PhD Defence Charlotte van der Vos

Breathing and metal artifact reduction in combined pet/ct imagingĀ 

Medical imaging has an established role in the clinical management of most cancers. One of these medical imaging techniques is positron emission tomography (PET). For a PET scan a radioactive-labelled tracer, a radiopharmaceutical, is injected intravenously in the patient. Depending on the used radiopharmaceutical, different biological processes can be visualized. At the moment, the most widely used PET-tracer is a radioactive glucose analogue (18F-fluorodeoxyglucose or FDG) that accumulates in cells with an increased glucose metabolism (for instance cancer cells). This radiopharmaceutical is also used for the clinical studies published in this thesis. FDG-PET can be used to detect and stage malignant disease, but can also be used to determine tumour volume and tumour boundaries, and to evaluate the treatment effect.

PET scanners are constantly being improved. An overview of the most recent developments is given in chapter 2. In this chapter both recently available hardware and software techniques are described. Two of these techniques are being addressed in more detail in the following chapters. These chapters focus on the reduction of errors (artefacts) caused by the combination of PET and computer tomography (CT) imaging.

Currently, PET is performed in combination with CT or magnetic resonance imaging (MRI). Hybrid imaging provides an important advantage since both anatomical and functional information is visualized. The additional advantage besides providing anatomical information is the ability to use these scans to correct for attenuation of photons. However, the combination of PET and CT can also cause artefacts, for instance respiratory and metal artefacts. Even though these artefacts influence the quality of the PET images, the ability to perform attenuation correction almost always outweighs such concerns. Nonetheless, it is preferred to reduce these artefacts to improve the accuracy of the interpretation and quantification of the PET/CT image.

Reduction of breathing artefacts

PET has relatively long scan times and patients are instructed to breathe freely during the acquisition. Due to the respiration of the patient the PET images can show blurring of the structures within the thorax and upper abdomen, resulting in inaccuracies in the interpretation of the images. Nowadays, it is possible to correct the PET images for the respiration of the patient. By implementing a respiratory gating technique a sharper image can be created. This improves the accuracy of the quantification, which is of importance to accurately determine the response to treatment.

Respiration does not only cause problems within the PET images, but can also result in complications with the registration of both images. In PET/CT scans performed for standard clinical care, the breathing instructions during the CT scan are similar to the breathing instructions provided during acquisition of the PET scan: patients are asked to breathe normally. The CT scan, however, is made relatively fast and can be considered an essentially motion free image. When the PET and CT images are matched, the possibility arises that the images do not represent the patient during the same moment of the respiratory cycle. This will lead to a mismatch of the location of anatomical structures between the PET and CT image. This not only causes difficulties in interpretation of the images, but can also lead to other inaccuracies. Since the CT images are used for the correction of the PET images, inaccuracies in the quantification of radiopharmaceutical uptake can arise. It is therefore important to acquire the PET and CT images at the same phase during the respiratory cycle. In this thesis three methods have been tested to reduce mismatch between PET and CT due to respiratory motion.

In chapter 3 the respiration at during the PET and CT acquisition have been synchronized by using simple respiratory instructions during the CT acquisition. For the study described in this chapter the patient cohort was divided into two separate groups. In the first patient group breathing instructions were provided during the CT scan, resulting in an expiratory breath-hold CT. The second group did not receive breathing instructions and was allowed to breath normally during the CT scan. The analysis of the overlap of PET and CT images showed that providing breathing instructions did not reduce mismatch significantly. This can be due to the general difficulties associated with the breathing instruction protocol, such as the interpretation and the ability of the patient to comply with the breath-hold instructions.

The second method to reduce mismatches is described in chapter 4. In this study also breathing instructions were used. However, these instructions were more personalized as compared to the previous study. In the current study, patients underwent one PET scan and two CT scans. The first CT scan was acquired without breathing instructions and served as control scan, while the second CT scan (acquired after the PET) was performed while using breathing instructions. During the PET scan and the second CT scan the respiration of the patient was recorded. The breathing instructions during the CT scan were timed to match the respiratory amplitude limits of the PET reconstruction. This method provides improved control over the exact starting point of the CT acquisition during the respiratory cycle. The personalized breathing instructions during the CT scan resulted in increased overlap between PET and CT.

In both studies, however, patients experienced difficulties with the breathing instructions. It is for some patients difficult, and in a few cases even impossible, to hold their breath during the CT scan. Therefore, in the third study, described in chapter 5, a different method to reduce respiratory motion artefacts was tested. In this study the respiratory signal was used to trigger the CT scan. Similar to the previous study, each patient underwent one PET and two CT scans. For the second CT scan a different scan protocol was used: a multiple step acquisition protocol instead of the conventional spiral protocol. This protocol used the optimized respiratory amplitude limits of the reconstructed PET images to determine the exact timing of these steps. For this study no breathing instructions were requested and patients could breathe normally during both CT scans and the PET scan. The results demonstrated that the overlap between PET and CT improved when this method was used.

Reduction of metal artefacts

In chapter 6 the influence of reduction of metal artefacts on PET/CT image quality was tested. Metal implants in the body can lead to severe artefacts on CT images. This applies to larger implants, such as hip and shoulder prostheses, but also to smaller metal implants (i.e. screws, dental implants), although in a lesser extent. In hybrid PET/CT images this could have impact on the interpretation of the scan by the nuclear medicine physician, but could also influence exact quantification of tracer uptake. There are methods available to reduce these metal artefacts on CT and recently they are also available for combined PET/CT scanners. In the study described in this chapter a metal artefact reduction method was applied to PET/CT scans of patients with different types of implants and different PET scan indications. The images reconstructed with and without the metal artefact reduction tool, were scored in two ways. First of all, quantitative differences were measured between the two image sets. Secondly, two nuclear medicine physicians scored the quality of the images, influence of metal artefact reduction on image interpretation and confidence level. The results of this study showed that the metal artefact reduction method improved the quality of the CT images and due to a reduction in attenuation correction artefacts also had a positive impact on the quality of the PET images. This did not only result in a more accurate quantification of radiopharmaceutical uptake, but also led to a higher confidence level during image interpretation.

In this thesis an overview of the most recent developments in the field of PET/CT is given. Furthermore, different artefact reduction methods and techniques have been described to reduce artefacts that are caused by the combination of the PET and CT scan. Innovations in artefact correction techniques will have a significant impact on the improvement of quality and accuracy of multimodality imaging techniques, now and in the future.