Somatic Mosaicism Patterns Define Clinical-Surgical Subtypes of Focal Cortical Dysplasia Through Cell-Type-Specific Expression

Focal Cortical Dysplasia Type II (FCDII) is a subtype of cortical malfunction and is the primary cause of drug-resistant epilepsy in children. Although somatic mosaicism and clonal expansion of brain cells have been identified as crucial factors in FCD cases, the overall genetic landscape and clinical implications of FCDII remain largely unclear due to a significant gap in translating genetic data to inform surgical approaches and prognostic evaluations of individual cases. We carried out deep exome sequencing and deep amplicon validation of surgical biopsies and matched blood samples from 14 FCDII patients with confirmed neuropathology. We further performed multiscale pathogenic validations and took advantage of existing single-nucleus RNA sequencing and spatial maps from developing human cortices to explore the functionality of potential pathogenic somatic variants. We identified novel somatic variants in several functional categories, like neurotransmission (TAAR2, GRM6, ZACN), structural regulation (TUBB2A, PLEC, COL18A1), cellular maintenance (IDO2, PARP4, P2RX5), and RNA processing (RBMX), mapping the expression of these genes back to the developing human brain demonstrated significant enrichment in neuronal cell types, especially excitatory neurons, further confirming their contributions in early brain development and phenotypic functions in dysmorphic neurons. Combining these genetic findings with clinical phenotypes, we found brain-specific mosaic variants with very high mosaic fractions (fraction of mosaic cells, MF, up to 99.5% on P2RX5) associated with different clinical phenotypes. FCDIIB, a more severe subtype that contains balloon cells, had higher MFs (>40%) for variants within resectable cortical layers (excitatory neurons in Layers 5 and 6). This allows potentially targeted resection and achieves better clinical outcome (87.5 % with Engel score I). FCDIIA subtype, on the other hand, displayed lower MFs (<5%) with diffuse distribution, and required hemispherectomy, with poor surgical outcomes (Engel score II/III). Our results suggest MF thresholds are high-definition biomarkers of surgical outcome estimate, with MF > 40% predicting viable focal resection and MF < 5% indicating network dysfunction that necessitates broad-spectrum resection. Combining genetic mapping with cellular localization thus offers a coherent solution to precision surgery in FCDII, translating molecular diagnosis to clinical practice.