failure. Of 18 patients who underwent deferred radical prostatectomy, all had a Gleason score of 7, 6 had positive surgical margins, and 2 had seminal vesicle involvement. Univariate analysis of the baseline variables showed that initial PSA level ( p < 0.001), free/total PSA ratio ( p < 0.001), PSA density ( p < 0.001), Gleason score ( p = 0.002), maximum percentage involvement of any core ( p = 0.002), % positive cores ( p = 0.03), clinical T stage ( p = 0.03), number of positive cores ( p = 0.04), and prostate volume ( p = 0.04) were associated with time to radical treatment (Table 3). On multivariate analysis, free/ total PSA ratio ( p < 0.001) and clinical T stage ( p = 0.006) remained statistically significant determinants of time to radical treatment (Table 3). The cohort was divided into three groups on the basis of initial PSA level and free/total PSA ratio. Although T stage was statistically significant, the large majority (86%) of the cohort was T1; therefore, we did not include T stage in these groups. Group 1 had both a PSA less than the median value (6.4 ng/ ml) and a free/total PSA ratio greater than or equal to the median (18%); group 2, had one, but not both, of these factors, and group 3 had both a PSA 6.4 ng/ ml, and a free/total PSA ratio < 18%. The actuarial rate of treatment at 3 yr was 0%, 27% (95%CI, 17 39%) and 55% (95%CI, 41 71%) for groups 1, 2, and 3, respectively (Fig. 1). No patient in group 1 has had histological disease progression. In group 2, 16 (28%) of patients showed histological progression; in group 3, 16 (35%) of patients histologically progressed. The distribution of PSA and free/total PSA ratios for patients who did or did not receive radical treatment is illustrated in Fig. 2. On the basis of the 154 patients with more than 24 mo of follow-up, ROC analysis for free/total PSA ratio as a predictor of radical treatment within 2 yr gave an area under the receiver operator characteristic curve (AUC) of 0.81 (Fig. 3). 4. Discussion Using data from a large prospective study of active surveillance for localised prostate cancer, we have evaluated baseline clinical variables with respect to time to subsequent radical treatment. The analysis generated the hypothesis that, in addition to the established clinical prognostic factors, baseline free/ total PSA ratio may be a useful marker of the likely progression to radical treatment in men with lowand intermediate-risk localised prostate cancer managed with an expectant policy. This finding Table 3 Analysis of time to treatment Variable Univariate analysis Multivariate analysis p value HR Lower 95%CI Upper 95%CI p value HR Lower 95%CI Upper 95%CI Initial PSA (ng/ml) < PSA density (ng/ml/ml) < % F/T PSA < < Gleason score Gleason 3+3 Gleason Max % any core <20% 20% % +ve cores T stage T1 T2a No. of +ve cores +ve cores = 1 +ve cores Prostate volume (ml) HR, hazard ratio; 95%CI, 95% confidence interval; PSA, prostate-specific antigen; % F/T PSA, percentage of free to total PSA; Max % core, maximum percentage involvement of any core; % +ve cores, percentage of positive cores.

5 european urology 54 (2008) Fig. 1 Time to radical treatment according to risk group. Chis, chi-square; df, degree of freedom. requires validation in independent datasets, but raises the possibility that the free/total PSA ratio could be used to inform the decision of whether to undergo immediate radical treatment at the time of diagnosis. The data also demonstrate the feasibility of active surveillance, with acceptable short-term clinical outcomes. There have been several prospective series of patients managed by active surveillance [12 14], but only one has looked at the prognostic significance of baseline characteristics [15]. Khan et al [15] analysed the baseline characteristics of 67 patients on active surveillance with respect to histological progression Fig. 3 Receiver operator curve for ratio of free to total prostate-specific antigen as a predictor of radical treatment within 2 yr (area under the receiver operator curve, 0.81). Fig. 2 The distribution of prostate-specific antigen (PSA) and free/total PSA ratio for patients who did, or did not, receive radical treatment (dotted line represents median free/total PSA ratio of 18%). on repeat biopsy. Free/total PSA ratio, PSA velocity, and prostate gland volume were statistically significantly associated with histological progression, although there were just 23 events for analysis. In addition, Meng et al [16] reported a retrospective population-based study of 457 men on watchful waiting in the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE) database. One hundred eighty-eight patients (41%) went on to

