Contemporary radiation therapy can result in cure of prostate cancer in a substantial proportion of cases. Deliberation continues regarding when it is best used rather than other radical treatments, in particular, RRP. Direct randomised evaluation of them would appear unlikely to ever happen, given that a highly anticipated and well funded international trial comparing radical surgery with seed brachytherapy recently closed due to poor accrual (ACOSOG Z0070 – the “SPIRIT” trial). Without such evidence, non-randomised data describing both tumour control and quality of life issues must be balanced against each patient’s concerns and overall health status. Multidisciplinary clinics should be regarded as the standard forum for helping patients balance these issues.
Irrespective of the perceived tumour biology, men under 60 years of age are offered surgical therapy in many centres based on, amongst other issues, the practical advantages of ascertainment of pathological tumour stage (discussed above), more accurate grading and a more dependable early PSA endpoint of success in most cases. Many patients, however, will be offered radiotherapy preferentially. This may be on the basis of advancing age, a high risk of not achieving surgical complete clearance or patient preference. Evidence to help determine appropriate therapy is now available for many clinical scenarios within this group of men.
In many cases, radiotherapy treatment decisions are now driven by the patient’s risk profile tumour containment. These profiles were typically developed using historical data, and either externally validated or confirmed as clinically relevant by stratification levels in prospective trials. The most simple of these is a three tiered system with (506):
Low risk: PSA≤10 ng/mL and Gleason Score 2-6 and stage cT1-cT2a
High risk: either cT3 or cT4 or PSA>20 ng/mL or Gleason Score 8-10
Intermediate risk falls between these levels
While nomogram models (507) have been shown to have more reliable discrimination of outcome, the three level model serves well in relation to known therapeutic option categories. These may involve combinations of external beam radiotherapy (EBRT), brachytherapy (BT) or androgen suppression (AS) therapy.
For those with low risk prostate cancer, modern EBRT techniques appear to be highly efficacious (508, 509). PSA control would be expected in over 85% of such cases presently. These treatments are typically delivered using conformal 3D techniques, or intensity modulated radiotherapy (IMRT) where these more complex treatments may aid in controlling toxicity, especially at higher doses. These excellent results are yet to have an additional demonstrated benefit when combined with therapies such as AS.
High risk prostate cancer traditionally has been a disease with a poor outcome when treated with radiotherapy alone (as with surgery). In the days before PSA detection of tumours, biochemical recurrence rates were typically over 80% using radiotherapy as primary therapy (510). With the risk of both local and metastatic progression being of concern, randomised trials investigated the role of adjuvant AS in what are now known to be typically high risk patients. Published trials of locally advanced disease showed AS to have a beneficial effect when given after radiotherapy for a duration of either 2 years (511), 3 years (375) or indefinitely (512). An overall survival benefit of 16% at 5 years was seen in the Bolla et al trial when compared with radiotherapy alone. Hence, most men today with advanced cancers will have this style of treatment offered as a minimum. Fit men in this group may also be offered therapy on a clinical trial, as the improved results in this group are continuing to be optimised.
Intermediate risk prostate cancer, however, appears to have a lesser reliance on treatment of sub-clinical metastatic disease present at diagnosis, as borne out in trials evaluating the value of local therapy intensity. Responses to increasing radiation doses (“dose escalation”) have now been shown in a randomised trial. For the subset of PSA 10-20 ng/mL, the freedom from failure rate at 6 years increased from 43% to 62% when the radiation dose was increased from 70Gy to 78Gy (513). This supports the results of other dose escalation observational series (509, 511, 514).
Additionally, short-term use of AS combined with EBRT has shown to be of benefit to some patients. The precise explanation for this improvement in results is less clear though. Observational studies suggested a benefit predominately to intermediate and possibly high risk men (515). A number of randomised trials also now exist which, to some extent, cover this group of men. The mature RTOG 86-10 trial examined the role of 2 months of complete AS before as well as during radiotherapy in patients with bulky primary tumours and was conducted prior to the wide availability of PSA. The subanalysis by grade showed significant gains in loco-regional control, cancer-specific and overall survival in the Gleason 2-6 tumours, also further reinforcing that bulky high-grade cancers should be approached as high risk disease. Another trial has recently reported a freedom from failure together with a survival advantage to having 6 months of neoadjuvant AS (NAAS) prior to 70Gy of EBRT in intermediate and high risk men (516).
