Stereotactic body radiotherapy as metastasis-directed therapy in oligometastatic prostate cancer: a systematic review and meta-analysis of randomized controlled trials Authors: Astrid E. Persson, Andreas Hallqvist, Louise Bjørn Larsen, Mette Rasmussen, Jonas Scherman, Per Nilsson, Hanne Tønnesen, Adalsteinn Gunnlaugsson Published: 2024 ABSTRACT BackgroundThe use of stereotactic body radiotherapy (SBRT) to definitively treat oligometastases in prostate cancer has drawn large clinical and research interests within radiation oncology. However, the evidence is considered in its early stages and there is currently no systematic review of randomized controlled trials (RCTs) in this field. We aimed to evaluate the efficacy and safety of SBRT as metastasis-directed therapy (MDT) in oligometastatic prostate cancer (OMPC) compared to no MDT reported in RCTs.MethodsMEDLINE, Embase, CINAHL Complete, and Cochrane Library were searched on October 28, 2023. Eligible studies were RCTs comparing SBRT as MDT with no MDT in extracranial OMPC, without restrictions on follow-up time, publication status, language, or year. Participant subsets fulfilling the eligibility criteria were included. Critical outcomes were overall survival and grade ≥ 3 toxicity, and additional important outcomes were progression-free survival (PFS), local control, grade 5 toxicity, health-related quality of life, and systemic therapy-free survival. Meta-analyses were planned. Risk of bias was assessed using the Cochrane risk-of-bias tool version 2, and the quality of evidence using the Grading of Recommendations Assessment, Development, and Evaluation.ResultsIn total, 1825 unique study reports were identified and seven phase II RCTs with 559 eligible participants were included. Four trials included multiple types of primary cancer. Outcome definitions were heterogeneous except for overall survival and toxicity. For overall survival, only one study reported events in both arms. Meta-analysis of the grade ≥ 3 toxicity results from two trials showed no difference (pooled risk ratio 0.78, 95% confidence interval 0.37–1.65, p = 0.52). Four trials reported significantly longer PFS, with a pooled hazard ratio of 0.31 (95% confidence interval 0.21–0.45, p < 0.00001). Risk of bias was of some concerns or high. Quality of evidence was low or moderate.ConclusionsPhase II trials have shown promising improvements in PFS for several OMPC states without excess toxicity. Overall survival comparisons are immature. In future confirmatory phase III trials, adequately large sample sizes, blinding of outcome assessors, and/or increased adherence to assigned intervention could improve the quality of evidence.PROSPERO registration number: CRD42021230131.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13014-024-02559-7. Background Comprehensive treatment of a few metastases, or oligometastases, using metastasis-directed therapy (MDT) has several theoretical benefits. It may prevent or delay metastasis-related complications, postpone the need for systemic therapy or a change of therapy by treating isolated nonresponding metastases [1], interfere with metastatic seeding [2], and be curative if all visible lesions represent the total tumor burden [3]. Stereotactic body radiotherapy (SBRT) is an appealing MDT by being noninvasive and delivered on an outpatient basis [4], and by potentially enhancing the anticancer immune response [5]. SBRT has shown encouraging safety and early efficacy across multiple primary cancer types, with one-year local control rates at approximately 95% [6]. There have been significant efforts in the radiation oncology community during the last years in conducting clinical trials on SBRT as MDT [7], as well as in harmonizing the nomenclature and differentiating clinical scenarios where oligometastatic disease (OMD) is encountered. In 2019, the International Commission on Radiation Units and Measurements (ICRU) published its report [8] on stereotactic radiotherapy (RT). In 2020, the European Society for Radiotherapy and Oncology (ESTRO)–European Organization for Research and Treatment of Cancer (EORTC) consensus recommendation [9] charted OMD states and was shortly followed by the ESTRO–American Society for Radiation Oncology (ASTRO) consensus document [10] specifying prioritized outcomes applicable to SBRT as MDT. Parallel to these efforts, understanding if SBRT as MDT is beneficial in a particular primary cancer type, and if so, determining its place in the current treatment armamentarium, are essential [11]. Prostate cancer (PCa) is one of the most studied primary cancer types [6, 7]. MDT with SBRT is increasingly offered for oligometastatic prostate cancer (OMPC) in routine clinical practice [12, 13], although the evidence is considered weak so far [14]. In response, ESTRO recently published recommendations [15] on patient selection, imaging, and SBRT delivery for clinical practice and acknowledged the lack of level I evidence. Several systematic reviews have covered MDT in OMPC, including both retrospective and prospective [16–23] and exclusively prospective [24–27] studies. Among the systematic reviews of prospective studies, some covered multiple types of MDT (i.e., also surgery) [24, 26] and SBRT alone [25, 27]. However, to our knowledge, there is not yet a systematic review on SBRT in OMPC limited to randomized controlled trials (RCTs), which provides the highest level of evidence [28]. Therefore, we conducted a systematic review of RCTs investigating the efficacy and safety of SBRT as MDT in OMPC. The main hypothesis was that this intervention would improve overall survival without increasing grade ≥ 3 toxicity compared to a control group (CG) receiving standard of care or no treatment. Methods The protocol was developed according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [29] and pre-registered in PROSPERO (CRD42021230131) [30]. The following report is written according to the PRISMA 2009 guidelines [31]. Eligibility criteria We included RCTs that investigated SBRT as MDT for extracranial OMPC. The participants were adults with OMPC diagnosed through biopsy of the primary tumor or metastasis and positron emission tomography (PET)-computed tomography (CT) or whole body magnetic resonance imaging (MRI) [32], or according to the study authors. Studies that included only patients with regional nodal metastases were ineligible. However, participant subsets fulfilling the eligibility criteria were included, such as those from a study that investigated several primary cancer types. All metastases, or all uncontrolled metastases as a minimum in oligoprogressive disease, had to be treated with SBRT without other concomitant MDTs in the intervention group (IG). SBRT could be referred to using related terms (e.g., stereotactic ablative RT or radiosurgery), with the reservation of dose per fraction being > 2 Gy according to the ICRU definition [8], and could be one of several investigated treatments. The CG was receiving standard of care (SOC), placebo, no treatment, or RT with lower intensity than the IG as part of the study design. Lower intensity RT was defined as standard palliative RT or equivalent dose in 2-Gy fractions (EQD2) of up to 50 Gy (α/β = 3 Gy) to any metastasis. Higher doses were allowed at the physicians’ discretion during follow-up. Each study report had to report at least one of the review outcomes (see Outcomes). There were no restrictions on follow-up time and publication status, language, or year of publication. Original study reports and other publication formats (e.g., editorials, comments, and letters) were eligible. Studies with insufficient data for qualitative or quantitative analysis were excluded. Outcomes The review outcomes were clinical outcomes from the ESTRO–ASTRO consensus document on metastasis-directed RT in OMD [10] and were grouped into critical, additional important, and other relevant outcomes. Critical outcomes were: overall survival, defined as time from randomization to death from any cause, or according to the study authors; and incidence proportion of grade ≥ 3 toxicity, defined as the proportion of participants with at least one event of grade ≥ 3 on the Radiation Therapy Oncology Group/EORTC toxicity scale [33] or Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 [34] after start of SBRT, or according to the study authors, at 5 years or longest follow-up. Additional important outcomes were progression-free survival (PFS), defined as time from randomization to disease progression according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (i.e., ≥ 20% increase in the lesions’ baseline sum diameter with a minimum absolute increase of 5 mm for measurable disease and unequivocal progression or progression warranting a change in therapy for nonmeasurable disease) [35], or according to the study authors; local control, defined as the proportion of participants without locally progressive disease according to RECIST 1.1 or according to the study authors, at 5 years or longest follow-up; incidence proportion of grade 5 toxicity, or lethality ascribed to treatment, after start of SBRT or as defined by the study authors, at 5 years or longest follow-up; health-related quality of life (HRQoL), assessed by the EuroQol Group’s EQ-5D-5L index value [36], 36-Item Short Form Health Survey (SF-36) physical or mental component summary [37], EORTC Quality of Life Questionnaire-Core 30 (QLQ-C30) version 3 summary score [38, 39] with or without the complementary module Quality of Life Questionnaire-Prostate 25 (QLQ-PR25) [40], Functional Assessment of Cancer Therapy-General (FACT-G) version 4 total score [41], Functional Assessment of Cancer Therapy-Prostate (FACT-P) version 4 