Differentiating high-flow vascular lesions from clinical and hemodynamic mimics in the head and neck: a pattern-based multimodality imaging approach.

PURPOSE: To present a pattern-based, multimodality imaging approach for differentiating true high-flow vascular lesions from their clinical, imaging, and hemodynamic mimics in the head and neck.

METHODS

This curated pictorial review draws on cases from two tertiary institutions in Singapore, encompassing true high-flow vascular lesions - including arteriovenous malformations (AVMs), arteriovenous fistulas (AVFs), and carotid-cavernous fistulas (CCFs) - as well as mimics, including carotid arterial wall pathologies (e.g. pseudoaneurysms, carotid blowout, dissections) and hypervascular tumors (e.g. paragangliomas). Imaging findings across ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), and digital subtraction angiography (DSA) were analyzed, with emphasis on structural and hemodynamic features.

RESULTS

True high-flow vascular lesions are defined by direct arteriovenous shunting, with key imaging features including early venous opacification, arterialized venous drainage, and the presence of a nidus (in AVMs). In contrast, mimics may demonstrate overlapping clinical presentations or apparent high-flow characteristics but lack true arteriovenous communication. Advanced imaging techniques - including 3D CT angiographic reconstructions and time-resolved contrast-enhanced MR angiography (e.g. TWIST) - improve delineation of vascular architecture and enable dynamic assessment of flow patterns. Digital subtraction angiography remains essential for definitive characterization and treatment planning.

CONCLUSION

Differentiation of true high-flow vascular lesions from their mimics relies on recognition of key hemodynamic and structural imaging features, particularly the presence of arteriovenous shunting. A pattern-based multimodality imaging approach facilitates accurate diagnosis, guides appropriate management, and reduces misclassification of clinically significant mimics.