Purpose of the Study: Thin-section MDCT images are conventionally reconstructed with an overlap of 50% resulting in an high number of images to be stored on PACS. We wanted to test whether reducing the number of chest CT studies images by wider reconstruction intervals (RIs) may influence either the image quality of multiplanar reformations (MPR) and 3D volume rendering reconstructions (3D-VR) or the performance of lung nodule computer-aided volumetry. Methods: Thin-section chest CT images of 30 patients with 54 lung nodules were reconstructed with the same slice thickness (2 mm) and different RIs, obtaining image overlap values of 50%, 40%, 30%, 20% and 10%, respectively. MPR, 3D-VR and lung nodules volumetry of each dataset were compared using the 50% overlap one as the reference standard by two radiologists. Results: MPR were considered diagnostic in every dataset, with only minimal step-ladder artifacts seen in the 20% and 10% overlap datasets. No significant differences in image quality were reported in 3D-VR images at any RI. Only lung nodules volumetry was significantly influenced by the RI values: datasets with lesser degrees of overlap provided smaller and larger volumes in 195/216 (90.3%) and 21/216 (9.7%) measurements respectively as compared to the reference standard. Mean volume variations were 3.39%, 5.83%, 7.59% and 8.82% in the 40%, 30%, 20% and 10% datasets, respectively. Conclusions: Wider RIs result in a considerably lower number of images to be stored on PACS. This has limited effect on image quality of MPR and 3D-VR but may influence lung nodule volumetry results. A trade-off is to keep standard 50% image overlap on datasets used for nodule volumetry and apply lower overlaps (till 10%) on datasets used for MPR and 3D-VR reconstructions. This approach would reduce images number of chest MDCT studies up to 30%.
Reducing the number of thin-section CT images of the chest by using wider reconstruction intervals: effects on image quality of multiplanar and 3D reformations and lung nodule volumetry results / Bartalena, T.; Rinaldi, M. F.; Braccaioli, L.; Sverzellati, Nicola; Mattioli, S.; Battista, G.; Zompatori, M.; Canini, R.. - In: JOURNAL OF THORACIC IMAGING. - ISSN 0883-5993. - 24:3(2009), p. 410.
Reducing the number of thin-section CT images of the chest by using wider reconstruction intervals: effects on image quality of multiplanar and 3D reformations and lung nodule volumetry results.
SVERZELLATI, Nicola;
2009-01-01
Abstract
Purpose of the Study: Thin-section MDCT images are conventionally reconstructed with an overlap of 50% resulting in an high number of images to be stored on PACS. We wanted to test whether reducing the number of chest CT studies images by wider reconstruction intervals (RIs) may influence either the image quality of multiplanar reformations (MPR) and 3D volume rendering reconstructions (3D-VR) or the performance of lung nodule computer-aided volumetry. Methods: Thin-section chest CT images of 30 patients with 54 lung nodules were reconstructed with the same slice thickness (2 mm) and different RIs, obtaining image overlap values of 50%, 40%, 30%, 20% and 10%, respectively. MPR, 3D-VR and lung nodules volumetry of each dataset were compared using the 50% overlap one as the reference standard by two radiologists. Results: MPR were considered diagnostic in every dataset, with only minimal step-ladder artifacts seen in the 20% and 10% overlap datasets. No significant differences in image quality were reported in 3D-VR images at any RI. Only lung nodules volumetry was significantly influenced by the RI values: datasets with lesser degrees of overlap provided smaller and larger volumes in 195/216 (90.3%) and 21/216 (9.7%) measurements respectively as compared to the reference standard. Mean volume variations were 3.39%, 5.83%, 7.59% and 8.82% in the 40%, 30%, 20% and 10% datasets, respectively. Conclusions: Wider RIs result in a considerably lower number of images to be stored on PACS. This has limited effect on image quality of MPR and 3D-VR but may influence lung nodule volumetry results. A trade-off is to keep standard 50% image overlap on datasets used for nodule volumetry and apply lower overlaps (till 10%) on datasets used for MPR and 3D-VR reconstructions. This approach would reduce images number of chest MDCT studies up to 30%.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