total score [42], or according to the study authors, at 3 months; HRQoL at 5 years or longest follow-up; and systemic therapy-free survival, defined as time from randomization to initiation of any systemic anticancer therapy or death from any cause, or according to the study authors. Five-year follow-up was used if reported for the outcomes defined at these timepoints; otherwise, the longest follow-up was reported. Timepoints of 1–6 months were eligible for HRQoL at 3 months. Other relevant outcomes were local PFS, defined as time from randomization to progression of metastases present at baseline according to RECIST 1.1 or according to the study authors; and distant PFS, defined as time from randomization to occurrence of a new metastatic lesion according to RECIST 1.1 or according to the study authors. Search strategy MEDLINE (through PubMed), Embase (through Embase.com), Cumulative Index to Nursing and Allied Health Literature (CINAHL) Complete (through EBSCOhost), and Cochrane Central Register of Controlled Trials (CENTRAL) were searched for study reports. Additionally, the ClinicalTrials.gov and International Clinical Trials Registry Platform (ICTRP) were searched for protocols to identify ongoing studies. Other sources were also screened for study reports: reference lists of relevant systematic reviews, abstract collections, backward and forward citations searching on included study reports using Web of Science, authors of the included studies, and selected vendor companies. Relevant systematic reviews were identified by applying systematic review filters in the database searches, except for CENTRAL where Cochrane Database of Systematic Reviews was searched instead. The first searches covered the period from inception until January 24, 2021, and were rerun on October 28, 2023, when also previously identified ongoing studies were checked for completion. The full search strategy is provided in Additional file 1. Study selection The identified study reports were imported into EndNote [43] where duplicates were removed according to the deduplication strategy described in Additional file 2. AEP and AG independently screened unblinded titles and/or abstracts in duplicate in Covidence [44] for potentially eligible study reports. Disagreements were resolved through discussion; otherwise, HT made the final decision. AEP screened other sources. Study reports were collated at the study level, and a main report was chosen. AH and LBL independently assessed unblinded full texts in duplicate for inclusion and decided the main reason for exclusion based on the order of importance listed in Additional file 3. Disagreements were resolved by discussion; otherwise, AG was consulted and could make the final decision. Study reports and duplicates could be included or excluded until manuscript submission. The study authors were contacted for clarification of the eligibility criteria, if needed. If no clarification could be obtained, the main report was used in case of conflicting information and if information was missing, the study was excluded. Ongoing and seemingly eligible studies without any reported outcome data were to be described in separate tables. Data collection Data were collected from unblinded study reports using Google Forms [45]. We used full reports, their supplements, and abstracts, as well as protocols when relevant for study definitions, and examined the latest protocol in the case of multiple versions. If results were reported only in figures, data were extracted using Engauge Digitizer [46]. AEP collected data on study characteristics, while AH and LBL collected outcome data using forms piloted on one included study. Discrepancies were resolved by discussion; otherwise, AG was consulted and could make the final decision. The study authors were contacted for missing or additional data. The critical and important outcomes were assessed for risk of bias using the Cochrane risk-of-bias tool version 2 (RoB 2) [47] that considers five domains: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of reported result, on the study and outcome level. AH and LBL independently performed the unblinded assessments in duplicate. Disagreements were resolved by discussion; otherwise, with AEP, HT, and AG until consensus was reached. The assessments were reported in a traffic light plot according to the Robvis tool [48]. The quality of evidence for the critical and important outcomes was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) [49] by applying the same process described above for risk of bias. The assessments were to be reported in the Summary of findings table generated in GRADEpro Guideline Development Tool [50]. Data items We extracted data on publication details (e.g., publication year and corresponding author), study details (e.g., design and phase, enrollment regions and country/countries, enrollment period, screening procedures, follow-up period and time, funding, and inclusion and exclusion criteria), diagnostic details (e.g., imaging modalities and OMD criteria), participant details (e.g., age, performance status, baseline prostate-specific antigen [PSA], controlled primary tumor, states of OMD, previous use and type of systemic treatment, previous use and type of MDT, and sites and number of treated metastases), intervention details (e.g., content, timing, methods of delivery, doses, number of and time between treatments, length of intervention, and intervention integrity), and outcomes. During the review process, we added data on randomization procedures and ratio, follow-up procedures, castration-resistant disease, PSA doubling time, and number of progressing metastases in the CG in case of oligoprogressive disease. Data analysis The hazard ratio (HR) was the effect measure for time-to-event outcomes (overall survival, PFS, systemic therapy-free survival, and local and distant PFS), the risk ratio (RR) for dichotomous outcomes (toxicity and local control), and the mean difference or standardized mean difference for continuous outcomes (HRQoL). Outcome data reported as intention-to-treat were used. Overall survival and incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up were going to be analyzed in meta-analysis using Review Manager 5.4.1 [51] and presented in forest plots. The random-effects model was chosen because of the expected large clinical heterogeneity among studies. If a study reported no events in neither the IG nor the CG, it was excluded from the meta-analysis. A meta-analysis would not be performed if the clinical or statistical heterogeneity were too high. For overall survival, a pooled HR with a 95% confidence interval (CI) was going to be estimated using the generic inverse-variance method based on the natural logarithm (ln) of the HR and the variance of the ln(HR) [52]. In case these statistics were not reported, they were estimated using direct or indirect methods described by Tierney et al. [53]. For incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up, a pooled RR with a 95% CI was going to be estimated using the DerSimonian and Laird inverse-variance method. The pooled effect estimates were to be tested for an overall effect using the Z-test with a significance level of 0.05. Statistical heterogeneity was to be assessed using the Chi2 test with a significance level of 0.10. The proportion of observed variation in effects between studies reflecting real statistical heterogeneity was to be estimated using I2 with a 95% CI and quantified using Tau2, the variance in true effects between studies, with a 95% CI. I2 of ≥ 50% was considered to represent substantial heterogeneity. Prediction intervals were to be calculated in case ≥ 10 studies were included in a meta-analysis [52]. Sensitivity analyses excluding studies with high risk of bias were planned. Formal assessment of publication bias was planned if ≥ 10 studies were included in the meta-analyses by generating funnel plots and testing for asymmetry using Egger’s test with a significance level of 0.10 [54]. All outcome data were reported narratively and in tables. The additional important outcomes were also to be reported using pooled effect estimates. Results Study selection We identified 5286 records from the database searches. After removing duplicates, protocols, systematic reviews, abstract collections, and book chapters, 1815 titles and/or abstracts were screened. After adding 10 records identified from other sources, 39 full texts were assessed for inclusion. Seven RCTs reported in 17 study reports were included. Three of the reports were identified from other sources: two by personal knowledge of the review authors and one by tracking a later report of a pilot study. The flow diagram of the study selection process is shown in Fig. 1.Fig. 1Modified PRISMA 2009 flow diagram. Flow diagram of study reports during study selection. CENTRAL Cochrane Central Register of Controlled Trials, CINAHL Cumulative Index to Nursing and Allied Health Literature, ICTRP International Clinical Trials Registry Platform. aPotentially eligible study reports from reference lists of relevant systematic reviews, abstract collections, backward and forward citations searching on included study reports, authors of included studies, and selected vendor companies bTwelve additional protocols were identified from other sources cReported in 17 study reports Modified PRISMA 2009 flow diagram. Flow diagram of study reports during study selection. CENTRAL Cochrane Central Register of Controlled Trials, CINAHL Cumulative Index to Nursing and Allied Health Literature, ICTRP International Clinical Trials Registry Platform. aPotentially eligible study reports from reference lists of relevant systematic reviews, abstract