Radical prostatectomy versus external beam radiotherapy with androgen deprivation therapy for high-risk prostate cancer: a systematic review

Radical prostatectomy (RP) and external beam radiotherapy (EBRT) combined with Androgen deprivation therapy (ADT) are widely used treatment modalities for high-risk localized prostate cancer (PCa). To date there is no consensus on which of both is the optimal treatment for men with high-risk PCa, as high-level evidence is lacking [1]. The only high-quality, well-known randomized controlled trial (RCT) comparing RP with EBRT is the ‘Prostate Testing for Cancer and Treatment’ (ProtecT) trial. The purpose of this trial, in which patients were enrolled between 1999 and 2009, was to compare oncological outcomes and side effects of RP, EBRT and active monitoring for, mainly low-risk localized, PCa detected by PSA screening. Only 2% of men included in the ProtecT trial had high-risk PCa [2, 3]. PCa-specific survival (PCSS) was excellent in all treatment groups (approximately 99% at 10 years) and there was no significant difference in PCa-related deaths per 1000 person-years. With respect to functional outcomes, the greatest negative impact was seen after RP and concerned in particular a decline in sexual function and urinary incontinence. Decreased bowel function and irritative urinary symptoms were more often reported following EBRT, but were usually temporary. A difference in general health-related quality of life (HRQoL) was not observed. The results of ProtecT cannot be generalized to high-risk patients, as treatment for high-risk PCa differs from treatment for low- to intermediate-risk PCa. In the latter group, nerve sparing surgery is often possible while this is generally not the case in high-risk PCa. In addition, in men with low- or intermediate-risk PCa treated with EBRT, no ADT or only short-term ADT is advised, while long-term ADT is recommended in case of high-risk PCa [4].

Two small RCTs compared RP with a radiation-based approach in men with localized-locally advanced PCa [5, 6]. Patients were recruited from 1989–1993 and from 1996–2001. No statistically significant differences in PCSS between both treatment groups were observed, however with fewer than 100 patients enrolled in each study, both studies were underpowered for oncological outcomes. In addition, treatment techniques have evolved, therefore results are not generalizable to contemporary practice. Currently, the ‘Scandinavian Surgery Versus Radiotherapy for Locally Advanced Prostate Cancer’ (SPCG-15) trial is the only randomized study comparing RP and EBRT in men with locally advanced PCa. The study is still recruiting and given disappointing accrual it will be some time before endpoints (including PCSS and HRQoL) will be reported [7, 8].

Thus, randomized studies comparing RP with a radiation-based approach are scarce and the existing trials either enrolled a different patient population or were underpowered and are outdated. Next to these randomized trials, multiple observational studies have been published comparing RP with a radiation-based approach in the treatment of high-risk PCa. We conducted a systematic review to summarize the results of recent evidence in terms of HRQoL, functional and oncological outcomes following RP compared to a radiation-based approach in high-risk PCa. In view of advances in surgical and radiation-based treatment of high-risk PCa, we focused on studies published from 2016 onwards.

For reporting the results of our review, we followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRIMSA) guidelines (supplementary information p. 1–2) [9] Medline, Embase, Cochrane Database of Systematic Reviews, Cochrane Controlled Trial Register and the International Standard Randomized Controlled Trial Number (ISRCTN) registry were systematically searched on 29 March 2021 for studies published from 2016 onwards. The search strategy is provided in the supplementary information (p. 3). Search results were combined and duplicate publications were removed. Comparative studies (RCTs, cohort and case–control studies) reporting on treatment with RP compared to dose-escalated EBRT and ADT for high-risk nonmetastatic PCa were included if at least 100 patients participated in the study. Patient series without comparison groups, editorials, reviews, commentaries, conference abstracts without publications and articles in languages other than Dutch or English were excluded.