6 1302 european urology 54 (2008) active treatment at a median of 1.7 yr after diagnosis. Baseline characteristics associated with progression to active treatment included younger age, level of formal education, initial PSA, and Gleason score. The free/total PSA ratio was not assessed. The free/total PSA ratio has been extensively studied as a potential marker for the presence or absence of prostate cancer. Roddam et al [17] performed a meta-analysis of the studies of free/ total PSA for the detection of prostate cancer in men with a total serum PSA between 2 and 10 ng/ml. They concluded that the free/total PSA ratio can be used to reduce the number of unnecessary biopsies while maintaining a high sensitivity for cancer detection. In addition, there is increasing evidence that the free/total PSA ratio may also be associated not just with the presence of prostate cancer, but also its clinical significance. Nam et al [18] recently published a nomogram that includes known risk factors for prostate cancer, and demonstrated that the free/total PSA ratio is a strong predictor of clinically significant disease, defined as the presence of Gleason grade 7 prostate cancer on biopsy. Raaijmakers et al [19] have shown that hk2 and percent free PSA have added prognostic value for the detection of minimal prostate cancer in screen-detected disease. Algorithms using hk2 and percent free PSA predicted minimal disease with an AUC of 0.82 [19]. Percent free/total PSA has been also been shown to predict adverse pathological features following radical prostatectomy [20]. The short-term clinical outcomes in the current analysis are consistent with those previously reported. Klotz et al [14] have reported the outcomes of active surveillance in 299 patients with low- and intermediate-risk disease. With a median follow-up of 5 yr, 198 patients (66%) were still on surveillance and 101 (34%) had had intervention, including 36 patients (12%) who elected to have radical treatment without having fulfilled any of the criteria for disease progression. Two patients had died of prostate cancer, and disease-specific survival at 8 yr was 99.2%. The outcome of 278 men with screendetected disease, identified through the Rotterdam arm of the European Randomized Study of Screening for Prostate Cancer (ERSPC), was described by Roemeling et al [21]: Deferred treatment was chosen by 82 (29%), and 8-yr cause-specific survival was 100%. Khatami et al [22] recently reported the results of 270 patients diagnosed in the Swedish section of the ERSPC and managed initially with surveillance. At a mean follow-up of 63 mo, there have been no prostate cancer deaths, and no patient has developed metastatic disease. Other smaller surveillance series [13,23] have reported similar findings. The long-term outcomes of active surveillance and their comparison with those of immediate radical treatment will be addressed by the National Cancer Institute of Canada PR11 (Surveillance Therapy Against Radical Treatment [START]) trial. This trial, which opened in 2007, aims to randomise 2100 men with low-risk localised prostate cancer between immediate radical treatment and active surveillance, with disease-specific survival as the main endpoint. This study has several limitations. Although it is a large active surveillance study, the duration of follow-up is modest, and the current analysis is based on just 65 events. The reported risk of treatment should be regarded as the short-term risk, with the risk of treatment in the long-term still uncertain. The decision to institute radical treatment was based on the definitions of biochemical and histological disease progression, which are, of necessity, not evidence-based. However, pretreatment PSA velocity has been shown to be an important determinant of fatal prostate cancer in several different settings [24 27],andGleasonscore is an established predictor of prostate cancer mortality in localised disease [2,28]. The choice of 1 ng/ml as the cut point to define biochemical progression was based on the distribution of PSA velocity in a previous active surveillance cohort so as to select approximately 20% of cases [23]. The threshold for defining histological progression was chosen as a slight modification of the criteria used by Choo et al [9] in the first report of active surveillance. Time to deferred treatment (TDT) cannot be considered a reliable measure of disease progression, because it is dependent on the arbitrary cut points in PSA velocity and Gleason scores. We have previously analysed histological upgrading alone [29] and PSA velocity alone [30] as individual end points. However, it is useful to analyse TDT in addition, not as a valid measure of disease progression, but rather as a pragmatic guide to indicate the probability that patients on this particular active surveillance protocol will receive radical treatment. The higher a man s probability of proceeding to radical treatment, the less attractive is active surveillance to him. Although the rate of radical treatment at 3 yr varied from 0% to as high as 55% between the three groups, the multivariate model explained only a moderate degree of the variation in the outcome. There remains a very real need for additional biomarkers of untreated prostate cancer behaviour that could be used to identify who does, or does not, need treatment.

