SBRT in recurrent oligometastatic ovarian cancer: An EORTC Y-ECI GCG systematic review

Highlights

• •SBRT is emerging as an effective treatment for recurrent oligometastatic ovarian cancer.
• •SBRT delays the need for systemic therapies, benefiting patients with limited disease.
• •Variations in SBRT protocols highlight the need for standardized treatment approaches.
• •Combining SBRT with PARP inhibitors may prolong the effectiveness of maintenance therapy.
• •Prospective studies are needed to refine patient selection and optimize SBRT use.

Abstract

Managing recurrent oligometastatic ovarian cancer with a limited disease burden is challenging, requiring localized treatments to achieve durable control while reducing systemic therapy exposure. Stereotactic body radiotherapy (SBRT) offers high precision and ablative dosing, minimizing toxicity to adjacent organs. This systematic review analyzed nine studies involving 426 patients and 809 lesions treated with SBRT. Patients had a median age of 61.2 years and underwent a median of three prior systemic treatments. SBRT, delivered at a median dose of 40 Gy, achieved a pooled median progression-free survival (PFS) of 10.32 months (95 % CI: 2.79–38.26) and a 2-year overall survival rate of 78 %. Local outcomes were promising, with a target lesion control rate of 95.9 % and a lesion response rate of 86 %. SBRT delayed systemic therapy initiation, with a median time of 9.45 months, and was well tolerated, with mostly mild toxicities (CTCAE grade ≤ 2). Rare grade 3 and 5 gastrointestinal toxicities were reported. Despite these encouraging results, the studies were limited by moderate to serious risk of bias, heterogeneity in treatment protocols, and variations in SBRT dose and fractionation. These findings highlight SBRT’s potential as an effective treatment for oligometastatic ovarian cancer, but further research is needed to refine strategies, standardize regimens, and identify patient subgroups most likely to benefit. Prospective trials are eagerly awaited to establish SBRT’s role in improving outcomes for this challenging clinical scenario.

Introduction

Epithelial Ovarian cancer (EOC) is the second cause of death of all the gynecological malignancies [1]. More than two-thirds are diagnosed in advance stages and recurrence rates are high [2]. Management of patients with advanced or recurrent EOC in the front-line includes cytoreductive surgery with the aim of no residual disease, platinum-based chemotherapy, and maintenance strategies that may include PARP inhibitors and/or bevacizumab [3]. Despite an initial high response rate to platinum-based chemotherapy, approximately 70 % of patients with high grade serous ovarian cancer will progress in the first 3 years [3].

In this evolving landscape, the role of radiotherapy (RT), which was once primarily seen as a palliative measure, has undergone a significant reassessment [[4], [5], [6]]. Over time, there has been a gradual yet steady shift toward viewing RT as a proactive and definitive treatment. It is now increasingly integrated into a multidisciplinary approach that includes surgery, conventional chemotherapy, and innovative options from targeted therapies [7].

In the past decade, as treatment techniques have improved, the use of radiotherapy in stereotactic (SBRT) conditions targeting metastatic lesions has made it possible to control these lesions as well as the prolongation of the systemic treatment-free intervals in various primary histologies [[8], [9], [10]]. The oligometastatic disease is defined as an intermediate state between localized and widespread metastatic disease in which local control may improve systemic control. The SABR-COMET randomized phase II trial, assessing SBRT in patients with controlled primary malignancy and 1 to 5 metastatic lesions, found that SBRT improves overall survival in patients compared with standard of care palliative therapy [11]. However, this study included various tumor types and only two patients with gynecological cancers.

The role of SBRT for oligometastatic EOC has been reported in small retrospective series as a potential efficient treatment, but it is not yet a standard treatment in EOC patients. We systematically reviewed the literature to assess the efficacy (progression free survival and local control) and safety (toxicity) of SBRT for oligometastatic EOC.

Methods and materials

Search strategy

We conducted a systematic review of the literature according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) checklist 2020 [12], where the primary sources of information were Pubmed, EMBASE, Cochrane and Medline. The search keywords were “ovarian cancer” OR “ovarian carcinoma” OR “ovarian tumour” OR “ovarian tumor” OR “Ovarian neoplasm” AND “radiosurgery” OR “stereotactic radiation therapy” OR “SBRT” OR “SABR” from all years to March 1st, 2023. These studies were introduced to COVIDENCE® to manage duplicates and subsequently screen the titles and full texts for inclusion, independently, by 4 reviewers (MK, CH, NS, AM). References were included according to the following criteria: 1) treatment of EOC metastases in an oligometastatic/oligorecurrent setting 2) reporting PFS specifically for patients treated for ovarian metastases and 3) using SBRT as a treatment technique. All types of publication except grey literature were included provided they were in English. Articles reporting on outcomes of SBRT in central nervous system EOC metastases and/or SBRT with a palliative intent were excluded. The study was registered in PROSPERO (ID CRD42023412324).

Data extraction and analysis

Two reviewers undertook data collection. The data source included the number of patients, age and BRCA status. Regarding systemic treatment, the number of lines of chemotherapy regimens prior to SBRT and any concomitant treatment were documented. In terms of radiotherapy, the total dose and number of fractions, number of lesions treated, and total dose reported in Biologically Effective Dose (BED) were collected. Information on treatment response was recorded as complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Disease outcomes were collected as local control (LC), progression free survival (PFS) and overall survival (OS), if reported. Acute and late toxicity data was collected when available as well as the grading scale.

