PhD Scientific Days 2023

Clinical Medicine III.

A new artificial intelligence-based method for the automatic adjustment of lumbar 3D spine models to standing radiographs

Előadó neve

Dr. Hajnal, Benjamin

Neptun code

mm14pu

Előadó munkahelye

National Center for Spinal Disorders

Előadó telefonszáma

+36705989752

Előadó e-mail címe

hajnal.beni@gmail.com

Az előadás címe

A new artificial intelligence-based method for the automatic adjustment of lumbar 3D spine models to standing radiographs

Szerző(k) neve és munkahelye

Benjámin Hajnal1,2, Áron Serebrenik1,3, Endre Péter Éltes1,4

1 In Silico Biomechanics Laboratory, National Center for Spinal Disorders, Buda Health Center, Budapest
2 Károly Rácz Doctoral School of Clinical Medicine, Semmelweis University, Budapest
3 Faculty of Informatics, Eötvös Lóránd University, Budapest
4 Department of Spine Surgery, Department of Orthopaedics, Semmelweis University, Budapest

Bemutatás módja

Szóbeli

Szekció

Clinical Medicine III.

Language of the presentation

Hungarian

Preferred session

Clinical Medicine

Összefoglaló szövege

Introduction: Virtual 3D lumbar spine models generated from CT scans are commonly used in computer-aided surgical planning and in silico studies. However, these models do not accurately reflect the biomechanically relevant standing spine configuration. Standing radiographs, which capture vertebral alignment in the standing position, are routinely used in preoperative spine surgical planning.
Aims: Our aim was to develop a method to automatically adapt the vertebral alignment of 3D lumbar spine models based on supine CT to standing radiographs, replacing the time-consuming manual method of rigid registration.
Methods: We used imaging data from a cohort of 110 patients with monosegmental degeneration of the spine. Manual segmentation was performed to produce 3D virtual models of the lumbar vertebrae and sacrum from CT images. Vertebral landmarks were manually marked on spine radiographs and CT slices, which served as training data for an artificial intelligence algorithm. The accuracy of the algorithm was validated on a test dataset. The transformation required for rigid registration was determined from the position of the vertebral corner points. The accuracy of the alignment was verified by comparison with X-ray images and with a model from a previous similar but manually performed and validated study.
Results: The new method significantly reduced the difference between the sagittal alignment of the X-ray images and the registered 3D spine models, and highly accelerated the registration procedure compared to the previous manual method (manual alignment: 60 minutes of work, automatic alignment: <1 minute run time).
Conclusion: We propose a fast, accurate, accessible, and reproducible method for generating patient-specific 3D geometries of the lumbar spine that accurately represents the alignment of the standing lumbar spine. This method can be applied to finite element-based in silico clinical simulation. We plan to integrate this method into a leading medical image processing software to create a fully automated, user-friendly workflow.
Funding: “Project no. KDP-12-10/PALY-2022 has been implemented with the support provided by the Ministry of Culture and Innovation of Hungary from the National Research, Development and Innovation Fund, financed under the KDP-2021 funding scheme.”

University and Doctoral School

Semmelweis University, Károly Rácz Doctoral School of Clinical Medicine

Supervisor

Dr. Peter Eltes Endre, Dr. Aron Lazary

Publication of my abstract

I do not give consent to the publication of my abstract on the website of the congress.

Kind

Szabad

Status

elfogadva

Accepted presentation method

szóbeli

Előadás fájl jóváhagyás

nem rendelkezett róla

Előadó

7389

Start

09:15

End

09:30

Authors (legacy)

Benjámin Hajnal1,2, Áron Serebrenik1,3, Endre Péter Éltes1,4

1 In Silico Biomechanics Laboratory, National Center for Spinal Disorders, Buda Health Center, Budapest
2 Károly Rácz Doctoral School of Clinical Medicine, Semmelweis University, Budapest
3 Faculty of Informatics, Eötvös Lóránd University, Budapest
4 Department of Spine Surgery, Department of Orthopaedics, Semmelweis University, Budapest