Since SFTs of the CNS are extremely rare, there is very limited evidence regarding treatment strategies. As mentioned in the introduction, recent studies and meta analyses have shown favorable results for PORT[7,8,9,10,11]. This retrospective patterns-of-care study aimed to explore outcomes based on patient characteristics and treatment modality. To our knowledge, this analysis represents the largest cohort of patients having received RT with protons and C12 in the literature.

Nevertheless, due to the still limited number of patients and the retrospective nature of our study, definitive causal conclusions cannot be drawn. Consequently, the evidence found here can only give limited instructive guidance for clinical practice. Herein, we put our data in the context of existing literature on the disease in an attempt to deduct meaningful conclusions for the treatment of our patients.

Our survival results are mostly in line with earlier studies which already support radiotherapy as an effective adjunct treatment modality in combination with surgery upfront[6, 9,10,11,12]. Specifically Lee et al.[21] report significantly longer LC and OS rates in patients receiving PORT versus no PORT (5-year-LC-rates 97% vs 44%; HR 0.05 p = 0.002; and 10-year-OS-rates 83% vs 25%; HR 0.2 p = 0.008).

The OS rates with 96.0% 5 years after treatment of a primary SFTs are excellent and other studies report similar survival rates[10, 21, 22]. On the other hand, our analysis shows that outcome of recurrent SFTs is rather poor, with low survival times (14.0% after 3 years). Due to scarce data regarding salvage treatment of recurrent SFTs, the ideal treatment method is hitherto unclear. Our results demonstrate that salvage RT appears to be an option for retreatment, as acute toxicity seems to be tolerable without any toxicities preventing or delaying radiotherapy. Especially toxicity during re-RT was well tolerated in our patient cohort with only 48.6% reporting grade I toxicity (41.2% fatigue and headache, 35.3% erythema, etc.).

There is literature demonstrating cumulative RT doses of over 60.0 Gy to be favored to improve local control rates[25], with Ghia et al. stating superior local control with PORT with ≥ 60.0 Gy (HR 0.12 [0.01–0.95] p = 0.045)[25]. Our study underlines this finding with a lower incidence of local recurrences with RT doses ≥ 60.0 Gy, without showing significance (p = 0.390) which might be due to the limited patient numbers. Interestingly, when analyzing recurrence rate data generally, including both out-field and in-field recurrences, radiation doses ≥ 60.0 Gy had a significantly lower risk for a recurrence than radiation doses < 60.0 Gy (OR = 0.145; CI [0.029; 0.742]; p = 0.027). This, in turn, may reflect better local control and more efficient treatment of the primary tumor, as well as a lower likelihood of metastases and recurrence both within and outside the treatment field.

A central question to our study was the efficacy of different radiation modalities, specifically particle radiotherapy with carbon ions and protons. Studies focusing on these radiation modalities are very scarce and limited to case studies[26, 27].

Particle RT, especially in the primary state, seems to yield excellent response rates likely due to dose escalation feasibility. Only 1 of 10 patients treated with protons developed an out-field recurrence and the one patient treated with C12-Ions did not experience recurrence. Local control appears to be significantly superior to standard photon therapy. To our knowledge there is only one case report investigating radiotherapy with protons in a recurrence state of an orbital hemangiopericytoma [26]. The significant advantage of protons and C12-Ions compared to photons is the steeper dose gradient which allows higher dose applications both in the cases of proximity to organs at risk (i.e., chiasm, brain stem, optic tract, …) and the re-RT situation [17, 18]. This is reflected in our data, where doses of at least 60.0 Gy could be reached in a higher proportion of patients with particle RT compared to conventional photon treatment.

