Differentiation Between Fibro-Adipose Vascular Anomaly and Intramuscular Venous Malformation Using Grey-Scale...

Area of Science:

• Medical Imaging
• Radiology
• Vascular Anomalies

Background:

• Distinguishing fibro adipose vascular anomaly (FAVA) from intramuscular venous malformation (VM) is crucial for appropriate clinical management.
• Ultrasound-based radiomics offers a non-invasive approach for characterizing soft tissue lesions.

Purpose of the Study:

• To develop and evaluate an ultrasound-based radiomics model for differentiating FAVA from VM.
• To compare the diagnostic performance of support vector machine (SVM) and random forest (RF) models in this differentiation.

Main Methods:

• Retrospective analysis of 65 patients with VM and 31 patients with FAVA.
• Feature extraction from ultrasound images, followed by dimensionality reduction using the least absolute shrinkage and selection operator (LASSO).
• Development of SVM and RF models for classification, with performance evaluated using receiver operating characteristic (ROC) analysis.

Main Results:

• A radiomics model with seven selected features was established.
• The SVM model achieved higher AUC (0.841), accuracy (96.6%), and specificity (100.0%) compared to the RF model (AUC 0.791, accuracy 93.1%, specificity 90.5%) in the testing group.
• The RF model exhibited higher sensitivity (100.0%) than the SVM model (88.9%).

Conclusions:

• An ultrasound-based radiomics model can effectively differentiate FAVA from VM.
• The SVM model demonstrates superior performance in terms of accuracy and specificity for this diagnostic task.
• This radiomics approach provides a valuable new tool for the clinical diagnosis of vascular anomalies.