Current echocardiographic devices provide only 2-dimensional views of the heart. To appreciate 3-dimensional structural relations, therefore, requires mental reconstruction of 2-dimensional views by an experienced observer. Our ability to answer new questions about the heart could be increased if 2-dimensional images could be combined to display 3-dimensional relations. Such 3-dimensional reconstruction would permit analysis of structures of unknown or complex shape and the noninvasive quantification of cardiac chamber size and function without making geometric assumptions. To overcome previous limitations, mechanisms have been developed for automated integration of images and positional data during routine echocardiographic scanning, thereby greatly enhancing the efficiency and application of image reconstruction. Refining the diagnosis of mitral valve prolapse has presented a uniquely 3-dimensional problem requiring information previously unavailable from the 2-dimensional technique. To date, 3-dimensional studies have demonstrated that the mitral valve is saddle-shaped in systole, so that apparent superior leaflet displacement in the mediolateral 4-chamber view, often seen in otherwise normal individuals, lies entirely within the bounds defined by the mitral annulus and occurs without leaflet distortion or actual displacement above the entire mitral valve. Other applications of 3-dimensional image reconstruction include calculation of ventricular volume and ejection fraction by transthoracic or transesophageal scanning without geometric assumptions; improving the standardization and accuracy of 2-dimensional measurements by improving spatial appreciation; and 3-dimensional reconstruction of vascular walls to guide interventions. In the future, systems for acquiring multiple views more rapidly by parallel processing and improving endocardial border extraction should allow more routine application of 3-dimensional methods as the next stage in the evolution of cardiac ultrasound, thereby expanding the range of questions that can be answered. Achieving these goals will depend, in large measure, on persistence in developing the necessary technology.