Prostate-specific antigen (PSA) testing became widely available in the United States in the late 1980s, and within a few years of widespread application in clinical practice, at least two new prostate cancer subsets had been "created": PSA failure post–definitive local therapy and PSA progression without overt radiologic evidence of metastases in the setting of castrate levels of testosterone (< 50 ng/ mL). Recognizing the dramatic impact of PSA on prostate cancer stage migration, Scher and colleagues described their "clinical states model" in 2000, providing an important framework to better understand this disease in terms of clinical progression and allowing for clinical research questions to be focused on the appropriate subset of patients with prostate cancer.[1]
Moul and colleagues have provided a crisp review of the landscape of prostate cancer with PSA progression. This is frankly a daunting task, given the significant heterogeneity of this patient population and the limited prospective data from which conclusions can be extracted. Their estimates of the number of patients with PSA-only progression may be a bit conservative; considering the number of patients undergoing definitive local therapy each year in the United States, a 20% to 30% PSA failure rate, and a 10- to 15-year natural history, there may in fact be nearly a million men in the US today with this clinical condition.
Failure: Is the Disease Local, Systemic, or Both
Moul et al provide an overview of the definition of PSA failure in the postprostatectomy and radiation therapy settings and discuss salvage interventions. One of the most difficult challenges faced by clinicians in managing initial PSA failure is to determine the potential utility of a salvage local therapy and then provide the patient with an appropriate discussion regarding risks and benefits. Given the well recognized inability of bone scan, computed tomography (CT), magnetic resonance imaging (MRI), and capromab pendetide (ProstaScint) imaging to provide meaningful data for the vast majority of patients with biochemical failure, clinicians are left with assessment of PSA kinetics and pathology (the latter even more problematic in radiotherapy-treated patients) to make therapy recommendations.
Without prospective evidence to guide decision-making, such discussions with patients can be considered a true expression of the "art of medicine." Novel markers and new insights into patients at high risk for systemic failure (circulating tumor or endothelial cells?) will be required before clinicians can distinguish high-risk from low-risk patients, enabling more rational management strategies.
Early Androgen Deprivation Therapy
Moul and colleagues have provided a review of the ongoing controversy regarding the application of androgen deprivation therapy in patients with biochemical failure only. Although in academic circles there has been a move toward use of this approach only in "high-risk" patients—those with PSA doubling times of less than 3 months—significant numbers of patients are still receiving hormonal therapy for PSA failure in the United States.[2]
Recent reports from Keating and D'Amico et al illustrate the harsh reality of androgen deprivation therapy–related toxicities.[3,4] Given the harm seen with merely short exposure to therapy, even proponents of intermittent therapy must remain vigilant regarding these emerging data.
Clinical Research in Patients With Biochemical Failure
Simply put, the challenge for investigators is to identify a nontoxic (or nearly nontoxic) therapy that works and that can be validated prospectively with hard endpoints. Consider that patients with biochemical failure (castrate or noncastrate) are typically asymptomatic and manifest a range of concerns about their status (blasé to PSA-induced hysteria) but are typically motivated to consider interventions as long as they are essentially without side effects. Investigators have learned the hard way about the vagaries and heterogeneity of PSA-only response endpoints, and the FDA has made it clear for now that only hard endpoints—time to progression and survival—are acceptable for drug approval.[5,6]
