Measurement of renal cortical volume using AI and prediction of renal function loss after donation
Sep 01, 2024
|
Professor Min Sang-il's team at Seoul National University Hospital Transplant Vascular Surgery (Professor Cho Eun-ah at Chung-Ang University Hospital, Professor Lee Joo-han at Severance Hospital, and CEO Kim Jin-sung at Onco Soft) published the results of a multicenter retrospective cohort study that measured renal cortical volume using AI-based CT images and analyzed the association with renal function decline after kidney transplantation.
As we enter an aging society, kidney transplant surgery for elderly donors is increasing. However, in the case of elderly males, renal function is likely to deteriorate due to changes in the kidney microstructure, such as glomerular hardening due to aging. Accordingly, there was a need for a method to predict renal function loss after kidney donation. Although previous studies have reported a correlation between renal cortical volume and renal function, existing measurement methods are complex and time-consuming, making it difficult to apply to actual clinical practice.
To solve these problems, the research team developed an 'AI-based automatic segmentation model'. This AI model is designed to analyze CT images before donation to automatically measure cortical volume in the kidneys. Since the renal cortex is an important part of the function of the kidney, accurate measurement of this volume is very important for renal function prediction.
To verify the accuracy of the model, the team compared the AI measured cortical volume with the results measured manually. As a result of the verification, the accuracy of the AI model was very high, and the Dice-like coefficient (an index to evaluate the overlap between two images, meaning that it is closer to 1) 0.97 and Hausdorff distance (measuring the maximum error between the predicted and actual boundaries, and the smaller the value, the more accurate it is). This means that AI models can measure the actual renal cortex volume with great accuracy.
The research team analyzed the association between renal cortical volume measured using AI models and renal function after donation (estimated glomerular filtration rate, eGFR). eGFR is an indicator of the kidney's filtration ability, meaning that renal function decreased as the value decreased. The generalized addition model (GAM) was used to analyze changes in renal function over time after donation.
As a result, older donors (over 60 years of age) tended to have a greater decrease in renal function after donation compared to younger donors (under 60 years of age). Specifically, the decrease in eGFR in older donors was statistically significantly greater (P = 0.041), which means older donors experienced more renal decline.
However, preoperative renal cortical bulky donors tended to have less renal function decline after donation. This difference was statistically significant (p<0.001), especially in older donors. This suggests that older donors with large renal cortical volumes are better able to maintain renal function after donation.
The research team emphasized that measuring renal cortical volume using AI can serve as an important indicator for predicting renal function loss after donation. In addition, if the model is integrated into the donor screening and evaluation process, it is expected to greatly contribute to increasing the success rate of kidney transplantation and enhancing donor safety, it added.
Professor Sang-il Min (transplant vascular surgery) said "The study, which demonstrated the clinical usefulness of measuring renal cortical volume using artificial intelligence, is considered to have made significant progress in the evaluation and prognosis of kidney donors"Especially for older donors, it is significant that we can support safer donation decisions by providing a more precise way to predict renal function loss than before."
The study was published online in the International Journal of Surgery 'International Journal of Surgery (IF=12.5)', a world-renowned journal in the field of surgery.
|
bellho@sportschosun.com
