Requirements and study design for the next proton FLASH clinical trials: an international multidisciplinary Delphi consensus.

Abstract

Purpose: The FLASH effect, defined as normal tissue sparing while maintaining tumor control with ultra-high dose rate (UHDR) irradiation, has been demonstrated preclinically in different tumors and tissues. Although biological mechanisms are unclear, there is a need for clinical trials investigating the value of proton FLASH irradiation (pFLASH). The purpose of this study was to establish an expert consensus regarding prerequisites, study design and endpoints for next clinical trials exploring the clinical potentials of pFLASH.

Methods and materials: Delphi methodology was used to develop a systematic expert consensus. An international expert panel was composed of 21 clinicians, physicists and biologists, well-balanced in expertise and geography, using predefined inclusion criteria. Statements were scored on a 5-point Likert scale in 2 rounds of online questionnaire voting. Definition of consensus was set a priori.

Results: Response rate was 100% in both rounds. Preclinical in vivo demonstration of the FLASH effect in normal tissue while maintaining tumor response is deemed essential before starting a clinical trial in a specific tumor site. The next clinical pFLASH trials are advised to include adult patients only, with a minimal expected overall survival of 1 year for palliative settings or, preferably, oligometastatic disease in the ablative setting. The pFLASH effect should be studied in a single treatment modality setting with toxicity reduction as the primary endpoint. Recommendations are described on the use of clinical targets and organs at risk constraints, requirements for evaluation and reporting and accuracy levels and pretreatment verification of dose rates. No consensus was reached on the use of multiple beams, multiple fractions and fraction dose.

Conclusions: There is a need for additional data regarding influence of fractionation and multiple beam planning. Results of this study can be used to develop roadmaps to guide future clinical trial design.