Abstract: TH-PO1121
Machine Learning-Guided Measurement of Visceral Fat Area in Computed Tomography and Delayed Graft Function After Kidney Transplantation
Session Information
- Transplantation: Clinical - Predictors of Outcomes - Biomarkers and Beyond
November 07, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Transplantation
- 1902 Transplantation: Clinical
Authors
- Kim, Ji Eun, Seoul National University Hospital, Seoul, Korea (the Republic of)
- Lee, Hajeong, Seoul National University College of Medicine, Seoul, Korea (the Republic of)
- Han, Seung Seok, Seoul National University College of Medicine, Seoul, Korea (the Republic of)
- Kim, Yon Su, Seoul National University College of Medicine, Seoul, Korea (the Republic of)
Background
Innovative machine learning can be applied to efficiently provide the information on medical imaging and be useful in the prediction model. Accordingly, the present study measured visceral fat area (VFA) with updated machine learning algorithm in kidney transplant recipients and evaluated its correlation with delayed graft function (DGF).
Methods
A total of 287 adult kidney recipients who examined abdominal computed tomography with full range of torso before transplantation were enrolled. VFA in the abdominal cavity was measured in cubic meters throughout machine learning algorithm. DGF was defined as the need for dialysis during the first transplantation week.
Results
The mean age was 47.8 ± 11.3 years and male was 66.2%. The mean body mass index was 24.6 ± 3.5 kg/m2. The VFA was 2.88 ± 1.92 m3, and the body surface area-adjusted value was 1.59 ± 0.95 m. The adjusted VFA had a linear relationship with the surgery time (β = 0.12; P = 0.040). The risk of DGF increased depending on an increase of 1 unit in adjusted VFA with an odds ratio of 1.80 (1.07–3.02). However, body mass index was not associated with DGF (odds ratio, 0.99 (0.83–1.19). The area under receiver operating characteristic curve of adjusted VFA was 0.73, which was greater than 0.50 in body mass index (P = 0.032).
Conclusion
Machine learning algorithm may efficiently provide information on VFA of kidney recipients. This issue will improve the predictive capacity of transplant outcomes such as DGF.