A vision system for patient positioning in radiation therapy

Abstract

PurposeThe paper outlines a new approach for positioning a patient on the treatment table for radiation therapy sessions. The vision approach utilizes lasers and cameras for positioning and has several advantages over the conventional methods.Design/methodology/approachThe positioning is accomplished by comparison of a set of computed tomography (CT) contours (acquired from the patient) with a set of corresponding contours acquired by a 3D vision system from the same region of the patient's body. The overall positioning error calculated by the iterative closest point algorithm is used to reorient the treatment table. Various issues related to the acquisition and generation of the 3D spatial data are discussed.FindingsPositioning is accurate and can detect small movement in the patient's position.Research limitations/implicationsTesting was done on a cast of a human torso and additional testing is required on in a hospital environment to fully test the efficiency of the approach.Practical implicationsThe method merges data readily available from standard CT imaging systems and 3D imaging systems. Therefore, the additional hardware requirements are minimal. The system integrates well with existing hardware, software and treatment practices.Originality/valueThe method introduces a new approach to patient positioning employing a combination of sensor technologies. The approach is accurate, reliable, consumes less time and most importantly prevents the use of X‐rays for patient positioning.