Evidence for the Use of Salvage Radiotherapy
Recent evidence from these three randomized trials suggests that early intervention with ART can lengthen biochemical disease-free, metastasis-free and overall survival in patients with pathologically advanced prostate cancer.[12-14] However, a disadvantage of routine ART is treating those who would never develop biochemical recurrence after RP, and unnecessarily exposing an increased number of patients to the side effects of RT.
In addition, there is some evidence that the use of ART may be associated with an increased risk of toxicity as compared to SRT. A retrospective multi-institutional analysis of 959 men treated with either adjuvant (19%) or salvage (81%) RT found a low rate of toxicity with a 5-year rate of late grade 2 or higher genitourinary (GU) toxicity of 12% and a late grade 2 or higher gastrointestinal (GI) toxicity of 4%. More serious toxicity was rare, with grade 3 GU toxicities in only 1% of all patients and grade 3 GI toxicities in 0.2% of all patients. Given the small number of events, there were no predictors that correlated with late GI toxicity, and there was no difference in GI toxicity between ART and SRT. However, on multivariate analysis adjuvant RT as compared to both salvage RT (16% vs 11%) and the use of hormonal therapy (19% vs 11%) predicted for increased risks of grade 2 or greater urinary toxicities. Therefore, the use of SRT might protect a significant portion of men who do not ever require radiotherapy, and in addition, even for those treated with RT may provide a modest reduction in GU toxicity. However, the cost of a strategy of using SRT in lieu of ART is that a certain portion of patients may have a lower chance of successful eradication of their disease with SRT. Whether an equivalent survival benefit can be attained with vigilant surveillance and early initiation of SRT upon PSA relapse is an unanswered question, and SRT cannot at present be considered to be equivalent to ART.
Given this uncertainty, two groups of investigators have attempted to define prognostic factors that predict the likelihood of obtaining a benefit from SRT. Trock et al. retrospectively analyzed 635 men, who either received no salvage treatment (n=397), SRT alone (n=160), or SRT combined with hormonal therapy (n=78). The authors found that 70% of all deaths during follow-up were from prostate cancer with 10-year rates of prostate cancer–specific survival of 86% in those treated with salvage RT as compared to 62% without RT. This represented a 3-fold increase in prostate cancer–specific survival compared to those who received no salvage treatment (hazard ratio [HR], 0.32; P < .001). The addition of hormonal therapy to SRT did not improve prostate cancer–specific survival. Also noteworthy was that when SRT was restricted to the population of patients with pT3 disease who would have been candidates for ART, the use of salvage RT also provided an OS benefit with 10-year OS of 98% vs 89%. Interestingly, the prostate cancer–specific survival benefit of SRT was only seen in men with a PSA doubling time of < 6 months, independent of pathologic stage or Gleason score. This runs counter to the more commonly held principle that a short doubling time is indicative of distant disease and, therefore, a lack of benefit to SRT. Moreover, patients who received SRT more than 2 years from the time of biochemical recurrence did not experience significant increases in prostate cancer–specific survival.
Further evidence for the use of SRT in prostate cancer comes from a retrospective study by Stephenson et al, in which they developed a model using a cohort of 1,540 patients. The authors described several prognostic features that should be considered when predicting improved biochemical control after SRT: These included PSA level < 2.0 ng/mL at time of SRT, Gleason score of 7 or less, PSA doubling time > 10 months, positive surgical margins, androgen-deprivation therapy before or during SRT, and the absence of lymph node metastasis. It was again demonstrated that SRT may significantly alter the natural course of the disease, as 60% to 70% of patients with disease recurrence develop metastasis within 6 years if they do not receive salvage therapy. In addition, SRT is recommended to patients with more favorable prognostic features, as they are thought to be at lower risk for widely disseminated disease. However, the Stephenson study, like the one by Trock et al., suggests that patients with unfavorable prognostic features may also benefit from SRT if treatment is initiated early after biochemical recurrence. Indeed the Trock study would suggest that patients with the shortest doubling time are at the greatest risk for prostate cancer–specific death. Although these patients may be less likely to have PSA control, given their greater risk of death from prostate cancer if they do achieve disease control, this translates into a cause specific survival benefit. In contrast, those with a longer PSA doubling time may be more likely to achieve PSA control with SRT, but given the lower clinical risk this does not appear to change the risk of prostate cancer–specific death.
