Beyond copy number: circulating cell-free mitochondrial DNA Fragmentomics and structural states in liquid biopsy

Circulating cell-free mitochondrial DNA (cf-mtDNA) is commonly treated as a concentration-based biomarker, yet the measured signal is a composite of biologically distinct molecular states that are differentially affected by blood collection, platelet activation, centrifugation, storage, extraction, amplification, and sequencing. This analytical heterogeneity has become more consequential as cf-mtDNA research moves from copy-number assays toward fragmentomics, topology, heteroplasmy, oxidative lesions, and carrier-resolved measurements. Recent multi-cohort studies indicate that fragment size, 5'-end composition, motif diversity, and regional breakage patterns can support cancer detection and tissue-of-origin inference, while clinical chemistry studies demonstrate that routine preanalytical choices can markedly alter the mitochondrial fraction recovered from plasma. This review develops a laboratory-medicine framework in which cf-mtDNA is considered a multidimensional analyte rather than a single abundance variable. We integrate mitochondrial release biology with the extracellular carrier states of mtDNA; explain how TFAM, membrane pores, mitophagy, extracellular vesicles, platelets, and nucleases shape the observed fragmentome; evaluate preanalytical and analytical sources of bias; and assess translational evidence across oncology, critical illness, cardio-renal disease, metabolic inflammation, transplantation, and neuroinflammatory disorders. Particular emphasis is placed on distinguishing true biological variation from procedure-induced redistribution of cf-mtDNA between soluble, vesicular, platelet-associated, and cell-free mitochondrial compartments. We propose minimum analytical descriptors, quality-control priorities, and validation steps required before cf-mtDNA fragmentomic signatures can become transportable clinical tests. The central implication is that the next phase of cf-mtDNA diagnostics will depend less on measuring more DNA than on measuring the correct molecular fraction with traceable preanalytics, orthogonal characterization, and clinically locked computational models.