A Semi-automatic Multi-seed Region-Growing Approach for Uterine Fibroids Segmentation in MRgFUS Treatment

摘要

Fibroids are benign tumors growing in the uterus. Most of fibroids do not require treatment unless they are causing symptoms. Traditional surgery treatments, like myomectomy and hysterectomy, are very invasive therapeutic approaches which not always preserves reproductive potential of the woman. MRgFUS, performed with Insightec ExAblate 2100 equipment, is a new and noninvasive technique for uterine fibroids treatment, not requiring hospitalization and recovery time for patients. An initial assessment of MRgFUS treatment is made by computing the ablated volume of uterine fibroid. In this paper a semi-automatic approach, based on region-growing segmentation technique, is proposed. The implemented approach gives a quantitative and qualitative evaluation of the treatment providing the volume and the three-dimensional (3D) model of the treated fibroid area. Considering these characteristics, the proposed approach can be used as a tool to integrate the information used by a Medical Decision Support System (MDSS). As step in the MRgFUS treatment evaluation, the achieved results improve the current methodology based on the manual uterine fibroid ROT segmentation.