With the increasing use of active surveillance, a greater proportion of men undergoing radical prostatectomy (RP) have high-risk disease, and as many as 50% of these men will develop biochemical recurrence (BCR) within 10 years.¹ Thus, an increasing number of men are being considered for postoperative radiotherapy (RT). Three randomized trials have demonstrated improved oncologic outcomes with adjuvant RT when adverse pathological findings are found at prostatectomy, specifically extracapsular extension, seminal vesical involvement and/or positive surgical margins (table 1).^2–4 However, the use of adjuvant RT has remained limited, at 10% to 15% at best.⁵ This low utilization is in large part due to the concern about overtreatment. About half of the patients in the control arms of these trials had no evidence of recurrence at 5 years. Thus, early salvage RT (given at the time of BCR) may be a preferable approach.

Three randomized trials comparing adjuvant RT to early salvage RT have recently been published (table 2).^6–8 RADICALS-RT recruited 1,396 patients in the United Kingdom, Denmark, Canada and Ireland (2007 to 2016); GETUG-AFU 17 recruited 424 patients in France (2008 to 2016); and RAVES recruited 333 patients in Australia and New Zealand (2009 to 2015). While the GETUG-AFU 17 and RADICALS-RT trials were designed to assess whether adjuvant RT was superior to salvage RT, the RAVES trial assessed whether salvage RT was noninferior to adjuvant RT. For all trials, patients randomized to receive adjuvant RT received it within 6 months after surgery. Salvage RT was triggered at low prostate specific antigen (PSA) level recurrences that differed slightly between the trials. The results of all 3 trials and a meta-analysis of the aggregate data (ARTISTIC)⁹ showed that the oncologic outcomes of early salvage RT were not inferior to adjuvant RT. In addition, all 3 trials reported increased side effects with adjuvant RT.

While these results support early salvage RT as the preferred approach, it is noteworthy that these trials included only a small proportion of patients with high-risk disease. Men with pathological Gleason score 8–10 and/or pT3a or higher disease comprised only 9% to 17% of those enrolled.^6–8 Since the majority had relatively favorable pathology, the potential benefit of adjuvant RT in high-risk patients may have been missed. In addition, fewer men underwent salvage RT than adjuvant RT in these trials. Since androgen deprivation therapy (ADT) was used concurrently with adjuvant RT in 2 of the trials, progression-free survival in the adjuvant RT group may have been favorably affected since ADT can delay time to progression. Immortal time bias, a well-recognized bias in observational studies, is another potential confounding factor. In the salvage RT arms of these trials, there was a period of time (salvage RT treatment planning and delivery) during which the outcome event (recurrence) could not occur. By scoring PSA failure on the salvage arm at a later time than the adjuvant arm, salvage RT could falsely appear superior to adjuvant RT.

If these trials were underpowered to address the adjuvant versus early salvage question in high-risk patients, could there still be a role for adjuvant RT? A recently published large retrospective analysis suggests a benefit to adjuvant RT in some cases. This analysis of 26,118 men from Germany and the United States, using propensity score matching to minimize treatment selection bias, found that in men with pN1, pathological Gleason score 8–10 and pT3a or higher disease, adjuvant RT was superior to early salvage RT with regard to all-cause mortality.¹⁰ In the salvage RT setting, multiple studies show better outcomes when radiation is administered at the lowest possible PSA level. Since PSA is proportional to the volume of residual disease, adjuvant RT would be delivered when there is the lowest microscopic disease burden (ie when PSA is undetectable).

The results of these trials establish early salvage RT as the standard of care in most cases, but adjuvant RT may still be preferable for some very high-risk patients. Prognostic nomograms such as the CAPRA-S scoring system (Cancer of the Prostate Risk Assessment–postsurgical) and/or the Stephenson nomogram can be used to accurately predict the probability of BCR.^11,12 If these nomograms predict a near certain chance of recurrence, adjuvant RT could be considered. Postoperative PSA nadir has also been shown to be an indicator of recurrence. In one study, BCR was unlikely in those with postoperative PSA nadir ≤0.02 ng/ml, but almost all cases with PSA nadir above this level eventually recurred, especially those with PSA ≥0.05 ng/ml.¹³ Prognostic nomograms and nadir postoperative PSA levels could be used together to select men for adjuvant RT.

New tissue-based biomarkers will undoubtedly be developed to assist in identifying the best candidates for adjuvant RT. There is some evidence from retrospective studies that a tissue-based genomic classifier score (Decipher Prostate Cancer Classifier), based on the expression of 22 genes, may help predict who might be best suited for adjuvant versus salvage RT.¹⁴ Dalela et al developed a risk-stratification tool incorporating the Decipher score and pathological features to identify patients who would benefit most from adjuvant RT.¹⁵ In patients with 2 or more high risk factors, adjuvant RT was associated with a 10.1% 10-year clinical recurrence rate compared to a rate of 42.1% in those managed with initial observation. While positron emission tomography/computerized tomography imaging has improved detection of residual disease after RP, the sensitivity is limited at the very low PSA levels at which adjuvant RT would be given. Thus, although we now have good evidence from randomized trials that an early salvage RT approach is preferable in most cases, the role of adjuvant RT after RP will likely evolve and continue to be debated.