Major Finding: INSS stage 4 patients had a 5-year survival of 8%, compared with 52% for those who were stage 1, 2, 3, or 4S.
Data Source: A retrospective analysis of children in the International Neuroblastoma Risk Group database.
Disclosures: The study was supported by the Little Heroes Pediatric Cancer Research Foundation, the Forbeck Foundation, and a grant from the National Institutes of Health. Dr. London and her associates reported no conflicts of interest.
CHICAGO — Time to relapse is highly predictive of overall survival post relapse in children with neuroblastoma, according to an analysis of 2,266 children in the International Neuroblastoma Risk Group database.
The analysis identified other factors prognostic of overall survival post relapse as well as a small proportion of relapsed patients who are salvageable, Wendy B. London, Ph.D., reported at the meeting.
Currently, clinicians do not know how to identify which patients are more likely to respond to postrelapse therapy and have difficulty in interpreting time to relapse because neuroblastoma is a heterogeneous disease, she said.
The median time to relapse among the 2,266 children was 13.2 months, with a wide range from 1 day to 11.4 years.
All told, 73% of children who relapsed were aged 18 months or older, 72% were International Neuroblastoma Staging System (INSS) stage 4, and 33% had amplified MYCN oncogene expression.
Overall survival at 5 years was 20%.
It was not possible to categorize time to relapse using a simple 1-year cutoff, said Dr. London, director of biostatistics at Children's Hospital Boston.
The risk of death was about the same for children who relapsed within the first 6 months as it was for those who relapsed at 18-24 months. The risk of death was highest in those who relapsed between 6 and 18 months.
All three groups had a significantly higher risk of death, compared with patients who relapsed after 36 months (P less than .001).
The association between time to relapse and overall survival appears to be driven by stage 3, 4, and MYCN-amplified patients, Dr. London said.
In a survival tree regression analysis that adjusted for time to relapse, disease stage was identified as the most highly significant variable for survival post relapse. INSS stage 4 patients had a 5-year survival of 8%, compared with 52% for those who were stage 1, 2, 3, or 4S.
“Aggressive disease that advances unhindered by treatment has a highly predictable time course and ultimate death; we already knew this, but now we're able to describe it quantitatively,” she said.
Upon further analysis, three cohorts emerged as salvageable after relapse:
▸ Patients who are stage 4, with nonamplified MYCN, and less than 18 months of age.
▸ Patients who are stage 1, 2, 3, or 4S with MYCN amplification.
▸ Patients who are stage 1, 2, 3, or 4S with nonamplified MYCN and undifferentiated grade histology.
Patients who had stage 4 disease and MYCN amplification had a 5-year survival of 4%, compared with 12% for stage 4 patients with nonamplified MYCN.
Time to first relapse as a predictor of survival is important for two reasons, said Dr. Andrew Pearson, chair of pediatric oncology at the Institute of Cancer Research and the Royal Marsden Hospital in London, who was invited to discuss the findings. It can be used to stratify and/or describe patients in early clinical trials and to identify a salvageable population post relapse.
“In the past, I'm sure that some agents have had a negative response in early clinical studies because a group of very poor prognosis patients were included,” he said.
The study findings will also be utilized by the International Neuroblastoma Risk Group, which is nearing completion of standardized international criteria for eligibility and response for phase II studies in neuroblastoma, he said.
In multivariable analysis, factors at diagnosis that were independently predictive of overall survival post relapse were stage 4 (hazard ratio, 6.9); stage 3 (HR, 4.3); stage 4S (HR, 3.5); MYCN amplification (HR, 2.4); age less than 18 months (HR, 1.6); and time to relapse less than 12 months (HR, 2.0) – all with a P value less than 0.0001, Dr. London said.
Time to relapse was predictive of survival post relapse in patients with stage 1, stage 2, or no MYCN amplification, but it was not independently predictive, she said.
My Take
Time to Relapse Useful for Determining Trial Eligibility
The observation that time to relapse predicts survival in neuroblastoma is made possible by the analysis of a large international collaborative database. Given the unique biology of neuroblastoma and the extreme clinical heterogeneity that impacts its natural history despite therapy and initial response to therapy, this finding will be important as new agents become available for investigation in this disease and especially when nontraditional end points such as time to progression and progression-free survival are considered.