Somewhat disparate though are the results from the Trans-Tasman Radiation Oncology Group (TROG) 96.01 trial which compared radiotherapy of 66Gy alone to the same dose in combination with either 3 or 6 months of neoadjuvant maximum androgen blockade. Looking at the risk subsets of men from this study shows that the benefit shown for the use of 6 months NAAS (particularly in terms of freedom from BF and cancer-specific mortality) was primarily limited to the high risk men. The power to detect a difference, however, was limited by the small proportion of intermediate risk patients (517).
Also somewhat confounding matters is a large 2x2 phase III trial of men with a calculated risk of nodal positivity >15% showing that 4 months of AS was a significant benefit only when given prior to and during radiotherapy (rather than adjuvantly), and furthermore, only when combined with whole pelvic radiation fields (rather than prostate only) (518). As most of these trials showing a benefit to AS in combination with radiotherapy have employed lower doses of radiation than used presently, the precise indications for combination therapy in this group are unclear, it does appear that a subgroup of men, who usually fit within the typical intermediate risk profile, will benefit from short term neoadjuvant therapy.
Toxicity: The major concerns with EBRT are damage to normal tissues at risk viz. the rectum, bladder/urethra as well as the neurovascularity involved with erectile function. Modern planning systems have capabilities to produce detailed descriptions of the dose to structures of interest (dose-volume histograms). Applying accurate constraints to these parameters is partly the explanation for the observation that serious toxicity from modern 3D EBRT is uncommon, with evidence suggesting that it is less common now than with previous EBRT planning and treatment techniques, even those previously using much lower doses (519).
During a typical radical course of EBRT, symptomatic urinary toxicity occurs in most men which can be a combination of so-called irritative or obstructive symptoms. Grade I-II toxicity typically not requiring any or only minor medical intervention happens in approximately 50%, while serious problems potentially requiring invasive intervention happen in less than 1%. Urethral stricturing is a major concern after radiotherapy in high doses, and occur in approximately <1% presently.
Towards the end of a course of radiotherapy, most men will be aware of some rectal urgency associated with frequency. Rectal bleeding during or soon after therapy occurs in 0-2% of men in modern series (519) with the late bleeding rate being approximately 1% also. A measurable, but not clinically problematic, change in bowel habit will be observed in 10-20% (519).
Changes in erectile function related to increases in radiation dose are yet to be adequately documented in long-term data from large scale randomised dose trials. Observational series do suggest that in those with adequate erections prior to radiotherapy, 50% will maintain erectile function for more than 2 years (520). The control arm (radiation alone to 66Gy) of a large Australian randomised trial showed that 1 year after radiotherapy, approximately half of the patients sexually active prior to treatment remained active (376). This trial also showed that the addition of AS to radiotherapy does not appear to increase the toxicity of radiotherapy beyond that seen with radiation alone, in agreement with other published results (521).
Brachytherapy (BT) for prostate cancer can involve 2 forms of interstitial implant therapy, either permanent low dose-rate (LDR) radioactive seed implant or a temporary high dose-rate (HDR) implant.
For men with early stage prostate cancer, the use of seed BT has increased dramatically over the past decade with potentially a quarter of all early cancers now being treated this way in the USA (522). The modern technique uses accurate imaging and guidance systems to place metal seeds, containing either Iodine125 or Palladium103 as the radioactive source, in a planned array within the prostate to deliver a conformal dose of radiotherapy.
