Publikationsrepositorium - Helmholtz-Zentrum Dresden-Rossendorf

During the session titled "The use of PET-CT for lung cancer" several topics will be addressed, such as the technical background of PET-CT and an overview of the possibilities offered by PET imaging in lung cancer. The aim of all PET imaging in treatment planning for tumours is an improved delineation of the Gross Tumour Volume (GTV). Knowing exactly the position and extent of the vital tumour should increase local control and lead to longer survival of patients. As long as the target is immobile as is the case in the brain fusion of the PET information into the planning CT seems to be effortless. But problems arise when PET imaging is to be used for moving targets such as in lung. Therefore, the issue to be discussed in this presentation is "Should we match on bone or on tumour?". Necessity of motion comprehension during PET-CT imaging: In PET-CT imaging two different motion-related effects have to be considered:

(1) the direct effect of motion on the CT and PET modalities separately, and
(2) the effect of improper attenuation correction caused by a mismatch in the temporal resolutions of the two data sets.Examples of direct effects are artefacts in the region between the lung and liver in 3D-CT data and motion-induced smearing of the activity concentration in 3D-PET data. These artefacts are unavoidable in 3D imaging but can be overcome with the introduction of 4D imaging (4D-CT and 4D-PET). The second effect (i.e. smearing) has to be considered in both 3D- and 4D-imaging. In 3D-imaging the 3D-CT artefacts described above can propagate into the PET data set due to attenuation correction. One example given by Beyer et al. shows 3D PET-CT data that appears to show soft tissue in the lung, but which in fact is part of the moving liver (Eur J Nucl Med Mol Imaging 2003). Additionally, the different temporal resolution of 3D-PET and 3D-CT can generate artefacts in attenuation corrected data sets although both 3D-CT and uncorrected 3D-PET are free of artefacts. One impressive example of this phenomenon was shown by Osman et al. (J Nucl Med 2003), where liver metastases appeared in the lung region. In 4D-imaging, artefacts due to improper attenuation correction can be avoided if the different datasets of the 4D-PET are corrected with corresponding datasets of the 4D-CT, as has been demonstrated by Pnisch et al. (Med Phys 2008).
Possibilities for motion compensation during treatment: Several technical approaches have been introduced into clinical practice to compensate for motion during radiotherapy.
Two questions have to be addressed: a) How can the motion be turned into a signal and b) how can this signal be used to compensate the motion? Answering question a), helpful techniques which can be mentioned include spirometry, pressure based systems or optical systems detecting the motion of points on the thorax or of the whole body surface. In all these techniques, the motion leads to a signal that represents more or less precisely the motion of the patient. However, it is not proven that any signal also exactly follows the motion of the tumour. Answering question b), two approaches can be mentioned: "Gating" and "Tracking". Gating uses the motion signal to trigger the linear accelerator according to a defined "gate", i.e. to turn the beam on and off in synchronization with the breathing motion. Tracking is a method in which the motion of the target is followed (tracked) either by motion of the leaves in the MLC or by motion of the table. Verification of motion in the chain from planning to treatment: 4D treatment planning including 4D-PET/4D-CT and breathing-synchronized radiotherapy for lung cancer patients need a carefully considered concept of verification.
In contrast to conventional verification procedures at the linear accelerator, where the position of the patient is verified only by bony anatomy (brain), target motion adapted methods require appropriate additional imaging in the treatment room. Either radioopaque markers or kV-CT images are necessary. For organs with mainly interfractional motion such as the prostate, correct positioning of the target before the start of the fraction is essential. The treatment of intrafractional moving targets imposes more complexity. Besides verification of the baseline, the amplitude of the motion is a second point to be considered. The most important in the verification of a breathing synchronized radiotherapy is the comparison with the 4D treatment planning dataset. The basic principle is the use of the identical motion signal both for 4D imaging and for the 4D treatment. Only then a possible shift between real tumour motion and visualized tumour motion can be neglected.
Conclusions: Should we match on bone or on tumour? The short answer should be ON TUMOUR. But the longer answer is, BUT NOT FOR EVERY PATIENT. The challenge is the finding of clinical values used as guidance as to which patients will benefit from the high workload of 4D treatment planning and delivery.