Segmentation of sacral curve on dynamic MRI for diagnosis of pelvic organ prolapse

Pelvic organ prolapse (POP) is a critical health condition for women. Dynamic magnetic resonance imaging (MRI) is currently used for assessing POP and to complement clinical examination. Current studies have shown some evidence on the association between the shape of the sacral curve and the development of POP. However, the sacral curve is currently extracted manually resulting in a time-consuming and subjective process. A new method is proposed to automate the identification and segmentation of the sacral curve on MRI. The proposed method identifies the region of interest without any user input by using our previously developed pelvic floor point identification model. Edges of the sacral structure are detected to identify points along the curve, which are then connected using a proposed adaptive shortest path algorithm. These points are used to finalize the segmentation of the sacral curve using smoothing curve fitting algorithm. Results show that the proposed method can achieve good accuracy for 80% of the dataset used in this study.