Dear Editor, We thank Dr Frauscher for his careful appraisal of our recent study [1]. We are delighted that there are positive remarks. However there are several aspects of the commentary we disagree with.The Edey et al. study [2] is misquoted. This study was an extension of a phase II multicentre trial (i.e. the current study) performed at a single site. It was therefore not the first study but a reported sub-series of patients (investigated with agent detection imaging, ADI)-an extension to the original large trial for which we are reporting. Edey et al. reported artefact only with the high doses with ADI and no other imaging technique. The primary aim of the paper was to see if ADI worked in the late phase following all the scheduled imaging as per protocol and to ascertain whether there was sufficient Sonazoid left around to perform ADI. Therefore there was no artefact described in our study with the low and high MI techniques, and this was not a secondary investigation within the definition of the protocol.Secondly, the Bernatik T et al. study [3] was of 1,349 patients, not 1,349 lesions. Our study was a detection study of lesions in the late phase while the Bernatik study was a characterization study of lesions during the complete vascular phase (arterial, portal and late). We have no doubt that were Sonazoid used for characterization, there would be improved diagnostic performance, and we agree that use of late phase imaging alone may influence accuracy of CEUS for liver imaging. Please note, however, that our results are relatively modest (78% sensitivity and 70% accuracy for 0.12 μL/kg dose group) in part because our per lesion analysis incorporated a relatively rigorous standard of reference (SOR) including the use of off-site reviewers. Incorrect lesion assignment to Couinard segments may also have led some true-positive lesions on US or CEUS being erroneously labelled as false positive on the SOR.We suggest that this dose-finding study, coupled with the very rigorous off-site blinded assessment, demonstrates the very robust nature of this technique using a versatile microbubble agent.The letter to which this is the reply can be found at http://dx.doi.