Radiotherapy of oligometastatic prostate cancer: a systematic review

Overall, 56 Studies from 2012 to 2020 were included. Study methods and designs are listed in Table 1. The vast majority of the studies were retrospective case series with median follow-up times between 6 and 70 months. Oligometastasis was inconsistently defined, with three and five metastases as the mostly used cut-off value. The inconsistent definition between the studies reflects the ongoing debate and suggests the difficulty of capturing the oligometastatic state by the sheer number of metastases alone. Of note, even though in most studies a maximum of five metastases was used in the inclusion criteria, the majority of patients had one or two metastases.

Whether the number of metastases also has prognostic value within the collective of oligometastatic patients remains unclear. While some studies - possibly underpowered due to small patient cohorts - could not show any influence, the number of metastases had an impact on the outcome in other studies [6‐12].

Data for local control (LC) and progression free survival (PFS) are shown in Table 1. LC rates ranged between 76 and 100% at 2 years. PFS was inconsistently defined, as biochemical progression, clinical progression or both. The reported PFS values ​​ranged from 38 to 100% at 1 year and 22–83% at 2 years and median PFS rates ranged from 7 to 63 month. The ORIOLE Trial, (a randomized phase II study) compared observation and MDT, and showed a significant difference in the median PFS with MDT (not reached vs. 5.8 months; hazard ratio, 0.30) [13]. Due to the large number of small case series, patient collectives, therapies and predictive factors differed substantially. Hereinafter, some of these factors and their predictive value will be discussed in detail.

The sites of treated metastases in the studies were mostly bone or lymph node. In the present review, twelve, seven and 37 studies with treatment of exclusively nodal metastases, bone metastases or both were included and investigated. In most studies including patients with nodal and bone metastases, the site of metastasis was not a predictive factor for the respective clinical outcomes [10, 12, 14‐25]. In contrast, Fodor et al. reported a higher risk for clinical relapse in patients with extra-pelvic lymph nodes metastases compared with pelvic lymph node lesions and in the studies of Schick et al. and Deek et al. a trend for better biochemical progression-free survival (BPFS) was shown in patients with lymph node metastases compared with those with bone metastases [6, 26, 27]. In addition, the largest study to date based on prospectively collected data based on patients treated on clinical trials, demonstrated that the presence of bone metastases was associated with a worse survival compared to lymph node metastases in MPC [28]. Hence, it is not surprising that in the recently published APCCC report, the majority of experts voted for the distinction of these two kinds of metastatic patterns [29]. However, since encouraging clinical outcomes of studies with exclusively bone metastases were reported, with 2-year LC and PFS rates of 76–100% and 27–38%, respectively, these patients may benefit from MDT and should not be excluded [8, 30‐35].

Due to the lack of predictive biomarkers, the definition of oligometastasis is currently based on the sheer number of metastases as determined by imaging, underscoring the critical importance of reliable imaging. Staging with ⁶⁸Ga-prostate-specific membrane antigen PET/CT (PSMA PET/CT) appears to show the highest detection rates of metastases compared to other imaging modalities till now [36]. High detection rates of 15–58%, 25–73% and 69–100% were reported for PSA ranges of 0.2–0.5 ng/ml, 0.5–1.0 ng/ml and 1–2 ng/ml, respectively [37‐41]. Compared to Choline PET/CT, PSMA PET/CT is substantially more sensitive, especially for low PSA values less than 2 ng/ml [42, 43].

Therefore, due to the lower detection rates in studies that did not use PSMA PET/CT as imaging, many patients may have been yet undiagnosed polymetastatic disease and were consequently understaged [44]. In fact, even staging with PSMA PET/CT cannot exclude this possibility, but it can be assumed that this modality comes closest to defining a “true” oligometastatic state.

Two of the included studies investigated staging with choline or PSMA PET/CT as predictive factor in univariate analysis but failed to detect any impact of imaging on LC, PFS, OS or treatment escalation [19, 22]. However, small case numbers may limit the statistic power to prove a significant difference.

Despite the absence of definitive evidence for superiority of PSMA PET/CT in the oligometastatic setting, there has been a remarkable increase in use of ⁶⁸Ga -PSMA PET/CT imaging in recent years. While 17% of the studies in this review published in 2017 used at least in part PSMA PET/CT, it was 47% and 78% of the studies in 2018 and 2019 [20‐23, 32‐34, 42, 45‐55]. Being in line with these data, the panelists of APCCC recommended PSMA PET/CT to confirm the diagnosis of an oligometastatic disease after radical treatment [5]. A PSA threshold of 0.3 to 0.83 ng/ml appears to be an optimal cut-off value for using PSMA PET/CT as staging [50, 51, 54].