collections, backward and forward citations searching on included study reports, authors of included studies, and selected vendor companies bTwelve additional protocols were identified from other sources cReported in 17 study reports Studies excluded during full-text assessment are shown in Table 1. Eighteen relevant ongoing studies were identified and are described in Additional file 4. Table 1Characteristics of excluded studies after full-text assessmentAuthor (Year)Reason for exclusionBazyar (2023)a [55]Ineligible outcomesBowden (2019) [56]Ineligible study designBroomfield (2014) [57]Not informing on original researchDe Bleser (2020)b [58]Ineligible outcomesDunst (2018) [59]Not informing on original researchDunst (2021) [60]Not informing on original researchGupta (2019) [61]Not informing on original researchHemmatazad (2021) [62]Not informing on original researchHermann (2021) [63]Not informing on original researchHorjeti (2023)a [64]Ineligible outcomesHrinivich (2019)a [65]Ineligible study designKumar (2021)d [66]Ineligible outcomesNguyen (2019)c [67]Ineligible participantsQu (2020)d [68]Ineligible outcomesQu (2021)d [69]Ineligible outcomesRaymakers (2021)d [70]Ineligible outcomesRyu (2019) [71]Ineligible participantsSiva (2020)c [72]Ineligible study designSprave (2018)c [73]Ineligible participantsSridharan (2016)e [74]Ineligible study designvan de Ven (2020) [75]Ineligible study designZelefsky (2021) [76]Ineligible control groupaInclude data from ORIOLEbInclude data from STOMPcInformation for eligibility assessment provided by personal correspondence with study authorsdInclude data from SABR-COMETeInclude data from TROG 03.04 RADAR Table 1Characteristics of excluded studies after full-text assessmentAuthor (Year)Reason for exclusionBazyar (2023)a [55]Ineligible outcomesBowden (2019) [56]Ineligible study designBroomfield (2014) [57]Not informing on original researchDe Bleser (2020)b [58]Ineligible outcomesDunst (2018) [59]Not informing on original researchDunst (2021) [60]Not informing on original researchGupta (2019) [61]Not informing on original researchHemmatazad (2021) [62]Not informing on original researchHermann (2021) [63]Not informing on original researchHorjeti (2023)a [64]Ineligible outcomesHrinivich (2019)a [65]Ineligible study designKumar (2021)d [66]Ineligible outcomesNguyen (2019)c [67]Ineligible participantsQu (2020)d [68]Ineligible outcomesQu (2021)d [69]Ineligible outcomesRaymakers (2021)d [70]Ineligible outcomesRyu (2019) [71]Ineligible participantsSiva (2020)c [72]Ineligible study designSprave (2018)c [73]Ineligible participantsSridharan (2016)e [74]Ineligible study designvan de Ven (2020) [75]Ineligible study designZelefsky (2021) [76]Ineligible control groupaInclude data from ORIOLEbInclude data from STOMPcInformation for eligibility assessment provided by personal correspondence with study authorsdInclude data from SABR-COMETeInclude data from TROG 03.04 RADAR Characteristics of excluded studies after full-text assessment aInclude data from ORIOLE bInclude data from STOMP cInformation for eligibility assessment provided by personal correspondence with study authors dInclude data from SABR-COMET eInclude data from TROG 03.04 RADAR Study characteristics The included trials were STOMP [77–79], SABR-COMET [80–83], ORIOLE [79, 84], ARTO [85–90], EXTEND [91], CORE [92], and STOP [93], which were published between 2017 and 2023. The trials included 559 participants with OMPC fulfilling our eligibility criteria: 200 in the IGs, 179 in the CGs, and an additional 180 participants in both groups in CORE. All were phase II multicenter trials conducted in high-income Western countries. Their characteristics are listed in Table 2 and further details are provided in Additional file 5. After contacting the authors of all included studies, we were able to receive supplementary data from the ARTO team. Table 2Characteristics of included studiesStudyAuthor (Year)Countries№ participants IG vs CG№ centersFollow-up time in months; medianImaging modalitiesAge in years; medianMetastases№ (Categories)Controlled primary tumorCastration-resistant diseaseState(s) of OMDaInterventionTotal radiation dose (Fractionation)ControlSTOMP [77–79]Ost (2017, 2020),Deek (2022)Belgium25 (subsetb) vs 316NICholine PET-CTNI1–3(Node, bone, or both)YesNoMetachronous oligorecurrenceSBRT30 Gy(10 Gy × 3)EQD23 78 GySurveillanceSABR-COMET [80, 83]Palma (2019, 2020),Olson (2019), Harrow (2022)Canada, the Netherlands, UK, Australia14 vs 2(both subsetsc)10NICT and bone scan or PET-CTdSpine MRI if vertebral metastasesNI1–5; ≤ 3 per organ(Node, bone, lung, liver, adrenal, pararenal)YesMixedPAnyPSBRT added to SOC16–60 Gy(5–20 Gy × 1–12)PEQD23 range 61–227 GySOCORIOLE [79, 84]Phillips (2020),Deek (2022)USA36 vs 18318.8(range 5.8–35.0)CT, MRI, and/or bone scanIG: 68(range 61–70)CG: 68(64–76)1–3(Node, bone, or both)NINoMetachronous oligorecurrenceSBRT19.5–48.0 Gy(5–12 Gy × 3–5)EQD23 37–144 GyObservationARTO [85–90]Francolini (2020, 2021, 2022a, 2022b, 2023a, 2023b)Italy75 vs 821624.9(IQR 17.1–35.8)CT and/or bone scan or choline, fluciclovine, or PSMA PET-CTIG: 74(IQR 68–79)CG: 74(68–79)1–3(Node, bone, or both)9% withouteYesAny with castration-resistanceSBRT16–40 Gy(6.5–16 Gy × 1–5)EQD23 61–108 GyAbiraterone acetate + prednisone, ADTAbiraterone acetate + prednisone, ADTEXTEND [91]Tang (2023)USA43 vs 44322.0(range 11.6–39.2)CT and bone scan or fluciclovine PET-CTIG: 67(IQR 63–72)CG:67(63–72)1–5 metastases(Node, bone ± node, other sites ± bone, node)YesfMixedIG: 9% CG: 7%AnySBRT (MDTg)Recommended, 12–70 Gy(2.3–27 Gy × 1–28)PEQD23 17–162 GyIntermittent hormone therapyIntermittent hormone therapyCORE [92, 115]Khoo (2023)UK, AustraliaTotal 180 (subsets in both armsh)30NICT and bone scan or choline/PSMA PET-CT or WBMRIPNI1–3 in 1–2 organsPYesPMixedPMetachronous oligorecurrenceSBRT added to SOCRecommended, 24–60 Gy(7.5–18 Gy × 3–8)PEQD23 53–227 GySOCSTOP [93, 116]Schellenberg (2023)Canada7 vs 2 (both subsetse,i)8NININI1–5 progressing; ≤ 3 per organPNI33% hormone-sensitiveeOligoprogressionSBRT to all progressing metastases added to SOCSOCReported in table as described by the study authors and denoted by superscript P if only reported in protocol. Equivalent dose in 2-Gy fractions (EQD2) for α/β = 3 GyADT androgen deprivation therapy, CG control group, CT computed tomography, EORTC European Organization for Research and Treatment of Cancer, ESTRO European Society for Radiation Oncology, Gy Gray, IG intervention group, IQR interquartile range, MDT metastasis-directed therapy, MRI magnetic resonance imaging, NI no information, OMD oligometastatic disease, PET positron emission tomography, PSMA prostate-specific membrane antigen, SBRT stereotactic body radiotherapy, SOC standard of care, vs versus, WBMRI whole body magnetic resonance imagingaAccording to ESTRO–EORTC consensus recommendationbSix additional participants in the IG were treated with surgery only as decided by the multidisciplinary teamcEighty-three additional participants with other primary cancer types were included in the studydPET-CT was required for solitary pulmonary nodules not pathologically verified, or could be done as decided by the treating oncologistseInformation provided by personal correspondence with study authorsfParticipants with untreated primary tumors received prostate radiotherapy on trialgSome participants may have received conventional fractionation only, e.g., due to regional nodal diseasehSixty-five additional participants with other primary cancer types were included in the studyiEighty-one additional participants with other primary cancer types were included in the study Table 2Characteristics of included studiesStudyAuthor (Year)Countries№ participants IG vs CG№ centersFollow-up time in months; medianImaging modalitiesAge in years; medianMetastases№ (Categories)Controlled primary tumorCastration-resistant diseaseState(s) of OMDaInterventionTotal radiation dose (Fractionation)ControlSTOMP [77–79]Ost (2017, 2020),Deek (2022)Belgium25 (subsetb) vs 316NICholine PET-CTNI1–3(Node, bone, or both)YesNoMetachronous oligorecurrenceSBRT30 Gy(10 Gy × 3)EQD23 78 GySurveillanceSABR-COMET [80, 83]Palma (2019, 2020),Olson (2019), Harrow (2022)Canada, the Netherlands, UK, Australia14 vs 2(both subsetsc)10NICT and bone scan or PET-CTdSpine MRI if vertebral metastasesNI1–5; ≤ 3 per organ(Node, bone, lung, liver, adrenal, pararenal)YesMixedPAnyPSBRT added to SOC16–60 Gy(5–20 Gy × 1–12)PEQD23 range 61–227 GySOCORIOLE [79, 84]Phillips (2020),Deek (2022)USA36 vs 18318.8(range 5.8–35.0)CT, MRI, and/or bone scanIG: 68(range 61–70)CG: 68(64–76)1–3(Node, bone, or both)NINoMetachronous oligorecurrenceSBRT19.5–48.0 Gy(5–12 Gy × 3–5)EQD23 37–144 GyObservationARTO [85–90]Francolini (2020, 2021, 2022a, 2022b, 2023a, 2023b)Italy75 vs 821624.9(IQR 17.1–35.8)CT and/or bone scan or choline, fluciclovine, or PSMA PET-CTIG: 74(IQR 68–79)CG: 74(68–79)1–3(Node, bone, or both)9% withouteYesAny with castration-resistanceSBRT16–40 Gy(6.5–16 Gy × 1–5)EQD23 61–108 GyAbiraterone acetate + prednisone, ADTAbiraterone acetate + prednisone, ADTEXTEND [91]Tang (2023)USA43 vs 44322.0(range 11.6–39.2)CT and bone scan or fluciclovine PET-CTIG: 67(IQR 63–72)CG:67(63–72)1–5 metastases(Node, bone ± node, other sites ± bone, node)YesfMixedIG: 9% CG: 7%AnySBRT (MDTg)Recommended, 12–70 Gy(2.3–27 Gy × 1–28)PEQD23 17–162 GyIntermittent hormone therapyIntermittent hormone therapyCORE [92, 115]Khoo (2023)UK, AustraliaTotal 180 (subsets in both armsh)30NICT and bone scan or choline/PSMA PET-CT or WBMRIPNI1–3 in 1–2 organsPYesPMixedPMetachronous oligorecurrenceSBRT added to SOCRecommended, 24–60 Gy(7.5–18 Gy × 3–8)PEQD23 53–227 GySOCSTOP [93, 116]Schellenberg (2023)Canada7 vs 2 (both