Diagnostic Accuracy of Photon-Counting Computed Tomography for Detecting Coronary Artery Stenosis: A Systematic Review and Meta-analysis (Preprint)

BACKGROUND Conventional coronary CT angiography (CCTA) loses specificity in heavily calcified or stented vessels. Photon-counting CT (PCCT), with superior spatial resolution and spectral capability, may mitigate these limitations. OBJECTIVE To evaluate the diagnostic accuracy of PCCT for coronary artery stenosis. METHODS A priori registration in PROSPERO (CRD420251126753). Following PRISMA, PubMed, EMBASE, Web of Science, Cochrane, and Scopus were searched through August 1, 2025. Prospective/retrospective studies were eligible; case reports, conference abstracts, and reviews were excluded. Two reviewers independently extracted data and performed QUADAS-2 risk-of-bias assessment. A bivariate random-effects model synthesized sensitivity/specificity; an SROC curve was generated. RESULTS Of 987 records, 32 articles underwent full-text review; 14 were qualitatively synthesized and 7 meta-analyzed. Pooled per-patient sensitivity was 94.6% (95% CI, 91.3–96.7%) and specificity 79.7% (70.0–86.9%). Per-vessel sensitivity/specificity were 94.7% (90.2–97.2%) and 89.3% (83.3–93.3%), and per-segment 96.3% (91.0–98.5%) and 94.6% (91.4–96.6%). The SROC AUC was 0.88 (0.84–0.90). Leave-one-out analyses indicated robust estimates; publication bias tests were non-significant. CONCLUSIONS PCCT demonstrates high diagnostic performance at patient, vessel, and segment levels, with reduced calcium blooming and improved specificity in heavily calcified disease. These findings support PCCT-CCTA as a more reliable gatekeeper to invasive angiography. Multicenter validation and protocol optimization (including UHR noise reduction and radiation-dose management) are warranted for broader clinical adoption.