Survival rates after childhood cancer have increased dramatically over the past fifty years such that currently more than 83% of children diagnosed with a malignancy will survive at least five years. This has resulted in a growing population of childhood cancer survivors which was estimated to number 420,000 in the United States in 2013 and which is expected to exceed half a million by the year 2020. However, the aggressive treatments required to cure children of their cancers in many cases also increases their risk of serious health conditions later in life. This is evidenced by the fact that more than half of survivors will develop at least one severe, disabling, life-threatening or fatal chronic health condition by the age of 50. Accordingly oncologists have made great efforts in recent decades to modify treatments with goal of trying to maintain high cure rates while reducing the risk of these devastating late effects such as chronic disease and death. However, the impact of these modifications on late outcomes in survivors is not yet well established and so the goal of our study was to examine whether changes in treatment over time have been translated to improved late health outcomes in survivors of childhood cancer.

To do this we utilised data from the Childhood Cancer Survivor Study, or CCSS, which is a federally funded retrospective cohort of children diagnosed with cancer between 1970 and 1999 at one of 31 contributing centres around the United States and Canada. Survivors were eligible for CCSS if they survived at least five years after diagnosis and were diagnosed with one of the common childhood cancers listed here – leukaemia, lymphomas, central nervous system malignancies, Wilms’ tumour, neuroblastoma or soft tissue or bone sarcomas.

CCSS collected detailed treatment data using medical records from participants and also followed survivors over time to collect a wide range of outcome information, primarily via questionnaires. Among the 23,601 survivors included in our analysis the median age at last follow up was 28 years and the median time since diagnosis was 21 years. The main outcome in our analysis was the fifteen year cumulative incidence of severe, disabling, life-threatening or fatal chronic conditions and we examined these within groups of survivors that were separated by the decade in which their cancer was diagnosed and treated.

So to interpret this, in the top part of this slide you can see that in survivors who were diagnosed between 1970 and 1979 12.7% of them went on to develop at least one severe, disabling, life-threatening or fatal chronic condition between five and fifteen years after diagnosis. Among survivors who were diagnosed in the 1980s this proportion dropped to 10.1% and it declined further for those diagnosed in the ‘90s to 8.9%. In addition to looking at survivors overall we also looked within specific diagnosis groups, those are shown in the plot here in the lower part of the slide where we’ve plotted the fifteen year cumulative incidence again with the black diamonds representing survivors treated in the 1970s, the red squares those in the 1980s and the blue triangles those in the 1990s. As you can see, we saw significant decreases in the incidence of serious chronic conditions in survivors of acute lymphoblastic leukaemia, astrocytoma, Hodgkin lymphoma, non-Hodgkin lymphoma, Wilms’ tumour and Ewing’s sarcoma. Furthermore, when we looked at the types of chronic conditions experienced by these survivors we saw that the reductions overall were driven by large decreases in the incidence of second cancers, endocrine conditions and neurological and gastrointestinal conditions.

Finally, we also wanted to examine whether these decreases in the incidence of serious chronic disease in survivors was mirrored by decreases in the treatment intensity that they received by decade of diagnosis. To do that we plotted our graphs such as shown here where on the left y-axis you can see we’ve plotted again the fifteen year cumulative incidence of these serious chronic conditions represented by the red dots in these figures. Then we have overlaid on that a treatment intensity score which is essentially a way to visualise the relative intensity of treatments that survivors received, whether they were diagnosed in the ‘70s, ‘80s or ‘90s. Those are represented by the black box plots in these figures. In general we did see that the decrease in the cumulative incidence of serious chronic conditions was mirrored by a decrease in treatment intensity over time and that’s shown nicely here in these examples for survivors of Hodgkin lymphoma, astrocytoma and acute lymphoblastic leukaemia.

In conclusion, in CCSS we found that more recent survivors of many childhood cancers have reduced incidence of serious chronic disease later in life. These changes were correlated with temporal changes in treatment and this demonstrates that the strategy of trying to reduce the intensity of therapy with the goal of reducing the occurrence of late effects, along with the changes in improvements in screening and early detection that occurred over the same timeframe have, in fact, translated to reduced incidence of serious late morbidity and improved late health outcomes in survivors of childhood cancer. Thank you.