Preoperative Differentiation of Non-Subungual Glomus Tumors from Other Superficial Soft Tissue Tumors Using a Clinical and Ultrasound-Based Model

<b>Objectives:</b> Non-subungual glomus tumors (NSGTs) are rare neoplasms arising outside the nail bed, typically presenting with disproportionate pain. Surgical excision usually achieves complete relief. Delayed or incorrect diagnosis of glomus tumors may also result in incomplete excision, which is the major cause of recurrence. Ultrasound is a well-recognized tool for preoperative evaluation, but the rarity of NSGTs complicates differentiation from angioleiomyomas (ALMs) and hemangiomas, which have alternative managements. This study aimed to establish an ultrasound diagnostic model and scoring system for NSGTs. <b>Methods:</b> A total of 58 NSGTs, 68 ALMs, and 67 hemangiomas confirmed pathologically between January 2014 and March 2025 were analyzed. Clinical and ultrasonographic features were evaluated, and significant variables were incorporated into a scoring time using the binary logistic regression. A total of 11 NSGTs, 10 ALMs, and 10 hemangiomas (April-August 2025) were used to validate the diagnostic performance of the scoring system. <b>Results:</b> Four independent predictors were identified: pain, upper extremity location, diameter < 10 mm, and vascular stalk sign (<i>p</i> < 0.05). The model differentiated NSGTs from ALMs and hemangiomas, yielding an AUC of 0.91, with sensitivity of 75.86% and specificity of 90.37%. The overall accuracy was 86.01%. <b>Conclusions:</b> Ultrasound proves valuable in diagnosing NSGTs. A logistic regression model based on pain, location, size, and vascular stalk sign shows high diagnostic performance and clinical utility in distinguishing NSGTs from other superficial soft tissue tumors.