ABSTRACT: Rising prostate-specific antigen (PSA) in nonmetastatic prostate cancer occurs in two main clinical settings: (1) rising PSA to signal failed initial local therapy and (2) rising PSA in the setting of early hormone-refractory prostate cancer prior to documented clinical metastases. Most urologists and radiation oncologists are very familiar with the initial very common clinical scenario, commonly called "biochemical recurrence." In fact, up to 70,000 men each year will have a PSA-only recurrence after failed definitive therapy. The ideal salvage therapy for these men is not clear and includes salvage local therapies and systemic approaches, of which the mainstay is hormonal therapy. Treatment needs to be individualized based upon the patient's risk of progression and the likelihood of success and the risks involved with the therapy. It is unknown how many men per year progress with rising PSA while on hormonal therapy without documented metastases. This rising PSA disease state is sometimes called, "PSA-only hormone-refractory prostate cancer." As in the setting of initial biochemical recurrence, evidence-based treatment options are limited, and taking a risk-stratified approach is justified. In this article, we will explore these prostate cancer disease states with an emphasis on practical, clinically applicable approaches.
In 2007, nearly 219,000 men are expected to be diagnosed with prostate cancer,[1] approximately 90% will undergo definitive local therapy,[2] and around 30% to 35% will develop biochemical (ie, prostate-specific antigen [PSA]) recurrence within 10 years.[3] Thus, approximately 60,000 to 70,000 men each year develop initial PSA recurrence. Similarly, almost 30,000 men die of prostate cancer annually. Presumably almost all these men progress through hormonal therapy and succumb to the disease in the hormone-refractory disease state
The natural history of the hormone-refractory nonmetastatic disease state is such that it may take 2 to 3 years between initial elevation in PSA on hormones to documented clinical metastatic disease. As such, there may be 60,000 to 90,000 men alive at any one time in this country with secondarily rising PSA nonmetastatic prostate cancer. In summary, this collective disease state is common, affecting many current era patients and the subject of many clinical encounters by urologists, radiation oncologists, and medical oncologists.
Definition of Primary PSA Recurrence
The goal of radical prostatectomy (RP) is to remove the entire prostate. Therefore, slight rises in PSA are used to indicate cancer recurrence, although the exact level that defines PSA recurrence is debatable. In general, PSA levels > 0.4 ng/mL[4] or > 0.2 ng/mL[5] are used in most studies. Recently, the American Urological Association published guidelines that establish the consensus definition of PSA recurrence after RP to be greater than 0.2 ng/mL and rising, as confirmed on a repeat test.[6] This definition is to establish recurrence for outcomes reporting; however, it may not be the appropriate cutpoint to initiate therapy.
Indeed, it is our practice in a patient with a consistent and clearly rising PSA, often based on ultrasensitive values, to occasionally begin salvage radiotherapy when the PSA is between 0.1 and 0.2 ng/mL. Because microscopic or focal benign prostate tissue can sometimes be left behind after RP and may produce some small amounts of PSA, it is clinically important to recognize that a PSA of 0.2 ng/mL may not always represent cancer recurrence. Therefore, in the majority of patients, we do wait until the PSA is > 0.2 ng/mL before beginning salvage treatments.
Unfortunately, defining recurrence following radiation therapy (RT) is more difficult. Unlike after RP, PSA does not fall to undetectable levels after RT. Rather, radiation induces a slow and not always steady PSA decline. The median time to PSA nadir is around 18 months and possibly longer following brachytherapy. Also, slight transient PSA upswings ("PSA bounce") are not uncommon. Finally, the concomitant use of hormonal therapy and the variable time period of return to normal testosterone can complicate the interpretation of PSA recurrence.
In 1997, the American Society for Therapeutic Radiology and Oncology (ASTRO) convened a consensus panel to develop a recurrence definition. The definition developed by the panel required three consecutive PSA increases after reaching a nadir, or a single rise so great as to trigger the initiation of hormone therapy.[7] The date of PSA failure is backdated to the midpoint between the PSA nadir and the first of the three rises. This backdating of the failure time introduces a bias that overestimates the success at shorter follow-up times (ie, PSA may be rising, but not enough follow-up has occurred to document three consecutive rises). In long-term studies, this definition introduces a bias in that no patient can recur "late" because all late recurrences are backdated to an earlier time, resulting in a leveling of the survival curve.
