- Email: [email protected]
The Cost-Effectiveness of Prostate Cancer Detection with the Use of Prostate Health Index
VALUE IN HEALTH 19 (2016) 153–157
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/jval
The Cost-Effectiveness of Prostate Cancer Detection with the Use of Prostate Health Index Eveline A.M. Heijnsdijk, PhD1,*, Dwight Denham, MBA2, Harry J. de Koning, MD1 1 Department of Public Health, Erasmus Medical Center, Rotterdam, The Netherlands; 2Global Health Economics and Reimbursement, Beckman Coulter Inc., Brea, CA, USA
AB STR A CT
Background: Clinical trial results suggested that prostate-specific antigen (PSA) screening can reduce prostate cancer mortality. Nevertheless, because the specificity of the PSA test for cancer detection is low, it leads to many negative biopsies. The Beckman Coulter Prostate Health Index (PHI) testing demonstrates improved specificity compared with the PSA-only screening and therefore may improve the costeffectiveness of prostate cancer detection. Objective: To examine the cost-effectiveness of adding PHI testing to improve cancer detection for men with elevated serum PSA. Methods: A microsimulation model, based on the results of the European Randomized Study of Screening for Prostate Cancer trial, was used to evaluate the effects of PSA screening and PHI reflex testing. We predicted the numbers of prostate cancers, negative biopsies, deaths, quality-adjusted life-years gained, and cost-effectiveness of both PSA (cutoff 3 ng/mL) and PHI (cutoff 25) testing methods for a European population, screened from age 50 to 75
years at 4-year intervals. Results: When the PHI test was added to the PSA screening, for men with a PSA between 3 and 10 ng/mL, the model predicted a 23% reduction in negative biopsies. This would lead to a 17% reduction in costs for diagnostics and 1% reduction in total costs for prostate cancer. The cost-effectiveness (3.5% discounted) was 11% better. Limitations found were the modeling assumptions on the sensitivity and specificity of PHI by tumor stage and cutoff values. Conclusions: Compared with PSA-only screening, the use of a PHI test can substantially reduce the number of negative biopsies and improve the cost-effectiveness of prostate cancer detection. Keywords: cost-effectiveness, modeling, PHI, prostate cancer, PSA, screening.
Introduction
In a clinical study of blood samples with PSA values between 2.0 and 10.0 ng/mL, it was found that at 95% sensitivity cutoff values, the specificity of PHI was between 22.7% and 30.5%, compared with a PSA specificity between 7.9% and 9.7% [6]. Several studies found that 15% to 30% of the biopsies could have been avoided with the use of PHI [7–10].A meta-analysis including eight studies also concluded that by using PHI it is possible to reduce the number of negative biopsies while maintaining a high cancer detection rate [11]. The cost-effectiveness of adding a PHI test to the PSA test, when the result is 2 to 10 ng/mL and 4 to 10 ng/mL, has been calculated for a US population [12]. It was predicted that adding the PHI test would save costs and increase total quality of life and therefore PHI may be an important factor in the recommendation to biopsy. Nevertheless, the number of PHI tests that will have to be used, the impact on quality of life, and the cost-effectiveness of the PHI test have not been determined for a European population. In the present study, the effects of selectively adding a PHI test after a PSA test was modeled for a European standard population by predicting the number of tests needed, the quality of life, and the cost-effectiveness. These effects were compared with the effects of using a PSA-only screening.
The European Randomized Study of Screening for Prostate Cancer (ERSPC) showed that prostate-specific antigen (PSA) screening reduced prostate cancer mortality [1,2]. Nevertheless, prostate cancer screening is also associated with high risks of overdiagnosis and overtreatment. The specificity of the PSA test for cancer detection is also low, leading to a large number of negative biopsies [3,4]. In the Rotterdam cohort of the ERSPC trial, the mean positive predictive value of a PSA of more than 3 ng/mL for a biopsy positive for prostate cancer in the screening arm in the first two rounds was 24.8% [5]. To better distinguish prostate cancer from benign prostatic conditions, new biomarkers are needed. One of the most promising biomarkers is the Beckman Coulter Prostate Health Index (PHI). PHI is an index calculated by a test analyzer from the combination of total PSA (tPSA), free PSA (fPSA), and [-2]proPSA (also referred to as p2PSA) assays. It is indicated for use in men older than 50 years with a total PSA between 2.0 and 10.0 ng/mL, accompanied by nonsuspicious digital rectal examination findings.
Copyright & 2016, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc.
Conflict of interest: D. Denham is an employee of Beckman Coulter. * Address correspondence to: Eveline A.M. Heijnsdijk, Department of Public Health, Erasmus Medical Center, P.O. Box 2040, 3000 CA, Rotterdam, The Netherlands. E-mail: [email protected]. 1098-3015$36.00 – see front matter Copyright & 2016, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jval.2015.12.002
154
VALUE IN HEALTH 19 (2016) 153–157
an impact on their life histories. This model has been described extensively before [3,13–15].
Methods The effects of screening using PHI and those using only PSA were compared using a microsimulation model. Models can extrapolate findings from trials to actual populations and allow for comparison of intervention strategies.
The Microsimulation Screening Analysis Model The MIcrosimulation SCreening Analysis (MISCAN) prostate cancer model [3,13–15] was used to simulate the effects of PHI testing and to calculate cost-effectiveness from a European societal perspective. The MISCAN model simulates individual life histories from birth to death, taking into account that some individuals will develop prostate cancer. Once an individual has prostate cancer, the cancer can progress to different preclinical states (Fig. 1). Tumors in the preclinical states are present without symptoms, but can be detected by screening. These preclinical detectable states are defined in 18 combinations of clinical Tstage (T1, T2, and T3), Gleason grade (Gleason score 2–6, 7, and 8– 10), and metastatic stage (locoregional and distant). From each preclinical detectable state, the cancer can progress to the clinical disease state, which implies that cancer is diagnosed because of symptoms. Detection of cancer leads to treatment (radical prostatectomy, radiation therapy, active surveillance, or palliative treatment) and may lead to death related to prostate cancer, with each treatment having treatment-specific survival rates. Using the model, screening is superimposed on life histories in the absence of screening. When screening tests are applied to a person in a preclinical disease state, this may result in detection and alteration of the life history of this individual, because a cohort of screen-detected individuals is cured from cancer. Individuals who have not had screening, however, will not see
No PC PRECLINICAL PC
T1 G<7 CLINICAL DIAGNOSIS
T1 G<7 T1 G=7 T1 G>7
T2 G<7 T2 G=7 T2 G>7
T1 G=7 T1 G>7 T2 G<7 T2 G=7 T2 G>7
T3+ G<7
T3+ G<7
T3+ G=7
T3+ G=7
T3+ G>7
T3+ G>7
SCREENING T1
T2
T3+
G<7,=7,>7
G<7,=7,>7
G<7,=7,>7
Death
Fig. 1 – Prostate cancer develops from no prostate cancer via one or more screen‐detectable preclinical stages to a clinically diagnosed cancer or screen-detected cancer. The arrows indicate the possible transitions. The stages are defined by clinical T‐stage (stage T1: impalpable; stage T2: palpable and confined to the prostate; stage T3þ: palpable with extension beyond the prostatic capsule) and differentiation grade (G o 7: Gleason score 4–6; G ¼ 7: Gleason score 7; G 4 7: Gleason score 8–10). Each state can be local or metastatic, but for simplicity this is not illustrated. Individuals with cancer in any state are at risk of death from other causes. PC, prostate cancer.
