By marko_laaksomaa 
-
-
- Fri Mar 18, 2022 11:17 am
#315
Hi everyone
Here some of our new results:
Background: H&N setup using AlignRT with Orfit 3-point open face hybrid mask (Group A, patient n= 18, 445 kV- image pairs) and laser setup with Orfit 5- point closed masks (Group L, patient n=25, 437 kV-image pairs) were evaluated. During the mask formation, CT scanning and at the linac it was checked that patient mouth was closed and arms were relaxed on the side. In group A, SRS faceROI (head positioning) was used together with postural video (chest wall+shoulder positioning). When needed, after systematic patient posture errors were corrected, we captured new reference surfaces for the upcoming setups in Group A. Posture correction was done based on kV/kV- images, at the beginning or during the treatment course. Online match was performed based on our matching guidelines to the c3 vertebrae. In both groups translational and rotational errors were daily corrected with 0 AL in the online match. Pitch or roll corrections were not done based on IGRT. Realized, offline matched errors to the sub-regions after daily couch correction based on online match are given. Residual errors to the vertebrae c1, c5/7, mandible, occiput bone and left and right clavicle (from the measuring point of 9 cm from the vertebral midline laterally) were retrospectively evaluated. Re-setups in both groups were rare and have no significant effect on comparison. Systematic (S), random (R) errors and margins (M) are given in this order together with percentages of the residual errors exceeding 3 mm. Two-tailed F-test was applied for systematic errors (test for equality of variances). The Wilcoxon rank sum test was applied for random errors (test for equality of means). A p-value ≤ 0.05 was considered statistically significant. Results: (mm) C1: (Group L) (AP) 0.9 mm, 0.9 mm, 2.9 mm, (4%), (CC) 0.4 mm, 0.6 mm, 1.4 mm, (1%), (LAT) 0.8 mm, 1.1 mm, 2.6 mm, (7%), (Group A) (AP) 0.7 mm, 0.8 mm, 2.4 mm, (0%), (CC) 0.4 mm, 0.8 mm, 1.5 mm, 1%, (LAT) 0.4 mm, 0,8 mm, 1.6 mm, (1%). C5/7: (Group L) (AP) 1.0 mm, 1.0 mm, 3.2 mm, (5%), (CC) 0.7 mm, 0.9 mm, 2.4 mm, (3%), (LAT) 0.8 mm, 1.2 mm, 2.9 mm, (7%), (Group A) (AP) 0.8 mm, 0.9 mm, 2.6 mm, (1%), (CC) 0.5 mm, 1.1 mm, 1.9 mm, (1%), (LAT) 0.7 mm, 1.1 mm, 2.6 mm, (5%). Mandible: (Group L) (AP) 1.2 mm, 1.4 mm, 4.0 mm, 17%, (CC) 0.8 mm, 1.1 mm, 3.1 mm, 8%, (LAT) 0.5 mm, 1.2 mm, 1.6 mm, (5%), (Group A) (AP) 1.1 mm, 1.2 mm, 3.3 mm (7%), (CC) 0.9 mm, 1.1 mm, 3.1 mm, (4%), (LAT) 0.5 mm, 0.8 mm (p=0.02), 1.7 mm, (0%). Occiput bone: (Group L) (AP) 1.1 mm, 1.3 mm, 3.6 mm, (9%), (CC) 0.8 mm, 1.0 mm, 2.6 mm, (3%), (LAT) 0.8 mm, 1.4 mm, 3.1 mm, (9%), (Group A) (AP) 1.0 mm, 1.2 mm, 3.4 mm, (4%), (CC) 0.7 mm, 1.1 mm, 2.5 mm, (4%), (LAT) 0.6 mm, 1.0 mm (p=0.003), 2.1 mm, (1%). Clavicle (R&L) (Group L) (30 patients+1692 clavicle match) (CC) 2.9 mm, 3.4 mm, 9.8 mm, exceeding of 6 mm 18%, exceeding of 1 cm 5%. (Group A) (16 patients+ 864 clavicle match) (CC) 1.9 mm (p=0.01), 3.7 mm, 7.4 mm, exceeding of 6 mm 13%, exceeding of 1 cm 3%. Conclusion: Based on this preliminary study, sufficient realized posture accuracy was found in both groups around large PTV with well- planned workflows in the whole process. 3- point open masks with AlignRT and FaceROI led to excellent accuracy of the mandible. In the clavicles in CC direction, with AlignRT and postural setup, systematic residual errors were smaller (p=0.01), but the repeatability needs improvements in both groups. SGRT- results are promising, more patients (Group A) and CBCTs are needed to create more detailed analysis.
Larger 42" monitor with AlignRT has been useful when using postural video for shoulder positioning inside treatment room.
Have a nice weekend.
