Prostate-specific antigen following radiation therapy

Abstract

Historically, conventional external beam radiation therapy (RT) has been an important therapeutic modality for early prostate cancer. Although the true therapeutic value of RT remains a point of controversy (1, 8), RT is the treatment of choice for many patients (5, 13). For example, for elderly patients (average age: 69 years) with Stage Tl and T2 disease treated with RT, the lo-year adjusted survival (AS) rate is 70% and the lo-year clinical local control rate is 88% (5). However, during the last 5 years, the prostatespecific antigen (PSA) has dramatically changed our understanding of clinically localized prostate cancer. First, for clinical Stage Tl, T2, and T3 disease treated with RT, the pretreatment PSA is a very strong predictor of ultimate long-term clinical outcome (11, 21). Based on the data from Kuban et al. (ll), RT patients with a pretreatment PSA 10 rig/ml. Of course, the same paradigm applies to radical prostatectomy (RP) patients. Surgical patients with a pretreatment PSA > lo-12 rig/ml have a 75-80% probability of being upstaged at the time of RP and these patients are at high risk for long-term clinical failure (4, 12). Second, the use of the posttreatment PSA to evaluate loyear bNED survival for both RT and RP documents that only a modest fraction of early prostate cancer patients may be ultimately cured with either modality (15, 17, 19). For RT patients with Tl-T3 disease, only lo-30% may be biochemically disease-free at 10 years (15, 17). For RP patients with favorable Tl and T2 disease, the lo-year bNED survival rate is only 41% (19). These data have stimulated interest in new methods that may safely increase the radiation dose to the prostate compared to conventional doses of 6,500 cGy to 7,000 cGy (at 180 cGy to 200 cGy per fraction). These new methods include transperineal prostate implant, dose escalation 3D conforma1 therapy, and neutron or proton beam radiotherapy (2, 6, 14, 16, 18). While the PSA has provided us with many new insights, it also has created new dilemmas for radiation oncologists. One of the most important problems which the PSA creates is how to interpret early posttreatment PSA profiles in terms of long-term clinical outcome (3). In this issue of the International Journal of Radiation Oncology Biology and Physics, Horwitz et al. (7) have successfully addressed many of the problems with the interpretation of the posttreatment PSA. First and most important, there is currently no standard definition of bNED. Horwitz et al. (7) highlight this problem by comparing five different definitions of bNED used at five different institutions. A variety of other definitions such as PSA l rig/ml, >50% increase, etc.), there are literally hundreds of possible definitions for bNED. Some authors have simply defined bNED based on a very low posttreatment PSA value such as <l rig/ml. This is reasonable because Kavadi et al. (10) have demonstrated a greater than 50% PSA recurrence rate at 5 years for RT patients with a nadir PSA level between 1 rig/ml to 2 rig/ml. However, even if one defines bNED according to a very low posttreatment PSA level, Kavadi et al. (10) have shown that the probability of relapse will still be a function of the pretreatment PSA level. In the RP data where bNED is defined as no detectable PSA, a 5-year bNED survival rate of 61% could not predict lo-year outcome because RP patients continue to fail at a rate of 4% per year producing a IO-year bNED survival rate of only 41% (19). Therefore, there is no definition of bNED, which ensures the absence of a subsequent clinical failure. In addition, the ability of any definition of bNED to accurately predict clinical outcome such as AS will be a function of Gleason score, pretreatment PSA, level of benign prostatic hypertrophy, patient age, and the number of years since completion of treatment. In short, the exact relationship be-