Translating iterative reconstruction for CT to the clinic

Abstract

Introduction: Increasing numbers of CT examinations has raised concerns about the possible risk of carcinogenesis due to radiation exposure. Therefore, CT examinations should only be performed when there is a proper indication. Furthermore, several developments in both CT hardware and software have been made to reduce the radiation dose. Iterative reconstruction (IR), an alternative reconstruction method to conventional filtered back projection (FBP), reduces the amount of noise thereby allowing to acquire images at a lower radiation dose level. Experimental assessment of IR showed that the radiation dose can be reduced while preserving objective image quality compared to a routine dose acquisition reconstructed with FBP. The potential radiation dose reduction with IR to submillisievert dose levels also offers the opportunity to replace conventional imaging with low dose CT for certain indications. Purpose: The first aim of this thesis was to investigate whether IR can be applied in a routine clinical setting by investigating the effect of IR on commonly used imaging-based quantification methods and on diagnostic accuracy. The second aim of this thesis was to investigate if IR should be applied in the field of imaging prior to cardiac surgery, by replacing the conventional preoperative chest radiograph by a low dose chest CT. Results:In this thesis, we showed in a prospective study that IR enables dose reduction to submillisievert levels in a clinical setting for coronary calcium scoring and pulmonary nodule assessment while image quality and interpretability are maintained. Furthermore, based on literature research we found an achievable dose reduction of around 50% for cardiothoracic CT using IR with preserved image quality. Although these results are promising, IR can also affect commonly used imaging-based quantification methods such as the coronary calcium score and pulmonary nodule volumetry. These unwanted effects are more prominent when using more advanced fully model-based IR algorithms.Furthermore, the potential replacement of conventional chest radiography prior to cardiothoracic surgery by low dose CT was studied. The routine preoperative chest radiograph infrequently affects the planned surgery or surgical approach. While, based on the current literature, a preoperative chest CT more frequently affects the surgical approach which ultimately results in a decreased postoperative stroke and mortality rate. A clinical trial in which patients are randomized between an additional low dose CT or conventional imaging only was proposed, which is currently ongoing. Conclusion: IR can substantially reduce patient radiation dose and is increasingly becoming embedded in routine clinical care. However, for each clinical indication the type and strength of the IR algorithm needs to be chosen with care to obtain accurate results with diagnostic image quality. IR potentially allows for replacement of conventional imaging by low dose CT for specific indications.