Abstract
Objectives: Methods for incorporation of PET-CT into radiotherapy are classified as direct, where the CT component of PET-CT is used for planning, or indirect by registering a previously acquired PET-CT to the treatment planning CT. In this study we compare the spatial accuracy of indirect radiotherapy planning using a standard staging PET-CT and dedicated planning PET-CT.
Methods: Five patients with head and neck malignancy referred for radiotherapy following PET-CT staging were enrolled with informed consent. All patients had an additional PET-CT scan acquired while immobilized in the treatment position. The accuracy of rigid (R) and non-rigid (NR) registration of the CT component of the staging and dedicated PET-CT to planning CT was ranked and scored (1-5) by two independent readers. The transformations derived from the CT were applied to the PET components of the staging and dedicated PET-CT scans. The PET data were ranked and scored (1-6) for alignment and image quality.
Results: There was substantial agreement in CT registration scoring and ranking (kappa 0.70 and 0.73 respectively). For registration with the planning CT, NR-CTdedicated was rated best (global misalignment <2mm), R-CTstaging was worst (rigid structure misalignment 2-5mm, >5mm in soft-tissue), NR-CTstaging and R-CTdedicated gave similar results (rigid and local soft-tissue misalignment <2mm). PET reader agreement was worse (kappa 0.31 and 0.12), NR-PETdedicated and R-PETdedicated (excellent global alignment, no artefacts) scored better than NR-PETstaging and R-PETstaging (good local alignment, no artefacts).
Conclusions: Radiotherapy planning on a PET-CT scan acquired in the treatment position offers greater accuracy than using a scan originally acquired for disease staging.
Methods: Five patients with head and neck malignancy referred for radiotherapy following PET-CT staging were enrolled with informed consent. All patients had an additional PET-CT scan acquired while immobilized in the treatment position. The accuracy of rigid (R) and non-rigid (NR) registration of the CT component of the staging and dedicated PET-CT to planning CT was ranked and scored (1-5) by two independent readers. The transformations derived from the CT were applied to the PET components of the staging and dedicated PET-CT scans. The PET data were ranked and scored (1-6) for alignment and image quality.
Results: There was substantial agreement in CT registration scoring and ranking (kappa 0.70 and 0.73 respectively). For registration with the planning CT, NR-CTdedicated was rated best (global misalignment <2mm), R-CTstaging was worst (rigid structure misalignment 2-5mm, >5mm in soft-tissue), NR-CTstaging and R-CTdedicated gave similar results (rigid and local soft-tissue misalignment <2mm). PET reader agreement was worse (kappa 0.31 and 0.12), NR-PETdedicated and R-PETdedicated (excellent global alignment, no artefacts) scored better than NR-PETstaging and R-PETstaging (good local alignment, no artefacts).
Conclusions: Radiotherapy planning on a PET-CT scan acquired in the treatment position offers greater accuracy than using a scan originally acquired for disease staging.
| Original language | English |
|---|---|
| Title of host publication | The journal of Nuclear Medicine |
| Pages | 258 |
| Volume | 55 |
| Edition | Supplement 1 |
| Publication status | Published - 2014 |