A Deep Learning-Based and Fully Automated Pipeline for Thoracic Aorta Geometric Analysis and Planning for Endovascular Repair from Computed Tomography

Saitta, Simone; Sturla, Francesco; Caimi, Alessandro Ciatti; Riva, A; Palumbo, Maria Chiara; Nano, Giovanni; Votta, Emiliano; Corte, Alessandro Della; Glauber, Mattia; Chiappino, Dante; Marrocco-Trischitta, Massimiliano M.; Redaelli, Alberto

doi:10.1007/s10278-021-00535-1

Cited by 17 publications

(14 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To the best of our knowledge, no AI application has yet been implemented for the automatic CILCA detection and classification of CT scans acquired for different clinical indications. A recent study [47] proposed a deep learning architecture feeding on high-resolution contrastinjected CT thoracic images and providing a comprehensive geometric analysis of the aorta for both standard and CILCA arch configurations, suitable to be used for unbiased quantitative analyses of geometric parameters in population studies for planning thoracic endovascular aortic repair. However, a single subject classifier based on the automatic classification of CILCA variants was not implemented, the above-mentioned method requiring instead the assessment of cut-off values for aortic metrics to define pathological versus normal patients.…”

Section: Discussionmentioning

confidence: 99%

A Combined Deep Learning System for Automatic Detection of “Bovine” Aortic Arch on Computed Tomography Scans

Secchi

Interlenghi²,

Alì

et al. 2022

Applied Sciences

Self Cite

View full text Add to dashboard Cite

The “bovine” aortic arch is an anatomic variant consisting in a common origin of the innominate and left carotid artery (CILCA), associated with a greater risk of thoracic aortic diseases (aneurysms and dissections), stroke, and complications after endovascular procedures. CILCA can be detected by visual assessment of computed tomography (CT) chest scans, but it is rarely reported. We developed a deep learning (DL) segmentation-plus-classification system to automatically detect CILCA based on 302 CT studies acquired at 2 centers. One model (3D U-Net) was trained from scratch (supervised by manual segmentation), validated, and tested for the automatic segmentation of the aortic arch and supra-aortic vessels. Three DL architectures (ResNet50, DenseNet-201, and SqueezeNet), pre-trained over millions of common images, were trained, validated, and tested for the automatic classification of CILCA versus non-CILCA, supervised by radiologist’s classification. The 3D U-Net-plus-DenseNet-201 was found to be the best system (Dice index 0.912); its classification performance obtained from internal, independent testing on 126 patients gave a receiver operating characteristic area under the curve of 87.0%, sensitivity 66.7%, specificity 90.5%, positive predictive value 87.5%, negative predictive value 73.1%, positive likelihood ratio 7.0, and negative likelihood ratio 0.4. In conclusion, a combined DL system applied to chest CT scans was developed and proven to be an effective tool to detect individuals with “bovine” aortic arch with a low rate of false-positive findings.

show abstract

Section: Discussionmentioning

confidence: 99%

A Combined Deep Learning System for Automatic Detection of “Bovine” Aortic Arch on Computed Tomography Scans

Secchi

Interlenghi²,

Alì

et al. 2022

Applied Sciences

Self Cite

View full text Add to dashboard Cite

show abstract

“…Random rotations, mirroring and affine transformations were applied to the processed patches. A combination of Lovasz-Softmax loss and focal loss [12] was used to trained the NN to perform simultaneous segmentation of the lateral and third ventricles. The remaining 20 scans (10%), including 5 cases of hydrocephalus, were used for testing.…”

Section: Automatic Segmentationmentioning

confidence: 99%

“…Detection of Foramen of Monro, which is the target for the tip of the catheter during EVD, was performed by a recursive thresholding approach [12].…”

Section: Automatic Segmentationmentioning

confidence: 99%

Mixed Reality and Deep Learning for External Ventricular Drainage Placement: A Fast and Automatic Workflow for Emergency Treatments

Palumbo

Saitta

Schiariti

et al. 2022

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

The treatment of hydrocephalus is based on anatomical landmarks to guide the insertion of an External Ventricular Drain (EVD). This procedure can benefit from the adoption of Mixed Reality (MR) technology. In this study, we assess the feasibility of a fully automatic MR and deep learning-based workflow to support emergency EVD placement, for which CT images are available and a fast and automatic workflow is needed. The proposed study provides a tool to automatically i) segment the skull, face skin, ventricles and Foramen of Monro from CT scans; ii) import the segmented model in the MR application; iii) register holograms on the patient's head via a marker-less approach. An ad-hoc evaluation approach including 3D-printed anatomical structures was developed to quantitatively assess the accuracy and usability of the registration workflow.

show abstract

“…4D flow MRI data were preprocessed using in-house Python code and following the workflow presented in figure 1: for each patient, a 3D binary mask of the aorta was extracted from PC-MR angiography (PC-MRA) images using semi-automatic tools available in the open source software ITK-SNAP [23]. To guarantee consistency of inlet plane location among all ATAA subjects, inlet planes were defined with respect to a commonly used anatomical landmark represented by the bifurcation of the pulmonary artery (PA) [24]. A triangulated mesh of the selected plane within the aortic lumen was generated; 4D flow velocity data were then probed at inlet plane nodal locations.…”

Section: Data Preprocessingmentioning

confidence: 99%

Data-driven generation of 4D velocity profiles in the aneurysmal ascending aorta

Saitta¹,

Maga²,

Armour³

et al. 2022

Preprint

View full text Add to dashboard Cite

Background and Objective: Numerical simulations of blood flow are a valuable tool to investigate the pathophysiology of ascending thoratic aortic aneurysms (ATAA). To accurately reproduce in vivo hemodynamics, computational fluid dynamics (CFD) models must employ realistic inflow boundary conditions (BCs). However, the limited availability of in vivo velocity measurements, still makes researchers resort to idealized BCs. The aim of this study was to generate and thoroughly characterize a large dataset of synthetic 4D aortic velocity profiles with features similar to clinical cohorts of patients with ATAA. Methods: Time-resolved 3D phase contrast magnetic resonance (4D flow MRI) scans of 30 subjects with ATAA were processed through in-house code to extract anatomically consistent cross-sectional planes along the ascending aorta, ensuring spatial alignment among all planes and interpolating all velocity fields to a reference configuration. Velocity profiles of the clinical cohort were extensively characterized by computing flow morphology descriptors of both spatial and temporal features.

show abstract

A Deep Learning-Based and Fully Automated Pipeline for Thoracic Aorta Geometric Analysis and Planning for Endovascular Repair from Computed Tomography

Cited by 17 publications

References 39 publications

A Combined Deep Learning System for Automatic Detection of “Bovine” Aortic Arch on Computed Tomography Scans

A Combined Deep Learning System for Automatic Detection of “Bovine” Aortic Arch on Computed Tomography Scans

Mixed Reality and Deep Learning for External Ventricular Drainage Placement: A Fast and Automatic Workflow for Emergency Treatments

Data-driven generation of 4D velocity profiles in the aneurysmal ascending aorta

Contact Info

Product

Resources

About