Hypofractionated Protontherapy in Chordomas and Chondrosarcomas of the Skull Base

Chordomas are rare, slow-growing tumors that develop from remnants of the embryonic notochord in the clivus, sacrococcygeal region, and mobile spine. Although the frequency of distant metastases is low, these tumors are locally aggressive and have an extremely high local recurrence rate. Similarly, chondrosarcomas have a potential for slow growth with a tendency to recur locally. They usually arise at the base of the skull or spine from mesenchymal cells or from the primitive cartilaginous matrix.

Although chordomas are considered clinically more aggressive than chondrosarcomas, their tendency to settle in similar anatomical locations and the high risk of local recurrence of both diagnostic entities have conditioned a similar therapeutic approach.

Standard treatment includes surgical resection that is as radical as possible; however, complete resection is feasible in less than 50% of cases, as it can be associated with significant postoperative morbidity and mortality, since these tumors frequently invade or contact critical structures (vascular, cranial nerves or spinal roots). Therefore, adjuvant or exclusive irradiation have a fundamental role in long-term local control. Therefore, optimization of the efficacy of radiotherapy represents a critical step in the management of these patients. Given their tendency to local recurrence, chordomas require the prescription of high doses for their local control, which are associated with potentially critical adverse effects if conventional photon irradiation techniques are used.

The α/β ratio, according to the linear-quadratic model, represents a measure of the sensitivity of a tumor to variable dose-per-fraction regimens. Tumors with a low coefficient (<4 Gy) are considered more sensitive to the effects of hypofractionated treatments, which involve the administration of higher irradiation doses per session in fewer sessions.

After analyzing historical studies and institutional experiences, different publications suggest that the α/β ratio of chordomas is 2.45 Gy and it is assumed to be very similar for chondrosarcomas. Therefore, the administration of hypofractionated schedules could be associated with an increase in the sensitivity of these tumors to radiotherapy treatment.

Conventional normofractionated radiation therapy administered adjuvantly after resection has historically been used with median doses of 60 to 66.6 Gy to 2 Gy per fraction, offering 5-year local control rates ranging from 23% to 50%. Technological advances in the field of image-guided intensity-modulated radiotherapy have led to an improvement in the precision of treatments, allowing the dose for chordomas to be scaled up to 76 Gy (the biological equivalent dose (BED) taking into account the α/ β of 2.45Gy is 138 Gy) and for chondrosarcomas up to 70 Gy (BED of 127 Gy), achieving and improvement in 5-year local control rates of 65% and 88%, respectively for each diagnostic entity.

Although promising, the use of high-dose photons is limited by the lower ability to protect nearby critical organs. Therefore, particle therapy (protons and carbon ions) has established its role as a standard technique, due to its potential to achieve greater conformation and better dose distribution. The main studies with proton therapy for these tumors appear to show statistically significant results in favor of increased survival when dose escalation above 70 Gy to 2 Gy per fraction (BED >127 Gy) is compared with techniques using irradiation with intensity modulated photons (IMRT). The 5-year local control rate is above 60% for chordomas and above 80% for chondrosarcomas with moderate toxicity and preservation of critical structures. However, the limited availability of these facilities together with the administration of very long treatment schedules, often of seven weeks or more, poses a significant problem for patients and remains an obstacle to the widespread adoption of this technique as a therapeutic standard.

Due to all these limitations and the important advances in terms of precision and dose distribution, the concept of hypofractionation has gained weight within radiation oncology thanks to its potential benefits in terms of reducing the duration of treatment and costs. The first publications on the hypofractionated treatment of chordomas, mainly of the skull base, go hand in hand with photon radiosurgery systems, using dedicated equipment such as the GammaKnife or CyberKnife. In the last 15 years, single-dose or hypofractionated treatment schemes have been explored as a therapeutic alternative to escalate the dose, improve the protection of organs at risk and reduce treatment time.

Some studies have evaluated single fraction stereotactic radiosurgery (SRS) in the management of chordomas and chondrosarcomas mainly located in the skull base with results comparable to proton therapy. The most relevant studies included series of 22 to 71 patients treated with median doses of SRS ranging from 12.7-24 Gy (BED 78.5-259 Gy). Five-year local control rates ranged from 21-85%, depending on the doses prescribed (higher doses, ≥15 Gy, were associated with better relapse-free survival).

However, based on the experience of Kano et al., for single-dose treatments, irradiation of volumes > 7 cc is associated with a significant worsening of tumor control. It is essential to emphasize that the tumor volumes that are usually treated in chordomas and chondrosarcomas, both at the base of the skull and, to a greater extent, in the spine, exceed, in most cases, 7 cc, since the entire clivus or affected vertebral bodies must be included in the majority of cases, in order to reduce the risk of marginal recurrence. For this reason, the role of single-dose radiosurgery loses weight in favor of hypofractionated stereotactic radiotherapy (HFSRT), which has theoretical advantages compared to single-dose treatment in volumes greater than 7 cc, including a lower risk of radiation-induced toxicity in nearby critical structures and the possibility of safely treating larger tumor volumes with multiple fractions (usually 5).