Tables were constructed to summarize the relevant information for each study. To estimate the pooled Target Lesion Response Rate (TLRR) and Target Lesion Control Rate (TLCR), the TLRR was calculated as the proportion of treated lesions showing a Complete Response (CR) or Partial Response (PR), according to RECIST criteria. The TLCR was estimated by including lesions that exhibited CR, PR, or Stable Disease (SD).

A meta-analysis was conducted to estimate the pooled median progression-free survival and the 2-year overall survival rates across studies. Given the skewed nature of time-to-event data, the median PFS and its associated confidence intervals were log-transformed prior to analysis to stabilize variance and normalize the distribution. The meta-analysis was performed using a random-effects model with the Hartung-Knapp adjustment. The Hartung-Knapp method was selected due to the small number of included studies and relatively small sample sizes, providing a more conservative estimate of between-study variability. The pooled results were then exponentiated back to the original scale in the case of median PFS for ease of interpretation in the forest plot. All these estimations were done using R version 4.3.3.

Publications requiring a full-text review for inclusion were assessed for bias using the Risk of Bias in Non-Randomised Studies of Interventions (ROBINS-I) protocol.[13].

Results

Eligible studies

The article selection is depicted in the PRISMA flow diagram (Fig. 1). In total 933 articles were assessed for abstract screening, and 30 for full text review. Consequently, nine articles were selected, as they evaluated the PFS in patients with oligorecurrent metastatic EOC treated with SBRT. As shown in Table 1 there were five retrospective studies [[14], [15], [16], [17], [18]], three case reports [[19], [20], [21]], one phase I clinical trial [22]. The retrospective study by Onal et al. included ovarian and cervical cancer metastases treated with SBRT. This study was included in the analysis as it provided data on the survival outcomes of the ovarian cohort (8 patients) but with limited information on the characteristics of this group, such as the number of lesions treated and the number of systemic treatments prior to SBRT [15].

Fig. 1.

PRISMA flow diagram.

Table 1.

Patient, tumour and systemic treatment characteristics, including chemotherapy and hormone therapy prior to SBRT as well as concomitantly.

Reference Year of publication	Article type	No of patients, n	Median age, years (range)	BRCA status	Median no of systemic line before RT,n (range)	Concurrent systemic treatment per lesion, n (%)	Pre SBRT-Platinum sensitivity n (%)
Lazzari et al [14] 2018	Retrospective study	82	60.4 (37.5–84.1)	Reported in 40.4 % of patients BRCA 1 mutation: 9.8 % BRCA 2 mutation: 8.5 %	3 (0 – 9)	29 (18.6)	Sensitive (PFI > 12 Mo) 36 (33) SBRT coursesPartially Sensitive (PFI 6–12 Mo) 36 (33) SBRT courses Sensitive (PFI < 6mo)37 (34) SBRT courses
Iftode et al [16] 2018	Retrospective study	26	62 (40–82)	NR	>2 = 56.8 % 2 = 43.2 %	NR	Sensitive: 26 lesions (59.1)Resistant: 18 lesions (40.9)
Hughes et al [19] 2019	Case report	1	85	NR	3	NR	NR
Onal et al [15] 2020	Retrospective study	8/29 (ovarian cancer)	50 (42–64)	NR	NR	NR	NR
Trippa et al [21] 2020	Case report	1	58	Negative	3	No	8 SBRT courses (100)
Macchia et al [18] 2020	Retrospective study	261	60 (28–85)	NR	2 (1 – 11)	No	NR
Kowalchuk et al [17] 2020	Retrospective study	35	62.8 (32.7–80.6)	BRCA 1 mutation: 14 %	NR	6 (6)	NR
Yavas et al [20] 2022	Case report	2	63.5 (57–70)	NR	3.5 (3–4)	No	2 SBRT courses (100)
Henke et al [22] 2022	Phase I prospective study	10	67.5 (46–80)	NR	2.5 (1–6)	No	NR

Abbreviations: Mo, months; NR, not reported; PFI, platinum free interval; RT, radiotherapy; SBRT, stereotactic body radiotherapy.

The risk of bias as per the ROBINS-1 tool, generated an overall risk ranging from moderate to serious (supplementary data).

Patient and treatment characteristics

The nine articles included 426 patients with a median age of 61.2 years (range 28–85 years). The patient characteristics and their treatment history are summarised in Table 1. The BRCA status was reported in three studies [14,17,21], one of them being a case report where the patient was BRCA1/2 wildtype [21], one retrospective study where BRCA status was known for 40.4 % of the population, of which 18.3 % harboured BRCA1/2 mutations [14], while in the third study by Kowalchuk 14 % were BRCA1 mutation carriers [17]. There was no further information if these were somatic or germline mutations and no subgroup analysis with regards to disease outcomes following SBRT.

Seven studies [14,16,[18], [19], [20], [21], [22]], reported that systemic treatment was administered prior to SBRT with a median of three (range 0–11) treatment lines of chemotherapy prior to irradiation. Concomitant systemic treatment was delivered in two retrospective studies [14,17]. Kowalchuk et al. reported that chemotherapy was administered in six cases [17] while Lazzari et al. reported that 9 (5.8 %) lesions were treated with concomitant chemotherapy and 20 (12.8 %) with hormone therapy [14], without mentioning the regimens used in any of the studies. No concurrent systemic therapy was used in four studies [18,[20], [21], [22]] (Table 1). SBRT was used in platinum-sensitive setting in 4 studies [14,16,20,21].

