Oral Presentation Advances in Neuroblastoma Research Congress 2016

Towards a model for risk stratification of high-risk neuroblastoma. A report from the HR-NBL-1/SIOPEN study. (#60)

Daniel A Morgenstern 1 , Ulrike Pötschger 2 , Lucas Moreno 3 , Vassilios Papadakis 4 , Cormac Owens 5 , Sue Burchill 6 , Shifra Ash 7 , Alberto Garaventa 8 , Claudia Pasqualini 9 , Fernando Carceller 10 , Gudrun Schleiermacher 11 , Dominique Valteau-Couanet 9 , Ruth Ladenstein 12
  1. Paediatric Oncology/Haematology, Great Ormond Street Hospital, London, United Kingdom
  2. Studies and Statistics on Integrated Research and Projects, St Anna Kinderkrebsforschung , Vienna, Austria
  3. Paediatric Haematology/Oncology, Hospital Niño Jesús, Madrid, Spain
  4. Paediatric Haematology/Oncology, Agia Sofia Children’s Hospital, Athens, Greece
  5. Paediatric Haematology/Oncology, Our Lady's Children's Hospital, Crumlin, Dublin, Ireland
  6. Paediatric and Adolescent Oncology group, University of Leeds, Leeds, United Kingdom
  7. Paediatric Haematology/Oncology, Schneider Children's Medical Center of Israel, Petah Tikva, Israel
  8. Paediatric Haematology/Oncology, Istituto Giannina Gaslini, Genova, Italy
  9. Paediatric Haematology/Oncology, Institut Gustav Roussy, Villejuif, France
  10. Paediatric Haematology/Oncology, Royal Marsden Hospital, Sutton, United Kingdom
  11. Paediatric Haematology/Oncology, Institut Curie, Paris, France
  12. Paediatric Haematology/Oncology, St Anna Kinderspital and St Anna Kinderkrebsforschung, Vienna, Austria

Introduction: Neuroblastoma is commonly classified as low, intermediate or high-risk according to criteria reported by the International Neuroblastoma Risk Group (INRG) project. As yet there are no established methods for stratifying patients with high-risk disease at diagnosis. The aim of this project is to explore potential criteria in patients treated on the HR-NBL-1/SIOPEN study and to develop models for risk stratification.

Methods: Data were extracted from the HR-NBL-1 trial database. The analysis was limited to patients with metastatic disease, aged ≥18 months at diagnosis and treated on study prior to the introduction of immunotherapy. The primary endpoint was 5-year event-free survival (EFS). For the statistical evaluation, the pseudo-value regression model was used to focus on the primary endpoint whilst eliminating the impact of short-term survival effects. A score was developed based on the estimated log-cumulative hazard (cHR) for 5-year EFS.

Results: Multiple different models – incorporating tumour MYCN-amplification status, serum LDH and ferritin (cut-off twice upper limit of normal, ULN), patient age (5 year cut-off) and involvement of one or multiple metastatic compartments – were developed. One model (incorporating data from 847 patients) included age>5 years (cHR 1.62; 95% CI: 1.28-2.04; 2 points in final score); LDH>2xULN (cHR 1.27; 1.13-1.61; 1 point) and involvement of >1 metastatic compartment (cHR 1.66; 1.16-2.24; 2 points). Using this model, we were able to identify a group of patients (with a score of 5), representing 12% of the total population, with a 5-yr EFS of only 8±3%; in contrast to patients with a score 0–1 who had 5-yr EFS >44%.

Conclusion: The development of models for risk stratification based on scores for criteria present at diagnosis is potentially useful both for predicting outcome and for treatment stratification in future high-risk neuroblastoma protocols. The work is being further developed to explore additional models, undertake validation and ultimately to include criteria such as blood/bone marrow mRNA levels or other biological parameters at diagnosis.