subsetse,i)8NININI1–5 progressing; ≤ 3 per organPNI33% hormone-sensitiveeOligoprogressionSBRT to all progressing metastases added to SOCSOCReported in table as described by the study authors and denoted by superscript P if only reported in protocol. Equivalent dose in 2-Gy fractions (EQD2) for α/β = 3 GyADT androgen deprivation therapy, CG control group, CT computed tomography, EORTC European Organization for Research and Treatment of Cancer, ESTRO European Society for Radiation Oncology, Gy Gray, IG intervention group, IQR interquartile range, MDT metastasis-directed therapy, MRI magnetic resonance imaging, NI no information, OMD oligometastatic disease, PET positron emission tomography, PSMA prostate-specific membrane antigen, SBRT stereotactic body radiotherapy, SOC standard of care, vs versus, WBMRI whole body magnetic resonance imagingaAccording to ESTRO–EORTC consensus recommendationbSix additional participants in the IG were treated with surgery only as decided by the multidisciplinary teamcEighty-three additional participants with other primary cancer types were included in the studydPET-CT was required for solitary pulmonary nodules not pathologically verified, or could be done as decided by the treating oncologistseInformation provided by personal correspondence with study authorsfParticipants with untreated primary tumors received prostate radiotherapy on trialgSome participants may have received conventional fractionation only, e.g., due to regional nodal diseasehSixty-five additional participants with other primary cancer types were included in the studyiEighty-one additional participants with other primary cancer types were included in the study Characteristics of included studies STOMP [77–79] Ost (2017, 2020), Deek (2022) Belgium 1–3 (Node, bone, or both) SBRT 30 Gy (10 Gy × 3) EQD23 78 Gy SABR-COMET [80, 83] Palma (2019, 2020), Olson (2019), Harrow (2022) Canada, the Netherlands, UK, Australia 14 vs 2 (both subsetsc) CT and bone scan or PET-CTd Spine MRI if vertebral metastases 1–5; ≤ 3 per organ (Node, bone, lung, liver, adrenal, pararenal) SBRT added to SOC 16–60 Gy (5–20 Gy × 1–12)P EQD23 range 61–227 Gy ORIOLE [79, 84] Phillips (2020), Deek (2022) USA 18.8 (range 5.8–35.0) IG: 68 (range 61–70) CG: 68 (64–76) 1–3 (Node, bone, or both) SBRT 19.5–48.0 Gy (5–12 Gy × 3–5) EQD23 37–144 Gy ARTO [85–90] Francolini (2020, 2021, 2022a, 2022b, 2023a, 2023b) Italy 24.9 (IQR 17.1–35.8) IG: 74 (IQR 68–79) CG: 74 (68–79) 1–3 (Node, bone, or both) SBRT 16–40 Gy (6.5–16 Gy × 1–5) EQD23 61–108 Gy Abiraterone acetate + prednisone, ADT EXTEND [91] Tang (2023) USA 22.0 (range 11.6–39.2) IG: 67 (IQR 63–72) CG: 67 (63–72) 1–5 metastases (Node, bone ± node, other sites ± bone, node) Mixed IG: 9% CG: 7% SBRT (MDTg) Recommended, 12–70 Gy (2.3–27 Gy × 1–28)P EQD23 17–162 Gy Intermittent hormone therapy CORE [92, 115] Khoo (2023) UK, Australia SBRT added to SOC Recommended, 24–60 Gy (7.5–18 Gy × 3–8)P EQD23 53–227 Gy STOP [93, 116] Schellenberg (2023) Canada Reported in table as described by the study authors and denoted by superscript P if only reported in protocol. Equivalent dose in 2-Gy fractions (EQD2) for α/β = 3 Gy ADT androgen deprivation therapy, CG control group, CT computed tomography, EORTC European Organization for Research and Treatment of Cancer, ESTRO European Society for Radiation Oncology, Gy Gray, IG intervention group, IQR interquartile range, MDT metastasis-directed therapy, MRI magnetic resonance imaging, NI no information, OMD oligometastatic disease, PET positron emission tomography, PSMA prostate-specific membrane antigen, SBRT stereotactic body radiotherapy, SOC standard of care, vs versus, WBMRI whole body magnetic resonance imaging aAccording to ESTRO–EORTC consensus recommendation bSix additional participants in the IG were treated with surgery only as decided by the multidisciplinary team cEighty-three additional participants with other primary cancer types were included in the study dPET-CT was required for solitary pulmonary nodules not pathologically verified, or could be done as decided by the treating oncologists eInformation provided by personal correspondence with study authors fParticipants with untreated primary tumors received prostate radiotherapy on trial gSome participants may have received conventional fractionation only, e.g., due to regional nodal disease hSixty-five additional participants with other primary cancer types were included in the study iEighty-one additional participants with other primary cancer types were included in the study STOMP, ORIOLE, and ARTO included only participants with PCa, whereas SABR-COMET, EXTEND, CORE, and STOP included several primary cancer types. In the included EXTEND study report, the PCa basket receiving intermittent hormone therapy was reported separately. In STOMP, ORIOLE, and CORE, only patients with metachronous oligorecurrence were included, and in STOP, patients with oligoprogression. OMD was defined as 1–5 (progressing) metastases detected using CT and/or bone scan, MRI, or PET-CT with choline, fluciclovine, or PSMA tracers, with some variations in restrictions on metastatic sites and the number of metastases per organ. It was uncertain if the participants had brain metastases in SABR-COMET, EXTEND, and STOP, but it was considered unlikely per personal correspondence with the study author for SABR-COMET and by the review authors for EXTEND and STOP. Only participants with castration-resistant disease were studied in ARTO and were eligible for enrollment in SABR-COMET, EXTEND, CORE, and STOP. STOMP investigated either SBRT or surgery as MDT, while the remaining trials investigated SBRT. There was some uncertainty as to whether some participants in EXTEND had received only conventionally fractionated RT, e.g., due to regional nodal disease; however, the proportion of participants with regional nodal disease (7%) was balanced between the arms. The fractionation schedules that were or could be used varied between total doses of 12–70 Gy, with fraction doses of 2.3–27 Gy in 1–28 fractions (EQD23 range 17–227 Gy). SBRT was delivered without systemic therapy in STOMP and ORIOLE, while the other trials either investigated or allowed systemic therapy in both arms: abiraterone acetate, prednisone, and androgen deprivation therapy (ADT) was added in ARTO; intermittent hormone therapy (ADT with or without a second-generation androgen receptor inhibitor for ≥ 6 months with planned break 4–8 months after enrollment) in EXTEND; and SOC in SABR-COMET, CORE, and STOP. Results of individual studies The outcomes fulfilling the review criteria and the corresponding data are listed in Table 3. In one study [84], crossover to the intervention arm was allowed at progression or 6 months, that is, the timepoint of the trial’s primary outcome. Therefore, we only included follow-up until 6 months in the analysis of this trial. Table 3Outcome data presented as effect estimates (95% confidence intervals), medians (95% confidence intervals), № (%)Systematic review outcomeSTOMP [77, 79]SABR-COMET [80–83]ORIOLE [79, 84]ARTO [85, 90]EXTEND [91]CORE [92, 115, 117]STOP [93, 116]Overall survivalOutcome nameOverall survivalOverall survivalOverall survivalOverall survivalOverall survivalNPOverall survivalDefinitionTime to deathPTime to deathTime to deathPTime to deathTime to deathPTime to deathPFollow-up––6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)–IGMedian not reached (i.e., < 50% with event)Median not reachedMedian not reachedCGMedian not reachedMedian not reachedMedian not reachedResultsModel does not convergeNo events in either groupa [79]HR 0.65(0.28–1.49)p = 0.302“Overall survival data were immature, […] similar between arms,” log-rank p = 0.21a [91]Incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-upOutcome nameToxicity grade ≥ 3Toxicity grade ≥ 3Adverse effects grade ≥ 3Adverse events grade ≥ 3Adverse events grade ≥ 3NPToxicity grade ≥ 3 related to RTPDefinitionCTCAE v4.0CTCAE v4CTCAE v4.0CTCAE v4.03CTCAE v4.0CTCAE v4 for each treated organPFollow-up––6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)–IG0/25 (0%)0/36 (0%)8/75 (11%)3/43 (7%)CG0/31 (0%)0/18 (0%)13/82 (16%)2/44 (5%)0/2 (0%)ResultsNANARR 0.67(0.30–1.53)p = 0.345bRR 1.53(0.27–8.74)p = 0.629bNAPFSOutcome namePFSPFSPFSPFSPFSPFSPFSDefinitionTime to PSA/clinical/ radiographic progression,c start of ADT, death, or study withdrawalTime to progression according to RECIST 1.0P or deathTime to PSA/clinical/ radiographic progression,c start of ADT, death, or study withdrawalTime to PSA/radiographic progression,d start of following treatment, or deathTime to PSA/clinical/radiographic progressione or deathTime to radiographic ± PSA/clinical progressionf or deathPTime to progression or deathPFollow-up––6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)––IGMedian not reachedMedian not reachedMedian not reachedCGMedian not reached17 months15.8 months(13.6–21.2)ResultsHR 0.09(0.01–0.65)p = 0.032gHR 0.21(0.07–0.61)p = 0.004a [84]HR 0.35(0.21–0.57)p < 0.001HR 0.25(0.12–0.55)p < 0.001HR 0.35(0.14–0.88)p = 0.025a [79]Local control at 5 years or longest follow-upOutcome nameSymtomatic or local progressionLesional control rateNPNPNPNPLesional controlDefinitionSee footnotehRECIST 1.0P–Follow-up––IG0/– occurrencesCG9/– occurrencesResultsNAIncidence proportion of grade 5 toxicity at 5 years or longest follow-upOutcome nameToxicity grade 5Toxicity grade 5Adverse effects grade 5Adverse events grade 5Adverse events grade 5NPToxicity grade 5 related to RTPDefinitionCTCAE v4.0CTCAE v4CTCAE v4.0CTCAE v4.03CTCAE v4.0CTCAE v4 for each treated organPFollow-up--6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)-IG0/25 (0%)–0/36 (0%)0/75 (0%)0/43 (0%)0/7 (0%)CG0/31 (0%)0/18 (0%)0/82 (0%)0/44 (0%)0/2 (0%)ResultsNANANANANAHRQoL at 3 monthsOutcome nameQuality of lifeQuality of lifeQuality of lifeQuality of lifeQuality of lifeNPNPDefinitionEORTC QLQ-C30, QLQ-PR25FACT-GBPI-SFEORTC QLQ-C30CES-D, MDASI, SF-12Follow-up3 months6 months3, 6 monthsP3 months3 monthsPIG–––Global Health Status: Mean 63.3 (13 participants)i–CGGlobal Health Status: Mean 78 (SD 18)a [77]Mean 63.3 (16 participants)iResults–No difference between groups, data not reported–HRQoL at 5 years or longest follow-upOutcome nameQuality of lifeQuality of lifeNPQuality of lifeQuality of lifeNPQuality of lifeDefinitionEORTC QLQ-C30, QLQ-PR25FACT-GEORTC QLQ-C30, BPI-SFCES-D, MDASI, SF-12FACT-GPFollow-up1 year6 years1 yearP22.0 months(range 11.6–39.2)5 yearsPIG–––––CGGlobal Health Status: Mean 75 (SD 20) Results–No difference between groups according to linear mixed-effectsmodellingSystemic therapy-free survivalOutcome nameADT-free survivalNPNPNAjNAjNPNPDefinitionTime to start of palliative ADT or deathFollow-up–IGCG13 months(80% CI 12–17)Results–Local PFSOutcome nameNPNPNPNPNPNPNPDefinitionFollow-upIGCGResultsDistant PFSOutcome nameRadiographic PFSTime to development of new metastasesRadiographic PFSNPTime to new lesion failureNPNPDefinitionTime to new nodal, bone, or visceral metastasis or deathTime to new nodal, bone, or visceral metastasis or deathTime to new lesion outside of the initial lesions presentPFollow-up––6 months until crossover2 yearsIG–Median not reacheda [91]CGResultsHR 0.33(0.11–1.0)p = 0.04, favoring the IGReported as described by the study authors and denoted by superscript P if only reported in protocol. Dash: No data/No informationADT androgen deprivation therapy, BPI-SF Brief Pain Inventory (Short Form), CES-D Center for Epidemiological Studies-Depression, CG control group, CI confidence interval, CT computed tomography, CTCAE Common Terminology Criteria for Adverse Events, EORTC European Organization for Research and Treatment of Cancer, FACT-G Functional Assessment of Cancer Therapy-General, HR hazard ratio, HRQoL health-related quality of life, IG intervention group, IQR interquartile range, MDASI MD Anderson Symptom Inventory, MRI magnetic resonance imaging, PFS progression-free survival, NA not applicable, NP not published, PSA prostate-specific antigen, QLQ-C30 Quality of Life Questionnaire-Core 30, QLQ-PR25 Quality of Life Questionnaire-Prostate 25, RECIST Response Evaluation Criteria in Solid Tumors, RR risk ratio, RT radiotherapy, RTOG Radiation Therapy Oncology Group, SD standard deviation, SF-12 12-Item Short Form Health Survey, v versionaExtracted from figurebEstimated from the groups’ resultscPSA progression: ≥ 25% and ≥ 2 µg/L above nadir confirmed after ≥ 4 weeks.P Clinical: symptomatic progression. Radiographic: ≥ 20% in soft tissue lesions’ sum diameter per RECIST 1.1 (applied to all lesions when a lesion did not meet criteria) on CT; ≥ 1 new lesion(s)P on bone scan; on MRI; or any evidence by sizedPSA progression: rises ≥ 1 week apart and ≥ 2 µg/L. Radiographic: 20% in sum of target lesions’ longest diameters; new soft tissue/visceral lesions; and/or unequivocal progression on CT/MRIPePSA progression: ≥ 25% and ≥ 2 µg/L above nadir. Clinical: treatment change needed.P Radiographic: RECIST 1.1PfRadiographic progression: RECIST 1.1. Bone scan and PSA were considered. When progression could not be measured, clinical evidence was usedPgEstimated from HR and CIhProgression of soft tissue lesions: ≥ 20% and ≥ 5 mm in longest diameter. Bone: ≥ 25% on CT [118]iData provided by personal correspondence with study authorsjTrial treatment included systemic therapy Table 3Outcome data presented as effect estimates (95% confidence intervals), medians (95% confidence intervals), № (%)Systematic review outcomeSTOMP [77, 79]SABR-COMET [80–83]ORIOLE [79, 84]ARTO [85, 90]EXTEND [91]CORE [92, 115, 117]STOP [93, 116]Overall survivalOutcome nameOverall survivalOverall survivalOverall survivalOverall survivalOverall survivalNPOverall survivalDefinitionTime to deathPTime to deathTime to deathPTime to deathTime to deathPTime to deathPFollow-up––6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)–IGMedian not reached (i.e., < 50% with event)Median not reachedMedian not reachedCGMedian not reachedMedian not reachedMedian not reachedResultsModel does not convergeNo events in either groupa [79]HR 0.65(0.28–1.49)p = 0.302“Overall survival data were immature, […] similar between arms,” log-rank p = 0.21a [91]Incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-upOutcome nameToxicity grade ≥ 3Toxicity grade ≥ 3Adverse effects grade ≥ 3Adverse events grade ≥ 3Adverse events grade ≥ 3NPToxicity grade ≥ 3 related to RTPDefinitionCTCAE v4.0CTCAE v4CTCAE v4.0CTCAE v4.03CTCAE v4.0CTCAE v4 for each treated organPFollow-up––6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)–IG0/25 (0%)0/36 (0%)8/75 (11%)3/43 (7%)CG0/31 (0%)0/18 (0%)13/82 (16%)2/44 (5%)0/2 (0%)ResultsNANARR 0.67(0.30–1.53)p = 0.345bRR 1.53(0.27–8.74)p = 0.629bNAPFSOutcome namePFSPFSPFSPFSPFSPFSPFSDefinitionTime to PSA/clinical/ radiographic progression,c start of ADT, death, or study withdrawalTime to progression according to RECIST 1.0P or deathTime to PSA/clinical/ radiographic progression,c start of ADT, death, or study withdrawalTime to PSA/radiographic progression,d start of following treatment, or deathTime to PSA/clinical/radiographic progressione or deathTime to radiographic ± PSA/clinical progressionf or deathPTime to progression or deathPFollow-up––6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)––IGMedian not reachedMedian not reachedMedian not reachedCGMedian not reached17 months15.8 months(13.6–21.2)ResultsHR 0.09(0.01–0.65)p = 0.032gHR 0.21(0.07–0.61)p = 0.004a [84]HR 0.35(0.21–0.57)p < 0.001HR 0.25(0.12–0.55)p < 0.001HR 0.35(0.14–0.88)p = 0.025a [79]Local control at 5 years or longest follow-upOutcome nameSymtomatic or local progressionLesional control rateNPNPNPNPLesional controlDefinitionSee footnotehRECIST 1.0P–Follow-up––IG0/– occurrencesCG9/– occurrencesResultsNAIncidence proportion of grade 5 toxicity at 5 years or longest follow-upOutcome nameToxicity grade 5Toxicity grade 5Adverse effects grade 5Adverse events grade 5Adverse events grade 5NPToxicity grade 5 related to RTPDefinitionCTCAE v4.0CTCAE v4CTCAE v4.0CTCAE v4.03CTCAE v4.0CTCAE v4 for each treated organPFollow-up--6 months until crossover24.9 months(IQR 17.1–35.8)22.0 months(range 11.6–39.2)-IG0/25 (0%)–0/36 (0%)0/75 (0%)0/43 (0%)0/7 (0%)CG0/31 (0%)0/18 (0%)0/82 (0%)0/44 (0%)0/2 (0%)ResultsNANANANANAHRQoL at 3 monthsOutcome nameQuality of lifeQuality of lifeQuality of lifeQuality of lifeQuality of lifeNPNPDefinitionEORTC QLQ-C30, QLQ-PR25FACT-GBPI-SFEORTC QLQ-C30CES-D, MDASI, SF-12Follow-up3 months6 months3, 6 monthsP3 months3 monthsPIG–––Global Health Status: Mean 63.3 (13 participants)i–CGGlobal Health Status: Mean 78 (SD 18)a [77]Mean 63.3 (16 participants)iResults–No difference between groups, data not reported–HRQoL at 5 years or longest follow-upOutcome nameQuality of lifeQuality of lifeNPQuality of lifeQuality of lifeNPQuality of lifeDefinitionEORTC QLQ-C30, QLQ-PR25FACT-GEORTC QLQ-C30, BPI-SFCES-D, MDASI, SF-12FACT-GPFollow-up1 year6 years1 yearP22.0 months(range 11.6–39.2)5 yearsPIG–––––CGGlobal Health Status: Mean 75 (SD 20) Results–No difference between groups according to linear mixed-effectsmodellingSystemic therapy-free survivalOutcome nameADT-free survivalNPNPNAjNAjNPNPDefinitionTime to start of palliative ADT or deathFollow-up–IGCG13 months(80% CI 12–17)Results–Local PFSOutcome nameNPNPNPNPNPNPNPDefinitionFollow-upIGCGResultsDistant PFSOutcome nameRadiographic PFSTime to development of new metastasesRadiographic PFSNPTime to new lesion failureNPNPDefinitionTime to new nodal, bone, or visceral metastasis or deathTime to new nodal, bone, or visceral metastasis or deathTime to new lesion outside of the initial lesions presentPFollow-up––6 months until crossover2 yearsIG–Median not reacheda [91]CGResultsHR 0.33(0.11–1.0)p = 0.04, favoring the IGReported as described by the study authors and denoted by superscript P if only reported in protocol. Dash: No data/No informationADT androgen deprivation therapy, BPI-SF Brief Pain Inventory (Short Form), CES-D Center for Epidemiological Studies-Depression, CG control group, CI confidence interval, CT computed tomography, CTCAE Common Terminology Criteria for Adverse Events, EORTC European Organization for Research and Treatment of Cancer, FACT-G Functional Assessment of Cancer Therapy-General, HR hazard ratio, HRQoL health-related quality of life, IG intervention group, IQR interquartile range, MDASI MD Anderson Symptom Inventory, MRI magnetic resonance imaging, PFS progression-free survival, NA not applicable, NP not published, PSA prostate-specific antigen, QLQ-C30 Quality of Life Questionnaire-Core 30, QLQ-PR25 Quality of Life Questionnaire-Prostate 25, RECIST Response Evaluation Criteria in Solid Tumors, RR risk ratio, RT radiotherapy, RTOG Radiation Therapy Oncology Group, SD standard deviation, SF-12 12-Item Short Form Health Survey, v versionaExtracted from figurebEstimated from the groups’ resultscPSA progression: ≥ 25% and ≥ 2 µg/L above nadir confirmed after ≥ 4 weeks.P Clinical: symptomatic progression. Radiographic: ≥ 20% in soft tissue lesions’ sum diameter per RECIST 1.1 (applied to all lesions when a lesion did not meet criteria) on CT; ≥ 1 new lesion(s)P on bone scan; on MRI; or any evidence by sizedPSA progression: rises ≥ 1 week apart and ≥ 2 µg/L. Radiographic: 20% in sum of target lesions’ longest diameters; new soft tissue/visceral lesions; and/or unequivocal progression on CT/MRIPePSA progression: ≥ 25% and ≥ 2 µg/L above nadir. Clinical: treatment