Performance of Prediction Models for Diagnosing Severe Aortic Stenosis Based on Aortic Valve Calcium on Cardiac Computed Tomography: Incorporation of Radiomics and Machine Learning

Abstract

Objective We aimed to develop a prediction model for diagnosing severe aortic stenosis (AS) using computed tomography (CT) radiomics features of aortic valve calcium (AVC) and machine learning (ML) algorithms. Materials and Methods We retrospectively enrolled 408 patients who underwent cardiac CT between March 2010 and August 2017 and had echocardiographic examinations (240 patients with severe AS on echocardiography [the severe AS group] and 168 patients without severe AS [the non-severe AS group]). Data were divided into a training set (312 patients) and a validation set (96 patients). Using non-contrast-enhanced cardiac CT scans, AVC was segmented, and 128 radiomics features for AVC were extracted. After feature selection was performed with three ML algorithms (least absolute shrinkage and selection operator [LASSO], random forests [RFs], and eXtreme Gradient Boosting [XGBoost]), model classifiers for diagnosing severe AS on echocardiography were developed in combination with three different model classifier methods (logistic regression, RF, and XGBoost). The performance (c-index) of each radiomics prediction model was compared with predictions based on AVC volume and score. Results The radiomics scores derived from LASSO were significantly different between the severe AS and non-severe AS groups in the validation set (median, 1.563 vs. 0.197, respectively, p < 0.001). A radiomics prediction model based on feature selection by LASSO + model classifier by XGBoost showed the highest c-index of 0.921 (95% confidence interval [CI], 0.869–0.973) in the validation set. Compared to prediction models based on AVC volume and score (c-indexes of 0.894 [95% CI, 0.815–0.948] and 0.899 [95% CI, 0.820–0.951], respectively), eight and three of the nine radiomics prediction models showed higher discrimination abilities for severe AS. However, the differences were not statistically significant ( p > 0.05 for all). Conclusion Models based on the radiomics features of AVC and ML algorithms may perform well for diagnosing severe AS, but the added value compared to AVC volume and score should be investigated further.

Assessment of aortic valve area on cardiac computed tomography in symptomatic bicuspid aortic stenosis: Utility and differences from Doppler echocardiography

“…Observers segmented the AV calcium score by carefully including the region of interest in the AV leaflet and annulus and excluding calcium in the adjacent sinus of Valsalva, left ventricular outflow tract, or mitral annulus, and image noise or beam hardening artifact was excluded. From the segmented ROI, the AVC volume was measured, and an AVC score was calculated using the Agatston method ( 21 , 22 ). All CT analyses were independently performed by two radiologists blinded to clinical information, echocardiographic results, and CT analysis results of the other reader.…”

Stability selection for LASSO with weights based on AUC

“…We used radiomics data on aortic valve calcium 10 and speech signal data on Parkinson's disease 11 as examples to demonstrate the application of the proposed method in this study. Due to the large dimensionality of both data sets, appropriate methods for variable selection and model classifiers are necessary.…”

“…Other parameters that determine the severity of AS include AVC attenuation, shape, symmetry, or distribution. According to a previous study, the severity of AS is linked to the degree of AVC and the location of the valve 10 . www.nature.com/scientificreports/ Radiomics refers to the extraction of high-dimensional, quantitative information from medical images in a high-throughput manner 17 .…”

Radiomic analysis of enhanced CMR cine images predicts left ventricular remodeling after TAVR in patients with symptomatic severe aortic stenosis

“…Recently, R. Schofield et al found that texture analysis based on cine images is of significance for the identification of the etiology of left ventricular hypertrophy (15), and Elham et al made it clear that the radiomic signatures from cine images have the potential to detect myocardial ischemia, with the best area under the curve (AUC) of 0.93 (16). On account of computing power development and process standardization, the application potential of radiomics is being constantly mined in the cardiac field, including etiology determination, diagnosis confirmation, and prognosis prediction (17)(18)(19)(20)(21). For mechanical learning, a large amount of invisible biological information included in medical images was transformed into objective and quantitative digital information (22,23).…”

“…For mechanical learning, a large amount of invisible biological information included in medical images was transformed into objective and quantitative digital information ( 22 , 23 ). The research of Nam et al has explored the excellent diagnostic efficacy of radiomics based on calcified plaques of the aortic valve for severe aortic stenosis with the highest AUC of 0.921 ( 20 ). Previous studies had also investigated the value of texture analysis in detecting left ventricular remodeling in cardiac computed tomography and CMR T1 mapping images ( 24 , 25 ).…”