In this report, we re-evaluated our experience with IORT in combination with radical surgery as part of a multimodality treatment strategy for RGC. Our results demonstrated that the multimodal approach is well tolerated but still associated with poor survival. After a median follow-up of 22 months, the 5-year OS was 55%, the LRC was 56% and the DMFS was 49% (Figs. 1, 2 and 3). Jablonska et al. reported the similar results with a 14-year local control rate of 51%, DFS and OS rates of 15–20% [16]. A phase II study on perioperative HDR reported high local control at around 80% and comparable OS at 46% after 16 years in non-pre-irradiated patients [17]. In addition, the authors observed a local control rate of approx. 60% and an OS rate of 16% in ten pre-irradiated patients after 14 years [17]. In our cohort, 67% of the patients were pre-irradiated and showed a comparable LRC albeit after a short follow-up (Table 2; Fig. 2).

Kaplan-Meier curves regarding OS

Kaplan-Meier curves regarding LRC

Kaplan-Meier curves regarding DMFS

IORT as single-dose can achieve promising local control after optimal surgery and complete resection [6, 10, 18]. In our cohort, the final histopathologic findings detected in 24 patients (60%) no residual tumour, microscopic residual disease in five patients (12.5%), and unfortunately in eleven patients (27.5%) no information on resection status was available (Table 2). For this reason, we did not include the incompletely reported resection status in our univariate analysis and comparison of the survival curves.

Conversely, IORT alone cannot improve OS and local recurrence rates in an unfavourable constellation with need for extensive surgery, residual tumour, and pre-radiation [19]. In addition to IORT, postoperative EBRT can achieve a higher cumulative radiation dose locally and locoregionally and could have an oncological advantage [20], especially in patients who have not been previously irradiated. Sole et al. reported that in lymph node recurrence, the addition of EBRT to IORT and surgery can improve local control without worsening the acute and chronic toxicity profile [21]. Importantly, this combination of EBRT and IORT in lymph node recurrence showed a significant benefit in local control (HR = 4.11, p = 0.04) and disease-free survival (HR 2.76, p = 0.04) [21]. Furthermore, Sole et al. found that EBRT for the primary tumour region provided no significant benefit in this setting. Consistently, the combination of IOERT and perioperative EBRT resulted in an improvement in locoregional control, especially in the case of tumour fragmentation with R0 resection [22]. In our study group, only eleven (27.5%) patients had received adjuvant EBRT with a median dose of 47.6 (39.6–54) Gy (Table 2), but with no significant difference to the cohort without adjuvant EBRT (Table 3). Notably, in our non-pre-irradiated cohort, a curative dose of adjuvant EBRT of 47.6 (39.6–54) Gy was used, which is particularly relevant for the treatment of subclinical tumour residuals. However, in the case of full-dose pelvic pre-irradiation, Backes et al. used postoperative EBRT with a mean of 26 Gy (range 10 to 40 Gy) in addition to IORT, which is probably an insufficient dose for tumour control [19]. In fact, the negative result of Backes et al. is similar with our findings regarding adjuvant EBRT [19], but the study populations and the prescribed doses are not comparable. It is conceivable, that the proportion of patients with IORT and adjuvant EBRT in our study (27.5%) is too small and the follow-up too short to show a significant oncological benefit. The development of radioresistance in gynecological tumours is based on complex interactions [23]. The emergence of RGC is probably associated with varying degrees of radioresistance. Consequently, radioresistance may explain the inconsistency of the perioperative EBRT impact. Congruent to the above-mentioned studies, we did not observe any grade 4 toxicities (Table 4). Furthermore, in our cohort the pre-irradiated patients who also received adjuvant EBRT did not suffer from increased toxicity rates. In our study for EBRT, intensity modulated radiotherapy (IMRT) and daily image guidance were routinely used for precise and conformal delivery. IMRT reduces side effects by better sparing the surrounding organs at risk compared to conventional 3D-RT with comparable oncological outcome [24, 25]. This may explain the mild toxicity profile in our study compared to older studies prior the widespread adoption of intensity modulated RT in the community.

Particularly in the presence of local recurrence after prior radiotherapy, HDR-IORT offers local dose escalation in the tumour bed with simultaneous toxicity reduction due to the steep dose decrease in the surrounding organs at risk. Furthermore, the application of HDR-IORT enables optimal dose coverage both in the anatomically difficult to access concave and large tumour cavities (> 10 cm). In addition, the HDR-IORT dose is prescribed to the 5 mm depth from the applicator surface. This leads to a maximum dose of up to approx. 150% directly at the centre of the contact surface between the applicator and the tumour cavity. This can lead to localized necrosis and potential increased toxicity. In our cohort, univariate analysis revealed a non-significant trend in OS and LRC with (OS: HR = 3.76, CI 95%: 0.95–14.881, p = 0.059 and LRC: HR = 2.165, CI 95%: 0.916–5.114 p = 0.078) in favor of the HDR-IORT technique was observed (Table 3).

19 women (47.5%) in our cohort were treated with IOERT (Table 2). IOERT was applied to the round-oval tumour cavities in the easily accessible areas. Furthermore, the IOERT energy (range 6–15 MeV) was selected for optimal dose coverage based on the preoperative diagnosis, intraoperative situs and resection margins. Remarkably, the application of HDR-IORT in our study showed a significant benefit for OS and DMFS (p = 0.06 and p = 0.03, Additional file 1–3, Figure S1, S2 and S3).

The addition of new agents to the standard chemotherapy regimen provides promising results and underlines the individualized therapeutic approach. The phase III study showed that the combination of dostarlimab and carboplatin-paclitaxel significantly prolonged progression-free survival in locally advanced or recurrent endometrial cancer [26]. Progression-free survival at 24 months was 36.1% (95% CI, 29.3–42.9) in the dostarlimab cohort vs. 18.1% (95% CI, 13.0-23.9) the placebo cohort (hazard ratio, 0.64; 95% CI, 0.51 to 0.80; P < 0.001) [26]. Patients with mismatch repair-deficient and, microsatellite instability-high tumours had the greatest progression-free survival benefit. Remarkably, the proportion of included patients in the recurrence situation was approximately 50% [26]. In addition, dostarlimab has a robust antitumour effect with a moderate toxicity profile [27], which is a key prerequisite for patient compliance. A potential synergistic effect of combined IORT and immunotherapy should be evaluated prospectively.

Despite these concordant results, the limitations of this analysis should be noted. This retrospective study was conducted in only one institution and included a heterogeneous cohort with individualized treatment approaches. The choice of IORT dose was predominantly based on intraoperative positive frozen margins and sparing of surrounding organs. This may have led to undertreatment in the critical regions and contributed to in-field recurrence in one third of the women (Table 2). Furthermore, the development of field margin recurrence was observed in five women (12.5%) (Table 2), possibly due to underestimation of the actual extent of recurrence. As highlighted above, the choice of the appropriate applicator size for IOERT and flab size for HDR-IORT may have been incorrectly small. As a result, the cavity margins did not receive a sufficient dose for durable tumour control. In addition, only 17 patients (27.5%) were irradiated percutaneously with different doses. Therefore, in our small cohort the actual benefit of additional percutaneous dose saturation in terms of improved local control cannot be conclusively answered. This in turn limits the transferability to other patient groups outside our institution. The short follow-up in our cohort provides only limited evidence of late toxicity and oncological survival benefit occurring over a longer period.