Typically a TRUS unit is used to image the prostate in real-time (with CT or MRI used in some centres) and a template utilised to guide needles loaded with the seeds into the prostate transperineally. The procedure is usually relatively simple and brief from the patient’s viewpoint, often administered in an outpatient setting, making it attractive to many. A number of series have now reported both PSA and toxicity outcomes for many thousands of men treated this way (523, 524). Although the long-term results published for this procedure amount to beyond 10 years at only a few centres (525, 526), the tumour control rates appear to be similar to those of surgery.
Biochemical control rates for those in the low risk group are presently greater than 85% (524, 525, 527), and shown to be equivalent or better than results of similar EBRT treated cohorts. Moving beyond the low risk category, the use of seed BT as the sole treatment modality almost universally shows an increase in the biochemical failure (BF) rates. In these cases, recommendations range from using seed BT in combination with EBRT, in combination with AS, or using EBRT only. No randomised evidence exists to help with this decision, however, there is considerable experience using seed BT in combination with EBRT indicating potential for good outcomes in this group (528). High risk men typically do poorly with seed BT alone and are rarely considered suitable for this therapy.
Following a seed implant, the dose of radiation is delivered over some months, depending on the half-life of the isotope. The predominant toxicity of the implant is that of urinary irritation, and typically the symptoms peak at 2-10 weeks (isotope-related) and the majority will have returned to normal at 12 months (529). These initial symptoms can be troublesome, with most series reporting rises in the International Prostate Symptom Score of 7-12 units at their worst, indicating moderate bother (373, 530, 531). Rectal toxicity is uncommon, with bleeding or fistula formation occurring in <1% (532).
Erectile function maintenance has been reported in a number of prospective brachytherapy studies. The highly conformal dose distribution of seed BT theoretically may reduce the dose of radiation to the structures important for maintaining erectile function, and this seems to be demonstrated in the relatively early figures. The loss of adequate erectile function rates appears to occur in 30-50% of patients by 3 years (533, 534) and longer term results are awaited.
HDR brachytherapy of the prostate has been utilised in a number of centres now for well over a decade (535-541). Proponents of the technique are in favour of the highly conformal nature of the radiation dosimetry able to be achieved with HDR, as well as a theoretical biological advantage in prostate cancer (542). It is performed by placing a series of fine BT catheters in the prostate transperineally, using a technique similar to the guidance of the needles for LDR seed BT. These are subsequently loaded with a radioactive source which sits at various positions in the catheters for various durations to build up a dose of radiation as predetermined by a computer planning system. It is typically used in combination with a truncated course of EBRT, although there are experimental protocols in place for using it as monotherapy.
Results using this technique are encouraging, and would appear to justify the theoretical basis of its use. Series routinely report freedom from biochemical failure results of better than 90% and better than 80% for low and intermediate risk patients respectively (537, 539, 540). Late urinary complications (grade 3) appear to occur in about 5%, and are typically ‘obstructive’ symptoms. Significant rectal toxicity is uncommon; the maintenance of erectile function is poorly reported to date.
The definition of therapeutic failure based on the post-radiotherapy PSA profile (“biochemical failure”) is a more difficult prospect than following radical prostatectomy. After surgery, essentially all the PSA-secreting tissue is removed, and hence an undetectable level is considered desirable. Naturally, this situation does not exist following radiotherapy; with the prostate remaining in situ with variable residual function. Early after the inception of PSA testing, the publication of BF results was based on a variety of interpretations which were felt by individuals to describe a high likelihood of recurrent disease. In response to concerns regarding the potential lack comparability of such published results, an ASTRO consensus panel was convened in 1996 to discuss future BF publication criteria. The panel considered data from a number of clinical databases, all treated with external beam therapy. The subsequently published definition of BF, known as the ASTRO consensus definition (ACD), stipulated BF to have occurred when three consecutive PSA rises are seen and the date of failure to be retrospectively backdated to the midpoint of the PSA nadir and first rise. This was largely based on the finding that three consecutive PSA rises was followed by a further rise in 89% of cases, plus the clinically plausible argument that the recurrence actually commenced well prior to the confirmatory third rise. Relatively little was known at the time regarding the relationship between the ACD BF status and clinical endpoints (543).