Current Treatment—Defining the Surgical Bed
Although some authors have reported on the use of low-dose rate or high-dose rate brachytherapy for the treatment of prostate cancer that has recurred after RP, by far the most commonly used treatment modality is external beam radiotherapy (EBRT). Therefore, our discussion will concern EBRT only. External beam salvage radiotherapy typically involves 3D conformal or Intensity Modulated Radiation Therapy (IMRT) to the prostate bed alone, with radiation fields designed to treat areas at the highest risk for local recurrence. Radiation therapy treatment volumes are in principle identical to those used for ART; therefore, lessons from ART randomized trials and ART consensus statements apply.
The randomized trials mentioned earlier were conducted in the era before the widespread adoption of 3D conformal or IMRT techniques, and therefore involved 9 × 9 cm or 10 × 10 cm fields centered around the prostatic fossa.[12-14] However, 3D conformal and IMRT techniques allow for the targeting of the prostatic fossa, urethrovesical anastamosis, and surrounding tissues at risk, with relative sparing of the rectum, bladder, and penile bulb. Multiple consensus guidelines have been created for the definition of the clinical target volume (CTV), most significantly from the EORTC, RTOG, and RADICALS groups.[34-36]
All three consensus groups generally advocate for the treatment of the vesicourethral anastamosis (VUA) and surrounding periurethral tissue. However, they advocate therapy to different amounts of additional tissue such as the bladder and seminal vesicle beds. The RTOG and RADICALS groups recommend defining the VUA using the most inferior visualized urine in the bladder on sagittal reconstruction, while the EORTC defines the VUA as 15 mm cranial to the penile bulb. At the level of the pubic symphysis, anteriorly and posteriorly, all three consensus groups essentially cover the region from the pubic symphysis to the rectum, and laterally the medial border of the obdurator internus and levator ani muscles. The lateral borders were generally the pelvic fascia superior to the pubic symphysis. At the bladder wall, the EORTC has perhaps the most limited CTV definition, and in contrast to the RTOG and RADICALS groups, does not advocate for the inclusion of 1.5 cm of posterior bladder and bladder wall. In the rectovesical /seminal vesicle bed space, the EORTC and RTOG advocate for the coverage of the seminal vesicle beds if there is pathologic involvement of the seminal vesicles in the surgical specimen, but to otherwise largely spare the seminal vesicle beds (though they do say to cover where the base of the seminal vesicles used to reside). Any retained seminal vesicle remnants should be included if the seminal vesicles were involved pathologically. The superior border in the rectovesicular space is at or 5 mm above the level of the cut end of the vas deferens or at the level of the most superior surgical clips. Inferiorly, the RADICALS group recommends placing the border at 8-12 mm below the vesicourethral anastamosis, but not to include the penile bulb. There was some concern in the RTOG group that apical tumors could extend into the genitourinal (GU) diaphragm and inferior urethral sphincter, and this was the reason it was recommended that the inferior aspect of the CTV extend to a level just above the penile bulb.
Separately, Miralbell et al. recommend a cylindrical CTV centered 0.5 cm posterior and 3 mm inferior to the VUA, measuring 4 cm in height and 3 cm in diameter. This volume considerably spared the rectum, and may represent a way in which to limit radiation-associated toxicities and improve the quality of life of prostate cancer patients. This CTV recommendation was based on an MRI series of 60 men, and is consistent with another MRI study showing recurrences largely around the VUA. However, this very VUA-centric volume stands in contrast to another MRI study which showed more local recurrences in the rectovesicular space outside of the proposed CTV. Further studies regarding the optimal volume of treatment are necessary, and it is hoped that information from the RADICALS trial will shed more light.
Minimizing daily set-up error and ensuring reproducible localization of the prostate bed is a current area of study. Calypso beacon localization has been suggested as a useful tool for localization of the prostate bed, as has daily portal imaging with implanted gold fiducial markers or daily cone-beam imaging or kilovoltage imaging. These techniques attempt to minimize daily setup error and take into account any variation in the location of the VUA depending on day-to-day differences in rectal volume and bladder distension. A general consensus on the differential benefit of these techniques has not been found, though most authors agree that daily localization is important for reducing treatment margins and thus further reducing radiation to normal tissue.
