High-frequency ultrasound combined with superb microvascular imaging for differential diagnosis of basal cell carcinoma and seborrheic keratosis: a retrospective analysis with nomogram implementation

Objective To evaluate the diagnostic performance of high-frequency ultrasound (HFUS) combined with superb microvascular imaging (SMI) in differentiating basal cell carcinoma (BCC) from seborrheic keratosis (SK), and to develop a nomogram for individualized probability estimation. Methods This retrospective analysis included 208 patients (49 BCC, 159 SK) with histopathologically confirmed lesions and complete preoperative HFUS and SMI records. Grayscale ultrasound assessed morphological features (size, depth, shape, margins, echogenicity, etc.), while SMI evaluated microvascular flow patterns and grading. Univariate and multivariate logistic regression analyses identified independent predictors, and a nomogram was constructed to estimate the probability of BCC. Results Univariate analysis revealed significant differences between BCC and SK in terms of age, lesion location, size, depth, morphology, internal echotexture, posterior acoustic features, and microvascular flow patterns (all p 0.001). Multivariate logistic regression analysis identified larger tumor size, head location, nodular ultrasound pattern, and random vascular flow as independent predictors of BCC. The resulting multivariate logistic regression model demonstrated good discriminative performance, with an area under the receiver operating characteristic curve (AUC) of 0.861. At the optimal cutoff, the model yielded a sensitivity of 77.55% and a specificity of 82.38%. A nomogram was subsequently constructed based on these independent predictors to facilitate individualized probability estimation of BCC in clinical practice. Conclusion HFUS combined with SMI provides a valuable non-invasive approach for differentiating BCC from SK. The developed nomogram, based on key ultrasound and clinical predictors, demonstrates good discriminative performance and offers a practical tool to support clinical decision-making, potentially reducing unnecessary biopsies and facilitating earlier detection of malignant lesions.