Effect of Carbon-Ion Radiation Therapy on Maxillary Ameloblastic Fibrosarcoma

Introduction

Ameloblastic fibrosarcoma (AFS) is a rare malignant odontogenic neoplasm of the head and neck,¹ characterized by a benign epithelial component and a malignant mesenchymal component that exhibits sarcomatous features. While AFS is histologically similar to ameloblastic fibroma, its ectomesenchymal component demonstrates marked cellularity and pleomorphism. A systematic review reported that AFS typically affects younger patients, with a mean age of 27.7 ± 15.6 years, and primarily occurs in the mandible.¹ Currently, no standard treatment and clinical guidelines exist, and clinical reports are limited. The most common treatment strategy is surgery with or without radiation therapy. Radical resection was recommended based on recurrence rate.¹ The level of evidence for the efficacy of radiation therapy (RT) in cases contraindicated for surgery is low.

Recent advances in RT have increased the focus on particle therapy, which includes both proton therapy and carbon-ion radiation therapy (CIRT). CIRT is characterized by its high linear energy transfer and enhanced biological effectiveness,² which make CIRT a more potent and targeted treatment modality than conventional X rays or proton therapy.

This report presents the effectiveness of CIRT in a patient with maxillary AFS who opted against surgery. Considering the lack of detailed reports on AFS treatment using CIRT, this is an important step toward increasing treatment options. This case report has been prepared in accordance with the CAse REport guidelines.

Case Report

A 68-year-old woman patient presented to a hospital with nasal congestion, toothache with radiating pain in the left maxilla, and a tearful left eye. Maxillary sinusitis related to dental caries and benign or malignant neoplastic lesions of the maxillary sinus were considered. A biopsy was performed under general anesthesia for a suspected mass in the left maxillary sinus. The pathology results indicated AFS. Although radical surgery with removal of the entire maxilla was recommended, the patient refused surgery. The patient was referred to our hospital to undergo CIRT. Computed tomography (CT) showed a mass centered in the left maxillary sinus and resorption of the surrounding maxillary sinus bone (Fig. 1a, b). Magnetic resonance imaging (MRI) recorded 2 to 3 weeks later showed periapical invasion of the mass from the left maxillary sinus to the nasal cavity, with infiltration of the left orbital floor (Fig. 1c, d). Fluorodeoxyglucose positron emission tomography showed abnormal accumulation in the left maxillary sinus and no distant metastasis (Fig. 1e, f). This tumor grew rapidly, and the treatment-planning CT showed larger infiltration with bone destruction in the anterior maxillary sinus than the initial CT (Fig. 2). We adopted the previously reported protocol for CIRT of bone and soft-tissue sarcomas: a total dose of 70.4 Gy (relative biological effect, RBE) administered in 16 fractions.³ The XiO-N system (Elekta) was used for treatment planning. Gross tumor volume (GTV) was used to define the maxillary sinus-based tumor and was primarily delineated based on CT and MRI. The definition of clinical target volume (CTV) differs between the first (CTV1) and second (CTV2) halves of the treatment⁴: CTV1 margins were GTV+5 mm, whereas CTV2 margins were GTV+3 mm, adjusted to reduce tumor growth and the dosage to organs at risk (OARs) such as skin, brain, and orbit adjacent to the tumor. Occurrence of adverse events (AEs), particularly late AEs, is more likely when OARs receive doses exceeding 40 to 50 Gy (RBE).⁵ We aimed to reduce AEs by dividing the total dose into 2 halves: the second half (35.2 Gy RBE/8 fr) limits the dose to OARs while maintaining adequate coverage of the GTV. We attempted to reduce the dosage as per the OAR criteria, including eye, optic nerve, brain, brainstem, and masticatory muscles, as recommended previously6, 7, 8, 9, 10 (dose-volume histograms shown in Figs. 3 and 4). The dose constraints for the left eyeball, left masticatory muscle group (masseter, temporalis, and medial and lateral pterygoid muscles), and maxilla could not be achieved owing to the extent of tumor progression (Table 1). Planning target volume (PTV) 1/2 was defined as CTV1/2+2 mm margin. This 2-step approach with sequential margin reduction was employed to prioritize GTV coverage while minimizing the risk of high-dose delivery to adjacent OARs. Compared with a single-plan strategy with local underdosing, this method provides greater robustness against geometric uncertainties, such as setup errors or minor anatomic changes, by ensuring that the high-dose region (PTV2) remains securely within the subacute target volume (PTV1).^2,11 The radiation dose was prescribed to the PTV isocenter. The planning goal was to cover at least 95% of the PTV with 95% of the prescribed dose (V95% ≥ 95%). To prioritize safety constraints, however, this coverage was intentionally compromised in areas where the tumor progressed into critical OARs. In this case, clinical priority was given to maintaining CTV coverage while minimizing high-dose delivery to the most sensitive portions of the OARs. The patient was immobilized using thermoplastic shells (Shellfitter; Kuraray) and positioned in customized cradles (Moldcare, Alcare). To prevent swallowing and occlusal movement, a customized mouthpiece was secured.¹² Toxicity was assessed by the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0¹³ through medical examination and nasopharyngoscopy.

