Shockwaves have been used routinely for the treatment of urolisthesis for decades. Strong evidences suggest that cavitation bubble activity plays an important role on the stone breakage. Recently, shockwaves were also shown to be effective in the treatment of certain soft tissue disorders such as tendinosis. However, whether or not cavitation occurs in tight tissue and contributes to tendon healing as well as pain relief remain unknown. In order to understand the possible physical mechanism of the shockwave therapy, we focused on searching the evidences of cavitation in tissue during the shockwave treatment, and also the determining factors of bubble generation. It was hypothesized that the shock-wave induced cavitation was constrained by the space for bubble expansion, and thus occurred more easily in vessels of larger diameters than capillaries, intercellular space and inside cells. To prove our hypothesis, a series of studies was performed to detect the presence of inertia! cavitation in soft tissue blocks of various densities ex-vivo and elastic tissue phantoms embedded with vessels of different size in-vitro. We found that shock-wave induced bubble generation highly depended on the 'porosity' (equivalent to the perfusion status) of the ex-vivo tissue blocks. The bubble generation in muscle was substantially less than that in liver, and was undetectable in tendon. For vessel phantoms, cavitation occurred easily in vessels with diameters greater than or equal to 0.7 mm when tap water was pumped through. However, no trace of bubble signal was found in smaller vessels under the same acoustic conditions. Size threshold existed. After adding low concentrations of contrast agent (Definity®) in the flow-through fluid, cavitation threshold dropped substantially. For a vessel of 0.32 mm in diameter, bubble and signal enhancement could be easily obtained even after a single shot of shockwave. Furthermore, the presence of cavitation-induced bubbles significantly enhanced the signal-to-noise ratio of vessels in the B-mode image, comparing with the use of contrast agent alone. Concentration as low as 0.005% (v/v) could elevate the B-mode's averaged gray level in the target area up to 76 levels, much higher than the enhancement contributed by contrast agent alone (about 2 to 3 levels in this concentration).