The ACD has been widely embraced for the analysis of post-radiotherapy results. A large number of concerns about its performance have been raised however. Although BF would appear to occur rarely after 5 years using the ACD, the reliability of the results are known to be highly dependent on the duration of follow-up. Practically this has shown to potentially worsen the derived freedom from BF (FFBF) rate by approximately 25% when the same data is analysed at a median follow-up of 6 years against that obtained at 2 years (544). This creates difficulty when examining the results of more recent series (for example, a new treatment technique) with historical data; the contemporary data will be favourably biased. This phenomenon is known to be an artefact of the backdating in the ACD. Furthermore, the known independent prognostic value of ACD BF and overall survival (relative risk 1.27) is substantially weakened by backdating (545).
The ACD has also been criticised for having unpredictable results when applied to treatment modalities other than EBRT, for which it was not planned to be utilised. False positive results can be problematic with seed brachytherapy, where benign PSA fluctuations (“bounces”) are seen in approximately one-third of cases (546). Similarly, when the testosterone recovers after combined AS and radiotherapy there may be a concurrent PSA rise of a level sufficient to trigger ACD BF in up to approximately 20% of cases (547, 548). There is poor applicability to surgical data, with backdating particularly creating differences in the late risk of failure (468). A further surgical analysis suggested that the ACD may underestimate the risk of BF by as much as 30%, however this study illustrates the propensity for the ACD to be misinterpreted (549). By stipulating the nadir date to be the date of surgery, rather than the defined last non-rising PSA date, the backdating effect was exaggerated to the theoretical extreme, further reinforcing the finding that the there are multiple areas of potential misinterpretation in the ACD (550), and that original publication guidelines are poorly adhered to (551).
A number of analyses of very large and mature patient cohorts have been published which suggest alternative BF definitions may be superior to the ACD. These have focused on the correlation of PSA dynamics and subsequent clinical events in the aim of finding a BF definition which is both sensitive in detecting recurrence early, yet specific to those factors which are likely to be clinically relevant in the future – data which was unknown at the time of the original consensus meeting.
The most ambitious undertaking was that of Thames et al (2003) (552). This study assimilated PSA and clinical outcomes of 4839 T1-T2 prostate cancer patients from nine USA institutions treated with EBRT without planned AS. A total of 102 different definitions of BF were assessed using a variety of quantitative measures of prediction of clinical failure. A number of definitions were found to have substantially better performance than the ACD. In particular, nadir (the lowest level to date) plus either 2 or 3 ng/mL showed good sensitivity, specificity and predictive indices in relation to clinical failure without resorting to backdating and prospective in nature (that is, it used only the PSA data available prior to failure, as would happen in the clinic). The nadir + 2 ng/mL definition, for example, showed a sensitivity, specificity and relative risk of clinical failure of 0.74, 0.84 and 58.4 respectively, while the ACD had respective values of 0.61, 0.80 and 7.0. Similarly, Kestin et al (2004) (553) used the mature data of 727 men to show the nadir + 2 ng/mL definition had a 73% versus 3% rate of clinical failure at 10 years for those who were deemed as BF or not respectively, while the ACD had 64% against 14%. Further analysis of the sensitivity and specificity characteristics using receiver operating characteristics show that the optimum balance appears with the nadir + 2 ng/ml or an absolute level of 3 ng/ml definitions (554).
Analysis of the interaction of various BF definitions and AS therapy have been performed by Pickles et al (2003) (548). The nadir + 2 ng/mL performed better in terms of predicting clinical failure in their cohort of 1490 patients, and especially so in those previously receiving AS, and this was also confirmed by Zietman et al (2005) (547). It would also appear that the definition of the nadir + 2 ng/mL style will be relatively unlikely to be influenced by PSA bounces related to seed brachytherapy, given the median bounce level is 0.4-0.7 ng/mL (546) and less than 3% are of sufficient level to trigger a false positive BF with this definition (555). In view of the considerable advances in the understanding of PSA biology following radiotherapy, another consensus meeting to discuss an update to the ACD was convened in January 2005. Results from this meeting are expected soon.