Several publications evaluate HFSRT for chordomas and chondrosarcomas. The number of cases included in these series ranged from 9 to 24 patients. The median follow-up was 24 to 46 months. Most patients were treated with 5 fractions with a prescription dose of 24-43 Gy (BED 52.7-194 Gy), depending on histology and therapeutic setting (radical, adjuvant, or reirradiation). The best local control results at 3 and 5 years obtained were 90 and 60% for chordomas, respectively, and 100% for chondrosarcomas. The most widely used regimen was 37.5 Gy in 5 fractions of 7.5 Gy (BED 152.3 Gy equivalent to 80 Gy at 2 Gy per fraction) for chordomas and 35 Gy in 5 fractions of 7 Gy per fraction in Chondrosarcomas (BED of 135 Gy equivalent to 74 Gy at 2 Gy per fraction). The toxicity described in most of these studies does not register worse data than those published with conventional fractionation in proton therapy, when it comes to primary treatments.

This growing evidence, which is described as a justification for the implementation of hypofractionated regimens, demonstrates that the standard implementation of these therapeutic modalities in patients who meet the appropriate characteristics can suppose a great advantage to improve accessibility and comfort for patients, potentially reduce acute side effects during treatment and increase therapeutic cost-efficiency.

Proton therapy, today, remains a limited resource, with only 99 facilities currently in operation worldwide in 2021, of which two new centers are in Spain, active since 2020. In addition, many patients must travel to access to this technology, so reducing the time a patient is away from home and their support network can have significant financial and psychosocial implications. Added to all this are the aforementioned radiobiological advantages of high doses per fraction, in tumors with a low α/β coefficient, such as chordomas and chondrosarcomas. That is why the implementation of hypofractionation within proton therapy has gained weight in the last 10 years, increasing the number of publications in this regard, which reflects the interest in this treatment approach.

Cao et al. presented a dosimetric study comparing different hypofractionated stereotactic treatment schemes in the treatment of intracranial tumors > 3 cm in greatest diameter. Treatment plans with GammaKnife, Cyberknife and VMAT were generated compared with proton therapy plans with or without modulated intensity. The authors suggest that proton therapy represents a desirable alternative to advanced photon techniques for treating large, irregularly shaped volumes near critical structures, such as chordomas and chondrosarcomas.

For all these reasons, a fundamental and necessary challenge today consists of increasing the scientific evidence of hypofractionated schemes in the treatment of chordomas and chondrosarcomas, adapting them to protontherapy, to increase clinical experience and combine the benefits of high-precision hypofractionated treatments to the dosimetric advantages of protontherapy (ability to treat volumes > 7 cc with a homogeneity index close to 1 and a decrease in the integral dose in healthy tissue).

In summary, considering the growing scientific evidence available from other studies on different therapeutic entities, we have a solid basis to reinforce hypofractionation protocols with protontherapy in the treatment of chordomas and chondrosarcomas.

The main advantages of these protocols, as mentioned above, are aimed at optimizing the treatment of these patients by providing potential clinical benefits derived from the advantages associated with the biological response to high doses per fraction, reducing the likelihood of acute side effects, increasing the availability and accessibility of proton therapy units to a larger number of patients, and optimizing therapeutic cost-effectiveness.

As disadvantages, these protocols are intended for a limited subgroup of patients who meet strict criteria, including small-volume lesions located at the skull base and sufficiently distant from organs at risk that are sensitive to high doses per fraction, such as the brainstem and the optic pathway, in order to avoid the development of potentially disabling chronic side effects.

Based on all of the above, we decided to design a first prospective study which, with the support and recommendations of this committee, became a clinical trial entitled: "Low-intervention Clinical Trial of Hypofractionated Proton Therapy in Skull Base Chordomas and Chondrosarcomas", approved on 26/09/2023 (PIC128_22_QUIRON). We published our initial results from the prospective series, which included 11 patients treated using the same criteria applied in the clinical trial, along with a description of the trial. Since its approval, three patients have been recruited within the clinical trial, resulting in a total of 14 patients treated under the same conditions, with a median follow-up of 28 months.

Based on the follow-up data obtained, we decided to perform an interim analysis to evaluate outcomes. Local control was 100% throughout the entire follow-up period. One patient died from complications secondary to treatments 32 months after proton therapy (reoperation for basilar artery stenosis and cerebrospinal fluid fistula following surgery). Acute tolerance was excellent in all patients; however, regarding chronic toxicity, we observed a higher number of temporal lobe necrosis cases than initially described in normofractionated proton therapy series, occurring between 8 months and 2 years after treatment administration (median of 12 months after proton therapy).