The proper radiation dose that delivers a balance of optimal disease control while limiting side effects is not clear; however, it is thought that the use of increased RT doses may provide higher chances of cure. Until recently, there were only three retrospective studies with small sample sizes that showed that doses above 64.8 Gy are beneficial.[43-45] While doses ≥78 Gy are used for RT in the definitive setting, doses for ART or SRT are generally lower because it is assumed that the tumor burden is microscopic,[46-48] and the presence of bladder and rectum within the prostate resection fossa increases the normal tissues radiated. As noted previously, randomized ART trials delivered radiation in the range of 60-64 Gy to relatively large fields.[12-14] The RADICALS trial is testing a dose of 66 Gy in 33 fractions, or 52.5 Gy in 20 fractions. King et al. recommend at least 70 Gy based on a retrospective study showing a significant dose response between 60 and 70 Gy of radiation to the prostate bed. Specifically, King et al. analyzed 122 patients with pathologically negative lymph nodes with a median follow-up > 5 years. Thirty-eight patients received a median dose of 60 Gy to the prostate bed and 84 patients received a median dose of 70 Gy. Sixty-eight patients received four months of androgen suppression therapy and 72 patients received whole-pelvic RT. The authors observed a significant dose response from 60 to 70 Gy (25% vs 58% biochemical disease-free survival at 5 years, respectively; P < .0001). On multivariate analysis the two clinical factors that predicted improved biochemical-free survival were a pre-RT PSA level ≤ 1 ng/mL (HR 0.28, P <.0001), and no seminal vesicle involvement (HR 0.44, P = .009). Thus, this study suggests that higher doses may help increase the likelihood of optimal disease-free survival.
The dose of 70 Gy correlated with an increased dose of 6 Gy required for SRT vs ART, which King et al. argued in a separate manuscript was due to the additional disease burden carried by SRT patients vs ART patients. In the absence of evidence that this additional dose causes worse late toxicities in patients undergoing SRT, a radiation dose in the region of 70 Gy is reasonable. Currently, the American Society for Therapeutic Radiology and Oncology (ASTRO) advises the use of the highest dose of radiation that can be delivered with acceptable morbidity (at least 64 Gy at conventional fractionation) for SRT.
Hypofractionated radiotherapy (daily radiation doses of greater than 2 Gy) has been considered for SRT in a retrospective analysis of 50 patients. Hypofractionated therapy is potentially desirable due to its shorter overall treatment length and theoretically higher biologically equivalent dose. Though toxicity and 2-year biochemical control rate appeared equivalent to published data for standard fractionation, additional follow-up and greater numbers of patients are needed before widespread adoption of this technique.
The use of hormone therapy in combination with SRT is an area of controversy that will hopefully be clarified by three randomized trials: 1) The RTOG 96-01 trial, 2) The RTOG 05-34 SSPORT trial, and 3) The RADICALS trial. The RTOG 96-01 trial is a prospective randomized trial comparing postoperative RT with and without 2 years of bicalutamide(Drug information on bicalutamide) 150 mg/day which has completed and should be presented in 2010. The RTOG 0534 is an ongoing phase III trial of short-term androgen deprivation with pelvic lymph node or prostate bed–only radiotherapy (SPPORT) in prostate cancer patients with a rising PSA after RP. This 3-arm randomized trial is assessing prostate bed RT vs prostate bed RT with short-term androgen ablation vs pelvic and prostate RT along with short-term androgen ablation. As noted previously, the RADICALS trial is a prospective trial with two randomizations. The first randomization will investigate immediate ART versus delayed SRT at the time of biochemical recurrence. Patients receiving RT will then be further randomized to RT alone, RT + 6 months of hormones, or RT + 2 years of hormones. Although hormone therapy has been shown to improve overall survival in combination with EBRT for men with prostate cancer of intermediate- or high-risk disease, the value of hormone therapy has not yet been proven for men undergoing either ART or SRT.[53-55]
Side Effects and Toxicities Associated With Radiotherapy After Prostatectomy
Radiation treatment is the only potentially curative treatment available for most patients with biochemical failure after RP. However, some would argue that quality of life (QOL) is as important as survival. Despite the evidence in support of using RT in this setting, the decision to use it must take into account the side effects associated with treatment. There have been multiple reports of acute and late toxicities after post-operative radiation therapy in prostate cancer. Overall, RT appears to be well-tolerated in patients undergoing ART and SRT, and lessons drawn from patients undergoing ART are therefore broadly applicable to SRT.
