Automatic 3D analysis of the ex-vivo porcine lumbar interbody fusion based on X-ray micro computed tomography data

• Published in: Computers in biology and medicine 2022-06, Vol.145, p.105438-105438, Article 105438
• Main Authors: Laznovsky, Jakub, Brinek, Adam, Zikmund, Tomas, Boudova, Marketa, Vojtova, Lucy, Dorazilova, Jana, Trunec, Martin, Stastny, Premysl, Krticka, Milan, Planka, Ladislav, Ira, Daniel, Faldyna, Martin, Kaiser, Jozef
• Format: Article
• Language: English
• Subjects: Algorithms; Animal models; Back surgery; Bone grafts; Bone implants; Clinical medicine; Computed tomography; Datasets; Fractures; Grafting; Growth factors; Iliac crest; Image processing; Lumbar spinal surgery; Neural networks; Polymers; Polyphosphates; Regenerative medicine; Skin & tissue grafts; Spinal fusion; Spine; Spine (lumbar); Substitute bone; Swine; Tomography; Vertebrae; Vertebrae fusion
• ISSN: 0010-4825; 1879-0534
• DOI: 10.1016/j.compbiomed.2022.105438

Spinal fusion is a surgical procedure used to join two or more vertebrae to prevent movement between them. This surgical procedure is considered in patients suffering from a wide range of degenerative spinal diseases or vertebral fractures. The success rate of spinal fusion is frequently evaluated subjectively using X-ray computed tomography. The pig was chosen as an animal model for spinal fusion, since its spinal structure is similar to the human spine. Our paper presents an automatic approach for pig's spinal fusion evaluation in 3D. The proposed approach is based on the determination of the vertebral fused area, which reflects the fusion quality. The approach was applied and tested on microCT (μCT) data of fused porcine vertebrae ex-vivo. In our study, three types of implants were used to perform spinal fusion: the iliac crest bone graft used as the gold standard, and two types of novel scaffold implants based on the polymer/ceramic porous foam involving either growth factors or polyphosphates. The evaluation worked automatically for all three types of used implants, and the fusion quality was determined quantitatively. The calculation is based on the detection of the fused area and area of facies intervertebralis, so the percentual representation of the vertebral joint can be determined. Since this approach is versatile and is described in detail as a guide for image processing the data of vertebrae fusion, this methodology has the potential to establish a standard approach for evaluating the fusion quality in ex-vivo samples that can be tested on clinical data. [Display omitted] •An automatic evaluation of spinal fusion process (no inaccuracies caused by the user).•A robust algorithm which can objectively evaluate the quality of vertebrae fusion, after the Lumbar Interbody Fusion surgery.•The result is presented as a single accurate number, representing how successfully have both vertebrae fused.•The presented methodology is able to quickly compare the fusion quality in samples, where different types of intervertebral implants were used.•The methodology is based on 3D microCT images with high resolution in order to analyse the fusion quality non-destructively.