نبذة مختصرة : Objective: Speed-of-Sound (SoS) images, derived from minute misalignments between ultrasound images consecutively acquired using a motionrobust technique, were studied with multi-frame acquisitions to derive a lesion assessment metric for differential diagnosis of breast cancer. Methods: SoS elevation within inclusions is characterized as SoS contrast by comparing inclusions to their surrounding tissue. Malignancy is predicted as having larger SoS contrast compared to benign inclusions. In a study involving 100 patients with various breast lesion types, findings were compared to gold-standard diagnoses based on biopsy and expert consensus. Results: The proposed SoS-contrast metric achieved an overall malignancy classification with an AUC of 69.70% (CI: 57.4-81.2%), with results varying depending on lesion type. Accounting for breast density or lesion size improved the AUC to 77.40% (CI: 59.2–93.3%) and 78.09% (CI: 56.2–95.8%), respectively. Focusing on BI-RADS 4 lesions, where mammography and B-mode imaging are inconclusive, a sensitivity of 25% at 100% specificity or a specificity of 23% at 100% sensitivity was attained. Conclusion: The study highlights the feasibility of SoS contrast as an imaging biomarker for the differential diagnosis of breast lesions. It also indicates the need for further technical advancements in SoS imaging to strengthen its ability to effectively classify lesions and to enable interactive measurements during US examinations. ; ISSN:0301-5629
No Comments.