Please use this identifier to cite or link to this item: doi:10.22028/D291-27271
Title: Employment statistics and positioning accuracy of three different image-guidance systems; CTV-PTV margin calculation for two different populations
Author(s): Beyhs, Monica Elisa
Language: English
Year of Publication: 2015
Place of publication: Homburg/Saar
SWD key words: Positioniergenauigkeit
Free key words: CTV-PTV
DDC notations: 610 Medicine and health
Publikation type: Doctoral Thesis
Abstract: Purpose of this study: To evaluate the impact kilovoltage imaging, a new imaging technology at our department, has had on the treatment of our patients and to determine the preferred imaging modality and treatment machine for different situations. Additionally, based on a retrospective evaluation of the alignment data for two different patient populations, set-up errors for individual patients and for the two populations, as well as adequate treatment margins are calculated. Methods and materials: Three different linacs, the Siemens Arstiste ART1 and ART 2 and the Siemens Oncor ONC2, each offering different imaging modalities for IGRT, were used. The ART1 is equipped with the best imaging technique, kilovoltage imaging. The data from 89 patients, 36 prostate cancer and 53 head-and-neck cancer patients, receiving IGRT between January 2013 and December 2013, was analyzed. A total of 3061 radiation fractions and 1260 set-up images were retrospectively evaluated. Set-up errors were calculated for three dimensions (superior-inferior (S-I), left-right (L-R) and anterior-posterior (A-P)) based on the alignment data from the couch shifts performed, in order to account for set-up variation detected in the verification imaging prior to treatment delivery. CTV-PTV margins were calculated applying the set-up errors to common margin recipes: 2Σ+0.7σ (Stroom et al., 1999) and 2.5Σ+0.7σ (van Herk et al., 2000). Results: Linac employment statistics: The preferred treatment machine for head-and-neck cancer treatment was the linac ART2 (56%) followed by the ART1 (31%) and the ONC2 (13%). Prostate cancer patients were treated mostly at the ART1 (44%) and the ONC2 (37%). The ART2 was used the least (19%). Planar imaging was used more frequently than CBCT imaging. The frequency of CBCT imaging was higher for prostate cancer patients in comparison to head-and-neck cancer patients. At the ART1 the best imaging modality, kV imaging, was not always used. Especially when treating head-and-neck cancer IBL planar images were often preferred. Set-up errors and margins: Systematic set-up errors (μ) for the individual head-and-neck cancer patients range mainly from -0.15 to 0.2 cm (S-I), -0.1 to 0.3 cm (L-R) and -0.1 cm to 0.1 cm (A-P) and for the individual prostate cancer patients from -0.4 to 0.25 cm (S-I), -0.3 to 0.1 cm (L-R) and -0.3 to 0.2 cm (A-P). Random errors (r) for individual head-and-neck cancer patients range from 0.06 to 0.43 cm (S-I), 0.07 to 0.48 cm (L-R) and 0.09 to 0.40 cm (A-P) and for individual prostate cancer patients from 0.15 to 0.72 cm (S-I), 0.21 to 0.83 cm (L-R) and 0.11 to 0.51 cm (A-P). The population systematic set-up error (M) for both populations is approximately zero; in all three directions values are smaller than 1 mm. The variation of the population systematic error (Σ􀍿 is larger for prostate cancer patients with values of 0.16 cm (S-I), 0.18 cm (L-R), and 0.17 cm (A-P) in comparison to 0.10 cm in all three directions for head-and-neck cancer patients. The population random error (σ) for head-and-neck cancer patients is 0.19 cm (L-R and A-P) and 0.22 cm (S-I) and for prostate cancer patients 0.37 cm (S-I), 0.49 cm (L-R) and 0.35 cm (A-P). No clear differences are found in the calculation of set-up errors when using the data from the different imaging modalities separately. The calculated margins for head-and-neck cancer patients are 0.35 cm (S-I) and 0.33 cm (L-R and A-P) according to Stroom et al. (1999) or 0.4 cm (S-I) and 0.38 cm (L-R and A-P) according to van Herk et al. (2000) and for prostate cancer patients 0.57 cm (SI), 0.70 cm (L-R) and 0.58 cm (A-P) according to Stroom et al. (1999) and 0.66 cm (S-I) 0.79 cm (L-R) and 0.67 cm (A-P) according to van Herk et al. (2000). Conclusion: The study shows that at our department anatomic differences are taken into account when choosing the different linacs and imaging techniques. Generally, for IGRT better imaging is used for patients with larger set-up errors. However, the study has also revealed that especially at the linac ART1 there is still room for improvement. As shown, prostate cancer patients are subject to much greater geometric variability in comparison to head-and-neck cancer patients. The random population set-up errors (σ) are larger for prostate cancer in comparison to head-and-neck cancer patients. The study has demonstrated that there is practically no systematic set-up error (M) for either patient population. Therefore, the quality assurance measures in our clinic are shown to be successful. Additionally, this study has revealed that the choice of imaging modality used for IGRT does not have a clear effect on the calculation of set-up errors. According to the margin recipes presented by van Herk (2000) and Stroom et al. (1999) the CTV-PTV margins applied in our clinic for prostate and head-and-neck cancer patients are sufficient.