Five patients were diagnosed with radiological radionecrosis (35.7% of the series, compared with 15% reported in the literature for normofractionated regimens) (39). Of these, three presented symptoms clearly related to imaging findings: in two patients, symptoms were mild and resolved with corticosteroid therapy (grade I-II asthenia, grade I-II headache, and a single isolated seizure episode). In another patient, symptoms were more severe (seizures and memory impairment) and were associated with other post-surgical complications (basilar artery stenosis), which ultimately led to the patient's death. The factor most clearly associated with the higher incidence of radionecrosis appears to be the maximum dose received by the affected temporal lobes (greater than 34 Gy in 2% of the volume).

The greater the risk of creating "hot spots," the higher the predisposition of the local vasculature surrounding the target to increased vascular injury, one of the potential triggers of radionecrosis. Changes in microvascular permeability and loss of blood-brain barrier integrity are key features of radiation-induced injury in the central nervous system, as is the expression of vascular endothelial growth factor (VEGF). Inherent to the nature of particle beams, proton therapy exhibits greater relative biological effectiveness (RBE) than photon-based radiotherapy, which may increase the risk of radionecrosis and subsequently enhance the inflammatory response and vasculopathy. This effect is explained by a higher linear energy transfer (LET), resulting in greater local energy deposition. DNA damage increases in quantity and/or complexity, representing a greater burden on the cellular repair system. This increased efficacy of ionizing radiation is quantified as the RBE. Numerous studies have demonstrated a clear increase in RBE toward the Bragg peak, where protons stop (at the end of the beam) and where LET is comparatively high.

Histologically, radiation-induced brain necrosis is characterized by the presence of geographic eosinophilic necrosis accompanied by gemistocytic astrocytes representing gliosis with atypia. In addition, fibrinoid vascular necrosis associated with dystrophic calcification, perivascular lymphoplasmacytic infiltrates, and telangiectasias representing reactive vascular changes are also observed, although these findings are not highly specific to radiation-induced brain necrosis. There is significant interpatient variability in radiosensitivity. Initial evidence supporting a genetic basis for radiation toxicity came from studies in breast cancer patients, in whom approximately 90% of the variation in the development of radiation-induced telangiectasia was attributed to underlying genetic differences. Patients with ataxia-telangiectasia, Fanconi anemia, and Bloom syndrome-hereditary disorders involving DNA damage surveillance and repair genes-exhibit increased radiosensitivity. The diverse genetic abnormalities responsible for these disorders suggest that radiosensitivity is polygenic, involving multiple pathways, some of which are inherited. In recent years, a broader genome-wide approach has been adopted. A recent genome-wide association study revealed a variable risk of radiation-induced temporal lobe necrosis associated with different single-nucleotide polymorphisms in a glioblastoma cell line (U87) treated with X-rays and H₂O₂. This is the first study implicating a susceptibility gene for radiation injury (Cep128) and provides novel insights into the underlying mechanisms of radiation-induced brain injury.

Based on the above, we believe that the increased rates of radionecrosis have a multifactorial component. On the one hand, there may be a greater predisposition to vascular damage, potentially related to pre-existing vascular injury prior to treatment (tumoral, post-surgical, or hereditary). Furthermore, in skull base tumors such as chordomas and chondrosarcomas, the distal end of the beams is often oriented toward the temporal lobes in order to avoid more sensitive structures such as the optic pathway or the brainstem, thereby increasing LET and RBE in this location. This enhances the radiobiological effect which, together with the increased dose per fraction, could explain the higher incidence of temporal lobe necrosis.

For all these reasons, we would like to request a protocol amendment to optimize inclusion criteria and improve planning quality parameters, with the aim of reducing the risk of temporal lobe radionecrosis while allowing patients to benefit from the advantages of improved acute tolerance, reduced treatment duration, and improved local control. For patients who are not candidates for extreme hypofractionation schemes (5 fractions), we will add a moderately hypofractionated arm with an integrated boost protocol (27 fractions).

In summary, the research project is based on the performance of a phase II clinical trial with a low level of intervention, which consists of the prospective evaluation of the clinical and radiological response after the administration of a treatment using radical or adjuvant hypofractionated proton therapy in 5 or 27 sessions in patients diagnosed with chordoma or chondrosarcoma of the skull base.

In addition, a cross-sectional evaluation of the quality parameters of the dosimetry of the hypofractionated proton therapy treatments administered to the patients included in the study and an evaluation of the quality of life by means of specific questionnaires, 3 months after treatment, will be carried out.