Best regards Marko
Here some of our new results:
Background: H&N setup using AlignRT with Orfit 3-point open face hybrid mask (Group A, patient n= 18, 445 kV- image pairs) and laser setup with Orfit 5- point closed masks (Group L, patient n=25, 437 kV-image pairs) were evaluated. During the mask formation, CT scanning and at the linac it was checked that patient mouth was closed and arms were relaxed on the side. In group A, SRS faceROI (head positioning) was used together with postural video (chest wall+shoulder positioning). When needed, after systematic patient posture errors were corrected, we captured new reference surfaces for the upcoming setups in Group A. Posture correction was done based on kV/kV- images, at the beginning or during the treatment course. Online match was performed based on our matching guidelines to the c3 vertebrae. In both groups translational and rotational errors were daily corrected with 0 AL in the online match. Pitch or roll corrections were not done based on IGRT. Realized, offline matched errors to the sub-regions after daily couch correction based on online match are given. Residual errors to the vertebrae c1, c5/7, mandible, occiput bone and left and right clavicle (from the measuring point of 9 cm from the vertebral midline laterally) were retrospectively evaluated. Re-setups in both groups were rare and have no significant effect on comparison. Systematic (S), random (R) errors and margins (M) are given in this order together with percentages of the residual errors exceeding 3 mm. Two-tailed F-test was applied for systematic errors (test for equality of variances). The Wilcoxon rank sum test was applied for random errors (test for equality of means). A p-value ≤ 0.05 was considered statistically significant. Results: (mm) C1: (Group L) (AP) 0.9 mm, 0.9 mm, 2.9 mm, (4%), (CC) 0.4 mm, 0.6 mm, 1.4 mm, (1%), (LAT) 0.8 mm, 1.1 mm, 2.6 mm, (7%), (Group A) (AP) 0.7 mm, 0.8 mm, 2.4 mm, (0%), (CC) 0.4 mm, 0.8 mm, 1.5 mm, 1%, (LAT) 0.4 mm, 0,8 mm, 1.6 mm, (1%). C5/7: (Group L) (AP) 1.0 mm, 1.0 mm, 3.2 mm, (5%), (CC) 0.7 mm, 0.9 mm, 2.4 mm, (3%), (LAT) 0.8 mm, 1.2 mm, 2.9 mm, (7%), (Group A) (AP) 0.8 mm, 0.9 mm, 2.6 mm, (1%), (CC) 0.5 mm, 1.1 mm, 1.9 mm, (1%), (LAT) 0.7 mm, 1.1 mm, 2.6 mm, (5%). Mandible: (Group L) (AP) 1.2 mm, 1.4 mm, 4.0 mm, 17%, (CC) 0.8 mm, 1.1 mm, 3.1 mm, 8%, (LAT) 0.5 mm, 1.2 mm, 1.6 mm, (5%), (Group A) (AP) 1.1 mm, 1.2 mm, 3.3 mm (7%), (CC) 0.9 mm, 1.1 mm, 3.1 mm, (4%), (LAT) 0.5 mm, 0.8 mm (p=0.02), 1.7 mm, (0%). Occiput bone: (Group L) (AP) 1.1 mm, 1.3 mm, 3.6 mm, (9%), (CC) 0.8 mm, 1.0 mm, 2.6 mm, (3%), (LAT) 0.8 mm, 1.4 mm, 3.1 mm, (9%), (Group A) (AP) 1.0 mm, 1.2 mm, 3.4 mm, (4%), (CC) 0.7 mm, 1.1 mm, 2.5 mm, (4%), (LAT) 0.6 mm, 1.0 mm (p=0.003), 2.1 mm, (1%). Clavicle (R&L) (Group L) (30 patients+1692 clavicle match) (CC) 2.9 mm, 3.4 mm, 9.8 mm, exceeding of 6 mm 18%, exceeding of 1 cm 5%. (Group A) (16 patients+ 864 clavicle match) (CC) 1.9 mm (p=0.01), 3.7 mm, 7.4 mm, exceeding of 6 mm 13%, exceeding of 1 cm 3%. Conclusion: Based on this preliminary study, sufficient realized posture accuracy was found in both groups around large PTV with well- planned workflows in the whole process. 3- point open masks with AlignRT and FaceROI led to excellent accuracy of the mandible. In the clavicles in CC direction, with AlignRT and postural setup, systematic residual errors were smaller (p=0.01), but the repeatability needs improvements in both groups. SGRT- results are promising, more patients (Group A) and CBCTs are needed to create more detailed analysis.
Larger 42" monitor with AlignRT has been useful when using postural video for shoulder positioning inside treatment room.
Have a nice weekend.
Best regards Marko
and daily CBCT. It is our workflow nowadays to acquire daily CBCTs with 6D corrections of H&N patients and in addition to accurate radiation theapy, retrospective posture evaluation is now easier. If the individual bony structures are offline matched including rotations and the purpose is to evaluate and give data of the groups concerning the realized posture of the patient with different methods, I would prioritize in the evaluation the difference between the individual bony structures (difference between c1 and c7 for example), not the residual errors to the structures after online match. Due to several things, general margin calculation you named, is a bit blurred if the individual bony structures are offline matched with rotations and margins are such calculated based on residual errors of those. For example, online matching location should be exactly the same, which in practice is not completely possible, since different matching locations are emphasized based on differently located PTVs and based on soft tissue deformations. As well, in daily practice individual bony structures can not be rotated and to offline rotate those may give different translational residual error to those than the data without rotations. On the other hand, for the data evaluation concerning patient posture, it may remain unclear how accurately the clavicle is positioned for example, if its rotations are not included. To always perform the automatch (to create the small VOI) to the same location of the structure to all the patients is important and a challenge in the data collection.