Treatment characteristics and efficacy

In total, 734 lesions were treated with SBRT. Of note, Onal et al reported that eight patients with at least one lesion each had been treated with SBRT [15]. Median dose was 40 Gy (range 5–75 Gy) and the median BED_α/β10 delivered was 55.1 Gy (range 7.5–262.5 Gy). In addition, Macchia et al. treated 53 (11.8 %) lesions by single fraction radiotherapy (SRS) delivering a median dose of 24 Gy (range 5–30) and a median BED_α/β10 81.6 (range 7.5–120), while the other 396 (88.2 %) lesions were treated with SBRT with a median dose of 27 Gy (range, 18–75) and a median BED_α/β10 48 (range, 28–262.5). The reported outcomes were based on the entirety of lesions treated with both techniques [18].

In all studies, most patients were treated for a single lesion only, while in five studies there were also multiple lesions treated with SBRT either concurrent or metachronous (Table 2). None of studies report that the SBRT treatments included a re-irradiation of a previously irradiated lesion. However, Kowalchuk et al reports an adverse event occurring following SBRT on a “site” that had been irradiated earlier by SBRT [17].

Table 2.

Summary of treatment characteristics, local control and overall survival rates and toxicity outcomes following SBRT in oligometastases of ovarian cancer.

Reference Year of publication	No of patients, n	No of lesions, n	No of lesions/patient, n (%)	SBRT location (%)	Total dose/No fractions	Median BED (Gy) (range)	Median follow up months (range)	Toxicity Grade Scale	LC	PFS	OS
Lazzari et al [14] 2018	82	156	1: 40(49)2: 23 (28)3: 11 (13.4) ≥4: 8(9.6)	Ln (72 %)Soft tissue (10 %) Lung (9 %)Liver (9 %)	24–30 Gy/3–5 fx	43.2 Gy (28–112.5)	17.4 (2.2–51.4)	G1 − 2 30 % CTCAE v4	2y 68 %	2y 18 %	2y 71 %
Iftode et al [16] 2018	26	44	Median 1 lesion/patient (range 1–4)	Ln (64 %)Lung (4 %)Liver (32 %)	36–75 Gy/3–8 fx	NR	28.5 (6–86)	G2 11 % CTCAE v4	2y 92.9 %	2y 38 %	2y 92.7 %
Hughes et al [19] 2019	1	1	1	Spleen	50 Gy/10 fx	75 Gy	12	G2 Acute (fatigue) No late CTCAE v5	1y 100 %	1 yr 100 %	1y 100 %
Onal et al [15] 2020	8/29 (ovarian cancer)	NR	NR	Ln (62 %)Liver (37 %)Bone (1 %)	14–54 Gy/2–5 fx	72 Gy (23.8–151.2)	23.2 (9.3–95.2)	G2 50 % CTCAE v4	NR	2y 13 %	2y 83 %
Trippa et al [21] 2020	1	8	8	Ln (100 %)	30–40 Gy/3–5 fx	72 Gy (59.5–72)	108	No Acute/Late	100 % (at last FU)	2y 0 %	100 % (at last FU)
Macchia et al [18] 2020	261	449	1: 146 (55.9)2: 70 (26.8)3: 28 (10.7)4: 9 (3.4)5: 6 (2.3) 6–7: 2 (0.8)*	Ln (65 %)Parenchyma (35 %)	5–75 Gy/1–13 fx	50.7 Gy (7.5–262.5)	22 (3–120)	G1 − 2 20.7 % Acute G1 − 2 6.1 % (late) RTOG, EORTC/ CTCAE	2y 81.9 %	2y 15.4 %	2y 73.6 %
Kowalchuk et al [17] 2020	35	98	1: 17 (49)2: 6 (17)3: 2 (6)4: 3 (9)≥5: 7 (20)	LN (52 %)Spleen (2)Liver (20 %)Lung (2 %)Bone (2 %)	12–40 Gy/3–6 fx	38.40 Gy (16.80–84.38)	33.7 (NR)	G1 22 cases G2 5 cases Acute toxicity G5 1 case Late GI bleed CTCAE v5	2y 80 %	2y 12 %	2y 60 %
Yavas et al [20] 2022	2	2	1	Parenchyma (100 %)	30 Gy/5 fx	48 Gy	16 (12–20)	0 % Acute/Late	1y 100 %	1y 100 %	1y 100 %
Henke et al [22] 2022	10	17	1	LN (70.6 %)Soft tissue (29.4 %)	35 Gy/7 fx	59.5 Gy	17.1 (5.3–27.3)	G3 1 case Acute (duodenal ulcer) No ≥ G3 late toxicity CTCAE v4	1y 94 %	NR	1y 80 % estimated

Abbreviations: BED biologically effective dose; CTCAE Common Terminology Criteria for Adverse Events; G grade; LC local control; LN lymph node; NR not reported; OS overall survival; SBRT stereotactic body radiotherapy.

(*) Metachronous lesions.

