PSA screening, lives saved, and hormonal monotherapy: It’s time to face some quietly ignored facts, "Beyond the Abstract," by Ian Haines, MBBS, FRACP, FAChPM and George L. Gabor Miklos, PhD

BERKELEY, CA (UroToday.com) -

A commentary on the multiple articles in the Journal of the National Cancer Institute dealing with prostate-specific antigen screening trials, prostate cancer mortality, and hormonal therapies

The hypothesis that PSA-screening of asymptomatic men from the general population leads to a meaningful reduction in prostate cancer (PCa)-specific mortality currently rests on data from two randomized European clinical trials, ERSPC and GOTEBORG. Both are promoted as providing solid statistical evidence for reductions of 21% and 44% in PCa-specific mortality due to PSA screening.[1, 2] These trials underpin the popular understanding (hereafter termed the PSA-hypothesis), that PSA screening saves sufficient lives to offset the known harms from treatment with which every practicing urologist is acutely aware.[3] These frequently quoted figures are accepted in good faith by many physicians, who believe that PSA screening of asymptomatic men is best practice. It may come as a surprise to learn that these figures are far more fragile than their supporters claim.

On the basis of PCa-specific mortality differences between the screening and control arms of these two trials, we questioned the robustness of the PSA-hypothesis.[4] We suggested that androgen deprivation monotherapy (ADT), (which is known to select for a more aggressive cancer cell population), could increase the risk of PCa-specific mortality. If correct, excessive ADT monotherapy used in the control arms of these trials would lead to an artificially increased difference in PCa-specific mortality between the arms, and the figures of 21% and 44% would be inflated ones. In short, the impact of PSA screening would be less than believed, and the PSA-hypothesis would require revision. Our article in the Journal of the National Cancer Institute was accompanied by a commentary from Brawley[5] who provided a broad appraisal of the biases, benefits and risks of PSA screening and subsequent treatments, and pointed out that a small effect of ADT could be powerful.

Our ADT monotherapy-driven PCa mortality proposal (hereafter termed the ADT-hypothesis), was followed by articles in the same journal,[6, 7] which described the ADT-hypothesis as fundamentally flawed. Following these articles, there was an editorial supporting the reduction in PCa mortality seen in a non-randomized screening trial,[8] as well as a commentary,[9] with differing perspectives: one on the hazards of PSA screening, by Wilt, and the other on the merits of tailored PSA screening, by Scardino. These articles together with our responses,[10, 11] and the recently released Finnish data from ERSPC,[12] all in the same journal, provide an informative thread for urologists. They can be evaluated in the context of the FDA’s ongoing updates of ADT therapies[13] and the USPSTF and ASCO recommendations.[14, 15, 16]

The most striking aspect of the PSA screening debate is the lack of predictive scientific analysis, based on quantitation of data. The claim by Walsh,[6] and Carlsson, Roobol, Schroder, Hugosson and Auvinen,[7] that stage migration together with strict guideline-based ADT treatment, completely accounts for the differences to which we have drawn attention, is simply an assertion devoid of any quantitative support. Its veracity, or lack thereof, will only be settled by rigorous scientific testing. This process can begin with important data that have either been lost in the details of the trials, or wilfully ignored.

The first data set is from Finland, which reported its detailed findings separately from the pooled 7 country ERSPC analysis. The Finnish randomized trial revealed no statistically significant difference in PCa-specific mortality between the screening and control arms of this largest component of the entire ERSPC trial.[12] Since the Finnish data of over 80 000 men contributes nearly half of the total of 162 000 men from 7 European countries that make up ERSPC, and since the pooled ERSPC data are statistically significant while the large Finnish contribution is not statistically significant, extensive heterogeneity clearly exists in PCa-specific mortality between the different countries. In fact, data from 5 of these countries (Finland, Belgium, Italy, Spain and Switzerland), provide no support whatsoever for the PSA-hypothesis, while 2 outliers, Sweden and the Netherlands, do so, a point made by a number of authors.[4, 5, 10, 11, 14, 15, 17]

The second data set is from Sweden. It involves the Swedish contribution to ERSPC, which consisted of 60% of the data from the stand-alone GOTEBORG clinical trial, transferred into ERSPC. It is scientifically unacceptable to transfer only some data from a full stand-alone clinical trial to a recipient conglomerate. This dubious transfer allowed the pooled data to reach statistical significance, a point highlighted by the USPSTF,[15] but somehow overlooked by the field. Without this transfer, the statistical significance of the ERSPC trial vanished.

