Satoshi Goshima MD, Presenter: Nothing to Disclose

Toshiharu Miyoshi RT, Abstract Co-Author: Nothing to Disclose

Hiroshi Kondo MD, Abstract Co-Author: Nothing to Disclose

Yusuke Tsuge MD, Abstract Co-Author: Nothing to Disclose

Haruo Watanabe MD, Abstract Co-Author: Nothing to Disclose

Masayuki Kanematsu MD, Abstract Co-Author: Nothing to Disclose

To develop and evaluate a new image processing algorithm to reduce artifacts caused by metallic implants in phantom and human subjects.

This study consisted of two parts; a phantom and a clinical study. The algorithm consists of three steps: first, x-ray absorption areas caused by metallic object were extracted on a photon counts curve of each slice displayed on a CT console, and a threshold and channel width of the counts decreased by metallic objects were determined. Second, the decreased photon counts areas were interpolated by counts of adjacent areas, and third, the interpolated photon count data were processed by a conventional image reconstruction algorithm. For the phantom study, we prepared 24 algorithms with different four thresholds and six channel width, and then applied to a dose measurement phantom simulating the pelvis with metallic hip joints. For the clinical study, the most efficient algorithm was applied to raw data in 20 patients with metallic hip joints (7 male and 13 female, mean age; 67.6 years, range 42-86). We measured CT values (HU) in five ROIs inside of metallic objects on phantom and on clinical images. Two radiologists assessed the image quality.

With the phantom study, 7 of 24 algorithms were efficient in reducing metallic artifact. The most efficient algorithm increased the mean CT values by 49.5 HU (P < .001) and reduced metallic artifacts by 69.1% in terms of SD measurement (P < .001). The radiologists also assigned the highest scores in image quality with this algorithm (P < .01). With the clinical study, the algorithm was efficient in improving mean CT values of the five ROIs obtained in 17 of 20 patients and in reducing SD of the ROIs in all patients. Overall mean ± SD obtained in twenty patients were -34.4 ± 35.0 HU with the algorithm, and were -74.8 ± 109.3 HU without the algorithm. The radiologists also assigned higher scores in image quality with the algorithm than without (P < .05).

The new algorithm helps to improve CT image quality by reducing metallic artifacts and to provide more detailed radiographic information for soft tissues otherwise affected by implanted metals.

This algorithm was feasible in clinical study. Further evaluations were needed for other metallic objects, such as artificial teeth, vertebroplasty, and preplanning for prostate brachytherapy.

Goshima, S, Miyoshi, T, Kondo, H, Tsuge, Y, Watanabe, H, Kanematsu, M, Developing New Image Processing Algorithm for Metallic Artifacts Reduction with Multidetector Row CT: Phantom Study and Preliminary Clinical Application.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6010400.html