DEVELOPMENT AND EXTERNAL VALIDATION OF A NOMOGRAM FOR PREDICTION OF A FAVORABLE OUTCOME AFTER MAGNETIC SPHINCTER AUGMENTATION: A MULTICENTER STUDY
Shahin Ayazi*1, Sebastian F. Schoppmann2, Katrin Schwameis1, Ping Zheng1, Milena Nikolic2, Aleksa Matic2, Ali H. Zaidi1, Kristy L. Chovanec1, Blair Jobe1
1Esophageal Institute, Allegheny Health Network, Pittsburgh, PA; 2Department of Surgery, Medical University of Vienna, Vienna, Vienna, Austria
Introduction:
Magnetic sphincter augmentation (MSA) is increasingly utilized in the management of patients with gastroesophageal reflux disease (GERD). Identifying patients who are more likely to benefit from this procedure will guide patient selection and optimize the outcome. We designed the current multicenter study to develop a nomogram to predict the probability of a favorable outcome after MSA and to validate it in an external cohort of patients.
Material and methods:
A nomogram to predict a favorable outcome based on GERD-HRQL total score, DeMeester score, age and sex was developed from 294 patients who underwent MSA at a high-volume foregut surgical center. These four parameters were previously identified in a multivariable regression analysis as preoperative factors predicting a favorable outcome after MSA. Favorable outcome was defined as freedom from proton pump inhibitor (PPI) and 50% or more improvement in GERD-HRQL total score.
A separate group of 77 patients from another high-volume foregut surgical center was used as validation population. External validation of the model was then assessed in this population using C statistic (area under the curve), sensitivity, specificity and correct rate prediction.
Results:
GERD-HRQL total score >15 (odds ratio, 7.47; 95% CI: 3.32 - 16.81; p<0.001), age <45 years (odds ratio, 4.17; 95% CI: 1.14 - 15.23; p=0.0305), abnormal DeMeester score (odds ratio, 2.55; 95% CI: 1.14 - 5.70; p=0.0225) and male sex (odds ratio, 2.49; 95% CI: 1.09 - 5.66; p=0.0301) were independently associated with a favorable outcome after MSA. The external validation of the predictive model was associated with a C statistic of 0.77, sensitivity of 91.2%, specificity of 45% and correct rate of 79.2%.
The constructed nomogram (Figure) translates multivariate model parameter estimates into a visual scoring system where the estimated probability of a favorable outcome can be calculated.
Conclusion:
The constructed model in this study correctly predicts a favorable outcome after MSA in 79.2 % of the patients and has a high sensitivity (91.2%). Utilizing this predictive nomogram will allow clinicians to identify patients with a high likelihood of a favorable outcome after MSA and can be used in patient selection and preoperative counseling.
In this predictive nomogram, points are assigned for each variable by drawing a line upward from the corresponding variable to the "Points" line. The sum of the points plotted on the "Total Points" line corresponds with the prediction of favorable outcome after MSA.
Back to 2020 Abstracts