Figure 1.

Findings before carbon-ion radiation therapy. Computed tomography showed a mass centered in the left maxillary sinus and resorption of the surrounding maxillary sinus bone: (A) coronal and (B) axial sections. Magnetic resonance imaging 2 to 3 weeks after computed tomography showed periapical invasion of the mass from the left maxillary sinus to the nasal cavity, with infiltration of the left orbital floor: (C) coronal and (D) axial sections. Fluorodeoxyglucose positron emission tomography showed abnormal accumulation mainly in the left maxillary sinus and no distant metastasis: (E) coronal and (F) axial sections.

Figure 2.

Carbon-ion radiation therapy treatment plan. Total dose: 70.4 Gy relative biological effectiveness/16 fractions. This tumor grew rapidly, and on treatment-planning computed tomography, the area of bone destruction in the maxillary sinus was larger than that at the initial visit. (A) Level of maxillary sinus on axial and (B) coronal sections. (C) Level of orbital floor on axial and (D) sagittal sections. Gross tumor volume (GTV): red line, clinical target volume 1 (CTV 1): magenta line, clinical target volume 2 (CTV 2): cyan line.

Figure 3.

Dose-volume histograms of the orbit.

Abbreviations: CTV = clinical target volume; GTV = gross tumor volume; RBE = relative biological effectiveness.

Figure 4.

Dose-volume histograms of masticatory muscles.

Abbreviations: CTV = clinical target volume; GTV = gross tumor volume; RBE = relative biological effectiveness.

Table 1.

Doses for organs at risk and their limitation details in this case

Organ at risk	Maximum dose (Gy [RBE])	V30 (cm3)	V40 (cm3)	V50 (cm3)	Limitation details	References
Brain	67.45	3.48	2.04	1.17	V40 < 7.6 cm3 V50 < 4.6 cm3	Koto et al8
Brainstem	2.72	0	0	0	V30 < 0.7 cm3 V40 < 0.1 cm3	Shirai et al9
Left eye	67.34	2.24	1.45	0.82	Maximum dose < 54.75 Gy (RBE) V30 < 2.29 cm3 V40 < 0.83 cm3 V50 < 1.45 cm3	Nachankar et al6
Right eye	14.02	0	0	0
Left optic nerve	49.06	0.32	0.12	0	Maximum dose < 57 Gy (RBE)	Hasegawa et al7
Right optic nerve	4.23	0	0	0
Masseter muscle	70.2	1.96	1.44	1.03	Maximum dose < 44.0 Gy (RBE)	Musha et al10
Temporalis	69.55	0.37	0.18	0.11	Maximum dose < 39.6 Gy (RBE)
Medial pterygoid muscle	70.95	1.47	1.17	0.85	Maximum dose < 60.4 Gy (RBE)
Lateral pterygoid muscle	71.27	2.28	1.91	1.57	Maximum dose < 57.6 Gy (RBE)
Coronoid process	69.83	0.34	0.31	0.25	Maximum dose < 38.0 Gy (RBE)
Maxilla	72.05	17.05	15.95	14.69	V50 < 3.0 cm3	Sasahara et al20

Abbreviations: RBE = relative biological effectiveness. V30, V40, V50 = volume receiving ≥30, 40, 50 Gy(RBE).

Acute AEs such as grade 3 oral mucositis and grade 1 dermatitis observed during and immediately after treatment improved with conservative treatment. However, grade 2 sinusitis developed as a late AE. At 18 months after treatment, the tumor filling the left maxillary sinus disappeared, and a maxillary sinus fistula associated with tumor shrinkage developed in the same area (Fig. 5). The fistula could be sealed with dentures. Although no further local or distant lesions were observed, the patient died at 20.7 months posttreatment due to worsening renal function.

Figure 5.

Progress after carbon-ion radiation therapy. At 18 months after treatment, the tumor lesion filling the left maxillary sinus disappeared. Computed tomography: (A) coronal and (B) axial sections. Magnetic resonance imaging: (C) coronal and (D) axial sections.