It’s time to face the quietly ignored, but crucial facts, described above. The dubious transfer of some data from GOTEBORG and the invalid pooling of heterogeneous data within ERSPC, reveal the fragility of the popularly quoted figure of a 21% relative reduction in the risk of death from PCa. In truth, this figure was always fragile, even though it continues to be defended at all costs. For example, it is indeed astonishing to many that some authors argue that heterogeneity is a strength of the ERSPC study because it provides ”an opportunity to analyze how differences in screening algorithms relate to the screening effect.[18] In addition, applying different methods of analysis,[19] to already defective data is statistically futile. It cannot return the data to normalcy and it certainly cannot confirm that PSA screening reduces PCa mortality. These hyper-analyses are more than just a problem of failing to see the forest for the trees. They appear to be a problem of not even clearly seeing the trees owing to a preoccupation with the statistical undergrowth. From a clinical perspective, no amount of algorithmic fine tuning, sophisticated data processing, model building, adjustments for missing values, or the application of different methods of analysis, can overcome the statistical dangers inherent in transferring and pooling heterogeneous data and the potentially misleading clinical conclusions that follow.

Finally, while relative differences invariably appear clinically impressive, it is prudent to examine them at the absolute level.[20] The claim of a 21% reduction in PCa mortality is quite different when viewed in absolute terms. The PCa figures in ERSPC were 0.41% in the screening arm versus 0.52% in the control arm in an intention-to-treat analysis. It is this difference of 0.11% that is at the center of debate, a figure, furthermore, that is assumed to be a reduction in PCa mortality, but in terms of scientific equanimity, it can either be a reduction, or an increase in deaths in the control arm, or a mixture of both.

In contrast to ERSPC, the stand-alone GOTEBORG trial has been endorsed by some as the best conceived and executed of all randomized PCa clinical trials,[21] and it now becomes the valuable center point for testing the PSA-hypothesis and the ADT-hypotheses. In GOTEBORG, there was a massive imbalance in patient treatment, with over 3 times more ADT monotherapy in the control cohort than in the PSA screening arm.[4, 11] Crucially, the data on stage migration, treatment and PCa-specific mortality already exist for each of the different at-risk patient groups in GOTEBORG, (namely for low, medium, high, advanced and metastatic disease) in both the screening and control arms. However, these data are still to be published or publicly released. The prediction of the stage migration-based authors,[6, 7, 8] is that it will completely account for all differences between the different risk groups and their treatments in the screening and control arms and that the ADT-hypothesis will fail. In addition, these authors should also be able to predict the expected and observed levels of PCa mortality in the large Finnish and Netherlands components of the ERSPC trial, as these have yielded divergent differentials between their screening and control cohorts. The GOTEBORG data will also quantify the contribution to PCa-specific mortality made by ADT monotherapy. If it is small, our ADT-hypothesis fails on the basis of normal scientific scrutiny, not on opinion-based beliefs.

It therefore now only remains for the GOTEBORG authors to release the existing de-identified data for independent audit. However, although this trial was largely funded from the public purse, the following statement from Carlsson, Assel, Sjoberg, Ulmert, Hugosson, Lilja and Vickers is unambiguous; “Data from the Goteborg randomized trial are not available to outside investigators.”[22] A similar restriction on access to ERSPC data has come from Schroder and Roobol.

The pushback to avoid independent and timely evaluation of the ERSPC and GOTEBORG data, or to seriously consider any explanation of them outside of the PSA-screening-saves-lives area, flies in the face of normally accepted scientific protocol. Instead of scientific evaluations in which the robustness of data is the central issue, practicing urologists find themselves in an unsatisfactory situation where the evidence on ADT monotherapy is not solid enough to base decisions on what is best for their patients with increased PSA levels, yet vital evidence that could help in this regard is quarantined.

