Adjuvant Radiotherapy After Prostatectomy: Does Waiting for a Detectable Prostate-Specific Antigen Level Make Sense?

Abstract

Three very well conducted and remarkably similar randomized trials (European Organization for Research and Treatment of Cancer [EORTC] 22911 [1], Southwest Oncology Group [SWOG] 8794 [2], and [Arbeitsgemeinschaft Radiologische Onkologie] 96-02 [3]) have shown that adjuvant radiotherapy (RT) directed at the prostate bed improves prostate-specific antigen (PSA) relapse-free survival compared with observation among patients with high-risk features. In these three trials, high-risk patients were characterized as having either positive surgical margins, extracapsular extension (pT3a), or seminal vesicle involvement (pT3b). In principle, ‘‘adjuvant’’ therapy implies treatment in the absence of measurable disease, and for prostate cancer this means an undetectable postoperative PSA level. These trials, conducted during the time periods of 1992‐2001 for EORTC, 1997‐2004 for ARO, and 1988‐1997 for SWOG, used the PSA assays available at those times. Those assays had a threshold of detection of 0.2 ng/mL. In fact, many of the patients in these adjuvant trials did have measurable PSA levels. In the EORTC trial 9% of patients had a PSA >0.2 ng/mL; in the SWOG trial 35% had PSA >0.2 ng/mL; and in the ARO trial (which had access to assays with lower threshold) 20% of patients had PSA >0.05‐0.1 ng/mL, and 59% of patients had PSA >0.03‐0.1 ng/mL. Since then, the threshold for ‘‘undetectable’’ PSA has been readjusted with the development of ultrasensitive PSA assays now able to detect levelsas lowas 0.01ng/mL. Within this framework then, many, if not most of the adjuvant trial patients would be categorized today as having measurable disease. By extrapolation, the median PSA level in the adjuvant trials is estimated to fall within the >0.05‐0.1-ng/mL range. In the era of ultrasensitive PSA assays, the question now facing clinicians and patients is this: even in the setting of high-risk features, is there any benefit of adjuvant RT if the PSA is undetectable with an ultrasensitive assay (i.e., PSA <0.01 ng/mL)? Conversely, is there a downside in waiting until the PSA becomes barely detectable before proceeding with adjuvant, or more precisely in this case salvage RT, atthattime?Unfortunately,eventheseexcellentadjuvanttrials cannot answer these questions satisfactorily. One reason is that the PSA assays used were not sensitive enough; the other is that the ‘‘control’’ arm was an ‘‘observation’’ arm and not a ‘‘salvage RT’’ arm. To explore an answer to this question, we look at the relationship between effectiveness of postoperative RTand PSA level at the time of postoperative RT. A literature search identified available published studies of adjuvant and salvage RT, both retrospective (4‐13) and prospective (1‐3), that reported the 5-year bNED (biochemical no evidence of disease) control rates as well as the median pre-RT PSA level. These data are summarized in Table 1. Before looking at these data it is constructive to have a radiobiologic framework of the expected relationship, and we construct this as follows. We know that tumor control probability (TCP) after RT can be modeled by Poisson statistics and is commonly expressed as TCP = e NS , where N is the number of tumor