© The Author 2015. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. Objective. To produce consensus-based definitions of the US elementary lesions in gout and to test their reliability in a web-based exercise. Methods. The process consisted of two steps. In the first step a written Delphi questionnaire was developed from a systematic literature review and expert international consensus. This collated information resulted in four statements defining US elementary lesions: double contour (DC), tophus, aggregates and erosion. The Delphi questionnaire was sent to 35 rheumatology experts in US, asking them to rate their level of agreement or disagreement with each statement. The second step tested the reliability by a web-exercise. US images of both normal and gouty elementary lesions were collected by the participants. A facilitator then constructed an electronic database of 110 images. The database was sent to the participants, who evaluated the presence/absence of US elementary lesions. A group of 20 images was displayed twice to evaluate intra-reader reliability. Results. A total of 32 participants responded to the questionnaires. Good agreement (>80%) was obtained for US definitions on DC, tophus, aggregates and erosion in the Delphi exercise after three rounds. The reliability on images showed inter-reader κ values for DC, tophus, aggregates, erosion findings of 0.98, 0.71, 0.54 and 0.85, respectively. The mean intra-reader κ values were also acceptable: 0.93, 0.78, 0.65 and 0.78, respectively. Conclusion. This, the first consensus-based US definition of elementary lesions in gout, demonstrated good reliability overall. It constitutes an essential step in developing a core outcome measurement that permits a higher degree of homogeneity and comparability between multicentre studies.