Inputs in the Model The input we used was the same as described extensively earlier [15]. The birth table of the European standard population [16] was used to generate a population of all ages. The incidence of and transitions through the disease states were based on the ERSPC Rotterdam and Sweden data [17] and validated on all ERSPC centers [15]. Primary treatment assignment (radiation therapy, radical prostatectomy, and active surveillance) was based on observed treatments in the ERSPC Rotterdam cohort by age, stage, and Gleason score. In the model, all men having metastases and all men dying of prostate cancer were assumed to have received palliative treatment. It was assumed that 46% of men receiving active surveillance will also receive secondary treatment within 7 years of diagnosis, as observed in the ERSPC trial [18]. The baseline survival by stage and age was based on literature and Surveillance, Epidemiology, and End Results (SEER) data [19]. We modeled the effects of treatment by assuming a relative risk of dying from prostate cancer of 0.65 for men undergoing radical prostatectomy as compared with those under watchful waiting [20]. This effect was also assumed for radiation therapy.
PSA and PHI Testing The effect of screening was modeled as a cure proportion. If a man is cured, he will not die from prostate cancer; but if he is not cured, date and cause of death are not changed by earlier detection. This cure proportion was estimated by fitting the model to a prostate cancer mortality reduction of 29% after a follow-up of 9 years for men who attended at least one screening, corresponding to the mortality reduction of the screened men in the ERSPC trial [1]. A cutoff of 3 ng/mL was used as an indication for biopsy in the PSA screening scenario. In the PHI testing scenario, all the men were screened with PSA first, and the men with a PSA result between 3 and 10 ng/mL were reflextested with a PHI test (by measuring fPSA and p2PSA). A PHI cutoff of 25 was used as an indication for biopsy. The stage-specific sensitivity of a PSA test (0.79 for T1 Gleason 2–6 to 0.99 for T3 Gleason 8–10) was fitted on the ERSPC Rotterdam and Sweden data. Because PSA and PHI levels were not directly simulated in our model, we assumed that PSA levels between 3 and 10 ng/mL occurred only when the tumor was in stage T1. On the basis of further analysis from the pivotal trial data set using a PSA cutoff of 3 ng/mL [21], it was assumed that the sensitivity of the addition of the PHI test after a PSA test result between 3 and 10 ng/mL was 8% lower for T1 tumors and the same as the PSA sensitivity for the other tumors. Using ERSPC Rotterdam data, it was assumed that 18.5% of all men screened had a PSA between 3 and 10 ng/mL [5]. The biopsy compliance rate after a positive screen test result was assumed to be 90%, with a sensitivity of 90% as observed in the ERSPC Rotterdam data [4,22]. The total number of biopsies performed was calculated by multiplying the number of positive test results with the positive predictive value (PPV) of a biopsy: 35.8% for clinically detected tumors [23] and 24.8% for screen-detected tumors [5] in the PSA-only scenario. On the basis of the ERSPC Rotterdam data, it was assumed that 81% of the screen-detected tumors had a PSA between 3 and 10 ng/mL and 19% had a PSA higher than 10 ng/ mL [5]. For the PHI test scenario, a PPV of 31.1% (Beckman Coulter, data not shown) was assumed for the screen-detected tumors having a PSA between 3 and 10 ng/mL and 44.8% for the screendetected tumors having a PSA of more than 10 ng/mL [5]. In the base model, a screening protocol consisted of men aged between 50 and 75 years with 4-year intervals because most of
VALUE IN HEALTH 19 (2016) 153–157
the men in the ERSPC were screened with this interval. A screening attendance of 80% was assumed.
Quality of Life and Costs Quality-adjusted life-years (QALYs) were predicted by using utility estimates, ranging between 0 (death or worst imaginable health state) and 1 (full health). By multiplying the number of men in a state with the loss in utility and the duration of the health state, the loss in quality of life was calculated. Utility estimates and durations of health states were based on literature and the ERSPC or Dutch data (Table 1). Most of these utilities were patients’ preferences and measured with the standard gamble approach. Life-years gained (output of the model) minus the loss in quality resulted in the QALYs gained. Costs were based on costs in the Netherlands in 2008 [14] (Table 1). Additional costs of a PHI test were estimated to be €78. Costs and effects were calculated over a period of 100 years, corresponding to the whole life span of the population, and an annual discount rate of 3.5% was used [24]. The sample size in the runs was 10 million, which minimized the stochastic noise. A sensitivity analysis was performed by varying the PPV of a biopsy for a PHI of 50%, the additional costs of a PHI test between €30 and €90, and the costs for a biopsy between €200 and €1000. These variables show large differences between countries and therefore can influence the effectiveness of PHI in comparison with PSA.
Results
155
true presence or absence of cancer), 137 diagnoses of prostate cancer, and 35 prostate cancer deaths per 1000 men (Table 2). Assuming that the population was screened with PSA every 4 years at age 50 to 75 years, the model predicted 443 negative biopsies and 178 diagnosed cancers. Of these cancers, 100 would be screen-detected, of which 41 would be overdiagnosed. Eight fewer men would die from prostate cancer. Compared with PSA screening, PHI reflex testing reduced the number of negative biopsies by 23% to 340. One less cancer case would be detected and two less overdiagnosed cases would be detected. The reduction in mortality would be almost the same. PSA screening would lead to 57.0 life-years gained and 39.5 QALYs gained compared with no screening, and PHI testing resulted in fewer life-years gained (55.8) but the same QALYs gained. The calculation of the QALYs for both methods is presented in Appendix Table 1 in Supplemental Materials found at http://dx.doi.org/10.1016/j.jval.2015.12.002. PSA screening increased the total costs for prostate cancer care by 44% compared with the no-screening scenario. Most of this increase was due to higher treatment costs, with increases in diagnostics and screening tests accounting for a smaller amount. Compared with PSA screening, PHI testing reduced the costs for diagnosis by 17% and increased the costs for screening by 30%, leading to a small decrease of 1% in total costs. Using the 3.5% discount rate, the cost-effectiveness was €126,426/QALY gained for PSA screening compared with no screening and €112,979 for PHI testing (11% more cost-effective).
