Introduction Cardiac computed tomography (CCT) emerged as a valuable non-invasive tool for early identification of high-risk coronary atherosclerosis. However, no validated biomarkers or scores are available to identify patients who may benefit from CCT in primary prevention. Aim of the present study was to explore for the first time the potential role of facial features analysis using artificial intelligence (AI) for the prediction of high-risk coronary atherosclerosis at CCT. Material and Methods A consecutive cohort of patients with a clinical indication to CCT and without known cardiovascular disease were prospectively enrolled between January 2022 and August 2022. Before CCT, 10 facial photos were collected in each patient from different random angles. An AI model was then built to predict the presence of high-risk coronary atherosclerosis and was compared with traditional risk factors and scores. Results The study population included 100 consecutive patients with a mean age of 62±10.7 years (69% male) from whom 1,000 different facial images were recorded. All the models built considering the somatic features demonstrated higher predictive capability when compared to SCORE 2 among asymptomatic subjects regarding high non-calcified plaque volume (AUC=0.89 vs. AUC_SCORE2=0.66, p-value <0.001), high low-density plaque volume (AUC=0.94 vs. AUC_SCORE2 =0.64, p-value <0.001) and the presence of at least one high-risk plaque feature (AUC=0.79 vs. AUC_SCORE2 =0.65, p-value =0.002). Conclusions This prospective observational study demonstrates that an AI algorithm applied to facial features may serve as a new “biomarker” for early identification of patients with high-risk coronary atherosclerosis.
TO FACIAL FEATURES AS A NEW BIOMARKER FOR EARLY IDENTIFICATION OF HIGH-RISK ATHEROSCLEROSIS: A TRANSLATIONAL APPROACH
CONTE, EDOARDO
2024
Abstract
Introduction Cardiac computed tomography (CCT) emerged as a valuable non-invasive tool for early identification of high-risk coronary atherosclerosis. However, no validated biomarkers or scores are available to identify patients who may benefit from CCT in primary prevention. Aim of the present study was to explore for the first time the potential role of facial features analysis using artificial intelligence (AI) for the prediction of high-risk coronary atherosclerosis at CCT. Material and Methods A consecutive cohort of patients with a clinical indication to CCT and without known cardiovascular disease were prospectively enrolled between January 2022 and August 2022. Before CCT, 10 facial photos were collected in each patient from different random angles. An AI model was then built to predict the presence of high-risk coronary atherosclerosis and was compared with traditional risk factors and scores. Results The study population included 100 consecutive patients with a mean age of 62±10.7 years (69% male) from whom 1,000 different facial images were recorded. All the models built considering the somatic features demonstrated higher predictive capability when compared to SCORE 2 among asymptomatic subjects regarding high non-calcified plaque volume (AUC=0.89 vs. AUC_SCORE2=0.66, p-value <0.001), high low-density plaque volume (AUC=0.94 vs. AUC_SCORE2 =0.64, p-value <0.001) and the presence of at least one high-risk plaque feature (AUC=0.79 vs. AUC_SCORE2 =0.65, p-value =0.002). Conclusions This prospective observational study demonstrates that an AI algorithm applied to facial features may serve as a new “biomarker” for early identification of patients with high-risk coronary atherosclerosis.File | Dimensione | Formato | |
---|---|---|---|
phd_unimi_R13176.pdf
accesso aperto
Dimensione
1.53 MB
Formato
Adobe PDF
|
1.53 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/183409
URN:NBN:IT:UNIMI-183409