Zielsetzung: Evaluierung der Auswirkungen einer neuen Kilovolt-Bildgebungstechnologie für IGRT in unserer Klinik und Ermittlung der bevorzugten Bildgebungsmethode und Bestrahlungsmaschine für verschiedene Situationen. Berechnung von systematischen und randomisierten Lage-Fehlern in der Positionierung von einzelnen Patienten und Patienten Populationen anhand einer retrospektiven Auswertung der Verschiebungsdaten der Bestrahlungs-Liege, sowie Berechnung der adäquaten CTV-PTV Expansion. Methoden und Materialien: Drei verschiedene Bestrahlungs-linacs (Siemens Artiste ART1 und ART2, Siemens Oncor ONC2) mit unterschiedlichen Bildgebungstechnologien für IGRT wurden verwendet. Ausgestattet mit Kilovolt Energie für die Bildgebung liefert der Beschleuniger linac ART1 die beste Bildqualität. Die Daten von 89 Patienten, 36 Prostatakrebs und 53 HNO-Krebs Patienten, welche zwischen Januar und Dezember 2013 Bestrahlung erhielten, wurden analysiert. Insgesamt 3061 Bestrahlungsfraktionen und 1260 Bildgebungen wurden retrospektiv evaluiert. Anhand der Daten der einzelnen Verschiebungen der Bestrahlung-Liege nach Bildgebung wurden Lage-Fehler (set-up error) in drei Dimensionen berechnet (superior-inferior (S-I), links-rechts (L-R) und anterior-posterior (A-P)). Die Fehler wurden verwendet um "CTV-PTV-marigns" anhand herkömmlicher Formeln, 2Σ+0.7σ (Stroom et al., 1999) und 2.5Σ+0.7σ (van Herk et al., 2000), zu berechnen. Ergebnisse: Linac-Anwendungsstatistik: Der bevorzugte Beschleuniger für HNO Patienten war der linac ART2 (56%), gefolgt von dem ART1 (31%) und dem ONC2 (13%). Prostata Patienten wurden meist am ART1 (44%) und am ONC2 bestrahlt; der ART2 wurde nur selten verwendet (19%). 2D Achsen Aufnahmen wurden häufiger als 3D CBCT Aufnahmen verwendet, jedoch erhielten im Vergleich Prostata Patienten deutlich häufiger CBCT Aufnahmen als HNO Patienten. Nicht immer wurde am ART1 die Energie für die beste Bildqualität, Kilovolt, für die Bildgebung verwendet. Insbesondere für HNO Patienten wurden häufig IBL Achsen Aufnahmen bevorzugt. "Set-up" Fehler und "CTV-PTV Expansion": Systematische Fehler (μ) für einzelne HNO-Patienten lagen zwischen -0.15 cm und 0.2 cm (SI), -0.1 cm und 0.3 cm (L-R), sowie -0.1 cm und 0.1 cm (A-P), für einzelne Prostata Patienten zwischen -0.4 cm und 0.25 cm (S-I), -0.3 cm und 0.1 cm (L-R), sowie -0.3 cm und 0.2 cm (A-P).
Link to this record: urn:nbn:de:bsz:291-scidok-ds-272710
Advisor: Niewald, Marcus
Date of oral examination: 21-Sep-2016
Date of registration: 4-Jul-2018
Faculty: M - Medizinische Fakultät
Department: M - Radiologie
Collections:SciDok - Der Wissenschaftsserver der Universität des Saarlandes

Files for this record:
File Description SizeFormat 
Employment statistics and positioning accuracy.pdf2,46 MBAdobe PDFView/Open

Items in SciDok are protected by copyright, with all rights reserved, unless otherwise indicated.