PSA Testing Pitfalls & Utility Stanley H. Weiss, MD, FACP Professor of Preventive Medicine and Community Health, UMDNJ-New Jersey Medical School Director & PI, Essex County Cancer Coalition October 29, 2011

PSA Testing Pitfalls & Utility Presentation prepared in collaboration with: Daniel M. Rosenblum, PhD Assistant Professor, UMDNJ-New Jersey Medical School

© Stanley H. Weiss, Oct-29 3 Two uses for PSA test Screening for disease: –in asymptomatic men & –in symptomatic men Surveillance: –in follow-up of patients previously diagnosed with prostate cancer

© Stanley H. Weiss, Oct-29 4 PSA Test – Screening Pros & Cons In my opinion, the standard PSA blood test is not a very good screening test for prostate cancer. It has limited sensitivity, specificity, and predictive value.

© Stanley H. Weiss, Oct-29 5 PSA Test – Screening Pros & Cons Nevertheless: The issues are — –Can it save lives? –If it saves lives, is it cost-effective? Number needed to screen to save a life [NNS] Number needed to treat to save a life [NNT] –How can we try to improve screening? –How long will it be until we’ll “know” the proposed new approach(es) will work?

© Stanley H. Weiss, Oct-29 6 Probabilities of Death Due to PCa — Control vs. Screening Ratios & Differences

© Stanley H. Weiss, Oct-29 7 What Are the Data? How Should They Be Interpreted? ConceptFormulaProblems What is a ratio? N 2 / N 1 At small N 1 (denominator), can be misleading What is a difference? N 1 – N 2 At small N, trends may not be apparent

© Stanley H. Weiss, Oct-29 8 Probabilities of Death Due to PCa — Control vs. Screening Ratios & Differences

© Stanley H. Weiss, Oct-29 9 Screening: 3 criteria A good screening test: 1.must accurately identify persons with the condition (sensitivity — if you have the disease, the test says so) and those without (specificity — if you don’t have the disease, the test says that too); 2.must identify those with the disease early enough to make a difference in treatment outcomes (survival and/or quality of life); AND …

© Stanley H. Weiss, Oct Screening: 3 criteria (cont.) 3.preferably has high sensitivity (is very likely to indicate you have the condition if you really do), has high specificity (is very likely to indicate you don’t have the condition if you really don’t), and must be acceptable for the population screened, rapid, and ideally, noninvasive.

© Stanley H. Weiss, Oct Screening for Prostate Cancer: 3 testing approaches for detection 1.Digital Rectal Exam (DRE) 2.Prostate-Specific Antigen (PSA) blood test 3.Trans-rectal ultrasonography (TRUS)

© Stanley H. Weiss, Oct The Digital Rectal Exam (DRE) is a mainstay of PCa screening, but the examiner cannot feel the whole prostate nor small lesions

© Stanley H. Weiss, Oct Blood serum test – for Prostate- Specific Antigen (PSA) – takes a small sample of blood and tests for the amount of a substance that is mainly produced by the prostate gland 3.Trans-rectal ultrasonography (TRUS) – limited value for screening; history, exam or lab findings may indicate a need for TRUS evaluation Additional details about each follow the presentation in the supplemental slides

© Stanley H. Weiss, Oct What are the risks & benefits of screening? The value of screening can be proven only by showing a reduction in the chance of dying of prostate cancer without an unacceptable reduction in quality of life (from both the screening and increased use of treatments that can have negative side effects). Conclusive evidence for the value of prostate cancer screening has not yet accumulated. But some evidence does indicate that screening offers the possibility to diagnose early prostate cancer and to reduce deaths from this disease.

© Stanley H. Weiss, Oct Risks of Screening On the other hand, screening also detects cancers that do not threaten the patient’s life. Finding such cases cannot be avoided at present. When screening the general population for PCa by PSA, over 50% of the PCa’s detected will be minimal cancers (Draisma 2003). As immediate treatment of these has not been shown to be beneficial, detection and diagnosis of some tumors may be unnecessary and counter-productive, as in some patients there will be treatment-associated morbidity (and, rarely, even mortality).

© Stanley H. Weiss, Oct PSA: (prostate specific antigen) PSA A protein Almost exclusively produced by the epithelial cells of the prostate in normal and in pathologic conditions such as infection, urinary retention, enlargement of the prostate, and prostate cancer. Approximately 40% of patients with organ-confined prostate cancer show no elevation of serum PSA. Unresolved: At what PSA value should more invasive examinations — such as prostate biopsies — be conducted? Not yet clear…

© Stanley H. Weiss, Oct Issues in the Optimal Implementation of PSA Testing PSA – cutoff value issues: Should it be age and/or race dependent? Should clinical findings be incorporated into testing algorithms as well? PSA – clinical issues: (fix) (e.g., PSA rises with BPH and transiently with acute prostatitis) Distinction between two uses of PSA testing: –in symptomatic persons (routine clinical care) vs. –prevention by population screening (main concern of present discussion)

© Stanley H. Weiss, Oct Issues in the Optimal Implementation of PSA Testing PSA velocity issues Baseline value(s) – at what age? How often to screen? What is a “rapid” rise?

© Stanley H. Weiss, Oct The Clinician-Patient Conversation In order to have an evidence-based discussion of benefits, it is necessary to have data to support screening. In its absence, clinicians speculated loosely about: Part of this is paraphrased from Brett AS & Ablin RJ, “Prostate-Cancer Screening — What the USPSTF Left Out.” N Engl J Med online 10/27/2011. potential benefit (“We might catch prostate cancer early enough to save your life”) and potential harm (“Screening might result in burdensome interventions with serious complications”)

© Stanley H. Weiss, Oct The Clinician-Patient Conversation (continued) Needs to include a discussion of morbidity & mortality associated with screening, since: a positive test leads to additional diagnostic tests and possibly treatment that also carries associated risks, not just benefits

© Stanley H. Weiss, Oct Early Treatment – The Controversies A randomized trial has demonstrated that radical prostatectomy can decrease the chance of dying of prostate cancer as compared to delayed treatment. This benefit has been attributed in part to suppression of the male hormone, testosterone. However, even in the delayed treatment group 8 years later, only ~ 25% are at risk of developing metastatic disease. Thus, a majority of those with prostate cancer die with, not from, the disease.

© Stanley H. Weiss, Oct Early Treatment – The Controversies Still impossible to determine up front which cases will not progress. It has not been shown that the same favorable results of surgery can be achieved when prostate cancer has been detected by screening. Uncertainty therefore remains. Active Surveillance — which is more active than “Watchful Waiting” — is a treatment option for low-Gleason score PCa that reduces effects of aggressively treating every diagnosed case

© Stanley H. Weiss, Oct Benefits of Screening Screening can find potentially lethal cancers at an early, still curable stage as well as provide an opportunity for earlier, and possibly life-prolonging, treatment of additional tumors. Men who decide to be screened take a chance, and need to be informed about – and balance – the potential risks and benefits of screening and subsequent treatment. The decision currently must be individualized, and men who choose to be screened should not be denied the early diagnostic tests. Data concerning cost efficacy – an important determinant of public policy recommendations – are limited and controversial.

© Stanley H. Weiss, Oct European Randomised Study of Screening for Prostate Cancer (ERSPC) ERSPC established >10 years ago Largest randomized study (220,000 men in eight western European countries*) on screening for prostate ca Prostate cancer – 2 nd leading cause of cancer death in men in Western Europe and the U.S. * Netherlands, Sweden, Finland, Belgium, France, Spain, Italy and Switzerland

© Stanley H. Weiss, Oct European Randomised study of Screening for Prostate Cancer (ERSPC) Provides some evidence-based advice to the pivotal question – does screening lead to an improvement of cancer-specific survival? Initial results: 20 percent reduction in the rate of death from prostate cancer after first ten years of follow-up

© Stanley H. Weiss, Oct Caveats – ERSPC The 20% reduction in rate of death due to prostate cancer underestimates true effect since some controls were also screened (“contamination”) and some assigned to screening did not get screened (“noncompliance”), but … Model-based adjustment to correct for contamination only indicates that screening reduces mortality by 27%, and … Model-based adjustment to correct for both contamination and noncompliance indicates that screening reduces mortality by 31% Roobol MJ, et al., Eur Urol 2009:56(4);

© Stanley H. Weiss, Oct More Caveats – ERSPC Initial data analysis indicates need to screen 1410 men and treat 48 with cancer to save 1 life, but … Further modeling* indicates numbers needed to screen (NNS) & treat (NNT) to prevent 1 prostate cancer death decrease over time: Number of Years Elapsed NNSNNT 82, , * Loeb S et al., J Clin Oncol 29(4): , 2011 Definition: The number needed to screen (NNS) is the number of people who would need to be screened to prevent one person from dying. (Similarly for the number needed to treat, abbreviated NNT.) NNS is computed by taking the reciprocal of the difference between the fraction of people who die in the screening arm and the fraction of people who die in the control arm.

© Stanley H. Weiss, Oct More Caveats – ERSPC Limited # of biopsy samples Absence of A-A Limited follow-up

© Stanley H. Weiss, Oct Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial Objective: The PLCO hopes to provide some answers about the effectiveness of prostate cancer screening. Designed as a 17-year project of the National Cancer Institute (NCI). An initial report appeared in the New England Journal of Medicine online March 18, 2009 (in print – March 26 issue), coinciding with presentation of the ERSPC data at the European Association of Urology meeting in Stockholm, Sweden. PLCO: Six annual screenings for prostate cancer. FINDINGS: More diagnoses of the disease, but did NOT lead to fewer prostate cancer deaths. BUT there are limitations to the study…

© Stanley H. Weiss, Oct Caveats – PLCO Wrong PSA cutoff > 40% screened within 3 yrs of enrollment Half of “non-screened controls” really got PSA during trial, and 15% assigned to screening did not get screened, greatly REDUCING its power — only 33 percentage-point difference in screening between control and screening arms* Relatively short follow-up — indeed, 2009 publication of results to that date mandated by monitoring board, not by any achieved results * Schröder FH & Roobol MJ, Eur Urol 2010;58(1):46-52

© Stanley H. Weiss, Oct Caveats – PLCO Most men not biopsied when advised (limiting detection of early-stage PCa in screening arm) Limited & variable # of biopsies Variable care Very few A-A

© Stanley H. Weiss, Oct Caveat to Caveats – PLCO Despite the above weaknesses, in men with no or minimal comorbidities, after ten years number needed to screen was 723, number needed to treat was 5 to prevent one prostate cancer death But men with comorbidities who were diagnosed with PCa were less likely to receive curative treatment than men without comorbidities, and hence perhaps more likely to progress to death due to PCa, potentially reducing their benefit from screening and reducing the overall demonstrated value of screening in the PLCO Crawford ED et al., J Clin Oncol 29(4): , Note that these authors are mostly the same ones who initially reported the negative results of the whole PLCO for prostate cancer screening.

© Stanley H. Weiss, Oct Randomized Screening Trials

© Stanley H. Weiss, Oct Summary of Concerns There are limitations to the protocols that may have reduced the efficacy of screening Based on our knowledge of the epidemiology of prostate cancer, in each trial the follow-up period is too short to have expected to see the full potential benefits, and much too short to calculate cost/benefit ratios The US PLCO trial was significantly under-powered to have been able to demonstrate a positive result due to PSA screening of controls – reminiscent of the classic epidemiologic limitations of the negative cardiovascular “MRFIT” trial 2 decades ago

© Stanley H. Weiss, Oct The Göteborg Trial A smaller Swedish screening trial consisting of roughly 20,000 men designed with similar criteria as the ERSPC Current data were reported following a total of 14 years of screenings and follow-up Hugosson J et al., Lancet Oncol 2010(Aug); 11(8): –Some of the data from the trial were incorporated into the ERSPC

© Stanley H. Weiss, Oct The Göteborg Trial Final results indicate a roughly 50% decrease in mortality from prostate cancer in the screening group vs. the control group. –The benefit was greatest 10+ years from the beginning of the trial –Half of the attendees who died of PCa in the screening group were diagnosed in the first round of screening, and many of these men were 60+ years of age at entry –The number of men from the control group who may have received independent screening was not known or was not included – a bias that may underestimate the benefits of screening –293 screened and 12 diagnosed or treated to prevent 1 death from prostate cancer

© Stanley H. Weiss, Oct The Göteborg Results Cumulative Risk of Diagnosis: Cumulative Risk of Death: ▼ The approx ten-year point is marked by an arrow

© Stanley H. Weiss, Oct Probabilities of Death Due to PCa — Control vs. Screening Ratios & Differences

© Stanley H. Weiss, Oct-29 39

© Stanley H. Weiss, Oct-29 40

© Stanley H. Weiss, Oct Current Screening Guidelines from Major U.S. Organizations American Urological Assoc Best Practice Statement –Individual decision for those with 10yr life expectancy –Baseline PSA at 40 –PSA at subsequent intervals based on PSA level and risk factors American Cancer Society –Advises against routine screening –PSA should be offered as option Age 45 in those with risk factors (FH, AA) Age 40 in those at highest risk (multiple family members or a family member diagnosed at a young age) US Preventive Services Task Force –Do not screen routinely over age 75 –Inadequate evidence regarding younger ages –Current (October 2011) draft recommendation against PSA-based screening of any asymptomatic men

© Stanley H. Weiss, Oct Dr. Weiss’s Take on the BIG PICTURE PSA not for everyone Need more evidence to make good decisions –Studies evolving for whites, namely the studies we’ve just explored as well as others –No adequate studies commenced yet for other racial and ethnic groups!!

© Stanley H. Weiss, Oct Dr. Weiss’s Take on the BIG PICTURE Need improved guidance for clinicians as to what to do and in whom to do it There is some evidence of life- saving potential with PSA screening The costs (e.g., the NNS and NNT) may be much less than recent highly publicized estimates, and thus it may be cost-effective!