Similarly, there is only one case study in the recurrence situation of an extra cranial hemangiopericytoma treated with C12-RT [27]. In our study, most patients received C12-RT in the salvage setting with a 1-year LC rate of 68.6% and a median PFS of 12.0 months. Literature exploring the salvage setting mostly applies photon therapy; Sheen et al. [28] investigate SRS and report a 5 year LC rate of 76.0% and a median PFS of 21.0 months. The mean tumor volume in their study was 8.8 ml [28], whereas the mean GTV in our patients was 43.0 ml. This alone may point to a negative selection of our cohort and provide at least a possible explanation for our worse outcomes. Pretreatment of the patients was also not comparable to our study, with some patients having had received surgery only and SRS only after recurrence, some having had received SRS after initial definitive RT and some SRS after surgery and RT[28]. In our study all patients had received a surgical intervention and radiotherapy and only then re-RT with C12 or photons.

SRS is by definition only applicable in patients with small tumor volumes. C12-ions provide an option for re-RT for its higher RBE because of its higher linear energy transfer (LET) and – similar to protons – higher dose conformity with steeper dose gradients and therefore sparing of surrounding tissue[17, 27]. While our data shows a non-significant shorter median PFS of C12-RT vs photon RT in the salvage RTs of 12.25 vs 27.5 months we therefore still feel confident in at least further investigating ion radiotherapy especially in salvage situations to get a larger body of evidence for C12-RT.

As described in the backgrounds section there is general consensus that extent of resection (GTR vs STR) is a major contributor to survival times in SFT [4, 5]. This could not be confirmed in our study, both considering the extent of resection (EOR) as given by the surgeon as either resection or biopsy as well as our own assessment with pre-RT MRI imaging with either macroscopic tumor residual or none (GTR vs STR). Both did not show a significant correlation to tumor recurrence (p = 0.682; p = 0.450). A possible explanation could be our limited patient numbers. Regarding the pre-re-RT MRIs in the recurrence situation, we could identify that 11/12 patients had macroscopic tumor residuum/recurrence with 8 of them having had repeat surgery. This is not surprising since repeated surgery (especially after irradiation) is considerably more difficult and riskier to perform and has to be done more conservatively to avoid complications [29, 30]. There was no correlation between EOR in recurrent SFT regarding risk of tumor recurrence (p = 1.000). This could also be due to different tumor biology, already higher propensity for recurrence and the small patient sample size.

As with most high dose cerebral irradiations, there is a significant risk for radiation necrosis (RN) of brain tissue, especially after re-RT. Literature describes radiation induced contrast enhancement (RICE) as a term to morphologically describe said contrast enhancement in the brain tissue after RT, which can include blood–brain barrier lesions, pseudoprogression or even RN [14, 15]. Since only neurosurgery and pathology can safely distinguish between the different RICE lesions but is not always safely amenable in patients – especially after multiple surgeries already – definitive diagnosis and therefore treatment has proven to be difficult [14]. According to symptom severity antiedematous therapy with Dexamethasone is known to be alleviating of light symptoms [15] and RICE, with Bevacizumab (a VGEF-inhibitor) therapy offering a treatment option for RN [23, 24].

Over the course of re-RTs at our institution 4 patients developed RICE needing antiedematous treatment initiation. Dexamethasone as an initial treatment measure provided relief in 3 of 4 cases in our patients and symptoms didn’t prove to be severe. Bevacizumab therapy was not needed, and symptom alleviation and reduction of RICE could be achieved through Dexamethasone therapy only.

Our data suggest that women exhibit an overall higher risk of tumor recurrence than men. This finding is not yet understood. However, earlier surgical studies also show that female gender is associated with worse outcomes[22]. Potential biological or societal factors could predispose women for a worse outcome. Moving forward, we recommend that efforts should be undertaken to further examine gender differences in future studies.

Limitations to the current study include its retrospective character and the limited patient numbers as well as the heterogenous treatment techniques and applied doses. Of note, no new pathological evaluation had been performed as part of this retrospective review. This is a limitation of our study. In recent years new pathological findings have narrowed down and refined classification[1, 3]. We accepted the WHO grades without reclassification according to the then current WHO criteria. Our data supports the current consensus[8] and approaches significance in showing that WHO grade 3 SFTs have a higher risk for recurrence and especially in-field recurrences than WHO grade 2 tumors. The lack of significance is likely due to the limited patient numbers.