Sensitivity Analysis
Base Model In the no-screening scenario, the model predicted 246 negative biopsies (all biopsies having a negative result, irrespective of the
Table 1 – The utility estimates and duration of health states and the costs of interventions used in the model. Health state Screening attendance Biopsy Diagnosis RT RP Active surveillance 2 mo to 1 y RT 2 mo to 1 y RP Postrecovery period Palliative therapy Terminal illness
Utility estimate [15] 0.99 0.90 0.80 0.73 0.67 0.97 0.78 0.77 0.95 0.60 0.40
Duration 1 wk 3 wk 1 mo 2 mo 2 mo 7y 10 mo 10 mo 9y 30 mo 6 mo
Cost per unit (in euros) [14] PSA screening (including general practitioner) PHI testing (including PSA test and general practitioner) Diagnosis (biopsy) Staging RP RT Active surveillance Follow-up Advanced disease RP, radical prostatectomy; RT, radiation therapy.
Varying the PPV of a biopsy in the PHI testing scenario to 50% resulted in a large reduction in the number of negative biopsies and an increase in cost-effectiveness/QALY gained compared with PSA screening (Table 3). Decreasing the costs of the PHI test to €30 would lead to a 14% better cost-effectiveness, and increasing the costs to €90 would lead to a 10% better cost-effectiveness. Varying the costs of a biopsy between €200 and €1000 resulted in better cost-effectiveness of 6% to 13% compared with PSA screening.
48 126 700 [26] 200 11,800 14,178 1,588 150 12,276
Discussion PSA screening is associated with a high number of biopsies. In the ERSPC trial, more than 20,000 biopsies were performed after almost 140,000 PSA tests, which means that approximately 14% of PSA tests were followed by a biopsy [2]. When the PHI reflex test was done after PSA screening at 4-year intervals for men with a PSA between 3 and 10 ng/mL, the model predicted that 23% of negative biopsies could be prevented. The results are comparable with the results of Catalona et al. [7], who found that percent [-2]proPSA could prevent 19% of negative biopsies in the PSA range of 2 to 4 ng/mL and 31% in the PSA range of 4 to 10 ng/ mL. A prospective study found that at a PHI cutoff of 27.6 after a PSA between 2 to 4 ng/mL, 15.5% of the biopsies could have been avoided [8], compared with 17% (516 biopsies at PHI testing instead of 621 at PSA screening) in our study using different cutoffs for PSA and PHI. Previously, a cost-effectiveness study of annual screenings using PHI was performed in the United States [12]. Although the results are difficult to compare with our results, because of different natural history assumptions, follow-up, costs, and discount rates, the general conclusions are the same: adding PHI testing after a positive PSA screening leads to a small increase of less than 1% in QALYs and a decrease of approximately 10% in total costs for prostate cancer care. Therefore, the cost-effectiveness improves.
156
VALUE IN HEALTH 19 (2016) 153–157
Table 2 – Predicted effects of PSA and PHI testing (for a PSA between 3 and 10 ng/mL) at age 50 to 75 y at 4-y intervals, both compared with no screening*. Screening strategy Population screened at least once Screening tests No. of PHI tests Effects Negative biopsies Diagnosed cancers Screen-detected cancers Overdiagnosed cancers Prostate cancer deaths LY gained QALYs gained Cost in euros (1000) Screening Diagnosis Treatment Palliative care Total costs Cost-effectiveness Incremental costs/LY gained Incremental costs/QALY gained Cost-effectiveness, 3.5% discounted Incremental costs/QALY gained
No screening
PSA
PHI
Difference, PHI PSA (%)
– – –
867 3,720 –
867 3,723 685
0 (0) 3 (0.1) 685
246 137 – – 35 – –
443 178 100 41 27.3 57 39.5
340 177 98 39 27.5 55.8 39.5
103 (23) 1 (1) 2 (2) 2 (5) 0.2 (1) 1.2 (2) 0 (0)
0 268 1318 429 2016
181 435 1,942 335 2,893
235 361 1,926 337 2,860
54 (30) 74 (17) 16 (1) 2 (1) 33 (1)
15,387 22,172
15,116 21,371
271 (2) 801 (4)
126,426
112,979
13,447 (11)
LY, life-years; PHI, Prostate Health Index; PSA, prostate-specific antigen; QALY, quality-adjusted life-year. *Numbers presented are per 1000 men aged 0 to 100 y, over their entire lifetime. The attendance at screening is assumed to be 80%. Results are without discount, unless stated.
Table 3 – Results of PHI testing (for a PSA between 3 and 10 ng/mL) using various parameter values * . Scenario
Base model PPV biopsy 50% (PHI testing) Costs PHI €30 PHI €90 Biopsy €200 Biopsy €1000
Negative biopsies
Costs/QALY gained, 3.5% discounted
Increase in cost-effectiveness, 3.5% discounted (%)
340 244
122,979 99,318
11 21
340 340 340 340
109,151 113,936 103,724 118,531
14 10 6 13
PHI, Prostate Health Index; PPV, positive predictive value; PSA, prostate-specific antigen; QALY, quality-adjusted life-year. *The increase in cost-effectiveness has been compared with PSA screening using the same costs for biopsy as for the PHI testing.
Most of our predictions are robust to sensitivity analyses. Changing the positive predictive value of a biopsy or the costs of a PHI test had only a minor impact on the cost-effectiveness. Conversely, the costs of a biopsy had a larger impact on the costeffectiveness. In the models, large ranges were used for the costs of a biopsy (€200 to €1000), because large ranges are also found in literature (US $181–US $1923), because of different biopsy procedures used [25]. Strengths of our study are that by using microsimulation modeling techniques, the influence of changing individual parameters on the results can be evaluated without needing large randomized trials. The MISCAN model also allows us to estimate unobservable processes at the individual level, such as the natural history of prostate cancer, progression of the disease, and prediction of the rates of overdiagnosis.
