Ziying Zhang, Ying Jiang, Yu Gong, Hui Xie, Yaqian Han, Peng Chen
BACKGROUND: Circulating mitochondrial DNA is being investigated as a liquid-biopsy biomarker because of its high copy number and release during cellular stress. However, diagnostic estimates vary across tumor types and assays, and prognostic studies have measured both cell-free mtDNA and cellular blood-derived mtDNA, which are not analytically equivalent. Non-malignant tissue injury and inflammation may also increase circulating mtDNA, limiting disease specificity. METHODS: Following PRISMA 2020 guidance, seven databases were searched through December 2025. Diagnostic analyses synthesized predominantly cell-free mtDNA assays, whereas prognostic analyses synthesized the circulating or blood-derived mtDNA measures reported by the original studies. Random-effects models were used to pool diagnostic indices and hazard ratios across 31 unique studies comprising 8334 patients with cancer and 1286 controls. The protocol was registered in PROSPERO (CRD420261301732). RESULTS: Predominantly cell-free mtDNA assays showed a pooled sensitivity of 0.73, specificity of 0.80, diagnostic odds ratio of 15.41, and area under the summary receiver operating characteristic curve of 0.845, with substantial between-study heterogeneity. The mtDNA-79 subgroup showed a sensitivity of 0.79, specificity of 0.93, and area under the curve of 0.909. Higher circulating or blood-derived mtDNA measures were associated with poorer overall survival (HR = 1.70, 95% CI: 1.42-2.03) and relapse-free survival (HR = 2.04, 95% CI: 1.71-2.44), but not with progression-free survival or mortality. MT-CYB-based measures showed a stronger adverse prognostic association than MT-ND1-based measures. CONCLUSIONS: Circulating mtDNA measures show diagnostic and prognostic associations across malignancies, but current evidence does not support stand-alone clinical use. Heterogeneity in specimen matrix, pre-analytical handling, assay definition, and disease-control selection requires standardized multicenter validation. mtDNA-79 and MT-CYB warrant further study, while total mtDNA abundance should be interpreted as a cancer-associated rather than cancer-specific signal unless combined with tumor-informed genomic or fragmentomic features.