Only three studies reported the median PFS for their respective cohorts (14, 20, 22). Onal et al. also reported a median PFS, accompanied by its interquartile range (IQR) [15]. However, due to the limited sample size and the lack of a normal distribution assumption, this study was not suitable for inclusion in the meta-analysis. Consequently, it was excluded from the pooled estimate. The pooled median PFS for the three included studies was 10.3 months (95 % CI: 2.79–38.26; (Fig. 2). However, substantial heterogeneity was observed among the studies (Higgins I² = 83.2 %; p = 0.0026) for which caution should be exercised when interpreting these results. Three studies reported a median LC of 100 % (range, 94–100 %) at 12 months [19,20,22] while four studies reported an actuarial median LC of 81 % (range, 68–93 %) at 2 years [14,[16], [17], [18]] (Supplementary data). In the case report by Trippa et al, all eight lymph nodes metastasis treated over 9 years had completely responded to SBRT with evidence of outside field progression (pleural and lung metastases) at the last follow up treated with salvage chemotherapy[21]. Iftode et al. reported two (7.7 %) patients having experienced isolated local progression, one (3.8 %) with both local and distant progression while half of the patients had a distant progression [16]. Lazzari et al. reported that at the last follow up, 47 of 78 lesions (60 %) maintained the locoregional control, whereas disease progression was only regional in seven of 31 cases (22 %)[14]. Twenty-four (78 %) cases were restaged as having both regional and extra-regional progression. Distant relapse was also the most common type of recurrence (39 %) in the Kowalchuk et al. series, followed by regional (32 %) and local failures (17 %) [17]. The median objective response rate, was 83.5 % and the median clinical benefit rate, was 98.2 % (Table 3).

Fig. 2.

Forest plot summarizing the meta-analysis of median progression-free survival across included studies. The plot displays the study-specific effect sizes (Median PFS in months) with 95% confidence intervals (CIs) and random effects model (Hartung-Knapp method) pooled estimate. Weights are shown for each study based on the random effects model. Heterogeneity was assessed using Higgins I².

Table 3.

Summarising radiologic and/or FDG PET-CT response rates in lesions treated with SBRT only (i.e. no concomitant systemic therapy), progression free survival and factors affecting the local control and radiologically complete response rates.

Reference Year of publication	Median PFS (months)	Radiologic response lesions n (%)	PFS	ORR (CR + PR) CB (ORR + SD)	Factors associated with improved probability of CR rate	Factors associated with LC rate
Lazzari et al [14] 2018	5.6 (95 % CI 3.5–10.6)	CR 76 (61.8 %)PR 17 (13.8 %)SD 19 (15.5 %)PD 11 (8.9 %)	1y 34 % 3y 8 %, LPFS: 1y 82 % 3y 55 %	ORR 75.6 % CB 91.1 %		Stage III-IV vs I-II P = 0.0008; HR, 7.49; 95 % CI, 2.32–24.16);No of treated lesions (P = 0.0002; HR, 2.02; 95 % CI, 1.39–2.92);LN vs distant metastasis (P = 0.0008; HR, 3.37; 95 % CI, 1.66–6.84) Time from last chemotherapy > 7 months (P = 0.02; HR 2.39; 95 % CI 1.18–4.84)
Iftode et al [16] 2018	19	CR 26 (59.1 %)PR 9 (20.5 %) SD 6 (13.6 %)PD 3 (6.8 %)	1y 69.3 % 2y 38 % 5y 19 %	ORR 79.6 % CB 93.2 %		No factors associated:Platinum sensitivity (P = 0.78) No of lines of chemotherapy before SBRT (P = 0.45)No of lesions treated (P = 0.70)
Hughes et al [19] 2019		PR 100 %	1y 100 %
Onal et al [15] 2020	19.4 (range 7 – 150.1)	CR 5 (62.5 %)PR 2 (25 %)SD 1 (12.5)	2y 13 %	ORR 87.5 % CB 100 %	NR	NR
Trippa et al [21] 2020	NR	CR 100 %	NR	ORR 100 % CB 100 %	NR	NR
Macchia et al [18] 2020		CR 291 (65.2 %)PR 106 (23.8 %)SD 33 (7.4 %)PD 16 (3.6 %)	Outside field 2y 15.4 % 3y 12.7 %	ORR 89 %, CB 96.4 %	Multivariate analysis: Age ≤ 60 years (P = 0.027) PTV ≤ 18 cm3 (P = 0.005) Lymph node disease (P < 0.001) BED α/β10 > 70 Gy (P = 0.006)	Multivariate analysis. Achievement of CR (P < 0.001) Total dose > 25 Gy (P = 0.011) Univariate analysisOlder age (P = 0.048) Total dose > 25 Gy (P = 0.001)Achievement of CR (P < 0.001) BEDα/β10 ≥ 70 Gy (P = 0.023)
Kowalchuk et al [17] 2020		CR 52 (53 %)PR 20 (21 %)SD 12 (12 %)PD 1 (1 %)	2y 12 %	ORR 74 % CB 86 %	Treatment for recurrence (HR = 4.95)Higher BED (HR = 1.03)	BED ≤ 35 (p = 0.017)Treatment for recurrence (p = 0.029), Higher pre-treatment SUV (p = 0.026). Treatment of liver metastases (p = 0.025).
Yavas et al [20] 2022	16 (95 % CI 12–20)		1y 100 %	CB 100 %
Henke et al [22] 2022	10.9 (95 % CI 6.8–17.4)			CB 100 % at 3 months’ follow-up

Abbreviations: BED, biologically effective dose; CB, clinical benefit; SFI, systemic-therapy–free interval; CR, complete response; ORR, objective response rate; PD, progression of disease; PR, partial response; PFS, progression free survival; SD, stable disease.

