MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

3D, CT-Free Breast Treatment Planning Simulation for Low- and Middle-Income Countries: Proof of Concept

Y Roed1*, R Cardan2 , S Shaitelman1 , D Craft3 , R Howell1 , A Jhingran1 , R Popple2 , L Court1 , (1) UT MD Anderson Cancer Center, Houston, TX, (2) UAB University of Alabama, Birmingham, AL, (3) Mayo Clinic, Phoenix, AZ

Presentations

(Sunday, 7/14/2019) 1:00 PM - 2:00 PM

Room: 225BCD

Purpose: Many radiotherapy centers in low- and middle-income countries (LMICs) lack reliable access to CT imaging; therefore, dose is calculated to only a single point. We aimed to develop a low-cost, CT-free approach to patient-specific, 3D breast treatment planning.

Methods: We developed a process to generate simplified, synthetic CT images using three Microsoft Kinect cameras. These images are combined into a surface map from which orthogonal slices are constructed and converted into simplified, synthetic CT images. The CT images only contain information about the outer body shape without any internal anatomy. Camera positions were optimized to give the most reliable surface map. This system was tested using anthropomorphic breast (n=2) and chest wall (n=8) phantoms, including six in-house 3D-printed chest wall phantoms (based on actual patient CTs). To assess geometric accuracy, distances between landmarks (e.g., separation between midline and midaxillary line) were measured on synthetic and real CT images of these phantoms.

Results: The most reliable surface map was achieved when all cameras were positioned on the ipsilateral side of the phantom, facing the object at an angle of 30° from vertical, and stacked with the two outer cameras about 20 cm higher than the middle camera. All measured distances differed less than 5 mm between the actual and the synthetic CT (range: 0.5-4.8 mm, average: 1.8 ± 1.3 mm). Eighty percent of all measurements were within 2 mm.

Conclusion: This camera system could be an inexpensive solution for 3D, CT-free breast treatment planning. The system can reliably reproduce the phantom geometry within a few millimeters between actual and measured phantom geometry. The correct representation of the phantom geometry is an important, first step in developing a CT-free breast treatment planning system for LMICs.

Funding Support, Disclosures, and Conflict of Interest: Partially supported by Varian

Keywords

Breast, Treatment Planning, Computer Vision

Taxonomy

TH- External beam- photons: General (most aspects)

Contact Email