Suspicion for Sarcoma: Clinical Presentation, Multi-Modality Imaging Evaluation, and Ultrasound Artificial...

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Breast sarcomas are rare and aggressive tumors.
• Characterizing their presentation across imaging modalities is crucial for early diagnosis.

Purpose of the Study:

• To detail the clinical, histological, and imaging features of breast sarcomas.
• To evaluate the effectiveness of artificial intelligence decision support (AI DS) in detecting breast sarcomas.

Main Methods:

• Retrospective review of 18 patients with histologically proven breast sarcomas (2008-2024).
• Analysis of mammography, ultrasound, and MRI findings using BI-RADS lexicon.
• Retrospective application of AI DS to ultrasound images of breast masses.

Main Results:

• Mammographic findings were diverse, including masses, asymmetries, and skin thickening.
• Ultrasound commonly showed irregular masses with non-circumscribed, hypoechoic, and vascular features.
• AI DS identified 93.8% of ultrasound-detected breast masses as suspicious.

Conclusions:

• Breast sarcomas exhibit varied imaging appearances across modalities.
• AI DS can aid radiologists in diagnosing rare and aggressive breast sarcomas.