A Brave New World: Toward Precision Phenotyping and Understanding of Coronary Artery Disease Using Radiomics Plaque Analysis

HomeRadiologyVol. 299, No. 1 PreviousNext Reviews and CommentaryFree AccessEditorialA Brave New World: Toward Precision Phenotyping and Understanding of Coronary Artery Disease Using Radiomics Plaque AnalysisU. Joseph Schoepf , Tilman EmrichU. Joseph Schoepf , Tilman EmrichAuthor AffiliationsFrom the Department of Radiology and Radiological Science, Medical University of South Carolina, 25 Courtenay Dr, Charleston, SC 29425 (U.J.S., T.E.); Department of Diagnostic and Interventional Radiology, University Medical Center Mainz, Mainz, Germany (T.E.); and German Center for Cardiovascular Research (DZHK), Partner Site Rhine Main, Mainz, Germany (T.E.).Address correspondence to U.J.S. (e-mail: [email protected]).U. Joseph Schoepf Tilman EmrichPublished Online:Feb 16 2021https://doi.org/10.1148/radiol.2021204456MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In See also the article by Kolossváry et al in this issue.Dr Schoepf is a professor with appointments in radiology, medicine, and pediatrics at the Medical University of South Carolina in Charleston, SC. Dr Schoepf serves as the director of the Division of Cardiovascular Imaging, vice chair for research, and assistant dean for clinical research. His main scientific interest is the use of CT, MRI, image postprocessing, and artificial intelligence for diagnosis of thoracic disorders.Download as PowerPointOpen in Image Viewer Dr Emrich is a postdoctoral research fellow at the Division of Cardiovascular Imaging, MUSC, and a junior consultant specializing in cardiovascular imaging at the Department of Diagnostic and Interventional Radiology of the University Medical Center Mainz, Germany. His main scientific interests focus on multiparametric cardiac imaging in the general population and disease cohorts, advanced postprocessing of cardiac CT and MRI, and impact of image acceleration on cardiac imaging biomarkers.Download as PowerPointOpen in Image Viewer Atherosclerosis-related cardiovascular risk factors influence the development and progression of coronary artery plaques and subsequent coronary artery disease (CAD), leading to increased cardiovascular morbidity and mortality. Unlike conventional risk factors, such as hyperlipidemia, which causes lipid accumulation in the coronary artery wall, unconventional cardiovascular risk factors, such as HIV infection and cocaine use, affect plaque progression by means of a primarily inflammatory-driven pathophysiologic pathway (1). It is highly desirable to develop a better understanding of the different pathways in plaque development and progression, as it may improve personalized drug therapy (so-called precision medicine), resulting in enhanced individualized care of patients with CAD.Over the past few decades, coronary CT angiography has become an established noninvasive phenotyping tool, allowing for the visualization and quantitative assessment of CAD, even in a subclinical stage (2). While the effects of different cardiovascular risk factors can be identified and distinguished on a molecular level and detected as specific morphologic changes on a histologic level, current well-established coronary CT angiography biomarkers are unable to help differentiate between the various pathologic pathways of CAD. In addition to conventional imaging evaluation techniques, in the past decade, research initiatives explored the field of radiomics. In brief, radiomic analysis is a method of extracting a high number of imaging features from radiologic images. These features describe heterogeneity and spatial complexity of lesions (3) and can be applied to imaging, including coronary CT angiography.In this issue of Radiology, Kolossváry et al (4) investigated whether different cardiovascular risk factors contribute to the changes in plaque phenotype over time by using radiomic analysis of coronary plaques based on coronary CT angiography studies. The authors evaluated 300 randomly selected individuals with subclinical CAD confirmed with coronary CT angiography who also participated in a previous prospective epidemiologic study. They found that cocaine use, HIV infection, and conventional risk factors were associated with an increased risk of atherosclerotic cardiovascular disease in 23.7%, 1.3%, and 8.2% of coronary CT angiography radiomic features, respectively, over a mean follow-up time of 4.0 years ± 2.3 [standard deviation]. The authors established multivariable models adjusted for factors that are known to affect plaque morphology, such as inflammatory status, genetic predisposition, plaque-modifying medication (statins), and total plaque volume. Interestingly, the adjusted models did not reveal any overlap in radiomic features affected by conventional versus unconventional cardiovascular risk factors. However, there was overlap in radiomic features between both primarily inflammatory-driven unconventional risk factors (HIV infection and cocaine use), indicating those