The population of interest consisted of patients of any age, diagnosed with de novo high-risk nonmetastatic PCa. High-risk PCa was defined as ≥ cT2c, cN0/1, cM0, ISUP grade 4–5 and/or PSA > 20 ng/ml. This allowed both studies using the European Association of Urology (EAU) risk classification (high-risk: ≥ cT2c, ISUP grade ≥ 4 or PSA > 20 ng/ml) and studies using the National Comprehensive Cancer Network (NCCN) risk classification (high- or very high-risk: ≥ cT2c, ISUP grade ≥ 4 or PSA > 20 ng/ml) to be included. RP could be performed via an open, laparoscopic or robot-assisted surgical approach, as no approach is currently recommended over another [10]. Furthermore, RP could potentially be part of multimodality therapy with adjuvant RT and/or (neo)adjuvant ADT. Dose-escalated EBRT was defined as a biologically effective dose (BED) converted to 2 Gy fractions (EQD2) of at least 74 Gy. In addition, a brachytherapy boost, could be given [10]. There were no requirements with regard to the radiotherapy technique used (e.g. three-dimensional conformal radiotherapy and intensity modulated radiotherapy). In both treatment groups, pelvic lymph node dissection (PLND) could be performed for staging purposes. The primary outcome measures were HRQoL and functional outcomes. Secondary outcome measures included biochemical recurrence-free survival (BCRFS), clinical recurrence-free survival (cRFS), distant metastasis-free survival (DMFS), PCa-specific survival (PCSS) and overall survival (OS).

Title, abstract and full-text screening were performed independently by OLH and BLH. In case of different assessment, consensus was reached by discussion. For all included studies, details on study design, recruitment period, number of included patients, mean or median age, tumor characteristics (e.g. clinical T-stage, Gleason Score and PSA), treatment details (e.g. surgical approach and radiation dose), mean or median follow-up time and primary and secondary outcome measures were extracted by OLH and/or BLH. The Newcastle–Ottawa Scale was used to appraise the quality and risk of bias of included studies [11, 12]. A follow-up period of 3 years was considered sufficient for the primary outcome measures (HRQoL and functional outcomes), but in case only secondary outcome measures were reported, 5 years was considered the minimum acceptable follow-up length. Appraisal was done independently by OLH and BLH and once again disagreement was resolved by discussion. Studies with a total score of ≥ 7 were considered to have low risk of bias while studies with a score of ≤ 6 were considered to be at moderate to high risk of bias.

Curative treatment options currently recommended for localized high-risk PCa include RP, possibly as part of multi-modal therapy, and radiation based treatment combined with ADT [10]. There is substantial variation between individual hospitals in the utilization of both treatment options that is not explained by patient- and tumor characteristics or patient preferences [32]. The lack of high-level comparative evidence, absence of consensus regarding the optimal treatment for patients with high-risk PCa, the fact that neither treatment is recommended over the other in current guidelines and different definitions of high-risk PCa (e.g. EAU versus NCCN) contribute to this unwarranted clinical variation [33]. In this review, we have summarized the existing comparative evidence in terms of HRQoL, functional and oncological outcomes.

Several systematic reviews and meta-analysis have already been published on this topic, based on which treatment with RP appears to be more favorable in terms of OS and PCSS [1, 34‐37]. However, many studies included in these reviews were published in de late 1990’s or early 2000’s and the eligibility criteria used were less stringent (e.g. no requirements were set regarding the radiation dose). Consequently, results have been included from studies in which the treatment(s) used are now considered suboptimal. For example, technological advances in radiation treatment delivery have enabled dose-escalation, which is currently considered the standard-of-care. Dose-escalation and technological advanced are associated with improved BCRFS, DMFS, PCSS, OS and reduced toxicity [38‐42]. Regarding RP, the introduction of the robot-assisted procedure and centralization of care in high-volume hospitals are important developments. Although both developments are associated with improved perioperative outcomes, improvements in functional and oncological outcomes (e.g. DMFS, PCSS and OS) have not been demonstrated [43‐45].

