Applied Machine Learning for the Prediction of Growth of Abdominal Aortic Aneurysm in Humans

Lee, Regent; Jarchi, Delaram; Perera, Rafael; Jones, Amy; Cassimjee, Ismail; Handa, Ashok; Clifton, David A.; Bellamkonda, Kirthi S.; Woodgate, Felicity; Killough, Nicholas; Maistry, Niveshni; Chandrashekar, Anirudh; Darby, Christopher R.; Halliday, Anne; Hands, L; Lintott, Patrick; Magee, T.R.; Northeast, A.; Perkins, Jeremy; Sideso, Ediri

doi:10.1016/j.ejvssr.2018.03.004

Cited by 59 publications

(41 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the same way, Mocco et al [10] used the aneurysm morphology to demonstrate and predict that these ones can be good predictors. Finally, recent studies of Liu et al [17]and Lee et al [18]predicted the rupture risk using morphological parameters and healthy behaviors with feed-forward artificial neural network and SVR, respectively. The objective of this investigation is carried out a supervised machine learning algorithm (SVM) using the least amount of statistic significant morphologic and hemodynamic variables, simplifying the model and the compute-time of each iteration.…”

Section: Introductionmentioning

confidence: 99%

Untitled

2016

MLAIJ

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

Untitled

2016

MLAIJ

View full text Add to dashboard Cite

show abstract

“…Another study used ML techniques to predict the risk of AAA growth and the algorithm correctly predicted AAA diameter within 2 mm error in 85% and 71% of patients at 12 and 24 months, respectively. 7…”

mentioning

confidence: 99%

Prediction of Abdominal Aortic Aneurysm Growth and Risk of Rupture in the Era of Machine Learning

Lareyre¹,

Adam

Carrier

et al. 2020

Angiology

View full text Add to dashboard Cite

Abdominal aortic aneurysm (AAA) is a life-threatening disease and the indication to treat mainly relies on evaluating the risk of aneurysm growth and rupture. 1,2 Although the main risk factors for AAA rupture have been identified (eg, maximum AAA diameter, expansion rate, or wall stiffness), there is a need to improve this assessment. 2,3 Machine learning (ML) is a subfield of artificial intelligence where a computer program is given the ability to learn by combining sets of algorithms. 4,5 It can discover hidden patterns in large data sets and identify which variables are most relevant. Hirata et al aimed to use ML from computed tomography (CT) angiography data to predict AAA expansion. 6 In a data set of 50 patients with small aneurysms who had had at least 2 CT scans with a minimum of 6 months apart, they selected 9 anatomic features and developed a ML algorithm to predict the risk of AAA growth. The anatomic features included the major axis of AAA, minor axis of AAA, major axis of lumen without intraluminal thrombus (ILT), minor axis of lumen without ILT, AAA area, lumen area without ILT, ILT area, maximum ILT area, and maximum ILT thickness. The ML algorithm was a strong predictor of significant AAA expansion (>4 mm/y). Another study used ML techniques to predict the risk of AAA growth and the algorithm correctly predicted AAA diameter within 2 mm error in 85% and 71% of patients at 12 and 24 months, respectively. 7 These results represent a proof of concept for such tools to improve the evaluation of the risk of AAA growth and rupture. Machine learning could help surgeons assess the indications to treat and improve the surveillance of small asymptomatic AAAs. Nevertheless, the field is still in its infancy and further studies are necessary before their routine use. The results need to be validated in large multicenter cohorts to confirm their generalizability. Efforts should be directed toward developing large multicenter registries and resolving many concerns including data privacy and protection, data sharing, standardization of data collection, and storage. Such studies require institutional and financial support as well as close collaboration between health professionals, engineers, and researchers. Although major challenges remain, we believe that ML will become a powerful tool that will help surgeons to better assess the risk of AAA rupture and to decide the most appropriate timing for surgical repair.

show abstract

“…5 Machine learning algorithms are currently being developed to better assess the prognosis of patients with AAA, including the risk of AAA growth or rupture. 6,7 Further research should be oriented toward improving imaging analysis. Combining a detailed anatomical characterization of the AAA with the clinical and biological characteristics of patients would allow development of multivariable scores to better evaluate the risk of rupture in patients with AAA.…”

mentioning

confidence: 99%