Room: Exhibit Hall | Forum 6
Purpose: To investigate clinic impact of using newly-available iterative reconstructed (IR) MVCT for IGRT.
Methods: MVCT images acquired on a Radixact unit for both phantoms and patients were reconstructed with standard and IR algorithms. Phantom tests were performed to check image quality, imaging dose, IVDT table consistency, IGRT capability. Patient images were surveyed among 15 radiation oncology staffs. IGRT difference was assessed by comparing the centroid location difference of the rigidly transferred structures from registering the standard and IR images to planning images. Deformable image registration (DIR) was evaluated by utilizing Accuray’s PreciseArt module with planning structures transferred to the two rigidly aligned MVCT image sets and compared with manually drawn “ground truth,� and quantified with mean distance to agreement (MDA) and Dice similarity coefficient (DSC). Doses on both standard and IR images were calculated by applying the plan sinogram, and assessed with DVHs.
Results: Tests with phantom indicated that, with IR, image noise reduced by >50% and uniformity improved by >20 HU. Image survey demonstrated that standard MVCT was preferred (2/3 of surverors) for head and neck cancer but IR-MVCT was preferred exclusively for other sites. Difference between the centroid location of structures from IR and standard image can be as large as 4 mm. DIR with IR improves the DSC by 15% averagely for prostate patients, but not for head and neck, however, the MDA and DSC values are both within the recommended tolerance per AAPM TG132. With deformed structures, visible difference in DVHs were seen in IR and standard images.
Conclusion: The IR improves soft tissue contrast and substantially reduces image noise while keeping other features similar. With IR MVCT, patient alignment during IGRT is easier and more accurate as compared with the standard MCVT. DIR contours improved with IR, meaningful differences were seen in DVH plots.