As used in the literature, oligometastasis can be defined to be present if detected either synchronously at the time of diagnosis of the primary tumor or metachronously (at a later date). However, the former scenario was regarded by some experts simply as metastatic disease. Moreover, there is no consensus in literature on the exact interval between diagnosis of the primary tumor and detection of oligometastases to differentiate between metachronous versus occult synchronous disease. Nevertheless, a frequently used definition of metachronous disease is an interval of more than 6 months [56]. Although the vast majority of the studies included patients with recurrent, i.e. “metachronous” disease, the reported intervals between primary diagnosis and detection of metastases were often less than 6 months. These studies had therefore rather mixed populations with metachronous and synchronous metastatic disease.

The parameters “time between primary and detection of oligometastasis” or “time between primary and radiotherapy” were reported in 35 studies with a median time interval between 7 and 67 months (range 0–240 months). Only 13 studies evaluated and reported one of these parameters in univariate or regression analysis, nine of them found no impact on outcome [15, 19, 20, 22, 24, 26, 48, 49, 57, 58]. In contrast, Lépinoy et al. showed that a dichotomous division of patients by interval between primary and oligometastasis of more or less than 5 years was predictive for failure in both univariate and multivariate analyses with better outcome for intervals longer than 5 years [59]. Similarly, Ong et al. reported a better distant progression-free survival with longer time intervals and Kalinauskaite found an improved treatment failure free-survival in patients with time to first metastasis longer than 36 months [12, 23]. In accordance with this data, it seems rational that a longer interval between primary diagnosis and oligometastasis may indicate less aggressive tumor biology. Metachronous disease was also an inclusion criterion for the two major phase II studies STOMP and ORIOLE addressing MDT in metastatic prostate cancer [13, 60].

Since it is widely accepted that hormone-sensitive prostate cancer (HSPC) and castration-resistant prostate cancer (CRPC) are different entities in terms of tumor biology and prognosis, it is consequential that in most studies the status of hormone sensitivity was reported [7, 12, 13, 17, 19, 22, 30, 32, 34, 35, 44, 49, 50, 53, 54, 58, 61‐65]. In the study of Franzese et al. CRPC was an independent risk factor for inferior PFS compared to HSPC in multivariate analysis (HR 2.12, p = 0.021) [19]. This was confirmed by the data reported by Patel et al. (HR 8.43, p < 0.001) [34]. In addition, Guler and Deek reported a significant worse PFS in CRPC patients [27, 50]. The reasons could be the more aggressive tumor biology in CRPC or/and a worse response to MDT. In our opinion, HSPC and CRPC should be considered as two distinct subgroups for further studies of oligometastasis.

Little is reported about the influence of hormone sensitivity on LC rates. Deek et al. found a significant higher local failure rate in CRPC patients compared with HSPC patients and Franzese et al. confirmed CRPC as a predictive factor for worse LC in univariate analysis [19, 27]. However, this effect was no longer detectable in the multivariate analysis. LC rates in the mixed-group studies were similar to those in which only HSPC patients were included. The 2-year local control reported by Triggiani et al. was 92.8% and 90.2% for HSPC and CRPC, respectively, so that it can be concluded that SBRT was able to achieve an excellent LC rate in both CRPC and HSPC oligometastatic patients [18]. This is not surprising given the fact that most studies of RT palliation for bone metastases have reported high response rates [66].

The “standard of care” for MPC has been ADT alone until recently wherein combinations including other systemic agents such as abiraterone or docetaxel have been added [67‐69]. Even more recently local irradiation of prostate has been to standard systemic treatment and shown to improve overall survival for men with de novo metastatic prostate cancer with low metastatic burden [70]. However, some patients refuse systemic treatment primarily due to fears concerning their potential side effects and their comorbidity. Thus, androgen deprivation therapy free survival (ADTFS) was introduced by some authors in HSPC patients and reported in several studies, which ranged between 7 and 66 months [9, 12, 14‐18, 25, 27, 32, 46, 57, 58, 71, 72]. Similarly, in CRPC cohorts, systemic therapy-free survival and treatment escalation-free survival, ranging between 16 and 27 months, were also introduced in the management of prostate cancer and investigated in some studies [22, 24, 47]. In the case of newly developed oligometastasis after the initial metastasis-directed therapy, a second and further SBRT was allowed in some of these studies.

Of particular note is the prospective randomized STOMP study, which showed a prolonged ADTFS with MDT compared to observation after a medial follow-up of 3 years (21 vs. 13 month) [60]. LC and biochemical progression-free time were also improved in the MDT group with comparable quality of life. In the prospective single-arm trial reported by Siva et al., the ADTFS rate was 48% at 2 years. On the other side, there is a clear body of evidence showing improved overall survival with ADT and its combination therapies in metastatic disease [67, 73, 74]. Thus, omitting ADT may be associated with a worse survival while temporarily delaying side effects. This point should also be taken into account in decision-making of treatment and in the context of informed decision making with patients. Indeed 75% of the panelists of APCCC recommended adding MDT to systemic therapies, instead of replacing them [29].

