Last Friday we noted that a new paper by Vickers et al. has formally recommended that PSA velocity be removed from guidelines on the use of diagnostic data to decide whether a man should have a biopsy to assess his potential risk for prostate cancer. We expressed no opinion on the article at that time because we wanted to have the chance to read the full paper before we made any comment.
We have now been able to obtain and read the entire article (click here to see the abstract), as well as editorial comment on the article by Yao and Lu-Yao, which appears in the same issue of the Journal of the National Cancer Institute.
Vickers and colleagues have published a number of prior papers arguing that PSA velocity has very limited value compared to PSA alone in assessing a man’s risk for prostate cancer at a specific point in time. In the current paper, Vickers and his colleagues do three further things:
- First, they discuss why the National Comprehensive Cancer Network (NCCN) guidelines on the early detection of prostate cancer and the American Urological Association (AUA) best practice statement on use of PSA do, in fact, include a recommendation that a PSA velocity of > 0.35 ng/ml/year or > 0.4 ng/ml/year (respectively) is an appropriate indication for a biopsy.
- Second, they question the validity of that decision.
- Third, they use data from the Prostate Cancer Prevention Trial (PCPT) to show — again — that, by comparison with the most recent PSA value alone, PSA velocity has minimal impact on risk for a positive biopsy in a man with a PSA level < 2.5 ng/ml. Using PSA velocity in this way does, however, have a very significant impact on the number of men who are advised that they need a biopsy.
So, let us see if we can follow the proposition put forward by Vickers et al. without getting drowned in the statistics.
There is no argument about the recommendations of the NCCN and the AUA. They do, most certainly, state that a PSA velocity of > 0.35 ng/ml/year or > 0.4 ng/ml/year indicates the need for a prostate biopsy. In both cases, the primary source for the decision to recommend a biopsy is a paper by Carter et al. published in 2006. This paper states that, “PSA velocity measured 10-15 years before diagnosis (when most men had PSA levels below 4.0 ng/mL) was associated with cancer-specific survival 25 years later.”
Based on the fact that this association clearly exists, but that this association predates actual diagnosis of prostate cancer by at least 10 years, Vickers et al. argue that there are still no data to suggest that PSA velocity is any better than the actual PSA level at a specific point in time to indicate the need for a biopsy.
Vickers et al. then carry out a re-analysis of data from 5,519 participants in the placebo arm of the PCPT. These men all had serial PSAs and DREs performed, allowing calculation and analysis of the relative values of PSA alone or PSA velocity in the diagnosis of prostate cancer. They also all received a prostate biopsy at the conclusion of the trial, which therefore allowed correlation of PSA data at a point in time, DRE data at a point in time, and PSA velocity with the actual diagnosis of prostate cancer. (There were actually data on 5,562 men available, but essential data were missing on one or more predictors for 133 men, leaving 5,519 participants for analysis.)
No doubt it is possible to quibble with some of the assumptions and the statistical analyses made by Vickers and his colleagues, but — at least for us — the data show, with a compelling degree of evidenc e, that in men aged 55 and older, based on data from the PCPT, there is little good justification to think that PSA velocity adds predictive accuracy to PSA alone or other standard predictors of risk for prostate cancer (e.g., a positive DRE).
Vickers et al. conclude, after a formal evaluation of published guidelines on PSA velocity for prostate cancer detection, examination of several definitions of cancer, and examination of numerous different methods of calculating PSA velocity, that:
We found no reason to believe that implementation of the guideline would improve patient outcomes; indeed, its use would lead to a large number of unnecessary biopsies. We therefore recommend that organizations issuing policy statements related to PSA and prostate cancer detection remove references to PSA velocity.
This may be a hard pill for others in the prostate cancer community to swallow. However, we are not aware of any data published in the past 4 years that successfully refute any of the data presented by Vickers and his colleagues (in this or other previously reported articles) and that offer a compelling reason for the inclusion of PSA velocity as a standard risk factor for prostate cancer that requires immediate biopsy in a man with a PSA level < 2.5 or < 3.0 ng/ml. If a man has a PSA of ≥ 2.5 or ≥ 3.0 ng/ml, then a biopsy may well be necessary anyway (although the correct PSA cut-point for normal recommendation of a biopsy is another area of controversy in itself).
In their editorial comments on this article, Yao and Lu-Yao note that the use of a PSA velocity of > 0.35 ng/ml/year on men in the PCPT would have resulted in nearly one in seven men ultimately receiving a biopsy — as compared with one in 20 men if a PSA value of 4.0 ng/ml had been used as the cut-off. It may well be that 4.0 ng/ml is still too high to be used as an appropriate cut-off PSA value for men under (or over) 55 years of age. However, Yao and Lu-Yao go on to state that, “PSA velocity measurements take time to acquire, and recognizing that such data add relatively little information may help prevent inappropriate postponement of follow-up in affected patients. Avoiding the wait to acquire subsequent PSA values may also help reduce some of the anxiety associated with testing.”
There is one proviso. We are not saying that PSA velocity has no value at all in assessing the need for a biopsy in individual cases. The decision to have a biopsy is very much a matter for individual discussion between a man and his doctors. If PSA velocity data are available, it may well be worth taking these data into account. What we are saying, however, is that, like Vickers et al., we believe the inclusion of a specific PSA velocity in guidelines as indicating need for a biopsy is not justified by available data.
Somehow we suspect that this is not a finalized area of discussion. However, new and better data are certainly needed to justify the continued, standardized use of PSA velocity as an indicator for prostate biopsy – as opposed to just the PSA alone. We will need to wait to see if the argument put forward by Vickers and his colleagues is compelling to those who are responsible for developing guidelines issued by the NCCN, the AUA, and others.