SU‐E‐U‐10: Ultrasound Based Deformable Image Registration: Daily CT Images Derived From Daily IGRT Ultrasound
Purpose: CT‐images for every treatment fraction could be useful for daily dose recalculation. Acquisition of daily CT‐images imposes extra radiation burden and is usually not available in the treatment room. Nowadays, 3D Ultrasound (US) IGRT systems are available to assess absolute volumetric inform...
Gespeichert in:
Veröffentlicht in: | Medical Physics 2013-06, Vol.40 (6), p.375-375 |
---|---|
Hauptverfasser: | , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Purpose: CT‐images for every treatment fraction could be useful for daily dose recalculation. Acquisition of daily CT‐images imposes extra radiation burden and is usually not available in the treatment room. Nowadays, 3D Ultrasound (US) IGRT systems are available to assess absolute volumetric information of soft‐tissue on a daily basis. These intramodality US‐systems have a reference CT and US image acquired during planning stage. We are testing the hypothesis that, to generate a daily CT, deformable image registration (DIR) could be used. The deformation field (DF) calculated between the reference US and the daily US is applied to the reference CT. Methods: A deformable phantom was developed, consisting of two liquid filled balloons in a water phantom, mimicking a bladder‐prostate model. Four different balloon volume configurations were imaged with CT and US. DFs are computed between pairs of US‐images (US 1 and US 2 ) and then applied to CT1. This reconstructed CTdef is then compared to the real corresponding CT2 using the sum of squared differences (SQD) metric. The DFs are calculated using a DIR algorithm (REGGUI; morphons). The SQD metric was limited to the area where US information was available. 'Results: The DIR performed well visually in all cases. For all the image pairs, the SQD of CTdef was approximately half of the SQD of the non‐registered images (ratio 0.45–0.59). The DIR reduced the SQD differences between CTdef and CT2 to the level of SQD differences between images that only differed in noise content. Conclusion: A limitation of an US based DF is that the area of the CT on which one can perform the DF is limited to the area of which US data is available. Assuming the target organ is present in the US‐image, this method could reduce dose errors introduced by using the reference CT for the entire treatment. |
---|---|
ISSN: | 0094-2405 2473-4209 |
DOI: | 10.1118/1.4815163 |