The median time to salvage therapy after only SBRT treatment (i.e., no concomitant systemic therapy) was 7.4 months (range, 2.1 – 49.3) in the cohort by Lazzari et al. [14], 4.3 months in the Kowalchuk series [17], while Henke et al. the estimated the median systemic-therapy − free survival was 11.5 months (95 % CI, 8.2–17.4) [22]. The type of salvage treatment following a recurrence after SBRT is not reported.

The pooled 2-year overall survival rate from five studies was 73 % (95 % CI: 67 % to 79 %), indicating moderate and non-significant heterogeneity among the studies (Higgins I² = 50.1 %; p = 0.09) (Fig. 3).Radiologic response rate has been reported in 7 studies (14–19, 21) showing a complete or partial response in more than 74 % of the cases treated with SBRT (Supplementary data). Pooled TLRR was 86 % with 613 out of 713 evaluable lesions presenting with either CR or PR. This was improved in terms of TLCR with up to 95.9 % treated lesions (684/713) reaching either CR, PR or SD (Table 4). As summarised in Table 3, Macchia et al. was the only study to identify potential factors contributing to an improved CR rate [18]. They showed in a multivariate analysis that younger patients (≤60 years), smaller planning target volume (PTV ≤ 18 cm3), lymph node disease and biologically effective dose α/β10 > 70 Gy were associated with higher chance of CR. They also showed in a multivariate analysis that the two factors favouring a better LC were a CR (P < 0.001) and a total dose of > 25 Gy (P = 0.011). Lazzari et al. reported that in-field relapse risk was significantly associated with higher stage at diagnosis (III-IV vs I-II,P = 0.0008; HR, 7.49; 95 % CI, 2.32–24.16) [14]; with number of treated lesions (P = 0.0002; HR, 2.02; 95 % CI, 1.39–2.92); and stage of SBRT-treated lesion (distant metastasis including non-locoregional LNs versus regional LNs (P = 0.0008; HR, 3.37; 95 % CI, 1.66–6.84)). On the contrary, time interval from last chemotherapy > 7 months was associated with better LC (P = 0.02). Iftode et al. also tried to identify factors impacting on LC rate but none of the factors examined were statistically significant on Cox proportional hazard regression (platinum sensitivity [P = 0.78], number of lines of chemotherapy before SBRT [P = 0.45], number of lesions treated [P = 0.70]) [16]. Kowalchuk et al. reported potential predictive factors of local failure including BED ≤ 35 (p = 0.017), treatment for recurrence (p = 0.029), higher pre-treatment SUV (p = 0.026) and treatment of liver metastases (p = 0.025) (Table 3) [17].

Fig. 3.

Forest plot summarizing the meta-analysis of 2-year overall survival rates across included studies. The plot displays the study-specific proportions with their corresponding 95% confidence intervals (CIs), alongside the random effects model pooled estimate (Hartung-Knapp method). Weights for each study, as derived from the random effects model, are implicitly incorporated into the plot. Heterogeneity across studies is quantified using Higgins I², indicating moderate heterogeneity.

Table 4.

Summary of target lesion response rate (TLRR) and target lesion control rate (TLCR) from included studies displaying the study-specific response and control rates for target lesions, expressed as percentages (n/N).

Study	N	Number of radiated metastases	TLRR	TLCR
Lazzari et al., 2018 [14]	82	123	75.6 % (93/123)	91 % (112/123)
Iftode et al., 2018 [16]	26	44	79.6 % (35/44)	93.2 % (41/44)
Hughes et al., 2019 [19]	1	1	100 % (1/1)	100 % (1/1)
Onal et al., 2019 [15]	8	8	87.5 % (7/8)	100 % (8/8)
Trippa et al., 2020 [21]	1	8	100 % (8/8)	100 % (8/8)
Macchia et al., 2020 [18]	261	444	89.4 % (397/444)	96.8 % (430/444)
Kowalchuk et al., 2020 [17]	35	85	84.7 % (72/85)	99.8 % (84/85)
Yavas et al., 2022 [20]	2	2	NR	NR
Henke et al., 2022 [22]	10	17	NR	NR
Total	426	734	86 % (613/713)	95.9 % (684/713)

Abbreviations: Target lesion response rate, TLRR; Target lesion control rate, TLCR; Not reported, NR; Number of patients, N.

Toxicity

In all 9 articles SBRT was generally associated with mild acute and late toxicity, with mostly grade ≤ 2 adverse events reported, even in patients having up to eight lesions treated in the course of their disease (Table 2). There were only two serious adverse events described in the whole series. One grade 3 and one grade 5 events of acute and late toxicity respectively, in two patients with a single lesion treated at a time. The acute grade 3 event (duodenal ulcer) occurred on a periportal lymph node target, with no bevacizumab administered prior or after SBRT [22]. The late grade 5 event was a radiation related toxicity of a single lesion in porta hepatis treated with 18 Gy in 3 fractions a month following an earlier SBRT in porta hepatis, delivering 21 Gy in 3 fractions. In this case, it was not reported if concomitant systemic therapy was administered [17].

