Background and aims:  

The World Health Organization Global Initiative for Childhood Cancer identifies survival as the primary outcome for which to judge success. Accurate mortality and survival data for childhood cancers are lacking globally, particularly in the LMIC context.  

The SJCARES (The St. Jude Global Childhood Cancer Analytics Resource and Epidemiological Surveillance System) Registry is a Global Pediatric (HBCR) Network designed to support addressing this need with a strong focus on Quality Improvement (QI) 

Using data from this Registry, we analyzed four timepoints along the continuum of cancer diagnosis (from symptom onset to the initiation of therapy), described the primary cause of death (COD) reported, and characterized the temporality of these deaths. 

Methods:  

All cases in the SJCARES HBCR Network from September 2020 to January 2026 were included in this report. Sites with low quality data were excluded. For each patient, timelines were calculated using four dates in the registry: initial onset of cancer-related symptoms; first medical assessment; confirmed diagnosis; definitive treatment started. As patients are entered on a rolling basis, only records with complete pairwise intervals were included. Descriptive statistics are used to summarize patient demographics, diagnosis, the four timelines, and COD causes, with a focus on the first 90-days post-cancer diagnosis, stratified by 30-day increments. 

Results   

28,863 patient records with a confirmed cancer diagnosis from 60 institutions in 26 countries were included in the analysis. Median age at diagnosis was 6 years (IQR=8 years). 58% (n=16,786) cases were male, and 82% (n=21,188) of cases were diagnosed based on microscopic verification. Acute Lymphoblastic Leukemia was the most common diagnosis followed by Acute Myeloid Leukemia, representing 28% (n=7612) and 6% (n=1645) of cases, respectively .Median time from onset of cancer-related symptoms to first medical assessment was 22 days (IQR=40 days); first medical assessment to confirmed diagnosis was 7 days (IQR=18 days), confirmed diagnosis to definitive treatment started was 3 days (IQR=12 days) and onset of cancer-related symptoms to definitive treatment started was 44 days (IQR=64 days). From 17,465 patient records with known vital status, 27% (n=4,866) were reported dead. Neoplasm was the most frequent COD, accounting for 70% (n=3,425) of death, followed by infectious causes (21%, (n=1,058). 47% (n=2286) of all deaths occurred within the first 90 days of diagnosis. Among those early deaths, 67% (n=1538) occurred within the first 30 days. 

Conclusions:  

Most delays in median time from the onset of cancer-related symptoms to definitive treatment occurred before their first medical assessment, though large variations in the IQR were reported. Almost a half of the death events reported occurred in the first 90 days after diagnosis with most of early deaths reported within the first 30 days from diagnosis. Our data suggest that HBCR networks, with linked mortality and incidence outcomes, can provide surrogate global surveillance data, and guide near-real time QI initiatives by identifying preventable delays and early mortality patterns along the childhood cancer continuum at the hospital system level, and avert preventable deaths as part of the strategy to improve survival.