7 european urology 54 (2008) In light of these limitations, it will be important to attempt to validate our findings on an independent data set. If free/total PSA ratio were confirmed to be a strong, independent prognosticator of disease progression in untreated low- and intermediate-risk localised prostate cancer, then there would be a good case to incorporate it into treatment decision making. This could be done optimally by incorporating free/total PSA ratio, in addition to other significant prognosticators (such as PSA level, biopsy results and T stage), in a nomogram to provide individualised estimates of the risk of delayed treatment on active surveillance. Armed with this estimate, men with localised prostate cancer would be well placed to choose between immediate radical treatment on the one hand, or active surveillance on the other. 5. Conclusions In addition to the established prognostic factors, the free/total PSA ratio may be a useful predictor of time to radical treatment in untreated, localised prostate cancer managed by active surveillance. This possibility warrants further study. There remains a need for additional biomarkers of prostate cancer behaviour to identify who does, or does not, need treatment. Financial disclosures: The authors have no conflicts of interest to declare. This work was undertaken in The Royal Marsden NHS Trust, who received a proportion of its funding from the NHS Executive; the views expressed in this publication are those of the authors and not necessarily those of the NHS Executive. This work was supported by the Institute of Cancer Research, and the Cancer Research UK Section of Radiotherapy [CUK] grant number C46/A2131. Dr van As is funded by a grant from the Pelican Foundation. References [1] Bill-Axelsen A, Holmberg L, Ruutu M, et al. Radical prostatectomy versus watchful waiting in early prostate cancer. N Engl J Med 2005;352: [2] Albertsen PC, Hanley JA, Gleason DF, Barry MJ. Competing risk analysis of men aged 55 to 74 years at diagnosis managed conservatively for clinically localized prostate cancer. JAMA 1998;280: [3] Roemeling S, Roobol MJ, Postma R, et al. Management and survival of screen-detected prostate cancer patients who might have been suitable for active surveillance. Eur Urol 2006;50: [4] D Amico AV, Whittington R, Malkovicz SB, et al. Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer. JAMA 1998;280: [5] Zelefsky MJ, Lyass O, Fuks Z, et al. Predictors of improved outcome for patients with localized prostate cancer treated with neoadjuvant androgen ablation therapy and three-dimensional conformal radiotherapy. J Clin Oncol 1998;16: [6] Chism DB, Hanlon AL, Horwitz EM, Feigenberg SJ, Pollack A. A comparison of the single and double factor high-risk models for risk assignment of prostate cancer treated with 3D conformal radiotherapy. Int J Radiat Oncol Biol Phys 2004;59: [7] Warlick C, Trock BJ, Landis P, Epstein JI, Carter HB. Delayed versus immediate surgical intervention and prostate cancer outcome. J Natl Cancer Inst 2006;98: [8] Heidenreich A, Aus G, Bolla M, et al. EAU guidelines on prostate cancer. Eur Urol 2008;53: [9] Choo R, De Boer G, Klotz L, et al. PSA Doubling time of prostate carcinoma managed with watchful observation alone. Int J Radiat Oncol Biol Phys 2001;50: [10] Choo R, Klotz L, Danjoux C, et al. Feasibility study: watchful waiting for localized low to intermediate grade prostate carcinoma with selective delayed intervention based on prostate specific antigen, histological and/or clinical progression. J Urol 2002;167: [11] Terris MK, Stamey TA. Determination of prostate volume by transrectal ultrasound. J Urol 1991;145: [12] De Vries SH, Postma R, Raaijmakers R, et al. Overall and disease-specific survival of patients with screen-detected prostate cancer in the European randomized study of screening for prostate cancer, section Rotterdam. Eur Urol 2007;51: [13] Carter HB, Walsh PC, Landis P, Epstein JI. Expectant management of nonpalpable prostate cancer with curative intent: preliminary results. J Urol 2002;167: [14] Klotz L. Active surveillance for prostate cancer: for whom? J Clin Oncol 2005;23: [15] Khan MA, Carter HB, Epstein JI, et al. Can prostate specific antigen derivatives and pathological parameters predict significant change in expectant management criteria for prostate cancer? J Urol 2003;170: [16] Meng MV, Elkin EP, Harlan SR, Mehta SS, Lubeck DP, Carroll PR. Predictors of treatment after initial surveillance in men with prostate cancer: results from CaPSURE. J Urol 2003;170: [17] Roddam AW, Duffy Mj, Hamdy FC, et al. Use of prostatespecific antigen (PSA) isoforms for the detection of prostate cancer in men with a PSA level of 2 10 ng/ml: systematic review and meta-analysis. Eur Urol 2005;48: , discussion [18] Nam RK, Toi A, Klotz L, et al. Assessing individual risk for prostate cancer. J Clin Oncol 2007;25: [19] Raaijmakers R, De Vries SH, Blijenberg BG, et al. hk2 and free PSA, a prognostic combination in predicting minimal prostate cancer in screen-detected men within the PSA range 4 10 ng/ml. Eur Urol 2007;52: [20] Steuber T, Vickers AJ, Serio AM, et al. Comparison of free and total forms of serum human kallikrein 2 and