Baumann et al. examined the metabolic response rate in PSMA PET/CT after SBRT of bone metastases with 5 × 7 Gy [45]. 78% of the irradiated lesions showed a metabolic response, which correlated with the time interval between SBRT and the post-therapeutic PSMA PET/CT. The metabolic response rate was 100% when follow-up imaging was performed 5 months or longer after the radiation. Consequently, a time interval of at least 6 months was recommended for the post-therapeutic PSMA PET/CT as response evaluation.

The used fractionation schemes were highly variable (Table 1) ranging from single-dose SBRT with 24 Gy over total doses of 20–50 Gy (or more) in several fractions by moderately hypofractionated or normofractionated schedules. The most common fractionation scheme was 30 Gy in three fractions. Although in general the LC rate was high with acceptable toxicities, the optimal fractionation scheme remains undefined.

Some studies fail to show that radiation dose is predictive of outcome [17, 21, 24‐26, 33, 53, 75], however, Ost et al. found better local PFS in multivariate analysis with a biological effective dose (BED) > 100 Gy, using an α/β value of 3 Gy [16]. This cut-off dose was supported by another study in which a BED > 100 Gy resulted in prolonged systemic treatment-free survival in univariate analysis [24]. In addition, Hurmuz and colleagues showed a better progression free-survival with a BED > 108 Gy [11]. Muldermans et al. reported a higher 2-year LC rate for SBRT with ≥ 18 Gy compared to 16 Gy (95% vs. 58%, p = 0.001). In another study of 40 patients, a median single-fraction dose of 20 Gy was used. Local failure occurred only in two patients who were treated with a reduced SBRT dose, because of prior radiotherapy and/or vicinity to dose-sensitive organs related at risk [7, 31]. Additionally, Schick et al. found a significantly improved BPFS for SBRT with EQD₂ (equivalent dose in 2 Gy fractions) > 64 Gy using an α/β-value of 2 Gy (HR 0.37, p = 0.034) [6]. Although these observations may be considered “first hints” for defining the optimal dose, which should be taken into-account in the designing of further clinical trials, caution must be exercised in assuming that these post hoc studies are definitive due to issues related to major patient selection biases.

Regarding MDT of lymph node metastases, a distinction must be made between SBRT of the affected lymph nodes only and prophylactic elective nodal radiation therapy (ENRT) of the (loco)-regional lymph node station. ENRT usually involves using conventionally fractionated (i.e. 1.8–2.0 Gy) to imaging negative nodes to 45–50 Gy with a boost to the affected (i.e. PET positive) lymph nodes [6, 26, 76]. SBRT of lymph node metastasis was performed in a single fraction or hypofractionated with doses between 24 and 50 Gy in 3–10 fractions. Some studies reported a type of “involved field” irradiation without inclusion of the whole ipsilateral lymphatic drainage [21, 51, 54]. The doses used were 45–60 Gy with a boost up to total doses ranging from 63 to 74 Gy.

In two studies, the authors directly compared SBRT to ENRT plus Boost: Lépinoy et al. compared SBRT of affected lymph nodes mostly using 36 Gy in 5 fractions to conventionally fractionation extended field irradiation of the whole pelvis [59]. The use of ENRT was associated with a significantly longer failure-free time, albeit with a little more acute gastrointestinal toxicity. Their results were confirmed by De Bleser et al., who also reported fewer nodal recurrences and higher late toxicity in the ENRT group [48]. These findings and the pattern of progression described below support the hypothesis that in some cases, despite improved imaging sensitivity, the extent of metastasis, especially the spreading of microscopic cancer cells, is underestimated.

Distant/regional progression-free survival after MDT was 27–45% after 2 years [15, 51, 54, 62, 63]. Of the patients, who relapsed after the initial MDT, 50–91% relapsed again in an oligometastatic pattern (as defined in the initial definition of oligometastasis in each study) [15, 54]. A second, third and fourth course of SBRT was administered in some studies without increased toxicity [12, 15, 51, 54, 58, 62, 63, 65]. In the trials using SBRT for MDT, recurrences occurred mostly in the same organ system, in lymph nodes or bone, respectively [15, 17, 44]. Moreover, Nicosia et al. described that the majority of patients with nodal recurrence after SBRT suffered a lymph node relapse, which was out of but close to the radiation fields [53]. Soldatov et al. reported a shift from iliac lymph node metastases to retroperitoneal lymph node metastases or from retroperitoneal to distant lymph node metastases and bone metastases in patients with oligometastatic lymph nodes treated with ENRT [54]. This might be explained by the coverage of adjacent lymph nodes or elective lymph node stations. Moreover, the radiation dose for elective lymph node stations in the ENRT approach seems to be sufficient to eliminate the microscopic tumor cells, in principle favoring extended irradiation fields in this regard. However, less toxicities and the feasibility of repeated radiotherapy and possibly an enhanced immune response as shown in the ORIOLE trial potentially supporting the rationale for the use of SBRT alone [13].