In the SWOG 8794 study, no patients had to interrupt their RT secondary to side effects, although grade 2 or greater complications were more common in the ART group than in the observation arm (23.8% vs 11.9%, respectively; P = .002). Urethral strictures (17.8% vs 9.5%; P = .11), and rectal complications (3.3% vs 0%; P = .02) were the most frequent toxicities. In a companion health-related QOL study, 217 of 425 SWOG 8794 patients completed a questionnaire at baseline, 6 weeks, 6 months, and annually for 5 years. The 6-week assessment was included to record side-effects at their peak at the end or RT. Not surprisingly, patients being treated with RT had a greater likelihood of a decline in bowel QOL at the end of RT as compared to the observation arm, but after 2 years, there was no significant difference between the two groups in bowel QOL. With respect to genitourinary QOL, patients in the ART arm experienced significantly more urinary urgency than those in the observation arm. However, there was no statistically significant difference in erectile dysfunction (ED), but given that the SWOG trial was performed prior to adoption of nerve sparing RP, > 90% of patients in both the ART and observation arms had severe ED, limiting the ability to comment on the effect of RT on erectile function in this patient population. Most noteworthy was that although global health-related QOL was worse in the ART group initially, it became similar by year 2, and at 5 years, patients in the ART group reported an overall better QOL compared to those in the observation arm. This is not surprising when taking into account the increased risk for metastasis and death as well as the burden of salvage and hormonal therapies among the patients in the "wait-and-see" arm.
In EORTC trial 22911, radiation treatment was interrupted as a result of toxic effects in 3.1% of patients, consisting of diarrhea, urinary frequency, proctitis, cystitis and anal pain. Grade 2 or 3 late effects were significantly more numerous in the ART arm (P = .0005), but grade 3 toxicities were uncommon, with a 5-year rate of 2.6% in the observation arm and 4.2% in the ART arm (P = .0726). No grade 4 or higher late toxic effects were reported. In comparison to the EORTC 22911 and SWOG 8794 trials, the patients in the ARO 96-02/AUO AP 09/95 study had a significantly lower rate of severe (grade 3 and higher) toxicities at only 0.3%. This relatively low rate of complications is likely due to the use of three-dimensional treatment planning, which is known to reduce acute and late toxicities for RT for prostate cancer.
In addition to the toxicity data from these randomized ART trials, there have been several assessments of complications following SRT. In a phase II prospective study by Pearse et al., 75 patients with biochemical relapse or local recurrence after RP were evaluated for acute and late complications after SRT and 2 years of ADT. Twelve percent of patients had gastrointestinal (GI) dysfunction and 40% had genitourinal (GU) dysfunction prior to receiving RT. Median follow-up was 45.1 months. No patients interrupted treatment secondary to side effects. Overall, 94% of patients experienced acute complications, but grade 3 toxicities were rare and the cumulative incidences for severe GI and GU toxicities were 1.6% and 2.8% at 36 months, respectively. There were no late grade 4 complications. Patients with preexisting GU dysfunction and acute GU toxicity were more likely to have persisting late GU toxicity. In addition, the more severe the acute GU toxicity, the more likely it was to persist.
Peterson et al. reported on late toxicities (those occurring more than 90 days after completion of radiation treatment) in 308 postprostatecomy patients who had undergone salvage therapy. In the study, radiation dose ranged from 54.0 to 72.4 Gy with a median dose of 64.8 Gy and was given in 1.8-2.0 Gy fractions. Median follow-up from the end of treatment was 60 months. Thirteen percent of patients reported late complications, but only an estimated 0.7% (95% CI, 0.0–1.6%) of patients would experience severe (grade 3 or higher) toxicities by 5 years. Among those reported in the study were grade 3 cystitis and grade 4 rectal complications. These results are consistent with those of other reports, including data from the three recently randomized trials on ART.
Finally, as mentioned previously, Feng et al. reported on 959 patients who received ART or SRT, with a median dose of 64 Gy. At 5 years, grade 3 urinary complications were observed in 1% of patients and grade 3 bowel complications were only seen in 0.3%, indicating excellent tolerance to ART and SRT. Similar toxicity was seen in a series from UCSD and Germany, which showed resolution of acute urinary symptoms without grade 3 toxicity. Long-term toxicity was rare, and health related QOL changes were minor in comparison to baseline scores. Together, these studies support a low incidence of severe toxicities in patients receiving RT after RP.