risk factors potentially modify similar structural components of the coronary artery plaques. Based on their analysis, the authors were able to identify 13 unique radiomic clusters that allow for the separation of the structural plaque components that are affected by different risk factors.Overall, these intriguing results suggest that radiomic analysis of coronary artery plaques allows for the identification of unique morphologic patterns specific to different risk factors. Conventional cardiovascular risk factors, typically causing lipid accumulation in plaques, resulted in unique patterns of morphologic changes that were different from changes caused by unconventional inflammation-driven risk factors, such as cocaine use or HIV infection. In addition, Kolossváry et al were able to demonstrate that different cardiovascular risk factors may act specifically. For example, negative effects of cocaine use were present only in participants with a susceptible environment of existing conventional risk factors.To our knowledge, the study by Kolossváry et al is the first to demonstrate in vivo that radiomics, an advanced image analysis technique, may overcome limitations of current established imaging biomarkers to identify certain features in the pathomechanism of plaque development that are typically only explored on the molecular level. This study expands on the potential use of radiomics proposed by prior investigations that demonstrated the superiority of radiomic analysis to histogram-based and visual assessment of advanced coronary artery plaques on a histologic level (5). This means that radiomic analysis of coronary artery plaques not only enables better description of histologic plaque features but also may enable identification of underlying molecular effects of different risk factors and their subsequent manifestations in morphologic patterns. Therefore, radiomics may be useful to monitor the development and progression of coronary artery atherosclerosis, identify unknown pathways of disease progression, and allow for precision phenotyping of CAD based on clinical imaging data at levels previously deemed impossible. These efforts have the potential to create a better understanding of CAD progression and to identify, monitor, and treat plaques based on their specific morphologic pattern. This would ultimately result in a more individualized precision medicine approach in CAD.However, these remarkable initial and hypothesis-generating results and their promising future applications must be interpreted carefully. First, the authors only investigated a relatively small cohort of Black participants. Typically, replications of investigations targeted at the exploration of disease mechanisms require validation in larger internal or external study populations. Second, even if the method of radiomic analysis is considered relatively established (6), reproducibility and standardization of radiomic features must be further explored. The effects of segmentation, especially of small structures, such as coronary artery plaques (7), acquisition techniques, and image reconstruction algorithms (8), must be investigated and addressed before radiomic analysis can be implemented into longitudinal multivendor, multi-instrument, and multiobserver studies. Third, investigation of the underlying pathophysiology in different disease models is typically performed in the basic science realm and validated in preclinical and clinical studies thereafter. The reversed engineering approach proposed here lacks the controlled experimental environment given in a basic science setting, and the observed morphologic patterns must be translated back to the molecular level to understand the underlying pathophysiologic pathways of those observations. Additionally, it is essential that radiomics be evaluated with other "-omics" such as radiogenomics, proteomics, and metabolomics. Despite nearly 10 years of effort in radiomic research, radiomics has yet to be translated into a clinically feasible and meaningful method. The new insights in plaque characteristics proposed by Kolossváry et al still must prove their clinical relevance, benefits in CAD treatment, and prognostic implications.However, the study by Kolossváry et al sheds light into a brave new world, where advanced image analysis–based phenotyping may inspire new research avenues in basic science. Furthermore, already existing longitudinal data sets can be mined for data to identify new potential pathways and correlations that literally have been dwelling under the surface of our visual recognition. In addition, this expands the one-stop shop role of coronary CT angiography from a combined anatomic and functional test to a new field of precision medicine exploring mechanistic insights in diseases using radiomic signatures. A brave new world, indeed.Disclosures of Conflicts of Interest: U.J.