Session Item

The two main intentions of proton therapy referral are iso-toxic increase of target dose or iso-effective toxicity reduction of OARs. According to the Dutch Health Council, referrals aiming at toxicity reduction make up for 85% of expected referrals. Considering the volume of these referrals and the limited availability of proton installations, a model-based method of patient selection (MBS) has been proposed by Dutch centres. In this retrospective study, we apply the Dutch model to determine (a) how many of our head&neck patients would qualify for proton treatment planning and (b) if the photon dose calculation approach would affect patient selection.

A patient sample was first obtained by searching the ARIA database for all H&N patients that began their treatment in the first quarter of 2022 in all three centres of SLRON. Patients were removed from the results of the ARIA query if they were treated with 3DCRT plans or palliative intent, had distant metastases or previous H&N irradiation, had a primary tumour not in the larynx, pharynx or oral cavity, had unidentified primary tumour site or failed to complete treatment because of poor overall condition.

A total of 47 patients remained in the sample. Twelve plans were calculated in Eclipse (AAA v.15.6) and 35 plans were generated in Monaco (XVMC v.1.6). The Eclipse plans were recalculated with Acuros v.15.6 (dose to medium, same MUs) to assess the impact of dose calculation algorithm. Patient preselection was based on normal tissue complication probability (NTCP) for radiation induced xerostomia of Grade≥2 six months post treatment. Baseline xerostomia was not systematically documented. NTCPs for dysphagia and tube-feeding dependence were not considered as swallowing subunits are not routinely contoured in our centre. We used the mean dose to the contralateral parotid to calculate NTCP for the photon plans, except in bilateral cases, where we used the lowest dose among the two parotids.

We found that 34 out of 47 patients (72%) have a DNTCP=NTCP(photons)-NTCP(protons) for xerostomia above threshold (10%) and therefore qualify for proton treatment planning. Three patients had DNTCP between 10% and 20%, 14 patients between 20-30% and 17 patients above 30% (no parotid sparing in the photon plan). Mean parotid doses were reduced in the recalculated Acuros plans by 0.4 Gy-1 Gy, resulting in NTCP reductions between 0.5% and 1.7%, but none of these changes affected preselection. However, a recalculation with Acuros may be indicated for patients with NTCP values close to the selection threshold.

As NTCPs for dysphagia and tube-feeding dependence were not analysed in this study, our results are an underestimation of the number of H&N patients qualifying for proton planning. With 72% of our patients qualifying on xerostomia preselection alone, we concluded that a significant number of our patients currently not treated with protons would benefit from that treatment if it was more accessible in Ireland.