Development and validation of NTCP models for acute side-effects resulting from proton beam therapy of brain tumours.

BACKGROUND: The limited availability of proton beam therapy (PBT) requires individual treatment selection strategies, such as based on normal tissue complication probability (NTCP). We developed and externally validated NTCP models for common acute side-effects following PBT in brain tumour patients in effort to provide guidance on optimising patient quality of life.

METHODS: An exploration cohort including 113 adult brain tumour patients who underwent PBT was investigated for the following endpoints: alopecia, scalp erythema, headache, fatigue and nausea. Dose-volume parameters of associated normal tissues were used for logistic regression modelling. Statistically significant parameters showing high area under the receiver operating characteristic curve (AUC) values in internal cross-validation were externally validated on two cohorts of 71 and 96 patients, respectively.

RESULTS: Statistically significant correlations of dose-volume parameters of the skin for erythema and alopecia were found. In internal cross-validation, the following prognostic parameters were selected: V35Gy (absolute volume receiving 35 Gy) for erythema grade ≥1, D2% (dose to 2% of the volume) for alopecia grade ≥1 and D5% for alopecia grade ≥2. Validation was successful for both cohorts with AUC >0.75. A bivariable model for fatigue grade ≥1 could not be validated externally. No correlations of dose-volume parameters of the brain were seen for headache or nausea.

CONCLUSION: We developed and successfully validated NTCP models for scalp erythema and alopecia in primary brain tumour patients treated with PBT.