Kenichi Yokoyama MD, Presenter: Nothing to Disclose

Masamichi Imai, Abstract Co-Author: Nothing to Disclose

Toshiaki Nitatori MD, Abstract Co-Author: Nothing to Disclose

Misako Yorimitsu MD, Abstract Co-Author: Nothing to Disclose

Shigehide Kuhara MS, Abstract Co-Author: Nothing to Disclose

Shuhei Nitta, Abstract Co-Author: Nothing to Disclose

Tomoyuki Takeguchi, Abstract Co-Author: Nothing to Disclose

Nobuyuki Matsumoto, Abstract Co-Author: Nothing to Disclose

To determine the clinical feasibility of a prototype automatic slice-alignment method based on gated breath-hold axial multislice scout images to simplify cardiac scan planning.

Steady-state free precession sequences covering the range from the cardiac base to the apex were used to acquire multislice images in approximately 20 seconds. Some anatomical feature points of the heart were detected from these series of images using the knowledge-based technique which was designed to distinguish image patterns around the points, and all of the planes required for cardiac imaging were calculated based on them. 35 consecutive patients with cardiac diseases (myocardial infarction, angina pectoris, hypertrophic cardiomyopathy, dilated cardiomyopathy, apical ballooning, and other conditions) were included in this study. The images acquired with different scan conditions were visually scored by two cardiac radiologists using a 4-point scale, and were corrected if detected planes were obviously failed. In addition, the time required was measured and compared with that for the conventional method.

Automatic slice alignment was performed successfully in 35 (100%) of the 35 patients, and slice alignment was performed quickly and accurately for each of the short-axis, 4-chamber, 2-chamber, and 3-chamber images, on averaged scored 3.70±0.74, 3.57±0.84, 3.69±0.65, and 3.73±0.65, respectively. The average angular corrections of the each images were 1.30±3.92, 1.16±3.87, 0.17±1.01 and 0.00±0.00 degree, respectively. The patient with the worst score had pericardial tumor. It was assumed that knowledge-based algorithm could not work well since such pattern had not been learned. The time required from the start of examination to the completion of slice alignment was 1 minute and 50 seconds, which is significantly shorter than that for the conventional method.

This method is clinically useful for the evaluation of patients with various cardiac shapes and diseases. It is expected that knowledge-based recognition will provide more accurate setting results after a larger number of cases have been examined.

A new automatic slice-alignment method for cardiac MRI can simplify scan planning and employs knowledge-based recognition techniques to achieve higher accuracy for a variety of cardiac shapes.

Yokoyama, K, Imai, M, Nitatori, T, Yorimitsu, M, Kuhara, S, Nitta, S, Takeguchi, T, Matsumoto, N, Early Clinical Experience with New Automatic Slice Alignment Method for Cardiac Magnetic Resonance Imaging.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11001668.html