Radiation Physics

PV QA 3 - Poster Viewing Q&A 3

TU_15_3258 - Deformation-Guided Hybrid Approach for Interfractional Variations

Tuesday, October 23
1:00 PM - 2:30 PM
Location: Innovation Hub, Exhibit Hall 3

Deformation-Guided Hybrid Approach for Interfractional Variations
S. N. LIM1, Y. Zhang1, E. E. Ahunbay2, C. A. F. Lawton3, and A. Li1; 1Medical College of Wisconsin, Milwaukee, WI, 2Medical College of Wisconsin, Department of Radiation Oncology, Milwaukee, WI, 3Medical College of Wisconsin Department of Radiation Oncology, Milwaukee, WI

Purpose/Objective(s): It has been well documented that interfraction variations (primarily organ deformation) that cannot be fully accounted for by image-guided radiation therapy (IGRT) (i.e., patient repositioning) but can be corrected using online adaptive replanning (OLAR). However, OLAR is time consuming and labor intensive compared to the repositioning process. A hybrid approach that can uses either OLAR or repositioning would be practical as OLAR is not necessary for all fractions. The major issue is to rapidly and objectively determine if OLAR is needed after daily image is acquired. To address issue, we investigate whether the magnitude of deformation can be used to guide the selection of OLAR versus repositioning.

Materials/Methods: Daily CTs acquired using an in-room CT during radiation therapy for 10 representative patients with prostate cancer were deformably registered with reference CTs. The deformable registration was verified in accordance with the AAPM TG-132 report. The deformation vector field (DVF) was quantified using the Jacobian determinant matrix. Plans were created with a prescribed dose of 36.25 Gy to 95% of the planning target volume (PTV) of prostate plus a 3 mm margin in five fractions. For each daily CT, repositioning and OLAR plans were generated. Five fractional doses from either a repositioning or an OLAR plan as indicated by the DVF were deformed back to the planning CT and accumulated for the hybrid approach. Doses from five repositioning plans (all repo) and five replanning (all OLAR) were also accumulated for comparison.

Results: The Jacobian determinants measured the magnitude of deformation between the reference and daily CTs. Appropriate threshold on the amount of deformation to indicate the necessity of OLAR was determined. Depending on the amount of deformation, a hybrid approach including OLAR and repositioning was used. Table I presents a comparison of the plan qualities of the hybrid approach, all repo and all OLAR. It can be seen that the hybrid approach as determined by the magnitude of deformation can generate plans comparable to those from all OLAR treatments.

Conclusion: Analysis of interfraction anatomic deformation is a robust yet fast method to determine instances in which online adaptive replanning is necessary. This process can be completed without any segmentation, thus is fast (less than 1 minute), and allows us to quickly and objectively determine which fractions require adaptive replanning in a hybrid approach.
Table I

PTV V100 (%)

OAR V50% dose (%)

Hybrid (# repo) all OLAR all repo hybrid all OLAR all repo


95.0 90.6


15.0 15.0
P2 94.9 (1) 95.0 94.7 16.7 16.1 19.5
P3 95.0 (1) 95.0 95.7 15.0 15.0 19.0
P4 95.0 (1) 95.0 87.2 22.7 22.7 21.6
P5 95.2 (1) 95.2 92.3 23.8 23.8 22.2
P6 94.8 (2) 95.7 90.3 8.1 8.1 9.3
P7 95.0 (3) 95.0 95.1 5.1 5.1 5.7
P8 96.2 (3) 96.2 94.4 23.9 23.9 23.2
P9 95.5 (4) 95.0 95.5 22.3 21.1 23.1
P10 95.0 (4) 95.1 93.2 34.7 32.8 34.7
*OLAR required for all fractions

Author Disclosure: S.N. LIM: None. E.E. Ahunbay: None. C.A. Lawton: None. A. Li: None.



Send Email for SARA LIM


TU_15_3258 - Deformation-Guided Hybrid Approach for Interfractional Variations

Attendees who have favorited this

Please enter your access key

The asset you are trying to access is locked. Please enter your access key to unlock.

Send Email for Deformation-Guided Hybrid Approach for Interfractional Variations