Improvement of image quality of multislice spiral CT scans of the head and neck region using a raw data-based multidimensional adaptive filtering (MAF) technique.

Abstract

The purpose was to evaluate the potential of the multidimensional adaptive filtering (MAF) technique by investigating its effects on image noise and image quality in multislice spiral CT (MSCT) examinations of the head and neck region. Fifty patients with head and neck tumors were examined using MSCT with a high resolution protocol. Reconstructions were performed using dedicated reconstruction software with a standard algorithm both without and with MAF using different modification. In all reconstructions, we measured the noise in seven different anatomical structures. The image quality and image noise were rated on a five-point scale. There was a significant (P<0.05) reduction in mean pixel noise in the reconstructions using MAF in comparison to the standard reconstructions, but there was no significant difference between the different modification fractions. With MAF the mean reduction in noise level was 60%, depending upon body shape and anatomical region. Independently from the used modification fraction, MAF led to a significant (P<0.05) improvement of image quality. In direct comparison of the different filter strength, the optimal image quality was achieved in the investigations with 15% MAF. The use of MAF facilitates the distinction of anatomical and pathological structures from artifacts in the supraclavicular fossae and the upper mediastinum, whereas the image quality of the upper portions of the neck remained unchanged. MAF improved image quality by reducing the noise level and removing noise structures without loss of image sharpness. This technique offers new perspectives to reduce the patient dose.