3D Surface Imaging as a Viable Replacement for Permanent Tattoos for Patients Undergoing Breast EBRT - A Feasibility Study on a Phantom

Abstract

Cutaneous tattoos have historically been required for the setup and alignment of patients undergoing breast RT. This point-based approach yields low resolution of the underlying surface anatomy and leaves patients with permanent markings that can be psychologically distressing. Moreover, the validity of tattoo-based setup with DRR matching to internal bony landmarks can be challenging when used to align an external mobile target. The issue is particularly pronounced for patients with larger body habitus or pendulous breasts. The emerging 3D surface imaging can theoretically address the technical shortcomings of tattoos and simultaneously provide for an improved patient experience since no permanent markings are left behind. A feasibility study was therefore designed to investigate whether 3D surface imaging can replace permanent tattoos for breast RT. A phantom was immobilized on a Linac couch and five tattoos were made on the right breast per our institutional simulation protocol. These were aligned with in-room lasers. Couch coordinates at this position were recorded as the ground truth for the purpose of comparison. A reference surface image was then captured by a 3D surface imaging system. Five ROIs were designed for image registration and phantom shift calculation. ROI1 = RT breast; ROI2 = RT breast + 1/3 LT breast; ROI3 = RT breast + 1/3 LT breast + RT lateral side; ROI4 = RT breast + 1/3 LT breast + RT lateral side + RT superior side (up to inferior border of supraclavicular fossa) and ROI5 = RT breast + 1/3 LT breast + RT lateral side + RT superior side + RT inferior side. Except for ROI1, the others were asymmetric about the beam central axis. For each ROI, the phantom was repeatedly aligned to its pre-defined treatment position from a lowest random couch position with lasers off. A good alignment was defined as translations≤0.3 mm and rotations≤0.3°, as measured by the 3D surface imaging system. The elapsed time from start to an acceptable alignment and the final couch coordinates were recorded. Three measurements were made for each ROI. The same measurements were also performed for the tattoo cohort with lasers only. The mean setup times for each approach were: 86.3s (tattoos), 102.7s (ROI1), 86.7s (ROI2), 57.3s (ROI3), 55.7s (ROI4), and 48.7s (ROI5). For the 3D surface imaging cohort, the setup time was inversely proportional to the ROI size. Fitting a linear regression model yielded a linear coefficient of R2 = 0.892. The larger the ROI, the more robust and faster the setup became. The more asymmetric the ROI, the more effective and sensitive the ROI was in detecting rotational shifts. Post-setup residual errors as measured by couch coordinates were comparable in magnitude for each. With a properly-designed asymmetric ROI, 3D surface imaging may be able to replace tattoos as a viable and effective alternative for breast RT patient setup. The accuracy of this novel approach can be further enhanced by optimizing the topographical height of the skin surface.