Discussion

The rational of this work relies on the increasing role of SBRT in the management of oligorecurrent and oligoprogressive EOC, emphasizing its impact on PFS. To our knowledge, this is the first systematic review and metanalysis focused on recurrent oligo-EOC patients treated with SBRT.

In our study, we observed a pooled median progression-free survival of 10.32 months (95 % CI: 2.79 to 38.26) and a 2-year overall survival rate of 73 % (95 % CI: 67 % to 79 %). Nonetheless, comparison of the results of these studies remains challenging considering heterogeneity of available data and the high risk of bias of included studies (confounding, selection and outcomes reported bias) (Supplementary data). In 2020, Yegya-Raman et al. conducted a systematic review highlighting the evolving significance of SBRT across various oligometastatic gynecological malignancies settings (primarily ovarian). Their study comprehensively analyzed existing data concerning LC rates, PFS, OS as well as acute and late toxicities associated with treatment. They showed that SBRT was associated with a favorable local control (71–100 %) and low-rate toxicities (no grade ≥ 3), although they observed high rates of out-of-field progression (79 to 100 % of failures) with heterogenous PFS [8]. Despite the favorable LC rate, these results varied between the studies as out-field- progression was relatively common. Lazzari et al. also identified that pattern of failure was mainly out-field with 78 % of cases restaged as having both regional and extra-regional progression [14]. This pattern of progression is not only common to ovarian cancer since 97 % of progression events in the SBRT arm of the SABR-COMET trial had an out-field component [11]. For instance, distant relapse was the most common type of recurrence in the Kowalchuk et al. series accounting for 39 % of failure patterns followed by regional and local recurrences in 32 % and 17 % of the cases respectively [17].

In our work, the pooled response rate of metastatic lesions was 86 % and appeared relatively high after SBRT. These results are consistent with those from several recent and large oligometastatic studies in other cancer histologies [11,23,24]. However, there is no consensus regarding the use of SBRT in patients with oligometastatic EOC. Inconsistencies in outcomes have been documented, attributed to the diverse spectrum of RT doses (ranging from 5 to 75 Gy) and fractionation (ranging from 3 to 13 fractions), as well as the heterogeneity in the locations of metastases.

Predictive factors for treatment outcomes

We have identified several factors predictive of enhanced LC or radiographic response (when used as a surrogate for LC), documented by 3 studies, such as younger age (< 60 years old), smaller lesion size (PTV ≤ 18 cm³)[18], radiation dose escalation (BED _α/β10 > 70 Gy) [18] and lymph node disease [14,[16], [17], [18]]. Other studies have already documented the favorable influence of younger age (< 65 years old), patient performance status (Karnofsky performance status > 70), slow-growing disease and low disease burden in predicting CR [25,26].

Regarding the higher responsiveness in LNs compared to parenchyma, these data have also been reported with encouraging results in a variety of primary cancers [27,28]. Indeed, isolated LN recurrence rates fluctuate in the international literature, ranging from 12 % to 37 % [29], especially since their progression tends to be less aggressive compared to relapses in other sites, leading to a median post-recurrence survival of about 37 months and an OS of 109 months regarding Legge et al, in a cohort of 32 EOC patients [30]. Trippa et al. evaluated the outcomes of 11 patients treated with stereotactic body radiotherapy (SBRT) targeting 20 lymph nodes (LNs) affected by recurrent ovarian cancer. LN recurrence was confirmed using PET-CT, and treatment response was assessed with PERCIST criteria. At 3 months post-SBRT, PET-CT scans demonstrated a complete metabolic response in all treated lesions. The local control (LC) rates were reported as 74 % (+/- 12) at 2 years and 48 % (+/- 20) at 5 years, with a median LC duration of 57 months. In terms of survival outcomes, the overall survival (OS) rates at 2 and 5 years were favorable, reaching 78 % (+/-14) and 65 % (+/- 16), respectively. Furthermore, the study highlighted a higher LC rate in patients with supradiaphragmatic LN recurrence compared to those with subdiaphragmatic disease. This disparity may be attributed to the presence of synchronous microscopic peritoneal disease in subdiaphragmatic regions, which is not detectable by PET-CT [31].

Kowalchuck and al. also identified predictive factors of local failure such as insufficient radiation dose escalation (BED _α/β10 < 35 Gy) and higher pre-treatment SUV on planning PET CT [17]. The prognostic role of PET CT when considering SBRT treatments has been proved for other histologies, especially in non-small cell lung cancer in which a pre-treatment SUVmax value exceeding 3 was related to poorer survival outcomes and an increased risk of both local recurrence and distant metastasis [32]. These arguments present possibilities for monitoring patients with oligometastasis based on the probability of tumor response as well as and adjusting SBRT dose prescriptions based on SUV evaluation.

