Response Assessment in Paediatric Phase I Trials According to RECIST Guidelines: Survival Outcomes, Patterns of Progression and Relevance of Changes in Tumour Measurements.

INTRODUCTION: RECIST guidelines constitute the reference for radiological response assessment in most paediatric trials of anticancer agents. However, these criteria have not been validated in children. We evaluated the outcomes and patterns of progression of children/adolescents enrolled in phase I trials in two paediatric drug development units.

METHODS: Patients aged ≤21 assessed with RECIST (v1.0 or v1.1) were eligible. Clinico-radiological data were analysed using Mann-Whitney U and log-rank tests to correlate response categories and sum of longest diameters (SLD) with time-to-event variables and overall survival (OS).

RESULTS: Sixty-one patients (71 enrolments) were evaluated; median age: 12.7 years (range, 3.1-20.9). Overall, 7% achieved complete/partial response (n = 5) and 31% disease stabilisation (n = 22). Median (95% CI) OS (in months) was 29.1 (27.6-30.6) with complete/partial response, 8.9 (2.0-15.8) with stable disease and 2.8 (2.3-3.3) with disease progression (P < 0.001); 32.6% patients with measurable disease presented exclusive progression of existing non-target lesions and/or new lesions. The change in SLD at best response showed a linear correlation with duration of response (r = -0.605; P = 0.004) and time on trial (r = -0.61; P = 0.003), but the change in SLD at progression did not correlate with time to progression (r = -0.219; P = 0.206).

CONCLUSIONS: Response assessment according to RECIST correlated with OS in children/adolescents treated on phase I trials. The reduction in SLD at best response correlated with more prolonged responses. Tumour size did not constitute an optimal method to assess disease progression in one third of patients with measurable disease. Further refinement of current response assessment guidelines will enable the development of paediatric-specific radiological criteria.