EVALUATION OF A REGION GROWING ALGORITHM FOR SEGMENTING PELVIC COMPUTED-TOMOGRAPHY IMAGES DURING RADIOTHERAPY PLANNING

A clinical evaluation of a computer segmentation algorithm was perform ed to determine whether the incorporation of such an algorithm into a radiotherapy treatment planning computer would increase the speed of s egmentation and therefore increase user productivity. Six pelvic compu ted tomography (CT) data cubes were manually segmented akin to current radiotherapy practice on three occasions and the mean times for each data cube recorded to provide baseline measurements. The same images w ere then segmented using the new region growing algorithm, backed up b y manual segmentation and contour editing tools where necessary, and t he times compared with the baseline for each data cube. The results co nfirm that this algorithm can decrease segmentation times by a factor of 2.4 without compromising the quality of the final plan (p < 0.0001) . Most of this time gain is from a rapid segmentation of the pelvic bo nes and external contour rather than the soft tissues of the pelvis. W e conclude that such algorithms will be of value for segmenting large pelvic CT data cubes for radiotherapy planning by making the procedure less labour intensive.