Only one author among the studies included, reported a correlation between CR and improved PFS. Indeed, Onal et al. in a cohort including 21 patients with oligometastatic cervical cancer (72 %) and 8 EOC (28 %), reported a 2 year PFS and OS of 18 % and 62 % respectively, and pointed out a significant correlation on univariate analysis between CR and improved PFS (p = 0.02) and OS (p = 0.01) [15]. Reviewing the literature, Laliscia et al. in a retrospective study including 45 oligometastatic gynaecological patients, observed a 2.37-fold higher risk of progression and a 3.60-fold higher risk of death in patients who failed to achieve a CR compared with complete responders (P = 0.04 and P = 0.03, respectively) [33]. Similarly, Mills et al. conducted a retrospective analysis of 33 patients with EOC which underwent SBRT for a total of 90 extracranial lesions, and showed that achieving CR was associated with improved OS (p = 0.001, 5-year OS 54 % vs 30 %) and improved PFS (p = 0.014, 1-year PFS 22 % vs 7 %) [34]. Besides CR rate, two manuscripts identified other predictive factors of decreased PFS or OS such as Eastern Cooperative Oncology Group (ECOG) performance status > 0 (17), higher number of pre-SBRT systemic therapy lines or presence of residual tumor at first surgery [14].

Of note, only three manuscripts reported data on systemic treatment-free interval (SFI) ranging from a median time of 4.29 months for Kowalchuk et al. [17], 7.4 months (range, 2.1–49.3) for Lazzari et al. [14] to 11.5 months (95 % CI, 8.2–17.4) for Henke et al.[22]. SBRT of these low burden diseases may enable long term control and potentially extend the interval before initiating a new systemic therapy line, even in platinum-resistant patients and particularly in patients with limited tolerance to such treatments. The effectiveness of this approach has been already shown by Ost et al. in oligometastatic prostate cancer who observed a significant postponement in the initiation of androgen deprivation systemic therapy (ADT) in favor to the metastasis directed therapy (MDT) (13 months vs 21 months, respectively)[35]. However, these findings remain hypothesis-generating regarding OOC as there was heterogeneity in patient population across these studies and it was unclear whether patients achieving extended SFI harbored more indolent disease at baseline.

The recently published MITO-RT3/RAD Phase II trial evaluated the efficacy of SBRT in patients with LN recurrences from EOC. The study demonstrated an outstanding 1-year LC rate of 97.4 %, reinforcing the role of SBRT in achieving durable LN control. Importantly, CR after SBRT was strongly associated with improved SFI, with patients achieving CR showing significantly longer median SFI of 20 months compared to 11 months for non-responders (p = 0.004). In turn, a longer SFI was identified as an independent predictor of better OS, with multivariate analysis confirming its prognostic value (HR = 0.45, p = 0.01). Additionally, CR correlated positively with PFS, as patients achieving CR had a median PFS of 18 months versus 9 months for those without CR (p = 0.002). Another key finding was the relationship between CR and the PTV size, where smaller PTVs (<10 cc) yielded higher CR rates (88 %) compared to larger PTVs (>10 cc: 62 %, p = 0.03). Multivariate analysis further validated PTV size and SFI as independent predictors of PFS and OS. These results emphasize the importance of achieving CR and maintaining a longer SFI in improving survival outcomes. This Phase II trial further strengthens the findings presented in our systematic review, providing robust evidence of the potential benefits of SBRT in this clinical setting [36].

Finally, a prospective multi-institutional phase III trial is currently underway to evaluate the addition of SBRT to standard-of-care (SOC) therapy for recurrent ovarian cancer. The results of this highly anticipated study are expected to provide critical insights into the potential role of SBRT in improving outcomes for patients with recurrent disease [37]. Additionally, future phase III trials are eagerly awaited to establish standardized patient selection criteria, optimize fractionation schedules, and refine treatment timing, ultimately identifying the subgroups of patients who would benefit most from SBRT. If successful, these trials could redefine the therapeutic landscape by establishing SBRT as an integral component of management in this setting.

Combining SBRT and systemic therapy

The perception of radiation therapy only as a localized treatment has undergone significant transformation in recent years notably the low toxicity rates recorded with SBRT providing the basis for testing rational combinations of SBRT and systemic treatments. Kunos et al. reported data in a phase 1 trial that enrolled 12 patients, including 7 ovarian cancers, treated with chemotherapy (carboplatin and gemcitabine) followed sequentially with SBRT and showed an acceptable toxicity profile when using a dose prescription of 24 Gy in 3 fractions, with grade 3 and 4 toxicity rates mainly chemo-related (hematologic and metabolic toxicities) [38].

Moreover, it is recognized that RT can induce a systemic immune response, providing a strong rational for combining RT with immunotherapy [39,40]. Indeed, irradiated tumors can lead to exposure of tumor associated antigens which stimulates antigen-presenting cells (APCs) or with immune checkpoint inhibitors such as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) or programmed cell death protein (PD-1) to reinforce T-cell priming and leading to immunogenic cell death [[41], [42], [43]]. Verma et al. demonstrated that concurrent administration of immune checkpoint inhibitors targeting PD-1 or PD-L1, as well as CTLA-4, with SBRT in thoracic irradiation was safe [44]. Moreover, in a phase 1 trial on 73 patients with advanced solid tumors (including 9 EOC), treated with pembrolizumab and undergoing multisite SBRT on at least two distinct metastases, toxicity rates were generally low [45]. However, when going through this systematic review, immunotherapy was either rarely administered or data on concomitant administration was sparse.