change needed.P Radiographic: RECIST 1.1PfRadiographic progression: RECIST 1.1. Bone scan and PSA were considered. When progression could not be measured, clinical evidence was usedPgEstimated from HR and CIhProgression of soft tissue lesions: ≥ 20% and ≥ 5 mm in longest diameter. Bone: ≥ 25% on CT [118]iData provided by personal correspondence with study authorsjTrial treatment included systemic therapy Outcome data presented as effect estimates (95% confidence intervals), medians (95% confidence intervals), № (%) 24.9 months (IQR 17.1–35.8) 22.0 months (range 11.6–39.2) HR 0.65 (0.28–1.49) p = 0.302 24.9 months (IQR 17.1–35.8) 22.0 months (range 11.6–39.2) RR 0.67 (0.30–1.53) p = 0.345b RR 1.53 (0.27–8.74) p = 0.629b 24.9 months (IQR 17.1–35.8) 22.0 months (range 11.6–39.2) 15.8 months (13.6–21.2) HR 0.09 (0.01–0.65) p = 0.032g HR 0.21 (0.07–0.61) p = 0.004a [84] HR 0.35 (0.21–0.57) p < 0.001 HR 0.25 (0.12–0.55) p < 0.001 HR 0.35 (0.14–0.88) p = 0.025a [79] 24.9 months (IQR 17.1–35.8) 22.0 months (range 11.6–39.2) 22.0 months (range 11.6–39.2) No difference between groups according to linear mixed-effects modelling 13 months (80% CI 12–17) HR 0.33 (0.11–1.0) p = 0.04, favoring the IG Reported as described by the study authors and denoted by superscript P if only reported in protocol. Dash: No data/No information ADT androgen deprivation therapy, BPI-SF Brief Pain Inventory (Short Form), CES-D Center for Epidemiological Studies-Depression, CG control group, CI confidence interval, CT computed tomography, CTCAE Common Terminology Criteria for Adverse Events, EORTC European Organization for Research and Treatment of Cancer, FACT-G Functional Assessment of Cancer Therapy-General, HR hazard ratio, HRQoL health-related quality of life, IG intervention group, IQR interquartile range, MDASI MD Anderson Symptom Inventory, MRI magnetic resonance imaging, PFS progression-free survival, NA not applicable, NP not published, PSA prostate-specific antigen, QLQ-C30 Quality of Life Questionnaire-Core 30, QLQ-PR25 Quality of Life Questionnaire-Prostate 25, RECIST Response Evaluation Criteria in Solid Tumors, RR risk ratio, RT radiotherapy, RTOG Radiation Therapy Oncology Group, SD standard deviation, SF-12 12-Item Short Form Health Survey, v version aExtracted from figure bEstimated from the groups’ results cPSA progression: ≥ 25% and ≥ 2 µg/L above nadir confirmed after ≥ 4 weeks.P Clinical: symptomatic progression. Radiographic: ≥ 20% in soft tissue lesions’ sum diameter per RECIST 1.1 (applied to all lesions when a lesion did not meet criteria) on CT; ≥ 1 new lesion(s)P on bone scan; on MRI; or any evidence by size dPSA progression: rises ≥ 1 week apart and ≥ 2 µg/L. Radiographic: 20% in sum of target lesions’ longest diameters; new soft tissue/visceral lesions; and/or unequivocal progression on CT/MRIP ePSA progression: ≥ 25% and ≥ 2 µg/L above nadir. Clinical: treatment change needed.P Radiographic: RECIST 1.1P fRadiographic progression: RECIST 1.1. Bone scan and PSA were considered. When progression could not be measured, clinical evidence was usedP gEstimated from HR and CI hProgression of soft tissue lesions: ≥ 20% and ≥ 5 mm in longest diameter. Bone: ≥ 25% on CT [118] iData provided by personal correspondence with study authors jTrial treatment included systemic therapy Outcomes were measured from either date of informed consent, enrollment, randomization, or start of (systemic) treatment as reported in the study reports or protocols. Overall survival and toxicity were homogenously defined as time to death and by applying CTCAE version 4, respectively. The studies used different clinical, biochemical, and/or radiographic definitions of PFS and local control and HRQoL instruments. One study [77, 78] reported systemic therapy-free survival as time to start of ADT. No study has compared the local PFS. Four studies [79, 82, 91] reported distant PFS as time to new metastasis. Data were published exclusively as figures for overall survival and PFS until 6 months in ORIOLE [79, 84] and for HRQoL at 3 months in STOMP [77], and were carefully estimated from the respective figures. ORIOLE also reported distant PFS exclusively as a figure [79]; however, we refrained from extracting the data owing to uncertainties in estimation around the 6-month mark. EXTEND [91] reported overall survival exclusively as a figure with a log-rank test. The study authors described that “Overall survival data were immature, […] similar between arms,” with no events in the IG and two events in the CG, with a log-rank p of 0.21. Synthesis of results A meta-analysis for overall survival was not possible because only one out of three trials reported events in both arms. The trial [90] showed no significant differences between the arms (HR 0.65, 95% CI 0.28–1.49, p = 0.302). The pooled results for incidence proportion of grade ≥ 3 toxicity from two trials [90, 91] with events in both arms showed no difference between groups (pooled RR 0.78, 95% CI 0.37–1.65, Z = 0.65, p = 0.52, Fig. 2). Statistical heterogeneity was not detected.Fig. 2Meta-analysis for incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up. Pooled results for critical outcome incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up. CI confidence interval df degrees of freedom, IV inverse-variance, M-H Mantel–Haenszel method, SBRT stereotactic body radiotherapy Meta-analysis for incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up. Pooled results for critical outcome incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up. CI confidence interval df degrees of freedom, IV inverse-variance, M-H Mantel–Haenszel method, SBRT stereotactic body radiotherapy The pooled results from four trials [79, 83, 90, 91] showed significantly longer PFS after SBRT, with an HR of 0.31 (95% CI 0.21–0.45, Z = 6.21, p < 0.00001, Fig. 3). There was no evident statistical heterogeneity. A sensitivity analysis excluding trials with a high risk of bias was not performed because it was not detected in the trials for this outcome (see below).Fig. 3Meta-analysis for progression-free survival. Pooled results for additional important outcome PFS. The SABR-COMET trial results were for the prostate cancer subset. CI confidence interval, df degrees of freedom, IV inverse-variance, PFS progression-free survival, SBRT stereotactic body radiotherapy Meta-analysis for progression-free survival. Pooled results for additional important outcome PFS. The SABR-COMET trial results were for the prostate cancer subset. CI confidence interval, df degrees of freedom, IV inverse-variance, PFS progression-free survival, SBRT stereotactic body radiotherapy The critical and additional important outcomes are summarized in Table 4. Table 4Summary of findingsStereotactic body radiotherapy as metastasis-directed therapy vs no metastasis-directed therapy in oligometastatic prostate cancerPatient or population: patients with OMPCSetting: NA and EuropeIntervention: SBRT as MDTComparison: no MDTOutcomesAnticipated absolute effects* (95% CI)Relative effect(95% CI)№ of participants(studies)Certainty of the evidence(GRADE)CommentsRisk with no MDTRisk with SBRT as MDTOverall survivalfollow-up: median 24.9 months18 per 100a [90]12 per 100(5 to 26)HR 0.65(0.28 to 1.49)[Death]157(1 RCT) ⊕ ⊕ ⊕ ⊝ Moderateb2 additional studies had no events and 2 events in the CG only, respectivelyIncidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-upassessed with: CTCAE v4follow-up: range 6 to 39.2 months12 per 1009 per 100(4 to 20)RR 0.78(0.37 to 1.65)244(2 RCTs) ⊕ ⊕ ⊝ ⊝ Lowc,d2 additional studies had no events for in total 61 participants in the IGs and 49 in the CGsPFSfollow-up: range 5.8 to 39.2 months59 per 100e [79, 83, 90, 91]24 per 100(17 to 33)HR 0.31(0.21 to 0.45)[Progression]314(4 RCTs) ⊕ ⊕ ⊝ ⊝ Lowc,dLocal control at 5 years or longest follow-upNot reportedfIncidence proportion of grade 5 toxicity at 5 years or longest follow-upassessed with: CTCAE v4follow-up: range 6 to 39.2 months0 per 1000 per 100not estimable363(5 RCTs) ⊕ ⊕ ⊝ ⊝ Lowb,c5 studies had no eventsHRQoL at 3 monthsg2 studies reported no differences without further data and mean values only, respectively141(2 RCTs) − Quality of evidence assessment not performed due to absence of effect estimateHRQoL at 5 years or longest follow-up1 study reported no differences according to linear mixed-effects modelling87(1 RCT)–See aboveSystemic therapy-free survivalNot reported*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI)CI confidence interval, CTCAE v4 Common Terminology Criteria for Adverse Events version 4, GRADE Grading of Recommendations Assessment, Development, and Evaluation, HR hazard ratio, HRQoL health-related quality of life, MDT metastasis-directed therapy, OMPC oligometastatic prostate cancer, PFS progression-free survival, RCT randomized controlled trial, SBRT stereotactic body radiotherapyGRADE Working Group grades of evidenceHigh certainty: we are very confident that the true effect lies close to that of the estimate of the effectModerate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially differentLow certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effectVery low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effectaEstimated from number of events in the CGbDowngraded by 1 level for imprecision. Effect estimate based on a single small studycDowngraded by 1 level for risk of bias. Assessors were aware of received intervention and could have influenced assessmentdDowngraded by 1 level for imprecision. Few included participants and events; optimal information size criterion was not meteEstimated from number of events in the CGs in the studies included in the meta-analysis. Number of events was estimated from the