8 1304 european urology 54 (2008) prostate-specific antigen for prediction of locally advanced and recurrent prostate cancer. Clin Chem 2007;53: [21] Roemeling S, Roobol MJ, De Vries SH, et al. Active surveillance for prostate cancers detected in three subsequent rounds of a screening trial: characteristics, PSA doubling times, and outcome. Eur Urol 2007;51: , discussion [22] Khatami A, Aus G, Damber JE, Lilja H, Lodding P, Hogosson J. PSA doubling time predicts the outcome after active surveillance in screening-detected prostate cancer: results from the European randomized study of screening for prostate cancer, Sweden section. Int J Cancer 2007;120: [23] Hardie C, Parker C, Norman A, et al. Early outcomes of active surveillance for localized prostate cancer. BJU Int 2005;95: [24] D Amico AV, Chen MH, Roehl KA, Catalona WJ. Preoperative PSA velocity and the risk of death from prostate cancer after radical prostatectomy. N Engl J Med 2004;351: [25] Carter HB, Ferrucci L, Kettermann A, et al. Detection of life-threatening prostate cancer with prostate-specific antigen velocity during a window of curability. J Natl Cancer Inst 2006;98: [26] D Amico AV, Renshaw AA, Sussman B, Chen MH. Pretreatment PSA velocity and risk of death from prostate cancer following external beam radiation therapy. JAMA 2005;294: [27] Fall K, Garmo H, Andren O, et al. Prostate-specific antigen levels as a predictor of lethal prostate cancer. J Natl Cancer Inst 2007;99: [28] Cuzick J, Fisher G, Kattan MW, et al. Long-term outcome among men with conservatively treated localised prostate cancer. Br J Cancer 2006;95: [29] Venkitaraman R, Norman A, Woode-Amissah R, et al. Predictors of histological disease progression in untreated, localized prostate cancer. J Urol 2007;178: [30] Venkitaraman R, Norman A, Woode-Amissah R, et al. Prostate-specific antigen velocity in untreated* localized prostate cancer. BJU Int 2008;101: Editorial Comment on: Predicting the Probability of Deferred Radical Treatment for Localised Prostate Cancer Managed by Active Surveillance Jan Adolfsson CLINTEC, Karolinska Institutet, Stockholm, Sweden Oncological Centre, Karolinska University Hospital, Stockholm, Sweden jan.adolfsson@karolinska.se Very high disease-specific survival rates, even on national levels, have been reported in men with clinically localised prostate cancer [1], and active surveillance has been suggested as a remedy for overtreatment [2]. Predicting the outcome of active surveillance has thus become important. Van As et al [3] show that free/total prostatespecific antigen (PSA) ratio and local tumour stagewerepredictiveforfutureconversionfrom active surveillance to radical treatment, which was recommended when certain predefined criteria were met. The prediction of radical treatment depends on how these criteria are set. With different criteria, the outcome may have been different. Moreover, the value of conversion to radical treatment as a proxy for a variable such as death from prostate cancer remains to be shown. Risks for specific outcomes are often expressed in terms of relative risks, risk ratios, and so forth. Such risks usually relate to a specific group of patients who were analysed; risks are difficult to relate to the individual patient. Van As et al [3] note that even if free/total PSA ratio was predictive in their study, their model explained only a part of the total variance in outcome and that we need to find better predictive factors. Moreover, an increased risk does not necessarily go hand in hand with good performance in terms of sensitivity and specificity for predicting the outcome in question. For instance, in the study by Fall et al [4], PSAdoubling time could be coupled to a significantly increased risk for prostate cancer death although the sensitivity and specificity was low. Prediction of outcome for patients can be based on anything from good clinical judgment, to nomograms, to the use of neural networks. Most methods for prediction of clinical outcome have one thing common: They are not evidence-based in terms of knowing that using them will predict a better outcome for the patient than not using them [5]. References [1] Adolfsson J, Garmo H, Varenhorst E, et al. Clinical characteristics and primary treatment of prostate cancer in Sweden between 1996 and Data from the National Prostate Cancer Register in Sweden. Scand J Urol Nephrol 2007;41: [2] Kahn MA, Partin AW. Expectant management: an option for localized prostate cancer. Prost Cancer Prost Dis 2005;8:311 5.

9 european urology 54 (2008) [3] Van As NJ, Norman AR, Thomas K, et al. Predicting the probability of deferred radical treatment for localised prostate cancer managed by active surveillance. Eur Urol 2008;54: [4] Fall K, Garmo G, Andrén O, et al. Prostate specific antigen levels as predictor of lethal prostate cancer. J Natl Cancer Inst 2007;99: [5] Stapleton A, Pinnock C. Nomograms for prostate cancer is their use evidence-based? Nature Clin Pract Urol 2005;2: DOI: /j.eururo DOI of original article: /j.eururo