Tuesday, October 11, 2011

Screening for Prostate Cancer: The Task Force Report

The U.S. Preventive Services Task Force has issued its report regarding screening for prostate cancer. I will make a few observations here based upon the work I have done previously and other recent efforts. Let me begins by saying that not all prostate cancer is the same and at this time no one really knows how to determine the difference. This is a genetic pathway problem and the issue is that many elements of the pathway are yet to be identified and moreover the dynamics of the pathway are still unclear.

1. Does screening for prostate cancer with PSA, as a single-threshold test or as a function of multiple tests over time, decrease morbidity or mortality?

2. What are the magnitude and nature of harms associated with prostate cancer screening, other than overtreatment?

3. What is the natural history of PSA-detected, nonpalpable, localized prostate cancer?

Let us examines these questions based upon the Task Force’s approach and based upon our analyses and compare them.

1. Does screening for prostate cancer with PSA, as a single-threshold test or as a function of multiple tests over time, decrease morbidity or mortality?

Task Force:

No good- or fair-quality RCTs addressed this question. Two poor-quality RCTs with important flaws in design and analysis  do not show a mortality benefit from PSA screening independently or in a meta-analysis. We identified no RCTs that measured health outcomes from PSA screening by means other than single-threshold tests.

Now would one be willing to bet one’s life on 2 poor tests! Also as we have noted elsewhere the tests reported in NEJM a year or so ago were flawed for several reasons, mainly they used a fixed and out dated threshold, and in fact asked the wrong question, and also especially in the European study tests at too high a level and tested too infrequently.

The medical issue is that the PCa which is the most deadly is also the fastest growing and it should be tested at a lower level and more frequently. The remaining PCa is really indolent, no one denies that, but the determination of which is which is difficult without sophisticated genetic tests.

2. What are the magnitude and nature of harms associated with prostate cancer screening, other than overtreatment?

Task Force:

One cross-sectional and 2 prospective cohort studies of fair-to-good quality reported short- and long-term psychological harms from prostate cancer screening. Although abnormal screening results did not affect summary measures of anxiety or health-related quality of life, men with false-positive PSA screening test results were more likely to worry specifically about prostate cancer, have a higher perceived risk for prostate cancer, and report problems with sexual function compared with control participants for up to 1 year after the test. In 1 study, 26% of men with false-positive screening results reported moderate- to-severe pain during the prostate biopsy; men with false-positive results were also more likely to undergo repeated PSA testing and additional biopsies.

The counter to this is simple, many people are just terrified about death, yet it comes to all. Are there issue for some men, yes, but those most likely are the same men who smoke, drink in excess, are obese, and the list goes on. Is a prostate biopsy painful, it is a state of mind and medication. That is why God made the poppy!

This second issue seems to be a contrived issue. People have the same issue regarding colonoscopies, which have been clinically effective in reducing death from colon cancer. Yet we seem not to hear this issue. The Task force also had the same concern about mammographies.

Frankly why should this be an issue for the Task Force at all. The public will talk but let an educate patient decide, it is ultimately the patients choice.

3. What is the natural history of PSA-detected, nonpalpable, localized prostate cancer?

Task Force:

Three fair-quality cohort studies with small-to-medium sample sizes, highly self-selected elderly patients, and high drop-out rates show that some men with PSA detected, nonpalpable, localized (stage T1c) prostate cancer have good health outcomes up to 10 years after diagnosis. We did not identify any population-based studies in which patients with stage T1c prostate cancer were followed longitudinally with no intervention in order to determine health outcomes resulting from the natural progression of disease.

It appears that the Task Force has no answer here. In fact the asked the wrong question. They should have asked what genetic markers were prognostic of a virulent form of PCa? Simple question, but we as of yet really do not know. Assume we knew? Then What? Can we test every cell for these genes. What of the issue of a cancer stem cell, thus there being say just a few hundred or ten of them, and must we find them? Do they give off a measurable inter-cellular market to express their presence?

The answer seem to be we do not yet know. How will we find out? More studies with more men. Yes more biopsies where the samples are analyzed genetically in a large scale study. Yet with the admonitions given off by the Task Force that may soon become unlikely.

A recent study on the genetic level in Oncology states:


Arul M. Chinnaiyan and colleages sought to develop a new read-out for prostate cancer due to the non-specificity and unclear mortality benefit of PSA testing. The goal was to find a novel biomarker or biomarkers that can facilitate the individualization of PSA levels.  


The test the researchers developed is a clinical-grade, transcription-mediated amplification assay that can detect prostate cancer non-invasively in the urine. The read-out of the test is the quantitative measure of a TMPRSS2:ERG fusion transcript that is unique to prostate cancer. More than 50% of PSA-screen prostate cancer harbors this fusion between the transmembrane protease, serine 2 (TMPRSS2) and the v-ets erythroblastosis virus E26 oncogene homolog (avian) (EGR) gene.

Thus we are beginning to see such tests arising. Yet not a comment from the Task Force regarding this work. This is the essence of the Translational approach, the translating of science into the practice of medicine. This is another shortfall of the Task Force.

