Speed-of-sound (SoS) has large potential for tissue and pathology differentiation. We aim to develop a novel Ultrasound Computed Tomography (USCT) technique that can reconstruct local SoS in tissue on conventional ultrasound machines with hand-held linear arrays. Methods: A passive reflector is placed opposite the tissue sample as an echogenic reference to measure the time-of-flight (ToF) of ultrasound wavefronts. A Dynamic Programming algorithm provides a robust ToF measurements based on global optimization of all transmit-receive echo data. An Anisotropically-Weighted Total Variation (AWTV) algorithm allows sharp delineation of focal lesions based on limited-angle USCT data. Results: Inclusions, which are not visible in conventional ultrasound, could be delineated in SoS images. AWTV allows to reconstruct focal lesions with a contrast-ratio of 93.7% of their nominal value, compared to that of 31.5% with conventional least-squares based algebraic tomographic reconstruction. In fullwave simulations of realistic heterogeneous breast models, a high CR of 84.3% is observed, with the reconstruction filtering out background heterogeneity. In experiments, our proposed method quantifies SoS in a homogeneous background with an accuracy of 0.93 m s −1 , allowing to differentiate several tissue types. Conclusion: We validate our method using numerical simulations with ray-tracing and full-wave models, and phantom and ex-vivo data. Preliminary in-vivo results show the potential of this new technique to detect and differentiate malignant and benign lesions in the breast. Significance: Breast cancer is the most common cancer in women. Ultrasound B-mode only provides qualitative information about breast lesions, whereas USCT can provide quantitative tissue imaging biomarkers, such as SoS. The proposed method can potentially be implemented as a complementary modality to ultrasound for tissue and disease differentiation.