DEVELOPING RISK-BASED SELECTION CRITERIA FOR THE NEXT SIOP TRIAL OF "SIGHT-SAVING THERAPY" FOR CHILDREN WITH NF1-ASSOCIATED VISUAL PATHWAY GLIOMA (NF1-VPG) - A QUALITATIVE ANALYSIS OF A CONSENSUS SURVEY

Introduction A consensus survey involving 98 multi-disciplinary specialists for the first survey (10 cases) and 46 for the second survey (15 cases) in SIOP-E brain tumour group, European NF1 society and the Optic Pathway Glioma Working Group in North America were asked to identify case-based criteria for primary observation, treatment or randomisation in 25 representative, newly diagnosed, NF1-VPG childhood cases which had been the subject of a consensus workshop as part of the SIOP-LGG 2004 trial. Respondents provided justifying statements for their selection of preferred strategy which are the focus for this report Methods Qualitative analysis following grounded theory, using the comparative method categorized 808 comments to: 173 for primary observation, 426 for primary treatment and 209 for primary randomization. These were grouped into 8 themes by two reviewers and compared with the kappa statistic. The themes were: 1) risk of progression, 2) visual function, 3) age and sex, 4) site/dimension of tumour, 5) other symptoms, 6) inconclusive comments, 7) equipoise and 8) parental consent. The themes were analysed by strategy selected for each case. Result Good agreement was achieved for theme allocation between reviewers (k: 0.762). Comments selected for observation were supported by responses categorised as: unknown risk of progression (32%), good vision function (25%); for treatment as poor vision and intention to preserve-improve vision (38%); for randomization as unknown risk of progression (15%), tumour site/dimension (13%) and stable vision (12%). Conclusion This survey and its qualitative analysis identified sufficient uncertainty to propose a randomised trial design for observation versus treatment in a subgroup of newly diagnosed NF1-VPG. A pilot cohort with vision assessment date, site of tumour and age is needed to conduct a power calculation to design a future international randomised trial to study risk of progression and response to new treatments. Collaborations would be welcome.