Space-time clustering patterns in childhood leukaemia support a role for infection.

Area of Science:

• Epidemiology
• Paediatric Oncology
• Spatial Statistics

Background:

• Previous studies on space-time clustering in childhood leukaemia have yielded inconsistent results.
• Understanding geographical and temporal patterns is crucial for identifying potential causes of childhood cancers.

Purpose of the Study:

• To investigate space-time clustering of childhood leukaemia using robust statistical methods.
• To identify specific age groups and leukaemia subtypes exhibiting significant clustering.
• To explore potential aetiological factors, such as infectious agents.

Main Methods:

• Utilized Manchester Children's Tumour Registry data for childhood leukaemia cases.
• Applied Knox tests for space-time interactions with fixed spatial (<5 km) and temporal (<1 year) thresholds.
• Incorporated birth addresses and diagnosis locations, and used nearest neighbour distances.
• Employed a second-order analysis using K-functions for comprehensive assessment.

Main Results:

• Highly significant space-time clustering was observed, particularly for all leukaemias (0-14 and 0-4 years) and acute lymphoblastic leukaemia (ALL) (0-4 years) based on birth address and diagnosis time.
• Analyses using diagnosis location showed some significance but generally larger P-values.
• An excess of male cases was noted in space-time pairs, suggesting potential sex-based differences in susceptibility.

Conclusions:

• The findings strongly support an infectious aetiology for childhood leukaemia.
• Observed clustering patterns provide a foundation for future research into specific infectious agents.
• Sex-based differences in susceptibility to infection may influence leukaemia development.