Kaplan–Meier curve for one study. Another study had no information on number of events and was assumed to have occurred in 1 out of the 2 participants in the CGfTwo studies reported occurrences of progression with likely metastasis as unit of analysis instead of participantsgTimepoints 1–6 months allowed Table 4Summary of findingsStereotactic body radiotherapy as metastasis-directed therapy vs no metastasis-directed therapy in oligometastatic prostate cancerPatient or population: patients with OMPCSetting: NA and EuropeIntervention: SBRT as MDTComparison: no MDTOutcomesAnticipated absolute effects* (95% CI)Relative effect(95% CI)№ of participants(studies)Certainty of the evidence(GRADE)CommentsRisk with no MDTRisk with SBRT as MDTOverall survivalfollow-up: median 24.9 months18 per 100a [90]12 per 100(5 to 26)HR 0.65(0.28 to 1.49)[Death]157(1 RCT) ⊕ ⊕ ⊕ ⊝ Moderateb2 additional studies had no events and 2 events in the CG only, respectivelyIncidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-upassessed with: CTCAE v4follow-up: range 6 to 39.2 months12 per 1009 per 100(4 to 20)RR 0.78(0.37 to 1.65)244(2 RCTs) ⊕ ⊕ ⊝ ⊝ Lowc,d2 additional studies had no events for in total 61 participants in the IGs and 49 in the CGsPFSfollow-up: range 5.8 to 39.2 months59 per 100e [79, 83, 90, 91]24 per 100(17 to 33)HR 0.31(0.21 to 0.45)[Progression]314(4 RCTs) ⊕ ⊕ ⊝ ⊝ Lowc,dLocal control at 5 years or longest follow-upNot reportedfIncidence proportion of grade 5 toxicity at 5 years or longest follow-upassessed with: CTCAE v4follow-up: range 6 to 39.2 months0 per 1000 per 100not estimable363(5 RCTs) ⊕ ⊕ ⊝ ⊝ Lowb,c5 studies had no eventsHRQoL at 3 monthsg2 studies reported no differences without further data and mean values only, respectively141(2 RCTs) − Quality of evidence assessment not performed due to absence of effect estimateHRQoL at 5 years or longest follow-up1 study reported no differences according to linear mixed-effects modelling87(1 RCT)–See aboveSystemic therapy-free survivalNot reported*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI)CI confidence interval, CTCAE v4 Common Terminology Criteria for Adverse Events version 4, GRADE Grading of Recommendations Assessment, Development, and Evaluation, HR hazard ratio, HRQoL health-related quality of life, MDT metastasis-directed therapy, OMPC oligometastatic prostate cancer, PFS progression-free survival, RCT randomized controlled trial, SBRT stereotactic body radiotherapyGRADE Working Group grades of evidenceHigh certainty: we are very confident that the true effect lies close to that of the estimate of the effectModerate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially differentLow certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effectVery low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effectaEstimated from number of events in the CGbDowngraded by 1 level for imprecision. Effect estimate based on a single small studycDowngraded by 1 level for risk of bias. Assessors were aware of received intervention and could have influenced assessmentdDowngraded by 1 level for imprecision. Few included participants and events; optimal information size criterion was not meteEstimated from number of events in the CGs in the studies included in the meta-analysis. Number of events was estimated from the Kaplan–Meier curve for one study. Another study had no information on number of events and was assumed to have occurred in 1 out of the 2 participants in the CGfTwo studies reported occurrences of progression with likely metastasis as unit of analysis instead of participantsgTimepoints 1–6 months allowed Summary of findings Overall survival follow-up: median 24.9 months 12 per 100 (5 to 26) HR 0.65 (0.28 to 1.49) [Death] 157 (1 RCT) ⊕ ⊕ ⊕ ⊝ Moderateb Incidence proportion of grade ≥ 3 toxicity at 5 years or longest follow-up assessed with: CTCAE v4 follow-up: range 6 to 39.2 months 9 per 100 (4 to 20) RR 0.78 (0.37 to 1.65) 244 (2 RCTs) PFS follow-up: range 5.8 to 39.2 months 24 per 100 (17 to 33) HR 0.31 (0.21 to 0.45) [Progression] 314 (4 RCTs) Incidence proportion of grade 5 toxicity at 5 years or longest follow-up assessed with: CTCAE v4 follow-up: range 6 to 39.2 months 363 (5 RCTs) 141 (2 RCTs) 87 (1 RCT) *The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CI confidence interval, CTCAE v4 Common Terminology Criteria for Adverse Events version 4, GRADE Grading of Recommendations Assessment, Development, and Evaluation, HR hazard ratio, HRQoL health-related quality of life, MDT metastasis-directed therapy, OMPC oligometastatic prostate cancer, PFS progression-free survival, RCT randomized controlled trial, SBRT stereotactic body radiotherapy GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect aEstimated from number of events in the CG bDowngraded by 1 level for imprecision. Effect estimate based on a single small study cDowngraded by 1 level for risk of bias. Assessors were aware of received intervention and could have influenced assessment dDowngraded by 1 level for imprecision. Few included participants and events; optimal information size criterion was not met eEstimated from number of events in the CGs in the studies included in the meta-analysis. Number of events was estimated from the Kaplan–Meier curve for one study. Another study had no information on number of events and was assumed to have occurred in 1 out of the 2 participants in the CG fTwo studies reported occurrences of progression with likely metastasis as unit of analysis instead of participants gTimepoints 1–6 months allowed Risk of bias and quality of evidence The risk-of-bias assessment on the study, domain, and outcome levels are shown in Fig. 4, with further comments in Additional file 5.Fig. 4Risk-of-bias assessment for collated outcomes and per reported outcome. According to the Cochrane risk-of-bias tool version 2, for A Collated outcomes, B Overall survival, C Toxicity, D PFS, E Local control, F HRQoL, G Systemic therapy-free survival, and H Distant PFS. High risk of bias was considered due to the following: bias due to deviations from intended intervention—In STOMP and STOP, a part of the CG received the intervention treatment. In CORE, completion of SOC differed between the IG and CG; bias due to missing outcome data—In ARTO and EXTEND, HRQoL assessments were only available for a part of the participants, at 3 months in ARTO and at baseline in EXTEND; and bias due to measurement of the outcome—In EXTEND, HRQoL assessments were optional for participants and willingness to respond may have differed between the arms. CG control group, D domain, HRQoL health-related quality of life, IG intervention group, PFS progression-free survival, SOC standard of care Risk-of-bias assessment for collated outcomes and per reported outcome. According to the Cochrane risk-of-bias tool version 2, for A Collated outcomes, B Overall survival, C Toxicity, D PFS, E Local control, F HRQoL, G Systemic therapy-free survival, and H Distant PFS. High risk of bias was considered due to the following: bias due to deviations from intended intervention—In STOMP and STOP, a part of the CG received the intervention treatment. In CORE, completion of SOC differed between the IG and CG; bias due to missing outcome data—In ARTO and EXTEND, HRQoL assessments were only available for a part of the participants, at 3 months in ARTO and at baseline in EXTEND; and bias due to measurement of the outcome—In EXTEND, HRQoL assessments were optional for participants and willingness to respond may have differed between the arms. CG control group, D domain, HRQoL health-related quality of life, IG intervention group, PFS progression-free survival, SOC standard of care Overall, the risk of bias arising from the randomization process was low. One trial had some concerns due to no information on allocation concealment; however, only an abstract and trial registry entries were available at that time. Deviations from the intended interventions that could have affected outcomes were observed in three trials. Missing outcome data were considered a risk of bias in two trials, but only for HRQoL outcomes, owing to available data for only some of the trial participants. Bias in measurements of the outcomes was overall of some concerns, except for overall survival, because knowledge of the received intervention could have influenced assessment. However, this was considered unlikely except for HRQoL in one trial where the assessments were optional for the participants and the willingness to respond could have differed between arms. The quality of evidence assessment is shown in Table 4. The effect estimate for overall survival was considered to have moderate certainty after downgrading by one level for imprecision due to being based on a single study. Toxicity and PFS were considered to have a low certainty and were downgraded by one level each for risk of bias and imprecision. Risk of bias was considered because knowledge of the received intervention could have influenced assessments, and imprecision due to few included participants and events per the GRADE optimal information size criterion [94]. Discussion Summary of evidence This is, to our knowledge, the first systematic review approaching studies with high-quality designs using randomization and focusing solely on SBRT as MDT in OMPC. We identified seven randomized phase II trials, of which four investigated PCa among several primary cancer types and one investigated SBRT as one of two MDTs (together with surgery). The results from four studies, one with reported PCa participant subset data, show promising improvements in PFS in several OMD states with a pooled HR of 0.31 (95% CI 0.21–0.45). No excess toxicities or treatment-related