Phantom studies and clinical application of high resolution, image reconstruction using (18)F-fluoromethylcholine...

The aim of this study is to evaluate the impact of a high resolution (HR) image reconstruction with a voxel size of 2mm in comparison to the most routinely used standard reconstruction with 4mm voxels in patients suffering from prostate cancer having undergone (18)F-methylcholine PET/CT. Phantom studies were performed using a Jaszczak phantom and a custom made phantom containing small hot lesions (size 2-10mm). Clinical evaluation was performed on PET/CT scans of 50 patients. Images were reconstructed with 4mm and 2mm voxel size and analyzed quantitatively using AMIDE and MATLAB. Clinical images were judged by two observers concerning TNM staging, image quality and the correlation of PET and CT data. Phantom studies revealed increased SUVmean and SUVmax values in the HR images (P<0.01). The lower detection limit was approximately 3mm in the HR and 4-5mm in the conventional images. Lower FWHM values were found in the HR images. No significant difference was found concerning the image quality and the correlation of PET and CT (each P>0.5). For both reconstructions, a comparable total amount of lesions was reported (P>0.5) with no impact on the TNM staging. In conclusion, the HR PET reconstruction provides semi-quantitative advantages in the sense of an improved lower detection limit and increased semi-quantitative tumour-to-background ratios. In the setting of choline PET/CT for prostate cancer the high resolution reconstruction could be implemented clinically as there are no relevant qualitative differences between this and the conventional image resolution in terms of image quality, assessment confidence and lesion identification rate.