Now let us return an review what the Task Force said. The Task Force total basis of their conclusion seems to be as follows:

Effectiveness of Early Detection and Treatment: A meta-analysis of 2 poor-quality RCTs (random controlled trials) of population- based screening for prostate cancer using PSA and digital rectal examination found no reduction in prostate cancer mortality in men invited versus men not invited for screening (relative risk, 1.01 [95% CI, 0.80 to 1.29]). A recent RCT reported that men who received PSA screening had a decreased risk for receiving a diagnosis of metastatic prostate cancer. The USPSTF assessed the study as providing inconclusive evidence of benefit from screening because of a high likelihood of unequal outcome ascertainment and small absolute numbers of an imperfect intermediate health outcome (metastatic prostate cancer is an imperfect surrogate of prostate cancer mortality because of both high initial response rates to androgen deprivation therapy and competing causes of death). No RCTs have reported health outcomes from the variations of PSA screening that consist of multiple measurements over time (for example, measurements of PSA velocity, PSA slope, or PSA doubling time).

However that is the point. We examined many of these trials and we concluded:

1. PSA thresholds must be age dependent as well a family history dependent.

2. PSA velocity is an essential element of the analysis and it means that PSA tests should be performed annually starting say at 30 to 35. Thus PSA velocity can be reasonably determined.

3. PSA testing needs a better baseline as a test. There is substantial variation between PSA values based on different testing methods.

4. PSA measurements including % Free PSA are also useful and should equally be used.

5. PSA algorithms can assist in ascertaining what may be a cancer but a biopsy is usually required.

6. Biopsies

One of the conclusions is as follows:

How Does Evidence Fit with Biological Understanding? Prostate-specific antigen screening presupposes that most asymptomatic prostate cancer cases will ultimately become symptomatic cases that lead to poor health outcomes. However, the natural history of PSA-detected, nonpalpable, localized prostate cancer is poorly described. No prospective studies have followed a population-based cohort of patients with screening-detected cancer who have had no intervention in order to determine health outcomes resulting from natural progression of the disease. Evidence from small, selected cohorts of men with arbitrarily defined “favorable risk” (that is, with prostate cancer likely to be clinically indolent) suggest a good prognosis for some men with screening-detected cancer; however, the longest of these studies has reported health outcomes from 2 to 10 years after diagnosis only.

The statement is evidential hearsay at best. The problem is that prostate cancer, like so many cancers, is still somewhat of an enigma, but it is fair to say that it is a genetic pathway breakdown and this breakdown results in expanding cells that lose the ability to understand where they belong, thus metastasis. The Report totally fails to join this issue. The above is about as close as it appears to get.

Now the Press as usual is in the fray. The NY Times reports :

But doctors are divided about when to recommend watchful waiting. The decision can be guided by an indicator called the Gleason score, a measure of the aggressiveness of the cancer found in a biopsy, but there is often disagreement about how to care for men whose scores are in the middle — neither highly aggressive nor probably not aggressive. In addition, the biopsy process itself is imprecise; a standard “12-core biopsy” gives information about only one three-thousandth of the prostate, says Dr. Eric Klein of the Cleveland Clinic. According to research at Johns Hopkins, staging and grading mistakes occur in about 20 percent of specimens.

The Gleason score is a microscopic pathological marker. Simply put are the prostate glands normal, small opening with evenly distributed basal cells with luminal cells atop? Or are the glands reproducing, many small immature glands, then expansion of half formed glands, and then just a mass of uncontrolled cells, moving from Gleason 1 through 5 and Gleason 2 through 10 on its score. That may be useful but it totally fails to tell us the genetic makeup of the cancer stem cell controlling this entire process.

The goal is determining the genetic makeup, that is what should be the topic, not this foolishness.

As a final conclusion: one should look at the number of studies used, 3 for the first question, three for the second, and 3 for the third. Yes they scanned a thousand or so but 9 studies were used! If a doctoral student came to me with what they thought was an earth changing conclusion based on 9 disparate studies, some rated poor, I would wonder if that student were up to the challenge. I would not shout to the four winds the student's erstwhile world changing conclusion. Then again that is just me.

Welch, one of the authors of "Overdiagnosed" wrote today in the NY Times:

It’s a stark juxtaposition: screening is good for women and bad for men. But just how different are these two cancer screening tests? 

The answer is: not very. Neither is like the decision of whether or not to be treated for really high blood pressure. That’s an easy one — do it. Instead, both breast and prostate cancer screening are really difficult calls, and the statistical differences between them are only of degrees. Reasonable individuals, in the same situation, could make different decisions based on their valuation of the benefits and harms of screening. 

Personally, as a 56-year-old man, I choose not to be screened for prostate cancer (and, were I female, I believe I would choose not to be screened for breast cancer). Some of my patients have made the same choice, while others choose to be screened. That’s O.K., because there is no single right answer.
Screening is like gambling: there are winners and there are losers.
 I never gambled a penny in Las Vegas, I know the odds. However I have seen PCa patients with mets and DIC, not a pretty sight. The problem is that the question is NOT PSA. It is effective screening and what facilitates that. Somehow we have taken a crude but somewhat useful screening test and made it an all or nothing, a sine qua non. 
It is akin to throwing out the X Ray because we really cannot use a plain film to diagnose a block or bleed well enough in a stroke. Then along came CAT and MRI scans. The issue must not be PSA, the issue must be proper screening. We are not there yet, and we are wasting too much time arguing over what we know to have problems.