Outcomes of Patients Undergoing Transcatheter Aortic Valve Implantation With Incidentally Discovered Masses on Computed Tomography

• Published in: The American journal of cardiology 2020-10, Vol.132, p.114-118
• Main Authors: Ghotra, Amaninderapal S., Monlezun, Dominique J., Boone, David, Jacob, Robin, Poosti, Kassra, Loghin, Catalin, Garcia-Sayan, Enrique, Johnson, Sarah, Zhao, Yelin, Balan, Prakash, Nguyen, Tom C., Estrera, Anthony, Gregoric, Igor D., Loyalka, Pranav, Kar, Biswajit, Smalling, Richard W., Dhoble, Abhijeet
• Format: Article
• Language: English
• Subjects: Abdomen; Aorta; Aortic valve; Breast; Cancer; Computed tomography; Confidence intervals; Heart valves; Implantation; Learning algorithms; Lymphoma; Machine learning; Malignancy; Medical prognosis; Mortality; Neural networks; Prostate; Regression analysis; Statistical analysis; Thoracic surgery; Thyroid; Transplants & implants; Variables
• ISSN: 0002-9149; 1879-1913
• DOI: 10.1016/j.amjcard.2020.07.003

Routine preprocedural chest and abdomen computed tomography is done prior to transcatheter aortic valve implantation (TAVI), which, in turn, have led to the discovery of radiographic potentially malignant incidental masses (pMIM). It is largely unknown whether pMIM impact the outcomes of patients undergoing TAVI. In this retrospective cohort study from a single center, 1,081 patients underwent TAVI from 2012 to 2016, who had available computed tomographies, survived the index hospitalization, and also had 1 year follow-up data for review. Machine learning (backward propagation neural network)—augmented multivariable regression for mortality by pMIM was conducted. In this cohort of 1,081 patients, the mean age was 79.1 (± 9.0), 48.8% were females, 16.8% had a history of prior malignancy, and 21.1% had pMIM. One-year mortality for the entire cohort was 12.6%. The most common prior malignancies were prostate, breast, and lymphoma and the most common pMIM were present in the lung, kidneys, and thyroid. In a fully adjusted regression analysis, neither prior malignancy nor pMIM increased mortality odds. However, having both was associated with a higher 1-year mortality (odds ratio 4.02, 95% confidence interval 1.50 to 10.73, p = 0.006). In conclusion, presence of pMIM alone was not associated with an increased 1-year mortality among patients undergoing TAVI. However, the presence of pMIM and a history of prior malignancy was associated with a significant increase in 1-year mortality.