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: is a consultant for Bayer, Elucid BioImaging, General Electric, Guerbet, HeartFlow, and Siemens; institution received grants from Astellas, Bayer, Bracco, Elucid BioImaging, General Electric, Guerbet, HeartFlow, and Siemens Healthineers. Other relationships: disclosed no relevant relationships. T.E. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: receives a speakers fee and is reimbursed for travel by Siemens Healthineers. Other relationships: disclosed no relevant relationships.References1. Bäck M, Yurdagul A Jr, Tabas I, Öörni K, Kovanen PT. Inflammation and its resolution in atherosclerosis: mediators and therapeutic opportunities. Nat Rev Cardiol 2019;16(7):389–406. Medline, Google Scholar2. Baumann S, Kaeder F, Schoepf UJ, et al. Prognostic Value of Coronary Computed Tomography Angiography-derived Morphologic and Quantitative Plaque Markers Using Semiautomated Plaque Software. J Thorac Imaging 2020. 10.1097/RTI.0000000000000509. Published online April 3, 2020. Accessed November 19, 2020. Crossref, Medline, Google Scholar3. Kolossváry M, Kellermayer M, Merkely B, Maurovich-Horvat P. Cardiac Computed Tomography Radiomics: A Comprehensive Review on Radiomic Techniques. J Thorac Imaging 2018;33(1):26–34. Crossref, Medline, Google Scholar4. Kolossváry M, Gerstenblith G, Bluemke DA, et al. Contribution of Risk Factors to the Development of Coronary Atherosclerosis as Confirmed via Coronary CT Angiography: A Longitudinal Radiomics-based Study. Radiology 2021;299:97–106. Link, Google Scholar5. Kolossváry M, Karády J, Kikuchi Y, et al. Radiomics versus Visual and Histogram-based Assessment to Identify Atheromatous Lesions at Coronary CT Angiography: An ex Vivo Study. Radiology 2019;293(1):89–96. Link, Google Scholar6. Zwanenburg A, Vallières M, Abdalah MA, et al. The image biomarker standardization initiative: Standardized quantitative radiomics for high-throughput image-based phenotyping. Radiology 2020;295(2):328–338. Link, Google Scholar7. Kolossváry M, Jávorszky N, Karády J, et al. Effect of vessel wall segmentation on volumetric and radiomic parameters of coronary plaques with adverse characteristics. J Cardiovasc Comput Tomogr 2020. 10.1016/j.jcct.2020.08.001. Published online August 10, 2020. Accessed November 19, 2020. Crossref, Google Scholar8. Pinto Dos Santos D, Dietzel M, Baessler B. A decade of radiomics research: are images really data or just patterns in the noise? Eur Radiol 2020. 10.1007/s00330-020-07108-w. Published online August 7, 2020. Google ScholarArticle HistoryReceived: Nov 30 2020Revision requested: Dec 14 2020Revision received: Dec 16 2020Accepted: Dec 17 2020Published online: Feb 16 2021Published in print: Apr 2021 FiguresReferencesRelatedDetailsCited ByPhantom-based radiomics feature test–retest stability analysis on photon-counting detector CTAlexanderHertel, HishanTharmaseelan, Lukas T.Rotkopf, DominikNörenberg, PhilippRiffel, KonstantinNikolaou, JakobWeiss, FabianBamberg, Stefan O.Schoenberg, Matthias F.Froelich, IsabelleAyx2023 | European Radiology, Vol. 33, No. 7Introduction to radiomics for a clinical audienceC.McCague, S.Ramlee, M.Reinius, I.Selby, D.Hulse, P.Piyatissa, V.Bura, M.Crispin-Ortuzar, E.Sala, R.Woitek2023 | Clinical Radiology, Vol. 78, No. 2Medical RadiologyBaselYacoub, JosuaDecker, U. JosephSchoepf, TilmanEmrich, Jon F.Aldinger, AkosVarga-Szemes2022Comparison Study of Myocardial Radiomics Feature Properties on Energy-Integrating and Photon-Counting Detector CTIsabelleAyx, HishanTharmaseelan, AlexanderHertel, DominikNörenberg, DanielOverhoff, Lukas T.Rotkopf, PhilippRiffel, Stefan O.Schoenberg, Matthias F.Froelich2022 | Diagnostics, Vol. 12, No. 5Accompanying This ArticleContribution of Risk Factors to the Development of Coronary Atherosclerosis as Confirmed via Coronary CT Angiography: A Longitudinal Radiomics-based StudyFeb 16 2021RadiologyRecommended Articles HIV and Coronary Atherosclerosis: Research Separates Association from CausationRadiology2021Volume: 299Issue: 3pp. 581-582MRI of the Aortic Wall to Assess Cardiovascular Risk and PrognosisRadiology2022Volume: 304Issue: 3pp. 551-552Risk Scores versus Atherosclerosis Imaging: Time to Embrace What Is in Plain Sight!Radiology: Cardiothoracic Imaging2020Volume: 2Issue: 1Does the Use of Nitroglycerin at MR Angiography Help Diagnose Coronary Artery Disease?Radiology: Cardiothoracic Imaging2020Volume: 2Issue: 2CATCH the Wave of Coronary Atherosclerotic Plaque MRIRadiology2021Volume: 302Issue: 3pp. 566-567See More RSNA Education Exhibits Challenges In Diagnosis and Management of Spontaneous Coronary Artery Dissection: Pearls and Pitfalls on CT EvaluationDigital Posters2021Bringing MINOCA's Puzzle Pieces Together: The Role of Cardiac MRI (CMRI)Digital Posters2019Spontaneous Coronary Artery Dissection (SCAD): An Uncommon Diagnosis Of A Common Presentation—Primer For Cardiothoracic RadiologistsDigital Posters2021 RSNA Case Collection Acute thrombosed basilar tip aneurysmRSNA Case Collection2021Chronic myocardial infarctionRSNA Case Collection2020Anomalous origin of the right coronary artery from the left sinus of ValsalvaRSNA Case Collection2020 Vol. 299, No. 1 Metrics Altmetric Score PDF download