Evolution of the Diagnostic Paradigm for Giant Cell Arteritis: From Histopathology to Multi-Modal Imaging Integration Running title: GCA: Histology to Imaging

Giant cell arteritis (GCA) is a systemic vasculitis that predominantly affects mediumand large-sized arteries. Delayed diagnosis may result in irreversible blindness or stroke. Temporal artery biopsy (TAB), historically regarded as the diagnostic gold standard, has limited sensitivity (40–70%) due to the segmental distribution of inflammatory lesions and carries risks of procedural complications and diagnostic delay. This systematic review aims to:(1) compare the diagnostic accuracy of non-invasive imaging modalities with TAB;(2) assess the prognostic value of imaging findings; and(3) evaluate the implementation of imaging-first clinical pathways. In accordance with the PRISMA 2020 statement, PubMed and Embase,were searched for high-impact studies (n = 36) addressing diagnostic accuracy, guideline updates, and the effectiveness of the fast-track clinic (FTC) model. Colour Doppler ultrasound (CDUS) demonstrating the "halo sign" achieved a pooled sensitivity of 88–93%. Accordingly, the 2022 ACR/EULAR classification criteria assign CDUS findings diagnostic weight equivalent to a positive TAB. High-resolution MRI enables quantitative evaluation of cranial arterial wall thickening and contrast enhancement. 18F-FDG PET/CT is particularly useful for assessing systemic inflammatory burden and identifying large-vessel involvement associated with higher relapse risk, while CT angiography (CTA) delineates structural vascular damage. Implementation of FTC pathways reduces diagnostic latency to 24–72 hours and lowers the risk of permanent visual loss by 60–80%. Non-invasive, multimodal imaging has redefined the diagnostic paradigm of GCA. By enabling accurate diagnosis and risk stratification, it informs personalized management strategies. Future directions should emphasize standardized acquisition protocols and artificial intelligence–assisted analysis to reduce operator dependence and further enhance early detection.