Abstract:A fundamental challenge in the development of image-guided surgical systems is alignment of the preoperative model to the operative view of the patient. This is achieved by finding corresponding structures in the preoperative scans and on the live surgical scene. In robot-assisted laparoscopic prostatectomy (RALP), the most readily visible structure is the bone of the pelvic rim. Magnetic resonance imaging (MRI) is the modality of choice for prostate cancer detection and staging, but extraction of bone from MRI is difficult and very time consuming to achieve manually. We present a robust and fully automated multi-atlas pipeline for bony pelvis segmentation from MRI, using a MRI appearance embedding statistical deformation model (AE-SDM). The statistical deformation model is built using the node positions of deformations obtained from hierarchical registrations of full pelvis CT images. For datasets with corresponding CT and MRI images, we can transform the MRI into CT SDM space. MRI appearance can then be used to improve the combined MRI/CT atlas to MRI registration using SDM constraints. We can use this model to segment the bony pelvis in a new MRI image where there is no CT available. A multi-atlas segmentation algorithm is introduced which incorporates MRI AE-SDMs guidance. We evaluated the method on 19 subjects with corresponding MRI and manually segmented CT datasets by performing a leave-one-out study. Several metrics are used to quantify the overlap between the automatic and manual segmentations. Compared to the manual gold standard segmentations, our robust segmentation method produced an average surface distance 1.24±0.27mm, which outperforms state-of-the-art algorithms for MRI bony pelvis segmentation. We also show that the resulting surface can be tracked in the endoscopic view in near real time using dense visual tracking methods. Results are presented on a simulation and a real clinical RALP case. Tracking is accurate to 0.13mm over 700 frames compared to a manually segmented surface. Our method provides a realistic and robust framework for intraoperative alignment of a bony pelvis model from diagnostic quality MRI images to the endoscopic view.

A Semiautomatic Segmentation Method Framework for Pelvic Bone Tumors Based on CT‐MR Multimodal Images

Glioma Segmentation Using a Novel Unified Algorithm in Multimodal MRI Images

Male Pelvic Multi-Organ Segmentation Aided by CBCT-based Synthetic MRI

Automatic multiorgan segmentation in CT images of the male pelvis using region-specific hierarchical appearance cluster models

Automatic Segmentation of Juxta-Pleural Tumors from Ct Images Based on Morphological Feature Analysis

Synthetic MRI-aided Pelvic Multi-Organ Segmentation in Cone-Beam Computed Tomography.

A medical image segmentation method for rectal tumors based on multi‐scale feature retention and multiple attention mechanisms

A Semiautomatic Algorithm for Three-Dimensional Segmentation of the Prostate on CT Images Using Shape and Local Texture Characteristics.

A method framework of semi-automatic knee bone segmentation and reconstruction from computed tomography (CT) images

Fully Automated Pelvic Bone Segmentation in Multiparameteric MRI Using a 3D Convolutional Neural Network

A Semi-Supervised Multi-Region Segmentation Framework of Bladder Wall and Tumor with Wall-Enhanced Self-Supervised Pre-Training

Precise segmentation of multiple organs in CT volumes using learning-based approach and information theory.

TH-CD-206-02: BEST IN PHYSICS (IMAGING): 3D Prostate Segmentation in MR Images Using Patch-Based Anatomical Signature.

A Semi-Automatic Brain Tumor Segmentation Algorithm

Brain Tumor Segmentation in MR Slices Using Improved GrowCut Algorithm

A Framework of Fuzzy Information Fusion for the Segmentation of Brain Tumor Tissues on MR Images

Multi-organ Segmentation in Pelvic CT Images with CT-based Synthetic MRI

An Automatic Tumor Segmentation Framework of Cervical Cancer in T2-weighted and Diffusion Weighted Magnetic Resonance Images.

Modeling of the Bony Pelvis from MRI Using a Multi-Atlas AE-SDM for Registration and Tracking in Image-Guided Robotic Prostatectomy

An automatic segmentation algorithm for conglutinated bone fragments in 3D CT images of lower limb fractures based on morphology

Daily Cone-Beam CT Multi-Organ Segmentation for Prostate Adaptive Radiotherapy.