Automatic patient alignment system using 3D ultrasound. Issue 4 (20th March 2013)

• Record Type: Journal Article
• Authors: Kaar, Marcus; Figl, Michael; Hoffmann, Rainer; Birkfellner, Wolfgang; Hummel, Johann; Stock, Markus; Georg, Dietmar; Goldner, Gregor
• Is Part Of: Medical physics. Volume 40:Issue 4(2013)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2013-03-20
• Subjects: Registration -- Computed tomography -- Ultrasonographic imaging -- Dosimetry/exposure assessment -- Therapeutic applications; biological organs -- biomedical ultrasonics -- computerised tomography -- dosimetry -- image registration -- medical image processing -- radiation therapy -- tumours -- ultrasonic imaging -- ultrasonic therapy; radiation‐therapy -- ultrasound guidance -- tracking; Computerised tomographs -- Diagnosis using ultrasonic, sonic or infrasonic waves -- Radiation therapy -- Ultrasound therapy -- Digital computing or data processing equipment or methods, specially adapted for specific applications -- Image data processing or generation, in general; Ultrasonography -- Medical imaging -- Computed tomography -- Image sensors -- Calibration -- Dosimetry -- Image registration -- Radiation treatment -- Digital image processing -- Radiation therapy; Medical physics -- Periodicals; Medical physics; Geneeskunde; Natuurkunde; Toepassingen; Biophysics; Periodicals; Periodicals; Electronic journals; 610.153
• DOI: 10.1118/1.4795129
• Languages: English
• ISSNs: 0094-2405

Abstract

Abstract : Purpose: : Recent developments in radiation therapy such as intensity modulated radiotherapy (IMRT) or dose painting promise to provide better dose distribution on the tumor. For effective application of these methods the exact positioning of the patient and the localization of the irradiated organ and surrounding structures is crucial. Especially with respect to the treatment of the prostate, ultrasound (US) allows for differentiation between soft tissue and was therefore applied by various repositioning systems, such as BAT or Clarity. The authors built a new system which uses 3D US at both sites, the CT room and the intervention room and applied a 3D/3D US/US registration for automatic repositioning. Methods: : In a first step the authors applied image preprocessing methods to prepare the US images for an optimal registration process. For the 3D/3D registration procedure five different metrics were evaluated. To find the image metric which fits best for a particular patient three 3D US images were taken at the CT site and registered to each other. From these results an US registration error was calculated. The most successful image metric was then applied for the US/US registration process. The success of the whole repositioning method was assessed by taking the results of an ExacTrac system as golden standard. Results: : The US/US registration error was found to be 2.99 ± 1.54 mm with respect to the mutual information metric by Mattes (eleven patients) which revealed to be the most suitable of the assessed metrics. For complete repositioning chain the error amounted to 4.15 ± 1.20 mm (ten patients). Conclusions: : The authors developed a system for patient repositioning which works automatically without the necessity of user interaction with an accuracy which seems to be suitable for clinical application.