Automated detection of vertebral fractures in CT using 3D convolutional neural networks

Joeri Nicolaes, Steven Raeymaeckers, David Robben, Guido Wilms, Dirk Vandermeulen, Cesar Libanati, Marc DeBois

Onderzoeksoutput: Poster

Samenvatting

Objective: Develop a fully automated method to identify individual fractured vertebrae in computed tomography (CT) scans. Background: Despite their frequent occurrence and major associated burden, vertebral fractures remain underdiagnosed and patients undertreated.1 Spine-containing CT scans provide an opportunity to identify vertebral fractures, yet they commonly go unreported by radiologists.2 Automated detection of vertebral fractures would enhance medical care of patients with osteoporosis. Methods: We built a training database of 90 de-identified CT cases, acquired on three different scanners, containing 969 vertebrae scanned for various indications (average [range] age: 81 [70-101] years; 64% female).3 We developed a data-driven, automated vertebral fracture detection method that binarily classifies fractured or normal anatomy for each vertebra present in spine-containing CT images. Results: We performed a stratified 5-fold cross-validation experiment comparing automated predictions with ground truth read-outs from one radiologist resulting in an area under the Receiver Operating Characteristic (ROC) curve of 0.93±0.01. Conclusions: Our automated vertebral fracture demonstrated the potential for automated early identification of vertebral fractures in patients aged >50 years by opportunistically screening spine-containing CT images. Confirmatory analyses and additional methodological improvements (e.g. automatic Genant grading, fracture location) using more extensive datasets and method validation are ongoing. References: 1. Johnell O. Osteoporos Int 2006;17:1726-33; 2. Mitchell R. Arch Osteoporos 2017;12:71; 3. University Hospital Brussels Ethical Committee approval: “Designing deep learning algorithms for the automated detection of vertebral fractures”, B.U.N. 143201732477.
Originele taal-2English
StatusPublished - 20 okt. 2020
EvenementEuropean Calcified Tissue Society 2020 - Marseille, France
Duur: 21 okt. 202024 okt. 2020

Conference

ConferenceEuropean Calcified Tissue Society 2020
Land/RegioFrance
Periode21/10/2024/10/20

Vingerafdruk

Duik in de onderzoeksthema's van 'Automated detection of vertebral fractures in CT using 3D convolutional neural networks'. Samen vormen ze een unieke vingerafdruk.

Citeer dit