PhD Scientific Days 2017

Oral Presentations: Cardiovascular

Radiomic features of high risk coronary lesions

Előadó neve

Dr. Kolossváry, Márton

Előadó munkahelye

MTA-SE Cardiovascular Imaging Research Group, Heart and Vascular Center, Semmelweis University

Előadó telefonszáma

+36208250439

Előadó e-mail címe

marton.kolossvary@gmail.com

Az előadás címe

Radiomic features of high risk coronary lesions

Szerző(k) neve és munkahelye

Márton Kolossváry
MTA-SE Cardiovascular Imaging Research Group, Heart and Vascular Center, Semmelweis University

Szekció

Oral Presentations: Cardiovascular

Data of the presenter

Doctoral School: Basic Medicine
Program: 01. Cardiovascular Disorders: Physiology and Medicine of Ischaemic Circulatory Diseases
Supervisor: Pál Maurovich-Horvat
E-mail address: maurovich.horvat@gmail.com

Text of the abstract

INTRODUCTION
Positive remodeling (PR), low attenuation (LA), spotty calcium (SC) and the napkin-ring sign (NRS) are associated with major adverse cardiac events. PR, LA, SC are semi-quantitative plaque features, and require quantitative measurements, while the NRS is a qualitative parameter with distinct morphology. The NRS is described as a non-calcified plaque cross-section with a central area of low attenuation that is apparently in contact with the lumen, which is surrounded by a ring-like higher attenuation plaque tissue. The reproducible identification of NRS is challenging due to its qualitative nature. Radiomics is the process of extracting various different quantitative features from a given lesion to describe morphology.

AIMS
Therefore, our aim was to assess whether radiomic features are capable of identifing the NRS.

METHODS
We included 20 patients (mean age: 63.8±8.9, 20% female), 10 with NRS plaque and 10 matched patients with non-NRS lesions. We matched the patients and plaques based-on plaque localization, type and stenosis severity. Coronary lesions were segmented and voxels containing the vessel wall were exported as a DICOM dataset. Radiomic features were calculated on the exported datasets and compared between the two groups using Wilcoxon signed-rank test. Diagnostic accuracy was described using the Area Under the Curve (AUC) of Receiver operating characteristics calculations.

RESULTS
Overall, 15608 quantitative features were calculated for each lesion. 4619 (29.6%) parameters showed significant difference between the two groups at a p<0.05 level, 1270 (8.1%) at p<0.01. In total, 2834 (18.3%) of the calculated parameters had a AUC value greater than 0.8, 510 (3.3%) had a value greater than 0.9 and 83 (0.5%) had an AUC value greater than 0.95.

CONCLUSIONS
To the best of our knowledge, this is the first study to demonstrate the feasibility of calculating radiomic features on coronary lesions. Radiomics seems to be a promising new technique capable of quantifying lesion morphology and identifying high-risk coronary plaques.

Kind

Szabad

Előadás fájl jóváhagyás

nem rendelkezett róla

Előadó

1011

Authors (legacy)

Márton Kolossváry
MTA-SE Cardiovascular Imaging Research Group, Heart and Vascular Center, Semmelweis University