A detectable and rising PSA above the nadir value following radiation therapy may represent local or distant failure, or both. Determining the site or sites of tumour, although highly desirable, is notoroiously difficult with currently available diagnostic techniques with BF often preceding clinically evidence of metastases by several years. Consequently, a diagnosis of localized disease without metastases, can be only inferred. However, selecting the most appropriate therapeutic approach may be aided by reference to initial clinical staging, PSA and Gleason score from TRUS biopsies.
In addition to pre-treatment clinical and post-treatment pathological factors, PSA kinetics and, specifically, a pretreatment PSA velocity >2 ng/ml/year, an interval to PSA failure <3 years and a post-treatment PSA doubling time <3 months place a man at increased risk for metastases and subsequent prostate cancer-specific mortality, making these patients poor candidates for local-only salvage therapy. Men with any Gleason 8-10 tumour in their TRUS biopsy cores or those whose cores indicated seminal vesicle involvement are also very likely to have non-localised disease. As a result, Lee and D’Amico (2005) (556) identified an optimal candidate for local-only salvage therapy as one whose pretreatment PSA velocity was <2 ng/ml/year, interval to PSA failure exceeded 3 years, post-treatment PSA doubling time was at least 12 months and whose TRUS biopsy cores contained a Gleason score of less than 8 (556).
Touma et al (2005) (557) reviewed the current status of local salvage therapies following radiation failure for prostate cancer, reporting that salvage RRP offered 5-year biochemical relapse-free rates of 55 to 69%. They identified a life expectancy of at least 10 years, pre-radiation and pre-operative PSA <10 ng/ml and localized preoperative stage as desirable factors for RRP after radiotherapy failure. However, patients need to be aware that higher complication rates are to be expected compared with primary RRP, which include rectal injuries, bladder neck contracture and urinary incontinence (557).
As discussed above in Section VII, the most appropriate application of cryotherapy is for patients with bulky local disease and local recurrence after radiation therapy (315). Touma et al (2005) (557) noted that urinary incontinence, impotence, pelvic pain and urinary retention were the major side effects of salvage cryotherapy cited and considered salvage cryotherapy to be a valid option in hormonally naïve patients with preoperative PSA <10 ng/ml, Gleason score <8 and clinical stage <T3, being especially suitable for older men with some comorbidities but who are still considered to be reasonable anesthetic risks (557).
HIFU (also addressed briefly in Section VII) may have a role in these types of patients but relevant data to make this assertion are not yet available (313). Apart from local treatments, androgen suppression treatments may be considered (detailed in Section XIII, below).
Considerable interest exists regarding the treatment modality in early prostate cancer, and the impact it has on the patient’s quality of life (QOL). Once more, randomised data are lacking, although at least three large prospective series have now been published comparing treatment modalities (413, 558, 559).
EBRT was typically associated with the most bowel dysfunction, which settles over 1-2 years, and then tends to be stable. Acute urinary bother is mild, but significant worsening of symptoms at later stages (beyond 5 years) has been reported and requires further investigation (559). Sexual function consistently deteriorated over time in all series after EBRT, although notably, at the same rate as the untreated controls of one series.
Seed BT has initial elevation of urinary bother that settles to a large degree in the first 2 years. There is however, the suggestion that there may be subsequent deterioration of continence over the next 4 years. This has not been corroborated in other studies to date (560). Rectal morbidity, although typically minor, is documented to steadily settle with time. Erectile function deteriorates in the initial 1-2 years typically, but the longer term results of one series suggest little change in the QOL impact of this beyond that time.
Each of these series had a RRP control arm for comparison, which yielded consistent findings. RRP always led to a high QOL impact in terms of urinary continence and sexual dysfunction within the first 6 months post-therapy. These both improve over the first 2 years, but should be considered likely to remain stable beyond that time. Long-term follow-up comparisons in well designed trials will be required to ultimately decide the overall QOL trade-offs with each particular therapy.