Acute lymphoblastic leukemia (ALL), the most common malignancy in children and adolescents, arises from a heterogeneous and multifactorial etiology involving genetic and environmental factors. Studies of seasonal variation in ALL diagnosis have yielded inconsistent results, likely reflecting differences in study design and population characteristics. Here, we evaluated seasonal variation across ALL immunophenotypes, including two common genetic subtypes. We analyzed seasonal variation by ALL subtype in 1504 ALL patients diagnosed before the age of 18 between 1995 and 2017 using data from the National Cancer Register and the Swedish Childhood Cancer Registry. Subgroup analysis included 1305 B-cell precursor ALL (BCP-ALL) cases, including 422 high hyperdiploid (HeH) and 259 ETV6::RUNX1 fusion-positive cases, and 175 T-cell ALL (T-ALL) cases. For comparison, 214 acute myeloid leukemia (AML) cases and 1367 brain tumor cases, including 224 medulloblastomas, were analyzed. Cases were grouped into overlapping 3-month diagnostic periods and analyzed using a Bayesian GARIMAX model, an extension of the autoregressive integrated moving average (ARIMA) framework. A sensitivity analysis was performed restricted to children aged 1–17 years. Seasonal variation was observed in the overall ALL cohort, with peaks between June and October. BCP-ALL and T-ALL also showed informative seasonality, with August consistently included among the peak months. Similar results were obtained in the sensitivity analysis. No seasonal variation was observed in AML, medulloblastoma, or other brain tumors. Informative seasonal variation was not detected in the HeH or ETV6::RUNX1 -positive subgroups, although HeH showed peak quarters consistent with the overall ALL pattern. These findings support a role for seasonal triggers in ALL and warrant further investigation in larger, genetically stratified cohorts. • Seasonal variation is present in childhood BCP-ALL and T-ALL, but not in AML, medulloblastomas, or other brain tumors. • Bayesian GARIMAX model detects seasonal patterns in childhood ALL. • We observed late-summer peaks in childhood ALL, with August included in all peak quarters. • Our findings are consistent with infections and other environmental factors acting as potential triggers for ALL.
Bychkov et al. (Thu,) studied this question.