Click here to


Are you sure ?

Yes, do it No, cancel

Improving Patient Positioning Reproducibility During Head and Neck Re-Simulation for Re-Treatment Using a Surface Mapping System

S Liu1*, P Dubrowski2 , Y Fu3 , D Yang4 , Y Yang5 , A Yu6 , (1) Stanford university Scool of Medicine, Stanford, CA, (2) Stanford University, Palo Alto, CA, (3) Washington University in Saint Louis, St. Louis, MO, (4) Washington University in St Louis, St Louis, MO, (5) Stanford University Cancer Center, Stanford, CA, (6) Stanford University Cancer Center, Stanford, CA


(Sunday, 7/29/2018) 3:00 PM - 3:30 PM

Room: Exhibit Hall | Forum 4

Purpose: The purpose of this work is to study the potential improvement in patient positioning reproducibility during re-treatment simulation for H&N patients with the guidance of a commercial surface imaging system, Optical Surface Monitoring System (OSMS).

Methods: Two methods were performed to make the masks either with or without OSMS assistance. Method-1: Three masks were independently made by three therapists and computed tomography (CT) scans were performed. The first phantom scan with the first mask was served as the ground truth for the comparisons. The second and the third scans were to mimic the head and neck re-treatment in which therapists don’t have any additional guidance for making the mask. Therefore, the procedures for making the masks for head and neck re-treatment is no difference from the standard procedure. Method-2: The body contour of the first phantom scan from method-1 was exported to OSMS to guide the therapists. Two additional masks were independently made by two therapists with the assistance of OSMS and followed by CT scans. Both rigid and deformable registrations were implemented to fuse CT images to the first phantom scan of method-1and the average target registration errors (TRE) were calculated, which is the difference after registration between corresponding points not used in calculating the registration transform.

Results: The average translational TREs of rigid registrations are 1.6±0.5 and 0.9±0.4 mm for Method-1 and Method-2, respectively (n=3). For deformable image registrations, the average TREs are 1.3±0.6 and 0.82±0.2 mm for Method-1 and Medoth-2, respectively (n=3).

Conclusion: This work demonstrates the feasibility to accurately re-position H&N patients during a CD/re-treatment simulation with guidance of a surface imaging system. The overall improvement in positioning reproducibility would allow better dose mapping and treatment decisions for head and neck re-treatments, particularly for severe deformations of the head tilt and jaw positions.


Immobilization, Radiation Therapy


IM/TH- Image registration : CT

Contact Email