Reframing multiple reconstructed early dynamic PET images for smooth image rendering

Ziel/Aim Nuclear medicine PET list-mode acquisition allows a flexible data collection which can be rebinned as static, dynamic or gated images for different time intervals. For early dynamic processes, the first seconds and minutes after tracer injection are most important. Sometimes the critical information or the pathology is hidden inside an early dynamic timeframe. One approach to overcome this limitation is the method of overlapping reconstruction of list-mode data in high temporal resolution and a subsequent rearrangement of the frames, so-called reframing.

Methodik/Methods Data were acquired as list-mode using a standard PET/CT scanner (Biograph mCT, Siemens, Erlangen, Germany) and acquisition parameters were matrix 400, iterations 4, subsets 12 and Gaussian filter FWHM 5. Based on the image quality, the length of a single frame was determined to 7 seconds and the number of reconstructed frames to 10, depending on the clinical need. Subsequently, 6 additional reconstructions were calculated, each with one second time offset. The result was a data set comprising 70 frames of 7s length. Postprocessing and reframing was performed using PMOD (PMOD Technologies Ltd, Zurich, Switzerland).

Ergebnisse/Results A visually improved dynamic imaging was observed on different examples, e.g. an angiographic blood flow examination. For illustration of the pathological findings, the PET frames with reframing in comparison to the conventional reconstructions without overlapping provide much more information and dynamic smoothness.

Schlussfolgerungen/Conclusions Reframing produces smooth and improved dynamic PET imaging because short-termed or temporally limited processes can be assessed more accurately. However, this method seems too elaborate to be frequently used in clinical routine.

Publication History

Article published online:
08 April 2021