Logistic Regression as a Predictive Tool of Post-Operative Mortality in Hemiarch Surgery
Presented During:
Thursday, April 25, 2024: 5:38PM - 7:00PM
Sheraton Times Square
Posted Room Name:
Central Park
Abstract No:
P0193
Submission Type:
Abstract Submission
Authors:
Adam Carroll (1), Nicolas Chanes (1), Muhammad Aftab (1), T. Brett Reece (1)
Institutions:
(1) University of Colorado Anschutz, Denver, CO
Submitting Author:
Adam Carroll
-
Contact Me
University of Colorado Anschutz
University of Colorado Anschutz
Co-Author(s):
Nicolas Chanes
-
Contact Me
University of Colorado Anschutz
University of Colorado Anschutz
*Muhammad Aftab
-
Contact Me
University of Colorado Anschutz
University of Colorado Anschutz
*T. Brett Reece
-
Contact Me
University of Colorado Anschutz
University of Colorado Anschutz
Presenting Author:
Adam Carroll
-
Contact Me
University of Colorado Anschutz
University of Colorado Anschutz
Abstract:
Purpose:
Logistic regression algorithms have shown the potential to predict surgical outcomes in cardiac surgery. Although they have been used in aortic surgery, their application has been constrained by the extensive data needed before and after surgery to effectively employ these models. Applying logistic regression to hemiarch surgery to assess mortality would be of particular benefit, especially given disagreements in weighing risks of surveillance versus operative intervention. We sought to apply and develop a logistic regression model to predict post-operative mortality following hemiarch surgery, utilizing only data from patient presentation and intra-operative procedures performed.
Methods:
From our single institution prospectively maintained database, we identified a total of 602 adult patients who underwent hemiarch replacement between June 2009 and October 2022. These patients were randomly divided into training (80%) and testing (20%) sets and various logistic regression models were constructed to predict overall post-operative mortality. We considered 17 input parameters from the index hospitalization which were comprised of demographic and pre-operative characteristics. To assess model performance, we employed multiple measures, including accuracy, Brier score, and area under the receiver-operating characteristic curve (AUC-ROC). Furthermore, we calculated odds ratios and confidence intervals derived from the logistic regression model.
Results:
Post-operative mortality was noted in 56 patients (9.30%) who underwent hemiarch replacement. The final logistic regression model demonstrated a cross-validation accuracy of 91% and was well-calibrated as evidenced by the low Brier score of 0.09. The predictor also demonstrated strong performance on the testing set, achieving an accuracy of 86%. Our best performing overall post-operative mortality prediction model achieved an AUC-ROC of 0.70 both on the training and testing sets. A heightened mortality risk was linked to factors such as aortic dissection with malperfusion, the urgency of the procedure, adjunctive valvular repair, and concomitant CABG or root replacement. Factors that reduced risk included aortic dissection without malperfusion, elective procedures, and the performance of hemiarch surgery without any adjunctive interventions.
Conclusions:
Logistic regression algorithms can accurately predict mortality after hemiarch surgery, specifying key intra-operative procedures that lead to higher rates of mortality. Given the absence of risk models available for aortic surgery, logistic regression models may have the potential to serve as an excellent clinical tool to predict surgical outcomes.
Logistic regression algorithms have shown the potential to predict surgical outcomes in cardiac surgery. Although they have been used in aortic surgery, their application has been constrained by the extensive data needed before and after surgery to effectively employ these models. Applying logistic regression to hemiarch surgery to assess mortality would be of particular benefit, especially given disagreements in weighing risks of surveillance versus operative intervention. We sought to apply and develop a logistic regression model to predict post-operative mortality following hemiarch surgery, utilizing only data from patient presentation and intra-operative procedures performed.
Methods:
From our single institution prospectively maintained database, we identified a total of 602 adult patients who underwent hemiarch replacement between June 2009 and October 2022. These patients were randomly divided into training (80%) and testing (20%) sets and various logistic regression models were constructed to predict overall post-operative mortality. We considered 17 input parameters from the index hospitalization which were comprised of demographic and pre-operative characteristics. To assess model performance, we employed multiple measures, including accuracy, Brier score, and area under the receiver-operating characteristic curve (AUC-ROC). Furthermore, we calculated odds ratios and confidence intervals derived from the logistic regression model.
Results:
Post-operative mortality was noted in 56 patients (9.30%) who underwent hemiarch replacement. The final logistic regression model demonstrated a cross-validation accuracy of 91% and was well-calibrated as evidenced by the low Brier score of 0.09. The predictor also demonstrated strong performance on the testing set, achieving an accuracy of 86%. Our best performing overall post-operative mortality prediction model achieved an AUC-ROC of 0.70 both on the training and testing sets. A heightened mortality risk was linked to factors such as aortic dissection with malperfusion, the urgency of the procedure, adjunctive valvular repair, and concomitant CABG or root replacement. Factors that reduced risk included aortic dissection without malperfusion, elective procedures, and the performance of hemiarch surgery without any adjunctive interventions.
Conclusions:
Logistic regression algorithms can accurately predict mortality after hemiarch surgery, specifying key intra-operative procedures that lead to higher rates of mortality. Given the absence of risk models available for aortic surgery, logistic regression models may have the potential to serve as an excellent clinical tool to predict surgical outcomes.
Aortic Symposium:
Aortic Arch
Image or Table
