Tumorigenesis of well-differentiated and dedifferentiated liposarcoma: 12q amplicon architecture, oncogenic cargo, and evolutionary progression.

Well-differentiated liposarcoma (WDLPS) and dedifferentiated liposarcoma (DDLPS) are characterized by recurrent MDM2-centered amplification of chromosome 12q13-15, yet this diagnostic label reduces a complex structural lesion to a binary result. An architecture-resolved 12q amplicon platform requires an MDM2-centered dosage anchor plus direct evidence of nontrivial topology or carrier state; cargo, regulatory, and functional data define progressively higher-resolution features. When only MDM2 amplification is documented, the lesion is best described as diagnostic 12q amplification rather than a fully resolved platform. This review presents an evidence-graded architecture-to-function framework that separates direct WDLPS/DDLPS observations from mechanisms inferred from broader amplification biology. Disease-specific evidence supports heterogeneous 12q carriers, including ring, giant marker, rod-shaped, discontinuous, and neochromosome-like structures, as well as enhancer coamplification or altered 3D contacts. In contrast, chromothripsis, telomere crisis/BFB remodeling, ecDNA-like dynamics, and enhancer hubs are treated as extrapolated or testable mechanisms unless directly validated in liposarcoma. Conclusions are weighted by methodological resolution: cytogenetics/FISH define diagnostic amplification and visible carrier classes; MLPA or array-based assays define dosage and discontinuity; WGS with long-read or optical mapping is needed for junction-level architecture; and single-cell/spatial assays mainly resolve cellular-state or ecosystem context unless paired with structural readouts. MDM2-p53 and CDK4-RB remain central outputs, whereas selected cargo, lineage state, metabolism, and tumor-microenvironment programs are interpreted as evidence-ranked cooperating or downstream modules. Clinically, architecture-aware interpretation is proposed as a prospective framework for sampling, model design, biomarker development, and trial stratification, not as an established basis for routine management.