Our findings are also subject to certain limitations. First, we had to make important model assumptions in circumstances for which data were not available. We assumed that the sensitivity and specificity of the PHI test in a screening population could be extrapolated from the Beckman study population. Second, in the model a PSA of 3 ng/mL has been used as the lower cutoff to decide for a PHI test, although the PHI test is recommended for use above 2 ng/mL [6]. This has been done because the model is validated with a PSA cutoff of 3 ng/mL on the basis of results of the ERSPC trial. The differences in sensitivity and PPV between the PSA and the PHI test were calculated based on a PSA cutoff value of 3 ng/mL. When a PSA value between 2 and 10 ng/mL will be used as an indication for a PHI test, the sensitivity will increase and the PPV will decrease, leading to small differences from the reported results. Our estimate of the improved
VALUE IN HEALTH 19 (2016) 153–157
cost-effectiveness is conservative because it is assumed that the PHI sensitivity for high-risk tumors is not different from the PSA sensitivity. Studies suggest that PHI testing improves the detection of aggressive tumors [9,11] and therefore the overdiagnosis rate might be lower. We, however, have no validated model for the correlation between PSA or PHI level and the aggressiveness of the tumor, and PSA growth and PHI levels are not directly included in the model. In addition, we assumed in the model that treatment options will not differ for PHI result or PSA result. It is possible that in the future the decision for active surveillance, radiation therapy, or radical prostatectomy may be partly based on PHI test results. Further studies are warranted to evaluate the effectiveness of treatment decisions based on PHI test results.
Conclusions Our modeling results suggest that the use of a PHI reflex test after a positive PSA test result with benign digital rectal examination findings can reduce the number of negative biopsies and improve the cost-effectiveness of prostate cancer screening. Further research is needed to evaluate the performance of PHI, especially for various PSA and PHI cutoffs. Source of financial support: Beckman Coulter Inc. provided a grant for this study and editorial comments on the manuscript. The analyses were performed independently under the grant agreement, and E.A.M. Heijnsdijk and H.J. de Koning declare that editorial comments by Beckman Coulter did not alter the independent conclusions reached by the Erasmus MC team.
Supplemental Materials Supplemental material accompanying this article can be found in the online version as a hyperlink at http://dx.doi.org/10.1016/ j.jval.2015.12.002 or, if a hard copy of article, at www.valuein healthjournal.com/issues (select volume, issue, and article).
R EF E R EN CE S
[1] Schröder FH, Hugosson J, Roobol MJ, et al. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med 2012;366:981–90. [2] Schröder FH, Hugosson J, Roobol MJ, et al. Screening and prostate cancer mortality: results of the European Randomised Study of Screening for Prostate Cancer (ERSPC) at 13 years of follow-up. Lancet 2014;384:2027–35. [3] Draisma G, Boer R, Otto SJ, et al. Lead times and overdetection due to prostate-specific antigen screening: estimates from the European Randomized Study of Screening for Prostate Cancer. J Natl Cancer Inst 2003;95:868–78. [4] Schröder FH, van den Bergh RC, Wolters T, et al. Eleven-year outcome of patients with prostate cancers diagnosed during screening after initial negative sextant biopsies. Eur Urol 2010;57:256–66.
157
[5] Roobol MJ, Kranse R, Bangma CH, et al. Screening for prostate cancer: results of the Rotterdam section of the European Randomized Study of Screening for Prostate Cancer. Eur Urol 2013;64:530–9. [6] Jansen FH, van Schaik RH, Kurstjens J, et al. Prostate-specific antigen (PSA) isoform p2PSA in combination with total PSA and free PSA improves diagnostic accuracy in prostate cancer detection. Eur Urol 2010;57:921–7. [7] Catalona WJ, Bartsch G, Rittenhouse HG, et al. Serum pro-prostate specific antigen preferentially detects aggressive prostate cancers in men with 2 to 4 ng/mL prostate specific antigen. J Urol 2004;171:2239–44. [8] Lazzeri M, Haese A, de la Taille A, et al. Serum isoform [-2]proPSA derivatives significantly improve prediction of prostate cancer at initial biopsy in a total PSA range of 2-10 ng/mL: a multicentric European study. Eur Urol 2013;63:986–94. [9] Loeb S, Sanda MG, Broyles DL, et al. The prostate health index (phi) selectively identifies clinically-significant prostate cancer. J Urol 2015;193:1163–9. [10] Nordström T, Vickers A, Assel M, et al. Comparison between the fourkallikrein panel and prostate health index for predicting prostate cancer. Eur Urol 2015;68:139–46. [11] Filella X, Gimenez N. Evaluation of [-2] proPSA and prostate health index (phi) for the detection of prostate cancer: a systematic review and meta-analysis. Clin Chem Lab Med 2013;51:729–39. [12] Nichol MB, Wu J, Huang J, et al. Cost-effectiveness of Prostate Health Index for prostate cancer detection. BJU Int 2012;110:353–62. [13] Heijnsdijk EA, de Carvalho TM, Auvinen A, et al. Cost-effectiveness of prostate cancer screening: a simulation study based on ERSPC data. J Natl Cancer Inst 2015;107:36. [14] Heijnsdijk EA, der Kinderen A, Wever EM, et al. Overdetection, overtreatment and costs in prostate-specific antigen screening for prostate cancer. Br J Cancer 2009;101:1833–8. [15] Heijnsdijk EA, Wever EM, Auvinen A, et al. Quality-of-life effects of prostate-specific antigen screening. N Engl J Med 2012;367:595–605. [16] Office for National Statistics. European Standard Population. Newport, UK: Office for National Statistics. [17] Hugosson J, Aus G, Lilja H, et al. Results of a randomized, populationbased study of biennial screening using serum prostate-specific antigen measurement to detect prostate carcinoma. Cancer 2004;100:1397–405. [18] van den Bergh RC, Roemeling S, Roobol MJ, et al. Outcomes of men with screen-detected prostate cancer eligible for active surveillance who were managed expectantly. Eur Urol 2009;55:1–8. [19] Albertsen PC, Hanley JA, Fine J. 20-Year outcomes following conservative management of clinically localized prostate cancer. JAMA 2005;293:2095–101. [20] Bill-Axelson A, Holmberg L, Filen F, et al. Radical prostatectomy versus watchful waiting in localized prostate cancer: the Scandinavian prostate cancer group-4 randomized trial. J Natl Cancer Inst 2008;100:1144–54. [21] Catalona WJ, Partin AW, Sanda MG, et al. A multicenter study of [-2] pro-prostate specific antigen combined with prostate specific antigen and free prostate specific antigen for prostate cancer detection in the 2.0 to 10.0 ng/mL prostate specific antigen range. J Urol 2011;185:1650–5. [22] Postma R, Schröder FH, van Leenders GJ, et al. Cancer detection and cancer characteristics in the European Randomized Study of Screening for Prostate Cancer (ERSPC)—section Rotterdam: a comparison of two rounds of screening. Eur Urol 2007;52:89–97. [23] Otto SJ, van der Cruijsen IW, Liem MK, et al. Effective PSA contamination in the Rotterdam section of the European Randomized Study of Screening for Prostate Cancer. Int J Cancer 2003;105:394–9. [24] National Institute for Health and Clinical Excellence. The Guidelines Manual, January 2009. London: National Institute for Health and Clinical Excellence, 2009. [25] Ekwueme DU, Stroud LA, Chen Y. Cost analysis of screening for, diagnosing, and staging prostate cancer based on a systematic review of published studies. Prev Chronic Dis 2007;4:A100.
