Abstract
Background/Objectives: We aimed to characterize the microvascular imaging (MI) to demonstrate in hepatic hemangiomas in routine practice and to quantify the impact of lesion depth on MI signal detectability, and-when present-describe the distribution of MI appearances. Methods: In this single-center, retrospective study from January 2021 to December 2023, we screened 91 patients with 121 focal hepatic lesions on ultrasound. Lesions without typical hemangioma enhancement on dynamic MRI or dynamic CT were excluded. Two radiologists independently assessed MI signals and patterns using the Jeon classification, blinded to clinical and CT/MRI data; inter-observer agreement was quantified with Cohen's κ. Results: Of 121 screened lesions, 36 lacked typical enhancement and were excluded; 85 hemangiomas remained. A total of 13 were excluded for motion artifacts near the heart or pulsatile vessels, yielding 72 hemangiomas (61 patients) for analysis. No lesion showed flow on color or power Doppler. MI signals were detected in 68/72 hemangiomas (94.4%). Among signal-positive lesions (n = 68), the patterns were non-specific in 25.0% (17/68), nodular rim in 22.1% (15/68), strip rim in 17.6% (12/68), central dot-like in 16.2% (11/68), peripheral dot-like in 10.3% (7/68), and staining in 8.8% (6/68). Signal-negative lesions were deeper than signal-positive lesions (median depth: 85 mm vs. 41.5 mm; p < 0.05). The inter-observer agreement was very good (κ = 0.821, 95% CI 0.767-0.921). Conclusions: MI is a reproducible, contrast-free technique that demonstrates hemangioma vascularity with high detection rates, particularly in more superficial lesions. In this cohort, lesion depth rather than size was the primary determinant of MI signal detectability. MI should be considered complementary to CT/MRI and may be especially useful where contrast agents are unavailable or contraindicated.