BERKELEY, CA (UroToday.com) - The recently reported Prostate, Lung, Colorectal, and Ovarian Cancer Screening (PLCO) and European Randomized Study of Screening for Prostate Cancer (ERSPC) trials have led to a widely hyped but short-sighted and potentially dangerous conclusion about the use of prostate-specific antigen (PSA) screening-based screening for prostate cancer. Despite problems with design and protocol deviations, these trials have been used to recommend that PSA screening should not be performed routinely. We took exception to this decision.
|"...we believe that the implementation of PSA screening has resulted in significant reductions in metastatic disease in the prostate cancer population. This not only represents a reduction of human suffering but also a reduction of cost to the health care system. The recommendation to abandon routine PSA screening was wrong."|
Our major disagreement comes from the omission of one of the most important, certainly one of the most common, outcomes of the natural history of prostate cancer: the development of metastatic disease. Neither trial reported on this endpoint. This void in their handling of the data potentially obtainable from their efforts does a great disservice to the medical community and the patients they serve. The importance of metastatic disease development in a population can be understood when one considers the cost in terms of pain and discomfort to the patient and his family as well as the tremendous cost of treating such patients over the long natural history of this illness.
With these shortcomings in mind, we sought to understand the effect of PSA screening relative to prostate cancer-specific mortality (PCSM) as well as the development of metastatic disease. We did this by examining data from our institutional inception cohort of men treated for prostate cancer. Since we do not keep a database of all men in our population as they are screened, we could only look at patients definitively treated for prostate cancer. We divided patients into two groups: pre-screening (1986-1992, designated PRE) and post-screening (1993-1996, designated POST) relative to 1992 when the American Urological Association officially recommended routine PSA screening. We looked for variables affecting PCSM and metastatic disease.
One major confounder of the study is the effect of stage migration resulting from lead-time bias. We accounted for it by comparing the 10-year rates in the PRE arm to the 15-year rates of the POST arm. The reason we used a 5-year offset is because it was the most conservative estimate of the Japanese group who modeled lead-time bias. We then assessed various factors indicating disease virulence, follow-up intensity bias, and temporal group (PRE vs. POST) for their effect on PCSM and metastatic disease by comparing the rates at 10 years for the PRE group to those at 15 years for the POST group.
There were several characteristics of the dataset that were obvious, and expected, even from a cursory examination of the patient characteristics. First, disease aggressiveness, as assessed by Gleason score and clinical stage, was significantly higher in the PRE group, reflecting the anticipated stage migration. Second, there were many more patients in the POST group, likely reflecting a combination of the growth in our institution’s catchment area but also the large rise in prostate cancer diagnoses that were a result of PSA screening implementation (also noted in the US national cancer statistics). Third, at any one time, there were approximately 3 patients living with metastatic prostate cancer for every one patient who died of prostate cancer - which validated our decision to focus on metastatic disease as an endpoint.
As far as comparing the impact of screening implementation on metastatic disease, we noted significant associations. The 10-year PRE vs. 15-year POST rate of metastasis-free survival was 73.7 % vs. 81.7% (p < 0.0001; HR= 4.20; 95% CI= 3.08-5.73), respectively. This was in the face of accounting for disease virulence and follow-up bias. We noted that this difference persisted within risk groups and irrespective of androgen deprivation use.
Since the publication of the work, we have been asked to clarify a few issues. One is the potential effect of treatment changes over time. We assessed the effect of increasing radiation dose over time and noted it to be insignificant on multivariable analyses (p = 0.3371). To account for a possible effect from changes in radiation technique we also did a subset analysis of just the prostatectomy patients and found that POST patients still did better than the PRE patients (p < 0.0001; HR= 4.37; 95% CI= 2.81- 6.80). Another potential effect from changes in treatment over time was an increased use of post-operative radiation. We, in fact, found that the use of post-operative radiation decreased over time (adjuvant radiation PRE vs. POST = 5.5% vs. 3.1%, salvage radiation PRE vs. POST = 15.2% vs. 9.7%) likely reflecting the decrease in disease aggressiveness leading to fewer patients meeting criteria for post-operative radiation.
In the manuscript, we tried to make some assumptions in order to estimate the cost effectiveness of PSA screening relying on two sources of information: a cost-benefit analysis based on ERSPC and our data showing the proportion of PCSM to metastatic disease. Shteynshlyuger et al. assessed cost relative to PCSM, and we used our proportional difference to adjust their estimations. Essentially, the 3:1 ratio of metastatic disease to PCSM caused us to increase the event rate by a factor of 3 which yielded a two-thirds decrease in their number-needed-to-treat estimate from 48 to 16. The 16 from our adjustments was below their cut point of 21 patients, which they defined as a cost-effective threshold.
In summary, we believe that the implementation of PSA screening has resulted in significant reductions in metastatic disease in the prostate cancer population. This not only represents a reduction of human suffering but also a reduction of cost to the health care system. The recommendation to abandon routine PSA screening was wrong.
Jay P. Ciezki, MD 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.
9500 Euclid Avenue
Cleveland, OH 44195