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/jval
The Cost-Effectiveness of Prostate Cancer Detection with the Use of Prostate Health Index Eveline A.M. Heijnsdijk, PhD1,*, Dwight Denham, MBA2, Harry J. de Koning, MD1 1 Department of Public Health, Erasmus Medical Center, Rotterdam, The Netherlands; 2Global Health Economics and Reimbursement, Beckman Coulter Inc., Brea, CA, USA
AB STR A CT
Background: Clinical trial results suggested that prostate-specific antigen (PSA) screening can reduce prostate cancer mortality. Nevertheless, because the specificity of the PSA test for cancer detection is low, it leads to many negative biopsies. The Beckman Coulter Prostate Health Index (PHI) testing demonstrates improved specificity compared with the PSA-only screening and therefore may improve the costeffectiveness of prostate cancer detection. Objective: To examine the cost-effectiveness of adding PHI testing to improve cancer detection for men with elevated serum PSA. Methods: A microsimulation model, based on the results of the European Randomized Study of Screening for Prostate Cancer trial, was used to evaluate the effects of PSA screening and PHI reflex testing. We predicted the numbers of prostate cancers, negative biopsies, deaths, quality-adjusted life-years gained, and cost-effectiveness of both PSA (cutoff 3 ng/mL) and PHI (cutoff 25) testing methods for a European population, screened from age 50 to 75
years at 4-year intervals. Results: When the PHI test was added to the PSA screening, for men with a PSA between 3 and 10 ng/mL, the model predicted a 23% reduction in negative biopsies. This would lead to a 17% reduction in costs for diagnostics and 1% reduction in total costs for prostate cancer. The cost-effectiveness (3.5% discounted) was 11% better. Limitations found were the modeling assumptions on the sensitivity and specificity of PHI by tumor stage and cutoff values. Conclusions: Compared with PSA-only screening, the use of a PHI test can substantially reduce the number of negative biopsies and improve the cost-effectiveness of prostate cancer detection. Keywords: cost-effectiveness, modeling, PHI, prostate cancer, PSA, screening.
Introduction
In a clinical study of blood samples with PSA values between 2.0 and 10.0 ng/mL, it was found that at 95% sensitivity cutoff values, the specificity of PHI was between 22.7% and 30.5%, compared with a PSA specificity between 7.9% and 9.7% [6]. Several studies found that 15% to 30% of the biopsies could have been avoided with the use of PHI [7–10].A meta-analysis including eight studies also concluded that by using PHI it is possible to reduce the number of negative biopsies while maintaining a high cancer detection rate [11]. The cost-effectiveness of adding a PHI test to the PSA test, when the result is 2 to 10 ng/mL and 4 to 10 ng/mL, has been calculated for a US population [12]. It was predicted that adding the PHI test would save costs and increase total quality of life and therefore PHI may be an important factor in the recommendation to biopsy. Nevertheless, the number of PHI tests that will have to be used, the impact on quality of life, and the cost-effectiveness of the PHI test have not been determined for a European population. In the present study, the effects of selectively adding a PHI test after a PSA test was modeled for a European standard population by predicting the number of tests needed, the quality of life, and the cost-effectiveness. These effects were compared with the effects of using a PSA-only screening.
The European Randomized Study of Screening for Prostate Cancer (ERSPC) showed that prostate-specific antigen (PSA) screening reduced prostate cancer mortality [1,2]. Nevertheless, prostate cancer screening is also associated with high risks of overdiagnosis and overtreatment. The specificity of the PSA test for cancer detection is also low, leading to a large number of negative biopsies [3,4]. In the Rotterdam cohort of the ERSPC trial, the mean positive predictive value of a PSA of more than 3 ng/mL for a biopsy positive for prostate cancer in the screening arm in the first two rounds was 24.8% [5]. To better distinguish prostate cancer from benign prostatic conditions, new biomarkers are needed. One of the most promising biomarkers is the Beckman Coulter Prostate Health Index (PHI). PHI is an index calculated by a test analyzer from the combination of total PSA (tPSA), free PSA (fPSA), and [-2]proPSA (also referred to as p2PSA) assays. It is indicated for use in men older than 50 years with a total PSA between 2.0 and 10.0 ng/mL, accompanied by nonsuspicious digital rectal examination findings.
Copyright & 2016, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc.
Conflict of interest: D. Denham is an employee of Beckman Coulter. * Address correspondence to: Eveline A.M. Heijnsdijk, Department of Public Health, Erasmus Medical Center, P.O. Box 2040, 3000 CA, Rotterdam, The Netherlands. E-mail: [email protected]. 1098-3015$36.00 – see front matter Copyright & 2016, International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jval.2015.12.002
154
VALUE IN HEALTH 19 (2016) 153–157
an impact on their life histories. This model has been described extensively before [3,13–15].
Methods The effects of screening using PHI and those using only PSA were compared using a microsimulation model. Models can extrapolate findings from trials to actual populations and allow for comparison of intervention strategies.
The Microsimulation Screening Analysis Model The MIcrosimulation SCreening Analysis (MISCAN) prostate cancer model [3,13–15] was used to simulate the effects of PHI testing and to calculate cost-effectiveness from a European societal perspective. The MISCAN model simulates individual life histories from birth to death, taking into account that some individuals will develop prostate cancer. Once an individual has prostate cancer, the cancer can progress to different preclinical states (Fig. 1). Tumors in the preclinical states are present without symptoms, but can be detected by screening. These preclinical detectable states are defined in 18 combinations of clinical Tstage (T1, T2, and T3), Gleason grade (Gleason score 2–6, 7, and 8– 10), and metastatic stage (locoregional and distant). From each preclinical detectable state, the cancer can progress to the clinical disease state, which implies that cancer is diagnosed because of symptoms. Detection of cancer leads to treatment (radical prostatectomy, radiation therapy, active surveillance, or palliative treatment) and may lead to death related to prostate cancer, with each treatment having treatment-specific survival rates. Using the model, screening is superimposed on life histories in the absence of screening. When screening tests are applied to a person in a preclinical disease state, this may result in detection and alteration of the life history of this individual, because a cohort of screen-detected individuals is cured from cancer. Individuals who have not had screening, however, will not see
No PC PRECLINICAL PC
T1 G<7 CLINICAL DIAGNOSIS
T1 G<7 T1 G=7 T1 G>7
T2 G<7 T2 G=7 T2 G>7
T1 G=7 T1 G>7 T2 G<7 T2 G=7 T2 G>7
T3+ G<7
T3+ G<7
T3+ G=7
T3+ G=7
T3+ G>7
T3+ G>7
SCREENING T1
T2
T3+
G<7,=7,>7
G<7,=7,>7
G<7,=7,>7
Death
Fig. 1 – Prostate cancer develops from no prostate cancer via one or more screen‐detectable preclinical stages to a clinically diagnosed cancer or screen-detected cancer. The arrows indicate the possible transitions. The stages are defined by clinical T‐stage (stage T1: impalpable; stage T2: palpable and confined to the prostate; stage T3þ: palpable with extension beyond the prostatic capsule) and differentiation grade (G o 7: Gleason score 4–6; G ¼ 7: Gleason score 7; G 4 7: Gleason score 8–10). Each state can be local or metastatic, but for simplicity this is not illustrated. Individuals with cancer in any state are at risk of death from other causes. PC, prostate cancer.
