Radiation-Induced Non-DSB Clustered DNA Lesions: Chemistry, Repair, Mutagenesis, and Computational Modeling

Ionizing radiation (IR) generates chemically heterogeneous DNA damage whose spatial organization depends on radiation quality, including linear energy transfer (LET), the amount of energy deposited per unit path length. Among these lesions, non-double-strand-break (non-DSB) clusters comprise combinations of oxidized bases, abasic (AP) sites, and strand-break intermediates that do not initially constitute a double-strand break (DSB). Their biological effects depend on lesion identity, spacing, strand orientation, and the repair-state transitions that occur during base excision repair (BER). Prompt DSBs remain major determinants of immediate chromosome disruption and acute cytotoxicity, whereas non-DSB clusters represent a distinct processing-dependent hazard through lesion persistence, replication encounter, mutagenesis, and BER-mediated formation of repair-generated DSBs. This review integrates radiation chemistry, lesion-specific BER, defined-substrate studies, irradiation-based measurements, mutational outcomes, and track-structure modeling across the causal sequence from lesion formation to mutation. A structured, non-exhaustive evidence map showed that direct pairwise evidence occupied only 12.1-19.1% of lesion-pair categories across five inclusion scenarios; in the full 18-lesion ontology, 30 of 171 categories had direct pairwise support. These values describe coverage within the review-specific classification and search record, not biological lesion-pair frequencies. Defined-substrate studies provide configuration-specific mechanistic evidence, whereas irradiation-based assays retain biological context but generally lack complete lesion and repair-history resolution, and genome-wide sequencing records survival-conditioned mutational outcomes without identifying the initiating cluster. Together, the available evidence defines which lesion attributes, BER intermediates, contextual modifiers, and validation endpoints can currently constrain lesion-resolved state-transition models and where inference and uncertainty remain unavoidable. It also delineates the present limits of attributing cancer or other tissue-level health effects to specific non-DSB clustered configurations.