Discussion

Despite the lack of specific reports on CIRT for AFS, the role of conventional RT has been limited. The current standard of care is surgical resection; however, a systematic review reported a recurrence rate of 35%, with a mean time to recurrence of 22.2 months after surgery alone.¹ While postoperative RT is frequently employed,^1,14 late AEs such as xerostomia and trismus can manifest around 11 months posttreatment, highlighting the critical need for OAR dose reduction.¹⁵ To date, only 1 study has reported the use of proton beam therapy, but its follow-up period was limited to 6 months.¹⁶

In contrast, clinical studies of CIRT for radioresistant head and neck tumors have demonstrated promising results, including in osteosarcomas and soft-tissue sarcomas, which share sarcomatous features with AFS. Specifically, a 3-year local control rate of 91.8% and an overall survival rate of 74.1% have been reported.³ The successful local control observed in our case demonstrates the feasibility of CIRT for this rare histology, although findings from a single case with a 20.7-month follow-up are insufficient to establish definitive efficacy. Our experience suggests potential activity for CIRT in head and neck AFS, owing to its superior dose concentration and high biological effectiveness against low-radiosensitivity components.

In this case, the fistula was closed with dentures, causing no impairment in speech, eating, or swallowing. Koto et al⁸ reported that a brain volume receiving ≥40 Gy (RBE) (V40) <7.6 cm³ and a brain volume receiving ≥50 Gy (RBE) (V50) of <4.6 cm³ were tolerable dose constraints for radiation-induced brain injury (RIBI). In this case, V40 was 2.04 cm³ and V50 was 1.17 cm³, achieving the target doses. However, the maximum dose was 67.45 Gy (RBE), indicating that despite the small, irradiated volume, areas receiving significantly high doses exist. Although Koto et al reported no significant difference in RIBI incidence at the maximum dose, grade ≥2 RIBI developed in most cases, even when the maximum dose was within the 40 Gy (RBE) isodose line. Therefore, in cases where even a small high-dose area exists, careful long-term follow-up is necessary.

The primary justification for proceeding with treatment despite exceeding several OAR constraints was the prioritization of local tumor control for this advanced and radioresistant AFS. Strict adherence to all constraints would have compromised the radical dose to the CTV, significantly increasing the risk of treatment failure.

Regarding the clinical implications, a discrepancy was observed between the dosimetric predictions and clinical outcomes. Both the dose to the masticatory muscles, which is a risk factor, and the dose to the coronoid processes, which is a surrogate marker for trismus onset because of CIRT,¹⁰ failed to meet the target owing to the extent of tumor infiltration. The reported median onset time for trismus is 12 months (range, 10-22).¹⁰ Although trismus was anticipated in this case because of the high dose to the masticatory muscles, the patient maintained a stable quality of life without significant symptoms during the survival period. This might be attributed to our strategy of minimizing the volume of intermediate-dose radiation, despite the high maximum doses. Although several stretching methods and jaw mobilization devices have been reported,^17,18 there is no effective curative treatment once trismus develops. When tumor invasion around the masticatory muscles is limited, it is easier to reduce the radiation dose, potentially preventing its development. However, this dose reduction strategy becomes difficult when the tumor has invaded the structures surrounding the masticatory muscles, as in this case. Some novel CIRT techniques, such as the spot scanning method and linear energy transfer painting,¹⁹ could improve local control and AE onset rates.

Nevertheless, a critical limitation of our case is that the follow-up duration may be insufficient to fully capture late AEs. We did not observe severe RIBI or ocular complications. However, the reported onset ranges for these toxicities are broad: 16 to 132 months (median, 26 months) for RIBI⁸ and 15 to 47 months (median, 24 months) for ocular complications.⁶ With 20.7 months, the end of follow-up in our case corresponds to the beginning of the high-risk period (around the median onset times); thus, the absence of these toxicities should be interpreted with caution. The safety profile of CIRT for OAR-proximal AFS requires further validation through longer-term clinical observations.

Conclusions

This case demonstrates the feasibility of CIRT for head and neck AFS when surgery is contraindicated or organ preservation is prioritized. While the long-term local control and the risk of late AEs—such as RIBI, ocular symptoms, and trismus—remain unclear owing to the limited follow-up duration, the disease was controllable during the patient's survival period without the onset of severe, uncontrollable toxicities. Based on our experience, CIRT warrants further investigation as a potential therapeutic approach for this rare histology.

Disclosures

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.

Declaration of AI and AI-Assisted Technologies in the Writing Process

During the preparation of this work, the authors used ChatGPT, a language model developed by OpenAI, to improve the linguistic readability of the text. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication