Given that the cost of proton therapy is considerably higher than that of conventional radiotherapy with photons, it is necessary to establish whether these higher costs are worthwhile in light of the expected advantages2,3,4. Thus, clear evidence of the situations in which proton therapy outperforms conventional photon treatment is needed. The investigators therefore aim to demonstrate through an in silico trial that proton therapy decrease the amount of irradiated normal tissue and, consequently, the risk of side effects in the surrounding normal tissue as well as the risk of secondary tumors. The same overall treatment time and an equal number of fractions will be used for both treatment modalities wherever possible. The most optimal technique for each individual patient, based on objective criteria related to limiting dose to normal tissue, will be prescribed by the institution concerned for each treatment option.

The CTV to PTV margin will be determined by the individual institutes according to the treatment technique and treatment modality. For all treatment plans, modified ICRU 50 will be applied. The ICRU 50 recommendation is that 100% of the target volume (PTV) receives at least 95% of the prescribed dose. In this work, participants agree that D98 > = 95%, D2 < = 107%, cold (< 95%) and hot (> 107%) spots should be avoided in general, and that no hot spots will be allowed outside the PTV. The same overall treatment time (OTT) and an equal number of fractions will be used for all treatment modalities.

Treatment plans will be carried out using either photons, protons and C-ions for dose calculation for the indication H&N re-irradiation. A multicentric patient dataset, which can be found on a secured website5,6, will be used. The output databases with matrices including 3D delineation, the deliverable physical dose, RBE map, and BED/ EQD2 will be made available. To guarantee a similar set-up between the three treatments, similar dose distributions will be used. For C-ion treatment planning, a 20% uncertainty in the RBE values will be taken into account.

All treatment plans will be performed in centers that are already operating and have experience in treatment planning. Photon treatment plans will be carried out in Nijmegen en Maastricht, proton treatment plans at the University of Pennsylvania (UPENN) and the C-ion treatment plans at the University Hospital Marburg (UHM). For photons, all primary treatment plans were performed according to the planning procedure of each individual participating center by selecting the best available technique. The secondary photon treatment plan will be performed according predefined fractionation schedule and tumor dose. The proton treatment plans for H&N tumors will be planned with active beam delivery using IMPT and/or passive 4D beam delivery (PSPT). The C-ion treatment plans will be planned using a raster-scanning technique (IMIT).

All relevant OAR are delineated in the primary and secondary study set. GTV/CTV will be used accordingly to the actual treatment. New DVH's will be calculated for the added OAR. The secondary plan (i.e. re-irradiation plan) will be calculated again for the best currently available options for Photon, proton and C-ions. Between primary and secondary treatment there is a period of 1 to 5 years for all patients. To be on the save site a 30 % recovery of brainstem and spinal cord is assumed. There will not be a correction for fraction size. For practical reasons the location of the Dmax in the brainstem or spinal cord will not be taken into account. A maximal cumulative dose of 120 Gy will be accepted in the mandibular.

If the mandibular is part of the CTV in the first as well as in the second treatment, a higher dose will be accepted. A cumulative dose of max 120 Gy will also be accepted for the larynx , for the arytenoid is this 100Gy, unless the larynx is within the CTV for the first and second treatment than a higher dose will be accepted.

This means that for a total study population of 25 patients per tumor site and group at least 100 treatment plans will be carried out. All dose maps will be saved on the secure website as part of the database mentioned above.

Photon, proton and C-ion treatments will be compared based on dosimetric parameters on normal tissues such as mean parotic dose, etc. In addition, the NTCP for a fixed tumor dose or the same expected TCP will be determined. Cobalt Gy equivalent doses will be used when reporting the proton and C-ion dose. In the case of protons, a constant RBE value of 1.1 will be used for both the tumor and the normal tissues. The RBE of C-ions will be calculated based on the models used by the participating centers. The GSI in-house treatment planning system uses RBE values calculated on the basis of the local effect model (LEM). The LEM I (alpha/beta=2) is based on the radial dose distribution of each charged particle crossing into a cell nucleus, as well as on the radiosensitivity and repair capacity of the tissue. The TPSs used by UHM is also based on the LEM model. The model used at NIRS utilizes fixed RBE values that are dependent on the depth in the body, but independent of dose level or tumor type.

Inclusion Criteria:

• Primary treatment:

  1. ≥ 50 Gy with curative intent (with or without chemo)
  2. CTV includes primary tumor and at least level II-IV
  3. Severe dental scatter artifacts or metal implants of unknown density will be excluded
  4. IMRT plan available

• Secondary treatment after ≥ 1 year, dose 70 Gy

  a. Severe dental scatter artifacts or metal implants of unknown density will be excluded

• Overlap of at least one OAR in primary and secondary treatment

Exclusion Criteria:

• Patients referred for postoperative radiotherapy

patients with recurrent head and neck tumors