In the majority of studies included in the current review, a significant difference in oncological outcomes between treatment with RP and EBRT combined with ADT was not observed. In addition, five year OS and PCSS were generally good. Therefore, differences in functional outcomes and HRQoL are arguably important. Few studies reported these outcomes after treatment for high-risk PCa with RP compared to EBRT and ADT. Genitourinary toxicity and sexual dysfunction were reported more frequently after RP while gastrointestinal toxicity and reduced hormonal function were more common after EBRT combined with ADT. Results from studies comparing different surgical approaches (e.g. robot-assisted versus open RP), more often included functional outcomes. In studies specifically focusing on, or with a substantial proportion of, patients with high-risk PCa, erectile function recovery at 12–24 months after RARP was reported in 23–60% of patients with no erectile dysfunction at baseline. Erectile function recovery was defined as no or mild erectile dysfunction (International Index of Erectile Function-5 score ≥ 17) or erections sufficient for intercourse [46‐49]. Urinary continence recovery, in most studies defined as the use of 0–1 safety pad per day, was reported in 60.5–95% [46‐48, 50]. In patients with high-risk PCa the additional detrimental effect of adjuvant radiation therapy and/or ADT on functional outcomes should also be considered [51]. In trials comparing different radiation regimens, a cumulative 3- to 5-year incidence of grade ≥ 2 and ≥ 3 GU toxicity of 23–41.3% and 3.5–19% was observed after EBRT, respectively. The reported cumulative 3- to 5-year incidence of grade ≥ 2 and ≥ 3 GI toxicity was 12.2–23.4% and 1.4–3.3%. In addition, Rodda et al. reported a cumulative incidence of any pad use 5 years after treatment of 6.3% and retained or recovered erectile function in 45% of patients with adequate erections before treatment. Either the Radiation Therapy Oncology Group-European Organisation for Research and Treatment of Cancer (RTOG-EORTC) scoring criteria or the Late Effects of Normal Tissue-Somatic, Objective, Management, Analytic (LENT-SOMA) scale were used to score GU- and GI-toxicity and most patients included in these trials received (neo)adjuvant androgen deprivation therapy [52‐54]. Due to the limited number of studies directly comparing functional outcomes and HRQoL after RP versus EBRT combined with ADT and the use of different measurement methods across studies reporting these outcomes after either treatment, the magnitude of the effect of RP versus EBRT and ADT on functional outcomes and HRQoL remains largely unknown. Future research efforts, should focus on the effect of different treatment options on these outcome measures that are highly relevant to patients. In this regard, combination therapy of EBRT and brachytherapy should also be considered, as favorable oncological outcomes of this treatment combination have been reported [28, 55]. However, patients treated with EBRT and a brachytherapy boost were included in only one of the studies that evaluated functional outcomes and HRQoL after RP versus radiation based treatment [23].

Strengths of this review include the specific focus on functional outcomes and HRQoL after treatment for high-risk PCa. These outcome measures are currently under-reported in this patient group, which is confirmed by the current review. Furthermore the search strategy and eligibility criteria were chosen to provide a comprehensive summary of the available studies applicable to current clinical practice. Limitations include the fact that the studies included in the current review are, except for one, retrospective in nature (either using data retrospectively collected from medical records or using data from existing databases). In addition, the majority of studies were conducted at a single-institution and in many studies there were potential quality issues in the assessment of outcome measures. Although statistical methods were applied to control for potential confounders in most studies, residual and/or unmeasured confounding remains an issue. For example, patients with a better performance status and fewer comorbidities are more likely to be considered eligible for RP, which is supported by the generally younger age of surgically treated patients. Furthermore, inclusion criteria, definitions of high-risk PCa, applied surgical and radiotherapy techniques and use of adjuvant therapies varied within and across studies. Differences in methodology, outcome measures, and the information that was reported further contributed to the heterogeneity of data, precluding meaningful quantitative synthesis and preventing definitive conclusions regarding the optimal treatment for men with high-risk PCa.

High-level comparative evidence regarding surgery versus radiation-based treatment for high-risk PCa is lacking. Multiple, primarily retrospective, observational studies comparing RP with dose-escalated EBRT and ADT in this patient population have been published. In the majority of studies, no significant differences in oncological outcomes (e.g. DMFS, PCSS and OS) between treatment with RP and EBRT combined with ADT were observed. Studies reporting functional outcomes and HRQoL are very scarce and the magnitude of the effect of RP versus dose-escalated EBRT with ADT on HRQoL and functional outcomes remains largely unknown. Underlining the necessity for RCTs or well-designed observational studies investigating differences in functional outcomes, HRQoL and to a lesser extent oncological outcomes in the high-risk PCa population.