Inputs in the Model The input we used was the same as described extensively earlier [15]. The birth table of the European standard population [16] was used to generate a population of all ages. The incidence of and transitions through the disease states were based on the ERSPC Rotterdam and Sweden data [17] and validated on all ERSPC centers [15]. Primary treatment assignment (radiation therapy, radical prostatectomy, and active surveillance) was based on observed treatments in the ERSPC Rotterdam cohort by age, stage, and Gleason score. In the model, all men having metastases and all men dying of prostate cancer were assumed to have received palliative treatment. It was assumed that 46% of men receiving active surveillance will also receive secondary treatment within 7 years of diagnosis, as observed in the ERSPC trial [18]. The baseline survival by stage and age was based on literature and Surveillance, Epidemiology, and End Results (SEER) data [19]. We modeled the effects of treatment by assuming a relative risk of dying from prostate cancer of 0.65 for men undergoing radical prostatectomy as compared with those under watchful waiting [20]. This effect was also assumed for radiation therapy.
PSA and PHI Testing The effect of screening was modeled as a cure proportion. If a man is cured, he will not die from prostate cancer; but if he is not cured, date and cause of death are not changed by earlier detection. This cure proportion was estimated by fitting the model to a prostate cancer mortality reduction of 29% after a follow-up of 9 years for men who attended at least one screening, corresponding to the mortality reduction of the screened men in the ERSPC trial [1]. A cutoff of 3 ng/mL was used as an indication for biopsy in the PSA screening scenario. In the PHI testing scenario, all the men were screened with PSA first, and the men with a PSA result between 3 and 10 ng/mL were reflextested with a PHI test (by measuring fPSA and p2PSA). A PHI cutoff of 25 was used as an indication for biopsy. The stage-specific sensitivity of a PSA test (0.79 for T1 Gleason 2–6 to 0.99 for T3 Gleason 8–10) was fitted on the ERSPC Rotterdam and Sweden data. Because PSA and PHI levels were not directly simulated in our model, we assumed that PSA levels between 3 and 10 ng/mL occurred only when the tumor was in stage T1. On the basis of further analysis from the pivotal trial data set using a PSA cutoff of 3 ng/mL [21], it was assumed that the sensitivity of the addition of the PHI test after a PSA test result between 3 and 10 ng/mL was 8% lower for T1 tumors and the same as the PSA sensitivity for the other tumors. Using ERSPC Rotterdam data, it was assumed that 18.5% of all men screened had a PSA between 3 and 10 ng/mL [5]. The biopsy compliance rate after a positive screen test result was assumed to be 90%, with a sensitivity of 90% as observed in the ERSPC Rotterdam data [4,22]. The total number of biopsies performed was calculated by multiplying the number of positive test results with the positive predictive value (PPV) of a biopsy: 35.8% for clinically detected tumors [23] and 24.8% for screen-detected tumors [5] in the PSA-only scenario. On the basis of the ERSPC Rotterdam data, it was assumed that 81% of the screen-detected tumors had a PSA between 3 and 10 ng/mL and 19% had a PSA higher than 10 ng/ mL [5]. For the PHI test scenario, a PPV of 31.1% (Beckman Coulter, data not shown) was assumed for the screen-detected tumors having a PSA between 3 and 10 ng/mL and 44.8% for the screendetected tumors having a PSA of more than 10 ng/mL [5]. In the base model, a screening protocol consisted of men aged between 50 and 75 years with 4-year intervals because most of
VALUE IN HEALTH 19 (2016) 153–157
the men in the ERSPC were screened with this interval. A screening attendance of 80% was assumed.
Quality of Life and Costs Quality-adjusted life-years (QALYs) were predicted by using utility estimates, ranging between 0 (death or worst imaginable health state) and 1 (full health). By multiplying the number of men in a state with the loss in utility and the duration of the health state, the loss in quality of life was calculated. Utility estimates and durations of health states were based on literature and the ERSPC or Dutch data (Table 1). Most of these utilities were patients’ preferences and measured with the standard gamble approach. Life-years gained (output of the model) minus the loss in quality resulted in the QALYs gained. Costs were based on costs in the Netherlands in 2008 [14] (Table 1). Additional costs of a PHI test were estimated to be €78. Costs and effects were calculated over a period of 100 years, corresponding to the whole life span of the population, and an annual discount rate of 3.5% was used [24]. The sample size in the runs was 10 million, which minimized the stochastic noise. A sensitivity analysis was performed by varying the PPV of a biopsy for a PHI of 50%, the additional costs of a PHI test between €30 and €90, and the costs for a biopsy between €200 and €1000. These variables show large differences between countries and therefore can influence the effectiveness of PHI in comparison with PSA.
Results
155
true presence or absence of cancer), 137 diagnoses of prostate cancer, and 35 prostate cancer deaths per 1000 men (Table 2). Assuming that the population was screened with PSA every 4 years at age 50 to 75 years, the model predicted 443 negative biopsies and 178 diagnosed cancers. Of these cancers, 100 would be screen-detected, of which 41 would be overdiagnosed. Eight fewer men would die from prostate cancer. Compared with PSA screening, PHI reflex testing reduced the number of negative biopsies by 23% to 340. One less cancer case would be detected and two less overdiagnosed cases would be detected. The reduction in mortality would be almost the same. PSA screening would lead to 57.0 life-years gained and 39.5 QALYs gained compared with no screening, and PHI testing resulted in fewer life-years gained (55.8) but the same QALYs gained. The calculation of the QALYs for both methods is presented in Appendix Table 1 in Supplemental Materials found at http://dx.doi.org/10.1016/j.jval.2015.12.002. PSA screening increased the total costs for prostate cancer care by 44% compared with the no-screening scenario. Most of this increase was due to higher treatment costs, with increases in diagnostics and screening tests accounting for a smaller amount. Compared with PSA screening, PHI testing reduced the costs for diagnosis by 17% and increased the costs for screening by 30%, leading to a small decrease of 1% in total costs. Using the 3.5% discount rate, the cost-effectiveness was €126,426/QALY gained for PSA screening compared with no screening and €112,979 for PHI testing (11% more cost-effective).
Sensitivity Analysis
Base Model In the no-screening scenario, the model predicted 246 negative biopsies (all biopsies having a negative result, irrespective of the
Table 1 – The utility estimates and duration of health states and the costs of interventions used in the model. Health state Screening attendance Biopsy Diagnosis RT RP Active surveillance 2 mo to 1 y RT 2 mo to 1 y RP Postrecovery period Palliative therapy Terminal illness
Utility estimate [15] 0.99 0.90 0.80 0.73 0.67 0.97 0.78 0.77 0.95 0.60 0.40
Duration 1 wk 3 wk 1 mo 2 mo 2 mo 7y 10 mo 10 mo 9y 30 mo 6 mo
Cost per unit (in euros) [14] PSA screening (including general practitioner) PHI testing (including PSA test and general practitioner) Diagnosis (biopsy) Staging RP RT Active surveillance Follow-up Advanced disease RP, radical prostatectomy; RT, radiation therapy.
Varying the PPV of a biopsy in the PHI testing scenario to 50% resulted in a large reduction in the number of negative biopsies and an increase in cost-effectiveness/QALY gained compared with PSA screening (Table 3). Decreasing the costs of the PHI test to €30 would lead to a 14% better cost-effectiveness, and increasing the costs to €90 would lead to a 10% better cost-effectiveness. Varying the costs of a biopsy between €200 and €1000 resulted in better cost-effectiveness of 6% to 13% compared with PSA screening.
48 126 700 [26] 200 11,800 14,178 1,588 150 12,276
Discussion PSA screening is associated with a high number of biopsies. In the ERSPC trial, more than 20,000 biopsies were performed after almost 140,000 PSA tests, which means that approximately 14% of PSA tests were followed by a biopsy [2]. When the PHI reflex test was done after PSA screening at 4-year intervals for men with a PSA between 3 and 10 ng/mL, the model predicted that 23% of negative biopsies could be prevented. The results are comparable with the results of Catalona et al. [7], who found that percent [-2]proPSA could prevent 19% of negative biopsies in the PSA range of 2 to 4 ng/mL and 31% in the PSA range of 4 to 10 ng/ mL. A prospective study found that at a PHI cutoff of 27.6 after a PSA between 2 to 4 ng/mL, 15.5% of the biopsies could have been avoided [8], compared with 17% (516 biopsies at PHI testing instead of 621 at PSA screening) in our study using different cutoffs for PSA and PHI. Previously, a cost-effectiveness study of annual screenings using PHI was performed in the United States [12]. Although the results are difficult to compare with our results, because of different natural history assumptions, follow-up, costs, and discount rates, the general conclusions are the same: adding PHI testing after a positive PSA screening leads to a small increase of less than 1% in QALYs and a decrease of approximately 10% in total costs for prostate cancer care. Therefore, the cost-effectiveness improves.
156
VALUE IN HEALTH 19 (2016) 153–157
Table 2 – Predicted effects of PSA and PHI testing (for a PSA between 3 and 10 ng/mL) at age 50 to 75 y at 4-y intervals, both compared with no screening*. Screening strategy Population screened at least once Screening tests No. of PHI tests Effects Negative biopsies Diagnosed cancers Screen-detected cancers Overdiagnosed cancers Prostate cancer deaths LY gained QALYs gained Cost in euros (1000) Screening Diagnosis Treatment Palliative care Total costs Cost-effectiveness Incremental costs/LY gained Incremental costs/QALY gained Cost-effectiveness, 3.5% discounted Incremental costs/QALY gained
No screening
PSA
PHI
Difference, PHI PSA (%)
– – –
867 3,720 –
867 3,723 685
0 (0) 3 (0.1) 685
246 137 – – 35 – –
443 178 100 41 27.3 57 39.5
340 177 98 39 27.5 55.8 39.5
103 (23) 1 (1) 2 (2) 2 (5) 0.2 (1) 1.2 (2) 0 (0)
0 268 1318 429 2016
181 435 1,942 335 2,893
235 361 1,926 337 2,860
54 (30) 74 (17) 16 (1) 2 (1) 33 (1)
15,387 22,172
15,116 21,371
271 (2) 801 (4)
126,426
112,979
13,447 (11)
LY, life-years; PHI, Prostate Health Index; PSA, prostate-specific antigen; QALY, quality-adjusted life-year. *Numbers presented are per 1000 men aged 0 to 100 y, over their entire lifetime. The attendance at screening is assumed to be 80%. Results are without discount, unless stated.
Table 3 – Results of PHI testing (for a PSA between 3 and 10 ng/mL) using various parameter values * . Scenario
Base model PPV biopsy 50% (PHI testing) Costs PHI €30 PHI €90 Biopsy €200 Biopsy €1000
Negative biopsies
Costs/QALY gained, 3.5% discounted
Increase in cost-effectiveness, 3.5% discounted (%)
340 244
122,979 99,318
11 21
340 340 340 340
109,151 113,936 103,724 118,531
14 10 6 13
PHI, Prostate Health Index; PPV, positive predictive value; PSA, prostate-specific antigen; QALY, quality-adjusted life-year. *The increase in cost-effectiveness has been compared with PSA screening using the same costs for biopsy as for the PHI testing.
Most of our predictions are robust to sensitivity analyses. Changing the positive predictive value of a biopsy or the costs of a PHI test had only a minor impact on the cost-effectiveness. Conversely, the costs of a biopsy had a larger impact on the costeffectiveness. In the models, large ranges were used for the costs of a biopsy (€200 to €1000), because large ranges are also found in literature (US $181–US $1923), because of different biopsy procedures used [25]. Strengths of our study are that by using microsimulation modeling techniques, the influence of changing individual parameters on the results can be evaluated without needing large randomized trials. The MISCAN model also allows us to estimate unobservable processes at the individual level, such as the natural history of prostate cancer, progression of the disease, and prediction of the rates of overdiagnosis.
Our findings are also subject to certain limitations. First, we had to make important model assumptions in circumstances for which data were not available. We assumed that the sensitivity and specificity of the PHI test in a screening population could be extrapolated from the Beckman study population. Second, in the model a PSA of 3 ng/mL has been used as the lower cutoff to decide for a PHI test, although the PHI test is recommended for use above 2 ng/mL [6]. This has been done because the model is validated with a PSA cutoff of 3 ng/mL on the basis of results of the ERSPC trial. The differences in sensitivity and PPV between the PSA and the PHI test were calculated based on a PSA cutoff value of 3 ng/mL. When a PSA value between 2 and 10 ng/mL will be used as an indication for a PHI test, the sensitivity will increase and the PPV will decrease, leading to small differences from the reported results. Our estimate of the improved
VALUE IN HEALTH 19 (2016) 153–157
cost-effectiveness is conservative because it is assumed that the PHI sensitivity for high-risk tumors is not different from the PSA sensitivity. Studies suggest that PHI testing improves the detection of aggressive tumors [9,11] and therefore the overdiagnosis rate might be lower. We, however, have no validated model for the correlation between PSA or PHI level and the aggressiveness of the tumor, and PSA growth and PHI levels are not directly included in the model. In addition, we assumed in the model that treatment options will not differ for PHI result or PSA result. It is possible that in the future the decision for active surveillance, radiation therapy, or radical prostatectomy may be partly based on PHI test results. Further studies are warranted to evaluate the effectiveness of treatment decisions based on PHI test results.
Conclusions Our modeling results suggest that the use of a PHI reflex test after a positive PSA test result with benign digital rectal examination findings can reduce the number of negative biopsies and improve the cost-effectiveness of prostate cancer screening. Further research is needed to evaluate the performance of PHI, especially for various PSA and PHI cutoffs. Source of financial support: Beckman Coulter Inc. provided a grant for this study and editorial comments on the manuscript. The analyses were performed independently under the grant agreement, and E.A.M. Heijnsdijk and H.J. de Koning declare that editorial comments by Beckman Coulter did not alter the independent conclusions reached by the Erasmus MC team.
Supplemental Materials Supplemental material accompanying this article can be found in the online version as a hyperlink at http://dx.doi.org/10.1016/ j.jval.2015.12.002 or, if a hard copy of article, at www.valuein healthjournal.com/issues (select volume, issue, and article).
R EF E R EN CE S
[1] Schröder FH, Hugosson J, Roobol MJ, et al. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med 2012;366:981–90. [2] Schröder FH, Hugosson J, Roobol MJ, et al. Screening and prostate cancer mortality: results of the European Randomised Study of Screening for Prostate Cancer (ERSPC) at 13 years of follow-up. Lancet 2014;384:2027–35. [3] Draisma G, Boer R, Otto SJ, et al. Lead times and overdetection due to prostate-specific antigen screening: estimates from the European Randomized Study of Screening for Prostate Cancer. J Natl Cancer Inst 2003;95:868–78. [4] Schröder FH, van den Bergh RC, Wolters T, et al. Eleven-year outcome of patients with prostate cancers diagnosed during screening after initial negative sextant biopsies. Eur Urol 2010;57:256–66.
157
[5] Roobol MJ, Kranse R, Bangma CH, et al. Screening for prostate cancer: results of the Rotterdam section of the European Randomized Study of Screening for Prostate Cancer. Eur Urol 2013;64:530–9. [6] Jansen FH, van Schaik RH, Kurstjens J, et al. Prostate-specific antigen (PSA) isoform p2PSA in combination with total PSA and free PSA improves diagnostic accuracy in prostate cancer detection. Eur Urol 2010;57:921–7. [7] Catalona WJ, Bartsch G, Rittenhouse HG, et al. Serum pro-prostate specific antigen preferentially detects aggressive prostate cancers in men with 2 to 4 ng/mL prostate specific antigen. J Urol 2004;171:2239–44. [8] Lazzeri M, Haese A, de la Taille A, et al. Serum isoform [-2]proPSA derivatives significantly improve prediction of prostate cancer at initial biopsy in a total PSA range of 2-10 ng/mL: a multicentric European study. Eur Urol 2013;63:986–94. [9] Loeb S, Sanda MG, Broyles DL, et al. The prostate health index (phi) selectively identifies clinically-significant prostate cancer. J Urol 2015;193:1163–9. [10] Nordström T, Vickers A, Assel M, et al. Comparison between the fourkallikrein panel and prostate health index for predicting prostate cancer. Eur Urol 2015;68:139–46. [11] Filella X, Gimenez N. Evaluation of [-2] proPSA and prostate health index (phi) for the detection of prostate cancer: a systematic review and meta-analysis. Clin Chem Lab Med 2013;51:729–39. [12] Nichol MB, Wu J, Huang J, et al. Cost-effectiveness of Prostate Health Index for prostate cancer detection. BJU Int 2012;110:353–62. [13] Heijnsdijk EA, de Carvalho TM, Auvinen A, et al. Cost-effectiveness of prostate cancer screening: a simulation study based on ERSPC data. J Natl Cancer Inst 2015;107:36. [14] Heijnsdijk EA, der Kinderen A, Wever EM, et al. Overdetection, overtreatment and costs in prostate-specific antigen screening for prostate cancer. Br J Cancer 2009;101:1833–8. [15] Heijnsdijk EA, Wever EM, Auvinen A, et al. Quality-of-life effects of prostate-specific antigen screening. N Engl J Med 2012;367:595–605. [16] Office for National Statistics. European Standard Population. Newport, UK: Office for National Statistics. [17] Hugosson J, Aus G, Lilja H, et al. Results of a randomized, populationbased study of biennial screening using serum prostate-specific antigen measurement to detect prostate carcinoma. Cancer 2004;100:1397–405. [18] van den Bergh RC, Roemeling S, Roobol MJ, et al. Outcomes of men with screen-detected prostate cancer eligible for active surveillance who were managed expectantly. Eur Urol 2009;55:1–8. [19] Albertsen PC, Hanley JA, Fine J. 20-Year outcomes following conservative management of clinically localized prostate cancer. JAMA 2005;293:2095–101. [20] Bill-Axelson A, Holmberg L, Filen F, et al. Radical prostatectomy versus watchful waiting in localized prostate cancer: the Scandinavian prostate cancer group-4 randomized trial. J Natl Cancer Inst 2008;100:1144–54. [21] Catalona WJ, Partin AW, Sanda MG, et al. A multicenter study of [-2] pro-prostate specific antigen combined with prostate specific antigen and free prostate specific antigen for prostate cancer detection in the 2.0 to 10.0 ng/mL prostate specific antigen range. J Urol 2011;185:1650–5. [22] Postma R, Schröder FH, van Leenders GJ, et al. Cancer detection and cancer characteristics in the European Randomized Study of Screening for Prostate Cancer (ERSPC)—section Rotterdam: a comparison of two rounds of screening. Eur Urol 2007;52:89–97. [23] Otto SJ, van der Cruijsen IW, Liem MK, et al. Effective PSA contamination in the Rotterdam section of the European Randomized Study of Screening for Prostate Cancer. Int J Cancer 2003;105:394–9. [24] National Institute for Health and Clinical Excellence. The Guidelines Manual, January 2009. London: National Institute for Health and Clinical Excellence, 2009. [25] Ekwueme DU, Stroud LA, Chen Y. Cost analysis of screening for, diagnosing, and staging prostate cancer based on a systematic review of published studies. Prev Chronic Dis 2007;4:A100.