A model-based selection of oral cancer patient for passive scattering proton beam therapy

Clinical Snapshot

Key Findings

• P Value: Not reported in abstract; statistical significance of dosimetric differences not quantified in the abstract

• Effect Size: PSPT indicated (ΔNTCP exceeding Dutch threshold) in 24/40 patients (60%); PSPT reduced doses to spinal cord, pharynx, oral cavity, and contralateral salivary glands compared with VMAT

• Primary Outcome: NTCP-modelled risk of late adverse events (dysgeusia, xerostomia, osteoradionecrosis of the mandible) comparing PSPT versus VMAT plans in 40 OC patients

• Nnt Or Sensitivity: Not applicable in traditional sense; 60% selection rate for PSPT based on ΔNTCP thresholds. NNT cannot be calculated from NTCP model estimates without observed clinical event rates

• Confidence Interval: Not reported — no confidence intervals provided for NTCP estimates or ΔNTCP values

Clinical Application

The MBA framework using NTCP modelling is feasible at centres with proton therapy capability and access to validated treatment planning systems. However, implementation requires: (1) dual-plan generation (proton and photon) for each patient, increasing planning workload; (2) validated NTCP model parameters for the specific patient population; (3) institutional agreement on ΔNTCP thresholds. PSPT specifically is being replaced by pencil beam scanning at most modern proton centres, limiting direct applicability of these specific plans. Australia currently has no operational proton therapy centre (as of mid-2025), with the Sydney Proton Therapy Centre under development. The TGA has not specifically approved proton therapy as a device class requiring separate registration. RACGP and Cancer Australia guidelines do not currently include proton therapy in standard OC management pathways. PBS does not list proton therapy. Australian patients seeking proton therapy must travel overseas (Japan, USA, Europe), incurring substantial cost and logistical burden. When Australian proton facilities become operational, an MBA framework similar to that described — adapted with locally validated NTCP parameters and Australian population baseline risk data — would be an appropriate patient selection tool. The Dutch national protocol thresholds referenced in this study would require validation against Australian OC patient data before adoption. Adult patients with oral cavity cancer being considered for definitive or postoperative radiotherapy at centres with access to proton beam therapy, particularly those at elevated baseline risk of xerostomia, dysgeusia, or mandibular osteoradionecrosis

Abstract

BACKGROUND AND PURPOSE: To identify patients with oral cancer (OC) who benefit from proton therapy versus photon therapy using a model-based approach (MBA) to reduce late adverse events. MATERIAL/METHODS: Passive scattering proton beam therapy (PSPT) and volumetric modulated arc therapy (VMAT) plans were established for 40 patients with OC, consisting of a dose of 70 Gy (relative biological effectiveness) in 35 fractions. At least 95% of the planning target volume received 95% of the prescribed dose, and the dose to the organs at risk was restricted to be as low as possible. Normal tissue complication probability (NTCP) was used to estimate the risk of late adverse events, including dysgeusia, xerostomia and osteoradionecrosis of the mandible. Patients who were considered as candidates for PSPT were selected if the NTCP difference between both plans (ΔNTCP) exceeded the threshold defined in the Dutch national indication protocol. RESULTS: PSPT reduced the doses to the spinal cord, pharynx, oral cavity and contralateral salivary glands compared with VMAT. PSPT was also superior for the risk evaluation of each late adverse event calculated by incorporating dose metrics and patient baseline risk factors. For the MBA using ΔNTCP and the thresholds, PSPT was indicated in 24 of the 40 cases, and the remaining 16 cases were considered to be equivalent to VMAT. CONCLUSION: MBA was used to select OC cases that were candidates for PSPT. Compared with VMAT, the risk of expected late adverse events was significantly reduced with PSPT, and 60% of patients were considered eligible for PSPT.

References

1. 1.Narita, Y., Takayama, K., Kato, T., Komori, S., Seto, I., Nakasato, T., & Murakami, M. (2026). A model-based selection of oral cancer patient for passive scattering proton beam therapy. Journal of Radiation Research. https://doi.org/10.1093/jrr/rrag042