Practicing urologists know the well-described all-cause clinical consequences of ADT monotherapy, namely coronary heart disease, reduced insulin sensitivity, diabetes mellitus, decrease in muscle strength, anemia, bone fractures, and cognitive dysfunction, but none of us can yet quantify the degree to which ADT monotherapy contributes to hastening ADT resistance, metastatic spread, and PCa-specific deaths. The molecular evidence, however, points in one -- and only one -- direction. ADT monotherapy selects for a more aggressive cancer cell population.

So where do we go from here? Given the anxiety on the part of a patient with an elevated PSA level and proven localized prostate cancer who is not considered suitable for observation or for definitive radical local treatment, and the fear on the side of the physician that a therapeutic intervention may be missed, ADT monotherapy for other than metastatic disease looms as an appealing option, owing to its ease of administration and its immediate effects. PSA levels plummet after ADT monotherapy and can remain low for years, providing enormous comfort to a patient, and a sense of treatment efficacy for a physician. However, although the long term molecular consequences of ADT monotherapy are still being uncovered, its use nevertheless appears to be becoming more widespread outside of the FDA guidelines. We would advocate for more caution from physicians regarding its increasing use, while uncertainty about efficacy and risk continues.

We believe that in addition to the sympathetic perspective for patient treatment set out in the ASCO recommendations,[16] 3 issues deserve absolute priority. First, all men about to undergo PSA testing should be given written information on the facts about absolute benefits and risks of diagnosing early stage cancer. This would include the well-documented and extensive harms that ultimately befall so many asymptomatic men as a result of unnecessary overdiagnosis and overtreatment of low-risk disease.[3] Second, all men contemplating having a screening PSA test or treatment for proven early-stage prostate cancer should be alerted to the lack of evidence supporting ADT monotherapy for treatment of early stage disease and to the known downstream effects of hormonal monotherapies. They should also be informed of the proven inferiority of ADT monotherapy to radiation plus ADT for locally advanced high-risk T3 disease, a well-established finding derived from large randomized trials.[23, 24] Third, a randomized clinical trial needs to be fast tracked on ADT monotherapy in which the primary and immutable precondition is that all authors agree to release de-identified patient data into a publicly available constantly-updated database. This precondition already occurs in some areas of molecular medicine.

There is no justification whatsoever for quarantining taxpayer-funded clinical data from independent scientific scrutiny. The release of such crucial data is essential, not only for the treatment of prostate cancer patients, but for the provision of the best possible public health.

References:

