© 2008 The British Institute of Radiology
Early clinical evaluation of a novel three-dimensional structure delineation software tool (SCULPTER) for radiotherapy treatment planning
C A MCBAIN 1,
C J MOORE 2,
M GREEN 1,
G PRICE 2,
J S SYKES 2,
A AMER 2,
V S KHOO 1,
P PRICE 3
1 Academic Department of Radiation Oncology, Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, UK
2 North Western Medical Physics, Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, UK
3 Academic Department of Radiation Oncology. Christie Hospital NHS Trust, Wilmslow Road, Manchester M20 4BX, UK
 |
Abstract |
|---|
Modern radiotherapy treatment planning (RTP) necessitates increased delineation of target volumes and organs at risk. Conventional manual delineation is a laborious, time-consuming and subjective process. It is prone to inconsistency and variability, but has the potential to be improved using automated segmentation algorithms. We carried out a pilot clinical evaluation of SCULPTER (Structure Creation Using Limited Point Topology Evidence in Radiotherapy) — a novel prototype software tool designed to improve structure delineation for RTP. Anonymized MR and CT image datasets from patients who underwent radiotherapy for bladder or prostate cancer were studied. An experienced radiation oncologist used manual and SCULPTER-assisted methods to create clinically acceptable organ delineations. SCULPTER was also tested by four other RTP professionals. Resulting contours were compared by qualitative inspection and quantitatively by using the volumes of the structures delineated and the time taken for completion. The SCULPTER tool was easy to apply to both MR and CT images and diverse anatomical sites. SCULPTER delineations closely reproduced manual contours with no significant volume differences detected, but SCULPTER delineations were significantly quicker (p<0.05) in most cases. In conclusion, clinical application of SCULPTER resulted in rapid and simple organ delineations with equivalent accuracy to manual methods, demonstrating proof-of-principle of the SCULPTER system and supporting its potential utility in RTP.
