| Literature DB >> 1555957 |
R J Dowsett1, J M Galvin, E Cheng, R Smith, R Epperson, R Harris, G Henze, M Needham, R Payne, M A Peterson.
Abstract
Three-dimensional (3-D) treatment planning is a labor-intensive process with contouring of the target volume and critical normal tissues being a significant time-consuming component. The use of 3-D treatment planning on a routine basis may be limited by the time required to complete treatment plans. Despite the need to increase the efficiency of the process, there is little literature addressing the speed and accuracy of contouring systems. In an attempt to initiate systematic analysis of the contouring process, data sets consisting of 10 CT images each were developed on two patients with esophageal carcinoma. Nine different operators manually contoured structures (target volume, spinal canal, lungs) on the data sets using four different contouring systems present in our department. These included both commercially available systems and those developed by the authors. There was a wide variation in the hardware and software characteristics of these systems. The time required to contour the CT data sets was recorded and analyzed. The contouring accuracy was assessed by comparison with a standard template derived from the CT data set for each image. The contouring time was found to be dependent on the system design, previous contouring experience, and the type of drawing instrument (lightpen vs mouse). The mean contouring time ranged from 26 minutes per patient for the fastest system to 41 minutes for the slowest. Potential clinically significant errors in contouring were rare for the spinal canal and lungs but present at a greater rate for the target volume (30.3%). The implications of this finding are discussed.Entities:
Mesh:
Year: 1992 PMID: 1555957 DOI: 10.1016/0360-3016(92)90812-v
Source DB: PubMed Journal: Int J Radiat Oncol Biol Phys ISSN: 0360-3016 Impact factor: 7.038