Dynamic programming has previously been used when measuring Lumen Diameter (LD) and Intima-Media Thickness (IMT) in single frame ultrasound carotid artery images . An extension to multiframe detection of that procedure is proposed in this paper. Our new method transforms a three dimensional problem into multiple two dimensional problems that can again be solved by dynamic programming. First, several 'candidate boundaries' are detected in each image. Then, one boundary from each image in the sequence is selected in such a way that the sequence of boundaries is optimal among the considered possibilities. A model of the movements of the boundaries is used in the second step. Our proposed method shows promising performance on both synthetic and real ultrasound data.