Ionizing radiation also induces both physical and biochemical damage to DNA triggering cells to initiate mechanisms for DNA repair. One mechanism includes a pathway for single strand break (SSB) known as base excision including the activation poly (ADP-ribose) polymerase (PARP). Thus, the absence of PARP-1 or PARP-2 results in hypersensitivity to ionizing radiation and highlights the synergistic effect the combination of PARP inhibitors (PARPi) and radiotherapy [46]. Several studies have pointed out the effectiveness and safety of this combination in several malignancies [47,48] as well as in gynecological malignancies [49]. PARP inhibitors are now indicated as maintenance therapy for patients who respond to first-line platinum-based chemotherapy, with or without bevacizumab. They offer a particularly significant benefit for patients with BRCA mutations or those with homologous recombination deficiency (HRD) status [50,51]. Furthermore, multiple studies have demonstrated that disease progression on PARPi is often oligoprogressive [52,53] providing rational arguments for studying combinations of local therapy and PARPi in EOC.

In this context, the role of SBRT in managing oligoprogressive disease during PARP inhibitor therapy has emerged as a critical area of investigation. In a multicenter retrospective study involving 46 EOC patients who received SBRT and continued PARP inhibitor therapy for oligoprogression during PARPi maintenance, the authors reported a median interval of 12.4 months (95 % CI 8.3 to 19.5) before the need for subsequent systemic therapy, with a low incidence of toxicity [54]. Additionally, in a multi-center retrospective study (n = 74) continuation of PARPi following local therapy (mainly radiotherapy and surgery), reported a PFS of 11.5 months (95 % CI 7.4–17.2) [55]. Recently, the Epimetheo retrospective study evaluated the efficacy and safety of SBRT in 45 patients with EOC undergoing PARPi maintenance. A total of 68 oligoprogressive lesions were treated with SBRT while PARPi therapy was continued. The study demonstrated a 1-year LC rate of 90 % and a median PFS of 14 months, highlighting the effectiveness of SBRT in managing oligoprogressive disease. Notably, SBRT significantly delayed the need for a change in systemic therapy, with a median next systemic treatment change-free survival (NEST-FS) of 12 months, underscoring its role in preserving the benefits of ongoing PARPi maintenance. Additionally, the combination of SBRT and PARPi therapy was associated with improved systemic disease control, suggesting a potential synergistic effect. The toxicity profile was favorable, with no grade ≥ 3 adverse events reported and only transient grade 1–2 gastrointestinal and hematological toxicities in 20 % of patients. This study highlights SBRT as a valuable tool to extend the duration of effective PARPi maintenance in EOC, maintaining disease control while minimizing treatment-related toxicities [56]. This approach is currently under evaluation in the ongoing phase 2 non-comparative SOPRANO trial, which is assessing the use of stereotactic radiotherapy alone or in combination with subsequent niraparib for oligometastatic EOC following prior PARPi therapy (NCT05990192).

Brain metastasis

The incidence of brain metastases (BM) from EOC was approximately 1–3 % [57,58]. However, systemic agents cross the blood–brain barrier poorly which can lead to a significant increase in BM and consequently, a low survival rate (7 months) [59]. A variety of treatment modalities is currently available, including, resection, SRS, whole brain radiotherapy, supportive care [60]. In this systematic review of literature, we excluded studies treating on BM because of the different radiation doses delivered to metastases with SRS, the low impact of systemic treatments in central nervous system and thus the poor prognosis in these patients.

Limitations

There are several limitations to this systematic review. First, most included studies were retrospective, with case reports and only one phase I trial, generating an overall moderate to serious risk of bias of the studies included within this review (Supplementary data) [13]. A key limitation of this study was that some measures had to be estimated in the data extraction from reviewed articles. As retrospective studies and cases reports varied in style and structure, some did not include an explicit number of months between the radiotherapy treatment and the introduction of a new systemic line but rather implied a general timeline. This posed challenges in determining the true median time between SBRT and the introduction of a new systemic line, making it difficult to generalize data and draw quantitative conclusion for those parameters. All these biases could have negatively affected the external validity of these studies. Due to the limited sample size, lack of accuracy and data variability, high heterogeneity was observed among the pooled studies, making difficult to obtain robust results suitable for generalization. Despite these significant limitations, the observed results hold potential clinical value.

Conclusion

SBRT for EOC demonstrates promising response rates, durable local control, and the potential to extend systemic therapy-free intervals, all while maintaining a favorable toxicity profile.

This approach challenges the traditional view of radiotherapy as solely palliative in this context, supporting its integration into the therapeutic arsenal for selected patients with oligometastatic EOC.

Author contributions

Study concept and design: Constance Huck, Melpomeni Kountouri, Gloria Marquina, Ainhoa Madariaga, Noelia Sanmamed.

Acquisition of data: Manuel Espantaleon, Constance Huck, Melpomeni Kountouri, Noelia Sanmamed.

Analysis and interpretation of data: Constance Huck, Melpomeni Kountouri, Ramon Yarza.

Drafting of the manuscript: Constance Huck, Melpomeni Kountouri, Gloria Marquina, Ramon Yarza.

Critical revision of the manuscript for important intellectual content: Constance Huck, Melpomeni Kountouri, Gloria Marquina, Manuel Espantaleon, Ainhoa Madariaga, Judith Kroep, Fernanda Herrera, Noelia Sanmamed, Ramon Yarza.

Statistical analysis: Ramon Yarza.

Obtaining funding: None.

Administrative, technical, or material support: None.

Supervision: Ainhoa Madariaga, Noelia Sanmamed.

All authors contributed, read, and approved the final manuscript.

Funding statement

None.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: