Effect of different multiplanar reformation algorithms on image quality of intraoperative three-dimensional fluoroscopy.

Different multiplanar reformation (MPR-512 and -256) algorithms of intraoperative acquired 3-D-fluoroscopy data exist without recommendations for use in the literature. To compare algorithms, 3-D-fluoroscopic data sets of 46 radius fractures were blinded and processed using MPR-256 and -512 (Ziehm, Vision-Vario 3D). Each reformatted data set was analysed to evaluate image quality, fracture reduction quality and screw misplacements. Overall image quality was higher rated in the MPR-512 compared with the MPR-256 (3.2 vs. 2.2 points, scale 1-5 points), accompanied by a reduced number of scans that could not be analysed (10 vs. 19%). Interobserver evaluation of fracture reduction quality was fair to moderate (independent of the algorithm). In contrast, for screw misplacements MPR-depended ratings were found (MPR-256: fair to moderate; MPR-512: moderate to substantial). Optimization of post-processing algorithms, rather than modifications of image acquisition, may increase the image quality for assessing implant positioning, but limitations in evaluating fracture reduction quality still exist.