deaths were observed. Overall survival comparisons were immature, as only one of the three trials reported events in both arms. Our findings are in line with those of previous systematic reviews of prospective trials on OMPC that also found favorable disease control and low toxicity [24–27]. All of these systematic reviews included STOMP and ORIOLE, being two early RCTs in this field. Strengths and limitations To summarize the current evidence, we applied generous eligibility criteria allowing participant subsets and various definitions of OMD and SBRT, and performed a comprehensive search of bibliographic databases and other sources. By including eligible participant subsets and updating our searches in the end of 2023, we identified five additional trials with data on OMPC that were not included in previous systematic reviews (i.e., SABR-COMET, ARTO, EXTEND, CORE, and STOP). Reflecting the rapid development in this field, four of these trials were published in the preceding year. An important outcome of the review is the presentation of the backdrop against which the current evidence in OMPC is generated. The included studies addressed varying patient groups, treatment approaches (e.g., fractionation schedules and co-administered systemic therapies), and control treatments, which limit the conclusions that can be drawn from pooled effect estimates. The use of ADT and other systemic therapies varied between the trials, which need to be considered when interpreting the results. As more RCTs are published, future systematic reviews may be able to give a more granular picture of the treatment’s efficacy by restricting to individual OMD states and by sensitivity to hormonal therapy, a minimally required EQD2, or certain SOC comparators. Our results are also limited by aspects encountered in emerging fields, that is, evidence being so far generated in small exploratory phase II trials and heterogeneous outcome definitions. Combining studies in a meta-analysis is a strategy to overcome small sample sizes by increasing the statistical power [52]. However, small RCTs, even when pooled, present limitations. When small studies show positive results, they are more likely to exaggerate differences compared with larger trials owing to statistical imprecision [95]. Furthermore, balancing of prognostic factors between the trial arms from randomization can only be assumed if the sample sizes are sufficiently large. It may not be possible to discern if impressive results are due to a prognostic imbalance or a real treatment effect. Therefore, a pooled estimate would be equally uncertain [94, 96]. This was reflected in our quality of evidence assessment by downgrading for imprecision as the GRADE optimal information size criterion calls for meta-analyses of approximately 250 or more events [94]. The outcome definitions were heterogeneous except for overall survival and toxicity. The definitions of PFS included different event criteria and assessment approaches (e.g., different use of PSA, clinical, and radiographic criteria, imaging modalities, and instructions for repeated imaging), which limit comparability across trials. Our pooled effect estimate from the meta-analysis may have differed had one approach been applied in all studies. Furthermore, our quality of evidence assessment identified two additional challenges in RCTs: blinding of outcome assessors and adherence to assigned intervention. Blinding in trials of radiotherapeutic interventions is inherently difficult and could be considered for radiologists and data analysts. Issues of intervention adherence and crossover have become central as MDT is increasingly being offered in clinical practice. In two of the included studies [77, 93], parts of the CGs eventually received SBRT, and in one study [84], crossover was allowed after the primary endpoint at 6 months, precluding interpretation of most long-term outcomes in this review. Furthermore, one study [92] reported significantly lower completion of SOC in the IG, which could have underestimated the benefit of SBRT. Crossover to a yet unestablished treatment should be interpreted with caution because it may compete with other treatments. This could prolong systemic therapy-free survival, favoring the CG, or delay effective treatment, favoring the IG, and therefore cloud the comparison in the intention-to-treat population [97]. While the identified toxicity in the literature was low, there are known rare, severe toxicities that may be life-threatening. It is important to introduce these risks only after confirmatory trials have laid the foundation for risk–benefit assessments in comparison with other established palliative treatments. Currently, there is no evidence that MDT is curative for metastatic PCa, thereby necessitating a cautious approach. Future research We identified eight registered and ongoing confirmatory phase III trials that include patients with OMPC (ClinicalTrials.gov registration numbers NCT03143322 [98], NCT03721341 [99], NCT03862911 [100], NCT04115007 [101], NCT04983095 [102], NCT05209243 [103], NCT05717166 [104], and NCT06320067 [105]). One trial including multiple primary cancer types (SABR-COMET-10 [99]) has finalized recruitment, whereas all trials investigating only OMPC are currently recruiting. Of them, the largest is the multi-arm STAMPEDE2 platform [105] with a recruitment target of almost 2500 participants for the comparison of SBRT as MDT added to SOC vs SOC. SBRT as MDT has several important theoretical benefits (e.g., delaying the start of or a change in systemic therapy and enhancing of the anticancer immune response) that need confirmation through an overall survival impact and cannot be fully captured by other outcomes. Such improvements, together with disease control, could benefit patients by minimizing adverse effects, prolonging the usefulness of systemic therapy, and consequently improving HRQoL and optimizing the use of healthcare resources. Furthermore, the comparator or concomitant treatments become increasingly important in phase III trials for assessing benefits compared with other established treatments. There are several efficacious systemic therapy options for metastatic PCa at different stages of the disease (e.g., ADT, second-generation androgen receptor inhibitors, chemotherapy, poly-ADP ribose polymerase [PARP] inhibitors, and lutetium-177-PSMA-617 [106, 107]), some of which have only been introduced in the past few years, and is also a field of intense research [108]. Similarly, the participants’ previous treatment histories will become equally important to understand the clinical setting in which MDT is investigated. The differences in outcomes between oligometastatic hormone-sensitive and castration-resistant PCa was illustrated in a prospective phase II trial recruiting before the establishment of second-generation androgen receptor inhibitors [109]. After SBRT as MDT and at least two years of ADT, the three-year metastasis progressions-free survival was 67% (95% CI 53–77) in the cohort with hormone-sensitive disease and 26% (95% CI 7–51) in the cohort with castration-resistant disease. One way to address outcome heterogeneity is core outcome sets (COS), which are standardized sets of outcomes to be minimally measured and reported in trials for a certain health condition [110]. In the database of Core Outcome Measures in Effectiveness Trials (COMET) [111], an initiative promoting and facilitating the use of COS, three sets for research in advanced PCa are available: one with patient-reported symptoms [112], one for metastatic disease in general [113] (published after the start of this review), and one for castration-resistant disease [114]. However, they are not specific to OMD or MDT. There is a general agreement on important outcome domains, and the ESTRO–ASTRO consensus document [10] specifying outcomes for RT as MDT (used for outcome selection in this review) is a landmark step towards an OMD- and MDT-specific COS. In the future, harmonizing outcome definitions, measurements, and timing could clarify the interpretation of the results and improve trial comparability. Some of these measurements could include qualitative evaluations such as pain and toxicity and would benefit from involving patients, relatives, and healthcare providers during the development process. Fractionation schedules and EQD23 varied significantly within and between the included studies, introducing another aspect of heterogeneity. Our review highlights the need for careful consideration of the methods for SBRT planning and delivery, including reporting of dose prescription along with relevant dose-volume metrics for targets and organs at risk [8], in future assessments of ongoing research to ensure reproducibility in clinical practice. Conclusions Phase II trials have shown promising improvements in PFS for several OMD states in PCa without excess toxicity. Comparisons for overall survival are currently immature. Outcome definitions were heterogeneous, and harmonization of definitions, measurements, and timing will be essential to the interpretation of results and comparability across trials. The use of outcome measurements as defined in consensus documents is recommended. In future confirmatory phase III trials, adequately large sample sizes, blinding of outcome assessors, and/or adherence to assigned intervention could improve the quality of evidence. The PRISMA 2009 Checklist is found in Additional file 6. Supplementary Information Additional file 1.Additional file 2.Additional file 3.Additional file 4.Additional file 5.Additional file 6. Additional file 1. Additional file 2. Additional file 3. Additional file 4. Additional file 5. Additional file 6.