Because of these concerns, ASTRO convened a new consensus panel that developed new recurrence definitions: a PSA value higher than absolute nadir plus 2 ng/mL, or a PSA value higher than absolute nadir plus 3 ng/ mL.[8] This new definition is called the "Phoenix" definition in some circles because the consensus panel meeting was held in Phoenix, Ariz. Importantly, both definitions date the failure as the time the PSA rose above the threshold for recurrence (ie, failure is no longer backdated).
Natural History of Initial PSA Recurrence
The natural history of PSA recurrence is usually long but varied. Pound and associates[9] described 304 men with PSA recurrence following RP from Johns Hopkins Hospital who did not receive hormonal therapy until the time of metastasis. Moreover, very few received salvage radiation therapy. The median time from PSA recurrence to metastasis was 8 years, and from metastasis to death was 5 years. In a recent follow-up study by Freedland et al that included a slightly larger cohort, the median time from PSA recurrence to prostate cancer death was not reached after 16 years.[10] In the latter study, however, although rare, prostate cancer deaths were seen as early as 1 year after PSA recurrence.
Thus, although the natural history of recurrent prostate cancer is often one of a slowly progressing disease, in some men it can be very rapid. Moreover, patients today are younger than in the past, with a median age at diagnosis of 65 years.[2] In younger men with few competing mortality risks, even a slowly progressive cancer can ultimately lead to cancer death.[11]
Risk Factors for Clinical Progression and Prostate Cancer Death After Initial Rising PSA
Fifteen years ago, Carter and colleagues[12] showed that changes in PSA over time could predict the likelihood of being diagnosed with prostate cancer. Shortly thereafter, PSA kinetics were shown to predict the risk of distant vs local failure among men with PSA recurrence after RT[13] and RP.[14] These initial observations have been confirmed in later studies.[9,15,16] Recently, a rapid PSA doubling time (PSADT) has also been linked with prostate cancer death.[10,15,17]
The best cutpoint to define "rapid" PSADT is unclear. Various cutpoints have been described: 3 months,[17] 6 months,[18,19] 8 months,[20] 10 months,[9] and 12 months.[18] Given multiple cutpoints, it is likely that the association between PSADT and risk of poor outcome is on a continuum. Indeed, D'Amico and colleagues,[17] found that among men with a PSADT > 3 months after either RT or RP, PSADT as a continuous variable was significantly associated with prostate cancer death. More recently, this continuum between PSADT and risk was demonstrated by Freedland and coworkers[10] who identified three PSADT cutpoints separating men into four risk groups.
Another issue regarding the use of PSADT is how to calculate PSADT: How many PSA measurements are needed and over what time period? Unfortunately, no clear standard methodology exists. Prior studies found that in the early time period after PSA recurrence, PSA rises exponentially with first-order kinetics.[19,21] Thus, PSADT—which is based upon the natural log of PSA—is constant over time. Therefore, it is reasonable to calculate PSADT using two values as long as the two values are sufficiently spaced in time to avoid subtle variations in laboratory measurements from being interpreted as a rapid PSADT.
It is our practice to use the first two PSA measurements after recurrence (ie, ≥ 0.2 ng/mL) separated by at least 3 months before calculating a PSADT, with the exception of the patient with an extremely fast rise in PSA, in whom a shorter interval may suffice to determine PSADT. Whether a similar approach can be used for estimating PSADT using supersensitive PSA assays for values < 0.2 ng/mL is unknown.
Another unresolved issue is whether other variables, such as Gleason sum or time to PSA recurrence, add useful information to the PSADT. To a certain degree, the debate is academic in that all three variables are highly correlated: The patient with a rapid PSADT likely has recurred early with a high Gleason. Thus, if this patient has a poor outcome, should we ascribe it to the rapid PSADT, the early recurrence, or the high Gleason? While it is generally accepted that PSADT is the best prognostic factor currently available, some (but not all)[15,22] studies have found that the Gleason sum[9,10,23] and time to PSA recurrence[9,10,19,20] add information to the PSADT.
Although PSADT is a valuable prognostic factor, what about the absolute PSA level? In studies that attempt to estimate the risk of a future event (ie, metastasis or prostate cancer death), zero is often defined as the time of recurrence, and thus all men have similar small-sized tumors. However, if one does not define time zero as the time of recurrence, but rather as whatever the current time is, then not all men have equal-sized tumors (ie, some men are evaluated at the time of recurrence and others years later). In that case, the size of the tumor, as approximated by the absolute PSA level, becomes a significant predictor of time to metastatic disease,[24] or the likelihood that a bone scan or computed tomography (CT) scan done today will demonstrate metastatic disease.[16]
