Radiation Physics

PV QA 3 - Poster Viewing Q&A 3

TU_15_3260 - Biomechanical Model-Based Deformable Image Registration for Modeling Neck Flexion in Head and Neck Cancer Patients

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

Biomechanical Model-Based Deformable Image Registration for Modeling Neck Flexion in Head and Neck Cancer Patients
M. M. McCulloch1, B. M. Anderson2, A. S. Mohamed2, S. Volpe2, H. Elhalawani3, G. Cazoulat1, H. Bahig2, C. D. Fuller4, and K. K. Brock1; 1University of Texas MD Anderson Cancer Center, Houston, TX, 2The University of Texas MD Anderson Cancer Center, Houston, TX, 3Department of Radiation Oncology, University of Texas MD Anderson Cancer Center, Houston, TX, 4University of Texas Graduate School of Biomedical Sciences, Houston, TX

Purpose/Objective(s): A major challenge in modeling head and neck (HN) cancer patients undergoing radiation therapy is caused by the differences in patient positioning due to neck flexion. Accounting for the different HN flexion angles is a prerequisite for modeling the dose response of the salivary glands, and is possible using biomechanical model (BMM)-based deformable image registration (DIR). This dose response is crucial for predicting the normal tissue toxicity of patients. The hypothesis is that a BMM-based DIR algorithm with bony anatomy boundary conditions (BC) can be used to model the different angles of neck flexion observed in HN cancer patients.

Materials/Methods: 96 HN cancer patients with varying angles of neck flexion were retrospectively obtained. Using a script developed for the commercial treatment planning system, two CTs from each patient were rigidly registered with focus on skeletal anatomy, then deformably registered using both intensity-based and BMM-based DIR algorithms. For the BMM-based DIR, two sets of BC were applied. One set of BC (with only bony anatomy) included the mandible, C1 vertebrae, and C3 vertebrae. These were selected in particular since they are in close proximity to the salivary glands. The second set of BC included the mandible, C1 vertebrae, C3 vertebrae, and external bodies.

Results: The results of each method were compared using Dice similarity coefficient (DSC). Rigid registration alone yielded an average DSC of 0.48, while the DIR algorithms yielded average DSC of 0.72, 0.65, and 0.69 for the intensity-based, BMM-based with bony anatomy, and BMM-based with bony anatomy and external BC, respectively. For minimum DSC of all organs across all patients, the BMM-based DIR with bony anatomy and external BC outperformed the intensity-based algorithm at least 11% and up to 25%. The results of the BMM-based DIR algorithm had an average DSC of 0.08 less than intensity-based when the bony anatomy BC were applied, and 0.03 less when the bony anatomy and external BC were applied. When bony anatomy and external BC were applied, the average DSC was statistically insignificant (p=0.33) from the intensity-based method, based on a 2-tailed Student’s T-Test. The similarity of the results shows that the BMM-based algorithm can serve to model neck flexion in order to initialize a subsequent model of the dose response of the salivary glands.

Conclusion: This study shows that BMM-based DIR with bony anatomy and external BC can be used to describe the varying angles of neck flexion shown in images during radiation treatment, with indistinguishable results from the intensity-based algorithm. Modeling neck flexion can enable robust treatment planning by incorporating possible positions of the patient at the time of treatment planning. Additionally, BMM-based DIR can incorporate dose-volume response of normal tissue and have potential for the development of comprehensive HN models.

Author Disclosure: M.M. McCulloch: None. B.M. Anderson: None. A.S. Mohamed: None. S. Volpe: None. G. Cazoulat: None. C.D. Fuller: Research Grant; National Institutes of Health, National Science Foundation, Elekta AB. Grant funding; Elekta AB. Honoraria; Nederlandse Organisatie voor Wetenschappelijk Onde. Consultant; Elekta AB, Nederlandse Organisatie voor Wetenschappelijk Onde. Travel Expenses; Elekta AB, Nederlandse Organisatie voor Wetenschappelijk Onde. Reviewer; Radiological Society of North America. Associate Editor; Radiographics. Data Management Task Force Committee Member; MR-LinAc Consortium. Member; National Cancer Institute. Task Group Member; American Association of Physicists in Medicine. K.K. Brock: Research Grant; RaySearch Laboratories. Honoraria; RaySearch Laboratories. Royalty; RaySearch Laboratories.

Send Email for Molly McCulloch


Assets

TU_15_3260 - Biomechanical Model-Based Deformable Image Registration for Modeling Neck Flexion in Head and Neck Cancer Patients



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 Biomechanical Model-Based Deformable Image Registration for Modeling Neck Flexion in Head and Neck Cancer Patients