Feasibility of breathing-adapted PET/CT imaging for radiation therapy of Hodgkin lymphoma


Aim: Respiration can induce artifacts in positron emission tomography (PET)/computed tomography (CT) images leading to uncertainties in tumour volume, location and uptake quantification. Respiratory gating for PET images is now established but is not directly translatable to a radiotherapy setup. We investigate the feasibility of introducing a deep inspiration breath hold (DIBH) strategy in PET/CT imaging of Hodgkin lymphoma patients and its impact on image quantification parameters.  Methods: Three patients with suspicion of large mediastinal tumour burden were selected for this study and imaged in a Siemens Biograph 40 PET/CT before chemotherapy after injection of 400 MBq of FDG. The DIBH level reproducibility was ensured by visual coaching. Clinical free breathing (FB) scan was followed by a DIBH scan, acquired as 3 breathholds of 20 s each (1 PET field of view). Images were reconstructed with a TrueX algorithm (point spread function, 3 iterations 21 subsets, 2-mm full width half maximum Gaussian filtering). Active areas of uptake were delineated using automatic thresholds of 40% and 50% of maximum standardized uptake value (SUVmax).  Results: SUVmax is on average 20% higher in DIBH than in FB images (range 18-22%) due to the absence of breathing motion. In contrast, the volumes at 40% and 50% of SUVmax are smaller at DIBH than at FB (by up to 80 cm(3)).  Conclusion: DIBH is feasible and well tolerated in this patient group, and results in considerable changes in apparent uptake in PET/CT images. These results suggest that advanced therapies (such as SUV-based dose painting) will likely require breathing-adapted PET images and that the relevant SUV thresholds are yet to be investigated.
Original languageEnglish
JournalCancer Imaging
Pages (from-to)S117
Publication statusPublished - 2011


Dive into the research topics of 'Feasibility of breathing-adapted PET/CT imaging for radiation therapy of Hodgkin lymphoma'. Together they form a unique fingerprint.

Cite this