  1. Schroder FH, Hugosson J, Roobol MJ, et al., Prostate-cancer mortality at 11 years of follow-up. N Engl J Med. 2012; 366(11): 981-990.
  2. Hugosson J, Carlsson S, Aus G, et al., Mortality results from the Goteborg randomised population-based prostate-cancer screening trial. Lancet Oncol. 2010; 11(8): 725-732.
  3. Brawley OW, Goldberg P. How We Do Harm. New York: St Martin’s Press, 2012.
  4. Haines IE, Miklos GLG. Prostate-Specific Antigen Screening Trials and Prostate Cancer Deaths: the Androgen Deprivation Connection. J Natl Cancer Inst. 2013; 105(20): 1534-1539.
  5. Brawley OW. Prostate cancer screening: biases and the need for consensus. J Natl Cancer Inst. 2013; 105(20): 1522-1524.
  6. Walsh PC. Re: Prostate-specific antigen screening trials and prostate cancer deaths: the androgen deprivation connection J Natl Cancer Inst. 2014; Feb;106(2):djt432.
  7. Carlsson S Roobol M J Schroder F H Hugosson J Auvinen Are: Prostate-specific antigen screening trials and prostate cancer deaths: the androgen deprivation connection J Natl Cancer Inst. 2014; Advance Access, April, DOI:10.1093/jnci/dju079.
  8. Stampfer MJ, Jahn JL, Gann PH. Further evidence that Prostate-Specific-Antigen screening reduces prostate cancer mortality J Natl Cancer Inst. 2014; Mar 1;106(3):dju026.
  9. Wilt TJ, Scardino PT, Carlsson SV, et al., Prostate–Specific Antigen screening in prostate cancer: perspectives on the evidence. J Natl Cancer Inst.2014; Mar 1;106(3):dju010.
  10. Haines IE, Miklos GLG. Re: Prostate-specific antigen screening trials and prostate cancer deaths: the androgen deprivation connection. J Natl Cancer Inst.2014; Feb;106(2):djt433.
  11. Haines IE, Miklos GLG. Re: Prostate-specific antigen screening trials and prostate cancer deaths: the androgen deprivation connection. J Natl Cancer Inst. 2014; Advance Access, April, DOI:10.1093/jnci/dju081.
  12. Kilpelainen TP, Tammela TL, Malila N, et al., Prostate cancer mortality in the Finnish randomized screening trial. J Natl Cancer Inst.2013 105(10): 719-725.
  13. http://www.fda.gov/Drugs/DrugSafety/ucm229986.htm
  14. Moyer VA. Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2012; 157(2): 120-134.
  15. Chou R, Croswell JM, Dana T, et al., Screening for prostate cancer: a review of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med. 2011; 155(11): 762-771.
  16. Basch E, Oliver TK, Vickers A, et al., Screening for prostate cancer with Prostate-Specific antigen Testing: American Society of Clinical Oncology Provisional Clinical Opinion. J Clin Oncol. 2014; 30(24) 3020-3025.
  17. Miklos GLG. The dilemmas of prostate cancer screening. MJA. 2013; 199(9): 582.
  18. Hugosson J, Carlsson SV. The dilemmas of prostate cancer screening. MJA. 2013; 199(9): 583-584.
  19. Zappa M, Puliti D, Hugosson J, et al., Eur. Urol. 2014. Jan 7. pii: S0302-2838(13)01480-2.
  20. Dubben H-H. Trials of prostate-cancer screening are not worthwhile. Lancet Oncol. 2009; 10(3) 294-298.
  21. Catalona WJ. The United States Preventive Services Task Force Recommendation against Prostate-Specific Antigen Screening – Counterpoint. Cancer Epidemiology, Biomarkers & Prevention. 2012; 21: 395-397.
  22. Carlsson S, Assel M, Sjoberg D, et al., Influence of blood prostate specific antigen levels at age 60 on benefits and harms of prostate cancer screening: population based cohort study. BMJ. 2014; Mar 28; 348:g2296.
  23. Warde P, Mason M, Ding K, et al. Combined androgen deprivation therapy and radiation therapy for locally advanced prostate cancer: a randomized phase 3 trial. Lancet. 2011; 378(9809 ): 2104-2111.
  24. Widmark A, Klepp O, Solberg A, et al. Endocrine treatment, with or without radiotherapy, in locally advanced prostate cancer (SPCG-7/SFUO-3): an open randomized phase III trial. Lancet. 2009; 373(9660): 301-308. 

Written by:
Ian Haines, MBBS, FRACP, FAChPMa and George L. Gabor Miklos, PhDb as part of Beyond the Abstract on UroToday.com. This initiative offers a method of publishing for the professional urology community. Authors are given an opportunity to expand on the circumstances, limitations etc... of their research by referencing the published abstract.

aDepartment of Medical Oncology, Monash University at Cabrini Hospital, Malvern, Victoria, Australia 3144 and
bAtomic Oncology, Newport Beach, New South Wales, Australia 2106 (corresponding author)

All of this work has been carried out with personal funds. There have been no sponsors or external financial inputs. George Miklos does not accept funds from Federal or State governments, does not accept grants, gifts, gratuities or remuneration for speaking commitments from drug companies or their proxies, and does not stand to gain financially from any of this work. Furthermore, he does not hold shares in companies that relate to any of the PSA technologies mentioned in this work.

Prostate-specific antigen screening trials and prostate cancer deaths: The androgen deprivation connection - Abstract

More Information about Beyond the Abstract