© Stanley H. Weiss, Oct Supplemental Slides

© Stanley H. Weiss, Oct DRE: (digital rectal examination) Tendency to detect larger tumors with DRE Low chance of detecting clinically insignificant tumors with DRE, but risk depends strongly on the PSA level. Limitation: small multi focal lesions with aggressive biologic potential are not detected with DRE alone. The DRE is subjective = variable between different examiners. Several studies have questioned the use of DRE in screening programs and found little or no additional beneficial effect of a DRE in men with PSA levels ≥ 4.0 ng/ml (Catalona 1994, Rietbergen 1997). DRE may provide an additional value in detecting clinically significant cancer in men with a low [?“normal”} range of PSA (< 4.0 ng/ml) (Eastham 1999, Han 2004).

© Stanley H. Weiss, Oct TRUS: (transrectal ultrasound) Similar to the DRE, the interpretation of TRUS is highly dependent on the investigator. Several studies have shown that the value of TRUS has LIMITED value as a screening test to detect cancer, But is indispensable for guiding prostatic biopsies and assessing the prostate volume.

© Stanley H. Weiss, Oct PSA-related tests F/T PSA ratio: Objective: To (try to) increase the specificity of PSA as a screening tool derivates from PSA are studied. Total PSA consists of complex PSA (cPSA) and free PSA (fPSA). cPSA is serum PSA that is bound to circulating proteins. The proportion of circulating cPSA is higher in patients with carcinoma than in those with benign enlargement. Studies comparing the diagnostic efficacy of cPSA with total PSA and the free to total (F/T) ratio so far report inconsistent results.

© Stanley H. Weiss, Oct PSA-related tests (continued) proenzyme PSA (pro-PSA): Form of free PSA Elevated in cancerous prostate tissue Results from a multi-center study have validated proPSA as a detector of early stage prostate cancer. Findings suggest that proPSA may be associated with aggressive and significant prostate cancer, worthy of further investigation See: cebp.aacrjournals.org/content/19/5/1193.abstract?etoc

© Stanley H. Weiss, Oct PCA3 Screening PCA3 is a non-coding mRNA molecule that is believed to be prostate specific. –It is highly over-expressed in cancerous prostate cells relative to benign tissue –Present in urine (no blood test necessary) Potential to be used as supplement for PSA testing –PSA has a 21% specificity but a 87% sensitivity for prostate cancer –Conversely, a test for PCA3 was reported to have a sensitivity of only 49%, but a specificity of 78% –Additional studies are needed

© Stanley H. Weiss, Oct Loeb S, Vonesh EF, Metter EJ, Carter HB, Gann PH, Catalona WJ What Is the True Number Needed to Screen and Treat to Save a Life With Prostate-Specific Antigen Testing? Purpose The European Randomized Study of Screening for Prostate Cancer (ERSPC) reported a 20% mortality reduction with prostate-specific antigen (PSA) screening. However, they estimated a number needed to screen (NNS) of 1,410 and a number needed to treat (NNT) of 48 to prevent one prostate cancer death at 9 years. Although NNS and NNT are useful statistics to assess the benefits and harms of an intervention, in a survival study setting such as the ERSPC, NNS and NNT are time specific, and reporting values at one time point may lead to misinterpretation of results. Our objective was to re-examine the effect of varying follow-up times on NNS and NNT using data extrapolated from the ERSPC report. Materials and Methods On the basis of published ERSPC data, we modeled the cumulative hazard function using a piecewise exponential model, assuming a constant hazard of for the screening and control groups for years 1 to 7 of the trial and different constant rates of and for the screening and control groups, respectively, for years 8 to 12. Annualized cancer detection and drop-out rates were also approximated based on the observed number of individuals at risk in published ERSPC data. Results According to our model, the NNS and NNT at 9 years were 1,254 and 43, respectively. Subsequently, NNS decreased from 837 at year 10 to 503 at year 12, and NNT decreased from 29 to 18. Conclusion Despite the seemingly simplistic nature of estimating NNT, there is widespread misunderstanding of its pitfalls. With additional follow-up in the ERSPC, if the mortality difference continues to grow, the NNT to save a life with PSA screening will decrease. J Clin Oncol 29(4) (February 1, 2011)

© Stanley H. Weiss, Oct Cumulative hazard functions of prostate cancer-specific mortality from the ERSPC with a piecewise exponential model Loeb S et al. JCO 2011;29: ©2011 by American Society of Clinical Oncology

© Stanley H. Weiss, Oct Loeb S, Vonesh EF, Metter EJ, Carter HB, Gann PH, Catalona WJ What Is the True Number Needed to Screen and Treat to Save a Life With Prostate-Specific Antigen Testing? J Clin Oncol 29(4) (February 1, 2011) Table 1. Modeled Results Assuming a Piecewise Exponential Model Abbreviations: NNS, number needed to screen; NNT, number needed to treat.

© Stanley H. Weiss, Oct If You Choose to be Screened, What’s the Best Way? Risk calculator or nomogram …

© Stanley H. Weiss, Oct-29 54

© Stanley H. Weiss, Oct Example: