Personalized Prediction Model Generated with Machine Learning for Kidney Function One Year After Living Kidney Donation

Personalized Prediction Model Generated with Machine Learning for Kidney Function One Year After Living Kidney Donation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Personalized Prediction Model Generated with Machine Learning for Kidney Function One Year After Living Kidney Donation Rikako Oki, Toshihio Hirai, Kazuhiro Iwadoh, Yu Kijima, Hiroyuki Hashimoto, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5862042/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 01 Jul, 2025 Read the published version in Scientific Reports → Version 1 posted 10 You are reading this latest preprint version Abstract Living kidney donors typically experience approximately a 30% reduction in kidney function after donation, although the degree of reduction varies among individuals. This study aimed to develop a machine learning (ML) model to predict serum creatinine (Cre) levels at one year post-donation using preoperative clinical data, including kidney-, fat-, and muscle-volumetry values from computed tomography. A total of 204 living kidney donors were included. Symbolic regression via genetic programming was employed to create an ML-based Cre prediction model using preoperative clinical variables. Validation was conducted using a 7:3 training-to-test data split. The ML model demonstrated a median absolute error of 0.079 mg/dL for predicting Cre. In the validation cohort, it outperformed conventional methods (which assume post-donation eGFR to be 70% of the preoperative value) with higher R² (0.58 vs. 0.27), lower root mean squared error (5.27 vs. 6.89), and lower mean absolute error (3.92 vs. 5.8). Key predictive variables included preoperative Cre and remnant kidney volume. The model was deployed as a web application for clinical use. The ML model offers accurate predictions of post-donation kidney function and may assist in monitoring donor outcomes, enhancing personalized care after kidney donation. Health sciences/Nephrology/Kidney Health sciences/Health care kidney transplantation living donor machine learning kidney function post-donation prediction Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Careful screening of living donor candidates is critical to minimizing the risk of end-stage kidney disease (ESKD) and ensuring ongoing monitoring of renal function post-donation. Although rigorously screened living kidney donors were traditionally thought to have comparable risks of mortality and ESKD to the general population 1 , they may face a higher risk of ESKD compared to matched healthy non-donors. 2 Additionally, selection criteria for living donors have broadened to include medically complex individuals, such as those with advanced age, hypertension, obesity, or lower estimated glomerular filtration rate (eGFR). 3 These evolving trends underscore the urgent need for more precise and individualized prediction models to assess and monitor postoperative renal function in this population. Post-donation eGFRs are typically reach approximately 60–70% of the pre-donation levels, attributed to compensatory hypertrophy of the remaining kidney. 4 Grams et al. developed an online risk tool to estimate the long-term risk of ESKD for living kidney donor candidates, using meta-analyzed risk associations from seven general population cohorts. 5 Several reports have proposed predictive factors for post-donation kidney function that included donor age, sex, race, body mass index (BMI) and preoperative computed tomography (CT) volumetry of kidney. 6 – 9 These were based on observational study and traditional statistics which aims to identify significant risk factors among explanatory variables through model-driven regression using linear models. 10 Traditional regression models that fit data to pre-defined models are directly meaningful to clinicians, though it relies on liner predefined models and are constrained by strict assumptions. In contrast, symbolic regression (SR) via genetic programming (GP) is a machine learning (ML) algorithm where the goal is discovering an explicit mathematical formula that best describes a given dataset from the vast function space, enabling it to uncover complex, non-linear interactions directly from the data. 11 Genetic programming (GP) explores both the structure of the model and its parameters. This evolutionary methodology is based on the principles of mutation and natural selection, mirroring the way organisms evolve and adapt to their environments. In SR via GP, innumerable mathematical formulas evolve to fit the given dataset, progressively yielding better ones that can predict the target value from the explanatory variables. In the previous study, Ueno et al. utilized SR via GP to evaluate pathological factors associated with the development of glomerular hypertrophy (GH). 12 From a set of 60 variables, the SR model identified key factors such as inflammation, vascular damage, and obesity as significant predictors, while eGFR was ranked low (46th out of 60). This finding highlights the ability of SR to distinguish morbid GH from adaptive hypertrophy due to nephron loss. Collectively, these results suggest that SR can evaluate variables in an unbiased manner, providing valuable insights into the underlying mechanisms. The objective of this study was to develop a machine learning model for more accurate and personalized prediction of post-donation kidney function. We employed an SR model to predict post-donation creatinine (Cre) values using preoperative variables, including CT volumetry data for excised and non-excised kidney volumes. Given that excess visceral fat and reduced skeletal muscle mass—characteristic of sarcopenia—are recognized risk factors for the development of chronic kidney disease (CKD) 13 , 14 , 15 , we also implemented CT volumetry for visceral fat and skeletal muscle. The resulting model was integrated into a user-friendly interface to facilitate clinical application and improve decision-making in transplant practice. Results Patient’s background The clinical characteristics and laboratory data of the patients in the study cohorts are summarized in Table 1 . All the participants were Asian. The mean donor age was 59.9 years, and 35% were male. The median baseline Cre level before donation was 0.7mg/dl. There were no significant differences in the variables between the training and the validation cohort. Table 1 Comparison between training cohort and validation cohort. Variables all ( N = 204) Training cohort ( N = 143) Validation cohort ( N = 61) p Age at donation (y.o) 59.9 ± 8.78 60.3 ± 8.70 59.0 ± 8.96 0.363 Male [n (%)] 71 (34.8) 53 (37.1) 18 (29.5%) 0.338 Body weight (kg) 56.0 (51.0, 63.9) 57.0 (51.0, 64.0) 53.0 (51.0, 61.0) 0.098 Hight (cm) 159.0(154.5, 165.0) 160 (155, 165) 159 (154, 164) 0.698 BMI 22.4 (20.4, 24.2) 22.6 (20.6,24.4) 21.9 (20.2, 23.2) 0.053 Systolic blood pressure (mmHg) 129.9 ± 19.9 130.7 ± 19.7 127.9 ± 20.3 0.357 Diastolic blood pressure (mmHg) 76.7 ± 14.3 77.4 ± 13.0 75.1 ± 17.0 0.284 smoking [n (%)] 58 (28.4) 42 (29.4) 16 (26.2) 0.736 History of CVD [n (%)] 4 (2.0) 3 (2.1) 1 (1.6) 1.000 Antihypertensive agents [n (%)] 40 (19.6) 27 (18.9) 13 (21.3) 0.703 Lipid-lowering agents [n (%)] 21 (10.3) 12 (8.4) 9 (14.8) 0.209 uric acid lowering agents [n (%)] 4 (2.0) 4 (2.8) 0 (0) 0.319 WBC (/µl) 4955 (4265, 6055) 4930 (4270, 5995) 5190 (4240, 6180) 0.490 Hb (g/dl) 13.7 ± 1.24 13.8 ± 1.27 13.6 ± 1.18 0.468 Platelet (*10 3 /µl) 21.7 (18.2, 25.4) 22.0 (18.3, 25.3) 21.0 (18.2, 26.0) 0.846 Cre (mg/dl) 0.70 (0.62, 0.81) 0.69 (0.62, 0.84) 0.70 (0.62, 0.78) 0.612 eGFR (ml/min/1.73m 2 ) 71.9 (65.1, 80.3) 71.4 (65.0, 80.2) 72.7 (65.1, 80.2) 0.650 BUN (mg/dl) 13.8 (11.2, 15.7) 13.8 (11.2, 15.8) 13.8 (11.7, 15.6) 0.646 Na (mEq/L) 142 (141,143) 142 (141, 143) 141 (141, 143) 0.417 K (mEq/l) 4.2 (4.0, 4.4) 4.2 (4.0, 4.4) 4.2 (4.0, 4.4) 0.917 AST [IU/l] 20.0 (17.0, 24.0) 20.0 (17.0, 24.5) 20.0 (17.0, 24.0) 0.849 ALT [IU/l] 17.0 (13.0, 22.0) 17.0 (13.0, 22.0) 17.0 (14.0, 23.0) 0.583 TP (g/dl) 7.24 ± 0.36 7.24 ± 0.37 7.24 ± 0.34 0.989 CRP (mg/dl) 0.05 (0.04, 0.10) 0.05 (0.04, 0.11) 0.05 (0.04, 0.10) 0.460 T.chol (mg/dl) 211 (187, 233) 213 (186, 233) 205 (190, 233) 0.521 UA (mg/dl) 4.80 (4.10, 5.90) 4.90 (4.10, 5.90) 4.70 (4.10, 5.90) 0.524 Glucose (g/dl) 98.0 (92.0, 104) 98.0 (92.0, 104) 98.0 (92.0, 107) 0.905 HbA1C (%) 5.70 (5.50, 5.90) 5.70 (5.50, 5.90) 5.70 (5.50, 5.90) 0.772 Proteinuria (+) 2 (1.0) 1 (0.7) 1 (1.6) 0.510 Urinary occult blood (≥+) 14 (6.9) 11 (7.7) 3 (4.9) 0.561 Cre at 1-year post-donation (mg/dl) 1.06 (0.93, 1.23) 1.06 (0.93, 1.26) 1.06 (0.91, 1.18) 0.378 Results of CT findings Volume of excised kidney (ml) 143 (126, 161) 148 (129, 163) 140 (118, 157) 0.133 Volume of non-excised kidney(ml) 137 (120, 156) 138 (121, 159) 133 (114, 152) 0.165 Area psoas muscle at L3(cm 2 ) 11.3 (9.01,15.0) 11.5 (9.13, 14.9) 10.7 (8.92, 15.2) 0.501 Area skeletal muscle at L3(cm 2 ) 98.5 (85.8, 126) 100.5 (87.0, 126) 92.1 (82.5, 120) 0.139 Area visceral fat(cm 2 ) 86.4 (56.0, 125) 82.1 (58.2, 126) 90.2 (50.9, 115.7) 0.528 Area visceral Fat at L3 (cm 2 ) 77.2 (38.9, 127) 77.2 (46.1, 135) 81.1 (38.2, 118) 0.516 Continuous data are presented as mean ± SD or median (IQR): CVD, cardiovascular disease; WBC, white blood cell; Hb, hemoglobin; Cre, creatinine; eGFR, estimated glomerular filtration rate; BUN, blood urea nitrogen; Na, sodium; K, potassium; AST, aspartate aminotransferase; ALT, alanine aminotransferase; TP, total protein; CRP, C-reactive protein; T.chol, total cholesterol; UA, uric acid; HbA1C, hemoglobinA1C CT volumetry data There were no significant differences in CT volumetry data between training cohort and validation cohort (Table 1 ). The correlation between CT volumetry data and preoperative Cre level was investigated using Pearson’s correlation coefficient (r). The analysis revealed moderate positive correlations between area of psoas muscle/skeletal muscle/visceral fat and pre-operative Cre level (Supplementary table 1 A). Significant positive correlations were observed between CT volumetry data and pre-operative body weight (Supplementary table 1 B). Correlation between explanatory variables and post-donation Cre level All correlation coefficients between preoperative data and Cre level at 1 year post-donation are presented in Supplemental Fig. 1. Additionally, the point-biserial correlation coefficients (r values) for the top 10 variables with the strongest positive correlations are summarized in Table 2 . As with the pre-operative Cre level, the Cre level at 1 year post-donation demonstrated a positive correlation with preoperative skeletal muscle volume (r = 0.62, p < 0.001), along with weak positive correlations with the psoas muscle and visceral fat volumes. Table 2 Correlation between pre-operative variables and creatinine level at 1 year post-donation. Variables r (95% CI) p Cre 0.84 (0.80–0.88) < 0.001 Male 0.65 (0.56–0.72) < 0.001 Area skeletal muscle at L3 0.62 (0.53–0.70) < 0.001 Body weight 0.57 (0.47–0.65) < 0.001 UA 0.51(0.40–0.60) < 0.001 Height 0.47 (0.35–0.57) < 0.001 Area visceral Fat at L3 0.46 (0.34–0.56) < 0.001 Hb 0.40 (0.27–0.51) < 0.001 Area visceral fat 0.37 (0.24–0.48) < 0.001 Area psoas muscle at L3 0.25 (0.12–0.37) < 0.001 Cre, creatinine; UA, uric acid; Hb, hemoglobin Analysis of predictive variables for Cre level post-donation using ML By using 34 explanatory variables, Data Modeler automatically generated 1123 models that calculate Cre level at 1-year post-donation. Figure 1 show their distribution in the function space. We selected 98 models with lower complexities and lower 1-R 2 in each epoch to limit the number of models to 9% of all generated models. We ranked the frequencies of all explanatory variables that were used in the selected models (Fig. 2 ). The selected factors for generating models included age, male, BW, the history of CVD, BUN, Cre, HbA1C, and the volume of the non-excised kidney. Creating a predictive model for Cre level at 1-year To establish an optimized model for predicting Cre levels 1 year post-donation, ensemble learning was performed using the bagging method, which calculates the trimmed means of the selected models. The formula for the optimized DKF model, incorporating key variables extracted from the developed models, is provided in Supplemental Document 1. The R 2 , RMSE, and MAE values between the predicted Cre levels generated by the optimized DKF model and the measured values were 0.72, 0.1, 0.073mg/dl mg/dL in the training cohort (Fig. 3 A) and 0.69, 0.1, and 0.079 mg/dl in the validation cohort (Fig. 3 B), respectively. The impact of each variable in the generated formula We investigated factors that were used in the generated formula to determine which factors had higher impact on the target. A simulation was performed by changing each variable within its range while keeping other variables fixed at their median values. Table 3 lists the 8 most significant driver variables that influenced the result. Pre-Cre level and the volume of the non-excised kidney were the most influential factors for Cre levels at 1 year, with high ΔTarget values. Supplementary Fig. 2 illustrates the extent to which the target Cre value changes as the top 3 driver variables are varied. Table 3 The changes in the Cre level as each explanatory variable changes from its minimum to its maximum with other variables fixed at their median (ΔTarget). ΔTarget Variables 1 0.685 pre Cre level 2 0.450 volume of non-excised kidney 3 0.281 body weight 4 0.102 Male 5 0.062 BUN 6 0.047 Pre HbA1C 7 0.047 History of CVD 8 0.002 Age at donation Cre, creatinine; BUN, blood urea nitrogen; HbA1C, hemoglobinA1C; CVD, cardiovascular disease The development of a sparse model To enhance clinical applicability, we developed a simplified prediction formula (referred to as the sparse DKF model), where variables with smaller ΔTarget were fixed at their median values. Specifically, BUN and HbA1C were fixed at their median values, while the history of CVD was set to its mode, which was 0. The key driving factors for this simplified model included age, sex (male), body weight (BW), Pre-Cre, and the volume of the non-excised kidney. The developed model has been uploaded to GitHub and implemented as a web application for convenient calculations, accessible at the following address: [ https://donorcrcalculatoren-89wgze554kd9n2yjkcyubb.streamlit.app/ ]. Verification of accuracy of developed models To assess the accuracy of the developed models, we calculated the R 2 , RMSE, and MAE for the values predicted by the optimized DKF model, the sparse DKF model, and the conventional DKF model, comparing them to the measured eGFR values in the validation cohort (N = 61). (Table 4 ) The optimized DKF model achieved the highest R 2 and the lowest RMSE and MAE, indicating minimal prediction error and high predictive accuracy. The sparse DKF model, designed with clinical practicality in mind, also demonstrated significantly better predictive accuracy compared to the conventional DKF model. Consequently, both the optimized DKF model and the sparse DKF model showed superior data fit and can be considered more reliable than the conventional DKF model. Table 4 R-squared, RMSE and MAE in comparison of the predicted values by the optimized DKF model/the sparse DKF model/ the conventional DKF model and the observed eGFR values. Model R-squared Root Mean Squared Error Mean Absolute Error The optimized DKF model 0.58 5.27 3.92 The sparse DKF model 0.52 5.36 4.17 The conventional DKF model 0.27 6.89 5.80 DKF, donor’s kidney function Next, we analyzed the correlation between the predicted values from the optimized DKF model and the measured eGFR values (Fig. 4 A), as well as the predicted values from the conventional DKF model and the measured eGFR values (Fig. 4 B). Three cases with |Z| ≥ 2 were identified, and these cases completely overlapped between the optimized DKF model and the conventional DKF model. Supplementary Table 2 compares the patient characteristics among inliers (-2 < Z 2), and low outliers (Z < -2) in the prediction of Cre at 1 year post-donation. Although the small sample size limits statistical interpretation, cases with higher preoperative Cre levels appeared more likely to be classified as outliers in both the optimized DKF model and the conventional DKF model. Discussion In the present study, we generated a predictive model for Cre level at 1-year post-donation, using the SR via GP technique, which is one of the ML techniques. The developed model, referred to “the optimized DKF model” by ML was found to have higher accuracy, demonstrating higher R 2 values and lower RMSE and MAE, compared to the conventional DKF model which assumed eGFR post-donation was 70% of the preoperative value. Preoperative Cre level and the volume of the non-excised kidney were identified as the most influential predictive factors for Cre levels at 1 year post-donation. Predicting the kidney function post-donation will facilitate more rigorous management and help achieve optimal post-transplant outcomes. Statistical analysis revealed that there was a moderate correlation between the skeletal muscle volume and Cre levels at 1 year, as well as between the visceral fat volume and Cre levels at 1 year. The volume of non-excised kidney was not corelated to Cre level post-donation. Approximately 95% of the human body’s total creatine is located in skeletal muscle and serum Cre can be served as a surrogate marker of skeletal muscle mass even in CKD patients. 16 The previous study revealed that visceral adipose tissue detected by CT scan was associated with CKD when defined using cystatin C estimating equations but not when using a Cre-based estimating equation. 17 On the other hand, area of skeletal muscle and area of visceral fat were not identified as influential factors in the ML analysis. Volume of non-excised kidney was found to be a factor affecting Cre level at 1-year. The larger the volume of the non-excised kidney, the lower the Cre levels at 1 year tended to be. (Supplementary Fig. 2A) Shimada et al demonstrated that body surface area-adjusted preserved kidney volume calculated by the 3D reconstructed image was an independent risk factor for > 30% reduction of eGFR at 1-year post-donation (odds ratio, 0.93) 9 , which aligned with our findings. It is well known that adaptive hyperfiltration after donor nephrectomy is attributable to hyper perfusion and hypertrophy of the remaining glomeruli. 18 The increases in single-kidney renal plasma flow, cortical volume, and GFR continue through to the late post-donation period. 18 From the perspective of donor outcomes, it is suggested to determine which side of kidney to be excised based on the volume of the remaining kidney. As previously mentioned, the post-donation eGFRs are approximately 60–70% of the pre-donation values based on past reports. 4 To assess the generated model’s goodness-of-fit, we compared its R 2 , RMSE, and MAE values to those of the conventional prediction method. The optimized DKF model showed higher explanatory power and accuracy for predicting Cre level at 1 year compared to the conventional method. The eGFR was reported to increase by + 0.35 ml/min/ 1.73 m 2 per year from 6 weeks post-donation onward, with this increase leveling off by the fifth year. 19 It is reported that increase in GFR can begin as early as 8 hours post-donation, however, this acute compensation is less efficient in older donors. 20 Thus, predicting Cre levels at 1-year by ML may be more beneficial than the conventional prediction method in clinical practice. However, outliers also exist in the optimized DKF model. Since the cases with pre-operative Cre values deviating from the median were likely to become outliers, the prediction may not be applied in the cases with extreme Cre level before the donation. This model was developed by analyzing the characteristics of donors who were ultimately selected for kidney donation, excluding those who did not meet the criteria for a living kidney donor. Therefore, it is essential to note that even if a good Cre level was expected by the generated model, it does not guarantee the safety of prognosis for cases which don’t meet the criteria for a living kidney donor. It is important to adhere to the guidelines for evaluating eligibility for a living kidney donor. How to approach to donor selection has been discussed nowadays. While the Kidney Disease Improving Global Outcomes Guideline recommends the use of fixed cutoffs 21 , age and sex- based GFR cutoffs are commonly used in the British Transplantation Society, the European Renal Best Practice, and the Canadian Society of Transplantation, considering kidney function decline with healthy aging. 22 – 24 It is challenging to judge whether older candidates' kidney function is appropriate for their age because they tend to be more medically complex. In such cases, which include so-called marginal donors, the prediction of kidney function post-donation by the generated model might be useful in planning follow-up. In cases where a donor is acceptable for kidney donation but is predicted to have poor post-donation kidney function, careful follow-up should be warranted after donation. Additionally, as non-excised kidney volume was an influential factor for predicting Cre level at 1-year, it is possible to simulate the impact on kidney function based on CT imaging data. It can be used as a reference to consider the impact on kidney function depending on which kidney will be donated. Overall, although the optimized DKF model/the sparse DKF model cannot be used directly for donor selection, it might be helpful as a reference for donor selection, for simulating kidney function post-donation, and for planning follow-up. We acknowledge there are limitations in our study. This model was generated by data set from a single institution with a small sample size. There might be other factors affecting kidney function after donation beyond our dataset. Since the validation utilized a validation cohort included within the dataset, external data is required for further validation. Larger studies with more factors are warranted to verify our findings. However, this study indicated the strong potential of ML system for identifying unknown risk factors related to post-donated kidney function. In conclusion, the ML model achieved an effective predictive performance for predicting Cre level post-donation. It is meaningful for transplant physicians to pay sufficient attention to the care for donors after donation. Applying ML techniques to the clinical field has the potential to lead to better healthcare for patients. Materials and Methods Patients and data collection We retrospectively enrolled 204 patients who underwent enhanced CT prior to donor nephrectomy for living kidney transplantation at Tokyo Women’s Medical University Hospital between January 2012 and December 2016. Patients who were lost to follow-up after discharge or whose follow-up period after nephrectomy was less than one year were excluded. This study was approved by the Institutional Review Board of the Tokyo Women’s Medical University Hospital (#2021 − 0122). Since this is a retrospective study, informed consent was waived by the Institutional Review Board of the Tokyo Women’s Medical University Hospital. All methods of research procedures were performed in accordance with the Declaration of Helsinki. Basic information about the patients was obtained from donor medical records within 3 months before donation, including age, sex, medications for hypertension, hyperlipidemia, and hyperuricemia, smoking habits, systolic/diastolic blood pressure, body mass index, and laboratory data. The laboratory data comprised white blood cell (WBC) count, hemoglobin (Hb), platelets, AST (aspartate aminotransferase), ALT (alanine aminotransferase), blood urea nitrogen (BUN), serum creatinine (Cre), estimated glomerular filtration rate (eGFR), sodium (Na), potassium (K), total protein (TP), C-reactive protein (CRP), uric acid (UA), hemoglobin A1c (HbA1c), total cholesterol (T. Chol), urinary protein, and urinary occult blood as qualitative tests. Blood pressure was measured with a brachial sphygmomanometer at the first day of hospitalization prior to donation. CT volumetry Preoperative dynamic CT was performed using 64-,80- or 320-multidetector CT scanners (Aquilion 64, Aquilion Prime SP, Aquilion ONE; Canon Medical Systems Corporation, Otawara, Tochigi, Japan). Iodinated contrast media (Iohexol, Omnipaque 300; GE Healthcare Pharmaceutical Diagnostics, Tokyo, Japan) was injected using a power injector following an unenhanced CT scan. The nephrographic phase images were transferred and analyzed using dedicated software (SYNAPSE VINCENT;FUJIFILM Corporation Tokyo, Japan). The software semi-automatically generated a 3D model of the kidney by stacking 1-mm axial slices and calculated the kidney's volume (Supplementary Fig. 3A, B). Whole-body skeletal muscle mass was estimated from the trunk muscle area at the L3-4 level which is known to be highly correlated with the total body skeletal muscle volume (Supplementary Fig. 3C). 25 As for the visceral fat area at the naval level, a single axial slice at the L3-L4 intervertebral space was chosen for analysis, as it is frequently used as a surrogate for abdominal adiposity (Supplementary Fig. 3C). 26 Evaluation of renal function Renal function, measured by serum Cre levels, was evaluated pre-donation and at 1-year post-donation. The pre-donation serum Cre level was defined as the most recent result obtained within 3 months before donation. Serum Cre level at 1-year post-donation was employed as the primary outcome for this study. The eGFR was calculated using a formula specific to Japanese patients with CKD (eGFR [mL/min/1.73 m 2 ] = 194 x Serum creatinine(-1.094) × Age(-0.287) × 0.739 [if female]). 27 Statistical analysis Continuous data were described as mean ± standard deviation or median (interquartile range). Unpaired t-test or Mann–Whitney U test was used to compare continuous variables. The chi-square test or Fisher’s exact test was used to compare the categorical variables. A Pearson’s correlation test was performed on correlation between clinical factors and Cre level at 1-year post donation. A Pearson’s correlation test was also performed to evaluate the correlation between the eGFR predicted by the generated/conventional model and the observed eGFR measured in practice. Values for which p was less than 0.05 were inferred as significant. Model generation The data were randomly divided into a training and a validation data in a 7:3 ratio. By using training data, we conducted SR via GP to construct machine learning models for predicting Cre levels at 1-year post-donation using Data Modeler version 9.3 (Evolved Analytics LLC, Rancho Santa Fe, CA, USA) which runs on Mathematica version 12.1 (Wolfram Research Incorporated, Champaign, IL, USA). The explanatory variables included 34 pre-operative variables (Supplementary Table 3) such as gender, age, height, body weight, blood pressure, smoking history, disease history, laboratory data, and CT volumetry data. DataModeler was executed with 8 independent evolutions and 6 minutes modeling time. During the evolution, SR via GP automatically discards less important variables and selects more important ones, effectively performing dimensionality reduction. Then it generates hundreds of predictive models, selects a few to a few dozen of simpler and less erroneous models, and uses their trimmed mean as the optimized predictive model, just like a bagging method used in random forest. The optimized model created through the process is referred to as “optimized donor’s kidney function (DKF) model” in this paper. We also explored the driver variables in the optimized DKF model and their impact on the overall model, illustrating their effects on the target using an explore plot. In this plot, each variable was varied within its range while keeping all other variables fixed at their median values. Finally, we developed a simplified prediction formula from the optimized DKF model, in which variables with smaller effects on the target were fixed at their median values. This model is referred to as the “sparse DKF model.” As a conventional method for estimating renal function post-donation, a formula 0.7 × [eGFR at 1-year post-donation] was used and is referred to as the “conventional DKF model. Verification of accuracy of developed models To compare the accuracy among the optimized DKF model, the sparse DKF model, and the conventional DKF model, metrics such as R 2 , root mean squared error (RMSE), mean absolute error (MAE), and Pearson’s correlation coefficient between each predicted value and measured eGFR were utilized. Additionally, outliers in the optimized DKF model were identified using Z-scores, which represent the distance of a data point from the mean in terms of standard deviations. The standard cutoff values for defining outliers were Z-scores of ± 2 or more extreme. Based on the Z-scores, the validation cohort was divided into three groups: inliers (-2 < Z 2), and low outliers (Z < -2). This classification was used to investigate preoperative factors that are likely to contribute to outliers. Declarations Authorship RO, TH, and KI conceived the idea of the study. RO, TH, KI and YK developed and conducted the statistical analysis and machine learning. TB, KU, KO, TS, JH, TT, and HI contributed to the interpretation of the results. RO drafted the original manuscript and TH revised it. TH, KI and YK supervised the conduct of this study. All authors reviewed the manuscript draft and revised it critically on intellectual content. All authors approved the final version of the manuscript to be published. Data Availability Statement The data which support the findings of this study are available from the corresponding author, [T.H.], upon reasonable request. The sparse DKF model was published in GitHub (https://github.com/thirai-0813/Donor_Cr_Calculator_EN.git). Funding This work was supported by Funds for the Development of Human Resources in Science and Technology, Initiative for Realizing Diversity in the Research Environment, Tokyo Women’s Medical University. Disclosure The authors declare that they have no competing financial or other interest or personal relationship that could have influenced this paper or the study it describes. References Ibrahim, H. N. et al. Long-term consequences of kidney donation. N Engl. J. Med. Jan . 29 (5), 459–469. 10.1056/NEJMoa0804883 (2009). Muzaale, A. D. et al. Risk of end-stage renal disease following live kidney donation. Jama . Feb 12. ;311(6):579 – 86. (2014). 10.1001/jama.2013.285141 Textor, S. C. Medically Complex Living Kidney Donors: Where Are We Now? Kidney Int. Rep. 1 , 4–6 (2020). Mueller, T. F. & Luyckx, V. A. The natural history of residual renal function in transplant donors. J. Am. Soc. Nephrol. Sep. 23 (9), 1462–1466. 10.1681/asn.2011111080 (2012). Grams, M. E. et al. Kidney-Failure Risk Projection for the Living Kidney-Donor Candidate. N Engl. J. Med. Feb . 4 (5), 411–421. 10.1056/NEJMoa1510491 (2016). Augustine, J. J., Arrigain, S., Mandelbrot, D. A., Schold, J. D. & Poggio, E. D. Factors Associated With Residual Kidney Function and Proteinuria After Living Kidney Donation in the United States. Transplantation Feb . 1 (2), 372–381. 10.1097/tp.0000000000003210 (2021). Locke, J. E. et al. Obesity increases the risk of end-stage renal disease among living kidney donors. Kidney Int. Mar. 91 (3), 699–703. 10.1016/j.kint.2016.10.014 (2017). Lentine, K. L. & Patel, A. Risks and outcomes of living donation. Adv. Chronic Kidney Dis. Jul . 19 (4), 220–228. 10.1053/j.ackd.2011.09.005 (2012). Shinoda, K. et al. Pre-donation BMI and preserved kidney volume can predict the cohort with unfavorable renal functional compensation at 1-year after kidney donation. BMC Nephrol. Feb . 8 (1), 46. 10.1186/s12882-019-1242-0 (2019). Bzdok, D., Altman, N. & Krzywinski, M. Statistics versus machine learning. Nat. Methods Apr . 15 (4), 233–234. 10.1038/nmeth.4642 (2018). Schmidt, M. & Lipson, H. Distilling free-form natural laws from experimental data. Science . Apr 3. ;324(5923):81 – 5. (2009). 10.1126/science.1165893 Ushio, Y. et al. Machine learning for morbid glomerular hypertrophy. Sci. Rep. Nov . 9 (1), 19155. 10.1038/s41598-022-23882-7 (2022). Madero, M. et al. Comparison between Different Measures of Body Fat with Kidney Function Decline and Incident CKD. Clin. J. Am. Soc. Nephrol. Jun . 7 (6), 893–903. 10.2215/cjn.07010716 (2017). Tagliafico, A. S., Bignotti, B., Torri, L. & Rossi, F. Sarcopenia: how to measure, when and why. Radiol. Med. 127 (3), 228–237. 10.1007/s11547-022-01450-3 (Mar 2022). Wilkinson, T. J. et al. Association of sarcopenia with mortality and end-stage renal disease in those with chronic kidney disease: a UK Biobank study. J. Cachexia Sarcopenia Muscle Jun . 12 (3), 586–598. 10.1002/jcsm.12705 (2021). Patel, S. S. et al. Serum creatinine as a marker of muscle mass in chronic kidney disease: results of a cross-sectional study and review of literature. J. Cachexia Sarcopenia Muscle Mar. 4 (1), 19–29. 10.1007/s13539-012-0079-1 (2013). Young, J. A. et al. Association of visceral and subcutaneous adiposity with kidney function. Clin. J. Am. Soc. Nephrol. Nov . 3 (6), 1786–1791. 10.2215/cjn.02490508 (2008). Lenihan, C. R. et al. Longitudinal study of living kidney donor glomerular dynamics after nephrectomy. J. Clin. Invest. Mar. 2 (3), 1311–1318. 10.1172/jci78885 (2015). Lam, N. N. et al. Changes in kidney function follow living donor nephrectomy. Kidney Int. Jul . 98 (1), 176–186. 10.1016/j.kint.2020.03.034 (2020). Delanaye, P. et al. Outcome of the living kidney donor. Nephrol. Dial Transpl. Jan . 27 (1), 41–50. 10.1093/ndt/gfr669 (2012). Lentine, K. L. et al. KDIGO Clinical Practice Guideline on the Evaluation and Care of Living Kidney Donors. Transplantation Aug . 101 (8S Suppl 1), S1–s109. 10.1097/tp.0000000000001769 (2017). Andrews, P. A., Burnapp, L. & British Transplantation Society / Renal Association UK Guidelines for Living Donor Kidney Transplantation 2018. Summary of Updated Guidance. Transplantation Jul . 102 (7), e307. 10.1097/tp.0000000000002253 (2018). ERBP Guideline on the Management and Evaluation of the Kidney Donor and Recipient. Nephrol. Dial Transpl. Aug ; 28 Suppl 2:ii1–71. doi: 10.1093/ndt/gft218 (2013). Richardson, R. et al. Kidney Paired Donation Protocol for Participating Donors 2014. Transplantation Oct. 99 (10 Suppl 1), S1–s88. 10.1097/tp.0000000000000918 (2015). Shen, W. et al. Total body skeletal muscle and adipose tissue volumes: estimation from a single abdominal cross-sectional image. J Appl Physiol () . Dec 2004;97(6):2333-8. (1985). 10.1152/japplphysiol.00744.2004 Srikumar, T. et al. Semiautomated Measure of Abdominal Adiposity Using Computed Tomography Scan Analysis. J. Surg. Res. May . 237 , 12–21. 10.1016/j.jss.2018.11.027 (2019). Matsuo, S. et al. Revised equations for estimated GFR from serum creatinine in Japan. Am. J. Kidney Dis. Jun . 53 (6), 982–992. 10.1053/j.ajkd.2008.12.034 (2009). Additional Declarations No competing interests reported. Supplementary Files SupplementaryTableFiguredocumentation0119.pdf Cite Share Download PDF Status: Published Journal Publication published 01 Jul, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 04 Mar, 2025 Reviews received at journal 03 Mar, 2025 Reviews received at journal 22 Feb, 2025 Reviewers agreed at journal 19 Feb, 2025 Reviewers agreed at journal 01 Feb, 2025 Reviewers invited by journal 30 Jan, 2025 Editor assigned by journal 30 Jan, 2025 Editor invited by journal 22 Jan, 2025 Submission checks completed at journal 21 Jan, 2025 First submitted to journal 19 Jan, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5862042","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":405699688,"identity":"30fd63bb-5364-4f76-b235-de48b9500f1a","order_by":0,"name":"Rikako Oki","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Rikako","middleName":"","lastName":"Oki","suffix":""},{"id":405699689,"identity":"752c7879-03ad-4f91-9881-70b8f5444e63","order_by":1,"name":"Toshihio Hirai","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAklEQVRIiWNgGAWjYBACA2YgkQBiSTC2gRhycCnGBrxaDKBaEhiMeQhqgVMSDGwgzYk9hBxmzs778MMDhj9y8rOb2x48/GGTvl8iO02CocaOgXk2dmssm9mNJYAOMza4c7DdICEhLbdHInebBMOxZAbGOQewO+wwGwNIS+IGicQ2iYSEw7k90iAtbAcYGGck4NLC/AOkZf4MiJZ0HrCWf3i1sIFtabgB0ZIA1sLYhl+LRYKBsbEBWEtammHP/bebLRL7knlw+uX8MeabPyrk5ORnpD+T/GFjI8/ec3bjjQ/f7OQMcYQYVCO6ANBJPIYz8OjADuQlSNYyCkbBKBgFwxMAAGZQV/DT65xYAAAAAElFTkSuQmCC","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":true,"prefix":"","firstName":"Toshihio","middleName":"","lastName":"Hirai","suffix":""},{"id":405699690,"identity":"eeb52dbc-6d76-496f-b81c-feabe92271d7","order_by":2,"name":"Kazuhiro Iwadoh","email":"","orcid":"","institution":"Mita Hospital","correspondingAuthor":false,"prefix":"","firstName":"Kazuhiro","middleName":"","lastName":"Iwadoh","suffix":""},{"id":405699691,"identity":"ab45deef-ee8a-40d1-b472-9e985d5d2d23","order_by":3,"name":"Yu Kijima","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Kijima","suffix":""},{"id":405699692,"identity":"67dc50ad-c0bc-40fc-94f0-d2f6f2f0c79c","order_by":4,"name":"Hiroyuki Hashimoto","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hiroyuki","middleName":"","lastName":"Hashimoto","suffix":""},{"id":405699693,"identity":"938c9b67-a267-4dde-9735-0dd9adb0a4cb","order_by":5,"name":"Yasunori Nishimura","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yasunori","middleName":"","lastName":"Nishimura","suffix":""},{"id":405699694,"identity":"0a156052-f974-4737-a06d-f9d8cb007fa6","order_by":6,"name":"Taro Banno","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Taro","middleName":"","lastName":"Banno","suffix":""},{"id":405699695,"identity":"a396738b-0288-4bd4-983d-afbb5e16cfa2","order_by":7,"name":"Kohei Unagami","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kohei","middleName":"","lastName":"Unagami","suffix":""},{"id":405699696,"identity":"a0d8e200-d5bf-427e-945b-9d53d5369c56","order_by":8,"name":"Kazuya Omoto","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kazuya","middleName":"","lastName":"Omoto","suffix":""},{"id":405699697,"identity":"231ba0d0-3974-4259-86fe-cfd4f1c5fbdc","order_by":9,"name":"Tomokazu Shimizu","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Tomokazu","middleName":"","lastName":"Shimizu","suffix":""},{"id":405699698,"identity":"8a9d9d56-2c50-4530-9dd6-4deeb235c2af","order_by":10,"name":"Junichi Hoshino","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junichi","middleName":"","lastName":"Hoshino","suffix":""},{"id":405699699,"identity":"f1faab82-26d8-438f-ab63-e7fba0aa9b8f","order_by":11,"name":"Toshio Takagi","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Toshio","middleName":"","lastName":"Takagi","suffix":""},{"id":405699700,"identity":"21a5249c-58d8-4c9e-aa1d-11a7eef41f7f","order_by":12,"name":"Hideki Ishida","email":"","orcid":"","institution":"Tokyo Women's Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hideki","middleName":"","lastName":"Ishida","suffix":""}],"badges":[],"createdAt":"2025-01-20 02:53:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5862042/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5862042/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-02879-y","type":"published","date":"2025-07-01T15:57:34+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":74585915,"identity":"100543c2-4322-435a-ab13-4b990704df38","added_by":"auto","created_at":"2025-01-23 16:43:12","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":28670,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDistribution of functions generated using symbolic regression via genetic programming within the function space\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figure illustrates the distribution of complexities and errors of generated models in a function space. The horizontal axis represents complexity, while the vertical axis indicates error. Each dot corresponds to a single generated model. Red dots represent models on the Pareto Front, and the green dot marks the model positioned at the uppermost and rightmost edge of a rectangle used to select appropriate models that strike a balance between underfitting and overfitting.\u003c/p\u003e","description":"","filename":"Figure12.png","url":"https://assets-eu.researchsquare.com/files/rs-5862042/v1/7f0edabb1af98349e2e7dc93.png"},{"id":74585916,"identity":"a294f4a2-5fab-4878-8d4f-8d614fb08873","added_by":"auto","created_at":"2025-01-23 16:43:12","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":23743,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFrequently of variables utilized in the selected models.\u003cbr\u003e\n \u003c/strong\u003eThe frequency of variables used in the selected models is shown in descending order. The horizontal axis represents the frequency of appearance.\u003c/p\u003e","description":"","filename":"Figure22.png","url":"https://assets-eu.researchsquare.com/files/rs-5862042/v1/54dc5344b0bbd7eb4194f4c9.png"},{"id":74585917,"identity":"6514b065-6b59-405d-9e5e-23778b7daefc","added_by":"auto","created_at":"2025-01-23 16:43:12","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":43815,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelationship between calculated and measured creatinine values in (A) training cohort and (B) validation cohort.\u003c/strong\u003e\u003cbr\u003e\nRed dots represent the calculated creatinine values obtained from the optimized DKF model. Blue dots represent the measured creatinine values at 1 year post-donation. Arrows indicate the discrepancies between the calculated and measured values.\u003c/p\u003e","description":"","filename":"Figure33.png","url":"https://assets-eu.researchsquare.com/files/rs-5862042/v1/21d5a0f6287f88e55a980398.png"},{"id":74585919,"identity":"f294e2eb-a3b8-47bd-8427-7bb8e07e81c7","added_by":"auto","created_at":"2025-01-23 16:43:12","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":149240,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe correlation graphs with Z-score between observed eGFR and estimated eGFR\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe plots show the observed eGFR at 1 year post-donation (Y-axis) versus the estimated eGFR (X-axis) using the optimized DKF model (A) and the conventional DKF model (B). The color scale represents the Z-score. The green dashed line indicates the 75th percentile, while the orange dashed line indicates the 25th percentile.\u003c/p\u003e","description":"","filename":"Figure42.png","url":"https://assets-eu.researchsquare.com/files/rs-5862042/v1/51e9e7ddcf39b10966f054be.png"},{"id":86180098,"identity":"75583566-0c03-4ad3-acb7-0e832ef968ee","added_by":"auto","created_at":"2025-07-07 16:21:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1371506,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5862042/v1/bd54c9ce-4a0f-430e-aefb-3dcb430f01fe.pdf"},{"id":74586722,"identity":"16951edf-c9ff-4008-903a-df4c2b983483","added_by":"auto","created_at":"2025-01-23 16:51:12","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":501214,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTableFiguredocumentation0119.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5862042/v1/c604b8970247f6690795e5fb.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Personalized Prediction Model Generated with Machine Learning for Kidney Function One Year After Living Kidney Donation","fulltext":[{"header":"Introduction","content":"\u003cp\u003eCareful screening of living donor candidates is critical to minimizing the risk of end-stage kidney disease (ESKD) and ensuring ongoing monitoring of renal function post-donation. Although rigorously screened living kidney donors were traditionally thought to have comparable risks of mortality and ESKD to the general population\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e, they may face a higher risk of ESKD compared to matched healthy non-donors.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e Additionally, selection criteria for living donors have broadened to include medically complex individuals, such as those with advanced age, hypertension, obesity, or lower estimated glomerular filtration rate (eGFR).\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e These evolving trends underscore the urgent need for more precise and individualized prediction models to assess and monitor postoperative renal function in this population.\u003c/p\u003e \u003cp\u003ePost-donation eGFRs are typically reach approximately 60\u0026ndash;70% of the pre-donation levels, attributed to compensatory hypertrophy of the remaining kidney.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e Grams et al. developed an online risk tool to estimate the long-term risk of ESKD for living kidney donor candidates, using meta-analyzed risk associations from seven general population cohorts.\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e Several reports have proposed predictive factors for post-donation kidney function that included donor age, sex, race, body mass index (BMI) and preoperative computed tomography (CT) volumetry of kidney. \u003csup\u003e\u003cspan additionalcitationids=\"CR7 CR8\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e These were based on observational study and traditional statistics which aims to identify significant risk factors among explanatory variables through model-driven regression using linear models.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eTraditional regression models that fit data to pre-defined models are directly meaningful to clinicians, though it relies on liner predefined models and are constrained by strict assumptions. In contrast, symbolic regression (SR) via genetic programming (GP) is a machine learning (ML) algorithm where the goal is discovering an explicit mathematical formula that best describes a given dataset from the vast function space, enabling it to uncover complex, non-linear interactions directly from the data.\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e Genetic programming (GP) explores both the structure of the model and its parameters. This evolutionary methodology is based on the principles of mutation and natural selection, mirroring the way organisms evolve and adapt to their environments. In SR via GP, innumerable mathematical formulas evolve to fit the given dataset, progressively yielding better ones that can predict the target value from the explanatory variables. In the previous study, Ueno et al. utilized SR via GP to evaluate pathological factors associated with the development of glomerular hypertrophy (GH).\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e From a set of 60 variables, the SR model identified key factors such as inflammation, vascular damage, and obesity as significant predictors, while eGFR was ranked low (46th out of 60). This finding highlights the ability of SR to distinguish morbid GH from adaptive hypertrophy due to nephron loss. Collectively, these results suggest that SR can evaluate variables in an unbiased manner, providing valuable insights into the underlying mechanisms.\u003c/p\u003e \u003cp\u003eThe objective of this study was to develop a machine learning model for more accurate and personalized prediction of post-donation kidney function. We employed an SR model to predict post-donation creatinine (Cre) values using preoperative variables, including CT volumetry data for excised and non-excised kidney volumes. Given that excess visceral fat and reduced skeletal muscle mass\u0026mdash;characteristic of sarcopenia\u0026mdash;are recognized risk factors for the development of chronic kidney disease (CKD)\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e, we also implemented CT volumetry for visceral fat and skeletal muscle. The resulting model was integrated into a user-friendly interface to facilitate clinical application and improve decision-making in transplant practice.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatient\u0026rsquo;s background\u003c/h2\u003e \u003cp\u003eThe clinical characteristics and laboratory data of the patients in the study cohorts are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. All the participants were Asian. The mean donor age was 59.9 years, and 35% were male. The median baseline Cre level before donation was 0.7mg/dl. There were no significant differences in the variables between the training and the validation cohort.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison between training cohort and validation cohort.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eall\u003c/p\u003e \u003cp\u003e(\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;204)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTraining cohort\u003c/p\u003e \u003cp\u003e(\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;143)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eValidation cohort\u003c/p\u003e \u003cp\u003e(\u003cem\u003eN\u003c/em\u003e\u0026thinsp;=\u0026thinsp;61)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge at donation (y.o)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e59.9\u0026thinsp;\u0026plusmn;\u0026thinsp;8.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e60.3\u0026thinsp;\u0026plusmn;\u0026thinsp;8.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e59.0\u0026thinsp;\u0026plusmn;\u0026thinsp;8.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.363\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale [n (%)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e71 (34.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e53 (37.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18 (29.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.338\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody weight (kg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e56.0 (51.0, 63.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e57.0 (51.0, 64.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e53.0 (51.0, 61.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.098\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHight (cm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e159.0(154.5, 165.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e160 (155, 165)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e159 (154, 164)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.698\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e22.4 (20.4, 24.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.6 (20.6,24.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e21.9 (20.2, 23.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.053\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSystolic blood pressure (mmHg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e129.9\u0026thinsp;\u0026plusmn;\u0026thinsp;19.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e130.7\u0026thinsp;\u0026plusmn;\u0026thinsp;19.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e127.9\u0026thinsp;\u0026plusmn;\u0026thinsp;20.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.357\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiastolic blood pressure (mmHg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e76.7\u0026thinsp;\u0026plusmn;\u0026thinsp;14.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77.4\u0026thinsp;\u0026plusmn;\u0026thinsp;13.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e75.1\u0026thinsp;\u0026plusmn;\u0026thinsp;17.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.284\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003esmoking [n (%)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e58 (28.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42 (29.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e16 (26.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.736\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHistory of CVD [n (%)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (2.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (2.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (1.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAntihypertensive agents [n (%)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40 (19.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27 (18.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13 (21.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.703\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLipid-lowering agents [n (%)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21 (10.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12 (8.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e9 (14.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.209\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003euric acid lowering agents [n (%)]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4 (2.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4 (2.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.319\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWBC (/\u0026micro;l)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4955 (4265, 6055)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4930 (4270, 5995)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5190 (4240, 6180)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.490\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHb (g/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.7\u0026thinsp;\u0026plusmn;\u0026thinsp;1.24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.8\u0026thinsp;\u0026plusmn;\u0026thinsp;1.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.6\u0026thinsp;\u0026plusmn;\u0026thinsp;1.18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.468\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatelet (*10\u003csup\u003e3\u003c/sup\u003e/\u0026micro;l)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21.7 (18.2, 25.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22.0 (18.3, 25.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e21.0 (18.2, 26.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.846\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCre (mg/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.70 (0.62, 0.81)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.69 (0.62, 0.84)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.70 (0.62, 0.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eeGFR (ml/min/1.73m\u003csup\u003e2\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e71.9 (65.1, 80.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e71.4 (65.0, 80.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e72.7 (65.1, 80.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.650\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBUN (mg/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.8 (11.2, 15.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e13.8 (11.2, 15.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.8 (11.7, 15.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.646\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNa (mEq/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e142 (141,143)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e142 (141, 143)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e141 (141, 143)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.417\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eK (mEq/l)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.2 (4.0, 4.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.2 (4.0, 4.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.2 (4.0, 4.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.917\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAST [IU/l]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.0 (17.0, 24.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20.0 (17.0, 24.5)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e20.0 (17.0, 24.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.849\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eALT [IU/l]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e17.0 (13.0, 22.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17.0 (13.0, 22.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17.0 (14.0, 23.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.583\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTP (g/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7.24\u0026thinsp;\u0026plusmn;\u0026thinsp;0.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7.24\u0026thinsp;\u0026plusmn;\u0026thinsp;0.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.24\u0026thinsp;\u0026plusmn;\u0026thinsp;0.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.989\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCRP (mg/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.05 (0.04, 0.10)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.05 (0.04, 0.11)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.05 (0.04, 0.10)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.460\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT.chol (mg/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e211 (187, 233)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e213 (186, 233)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e205 (190, 233)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.521\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUA (mg/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.80 (4.10, 5.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.90 (4.10, 5.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4.70 (4.10, 5.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.524\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGlucose (g/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e98.0 (92.0, 104)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e98.0 (92.0, 104)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e98.0 (92.0, 107)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.905\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHbA1C (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.70 (5.50, 5.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.70 (5.50, 5.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.70 (5.50, 5.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.772\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eProteinuria (+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (1.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (0.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (1.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.510\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrinary occult blood (\u0026ge;+)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (6.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (7.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e3 (4.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.561\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCre at 1-year post-donation (mg/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.06 (0.93, 1.23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.06 (0.93, 1.26)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.06 (0.91, 1.18)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.378\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eResults of CT findings\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVolume of excised kidney (ml)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e143 (126, 161)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e148 (129, 163)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e140 (118, 157)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.133\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVolume of non-excised kidney(ml)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e137 (120, 156)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e138 (121, 159)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e133 (114, 152)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.165\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea psoas muscle at L3(cm\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11.3 (9.01,15.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11.5 (9.13, 14.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.7 (8.92, 15.2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.501\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea skeletal muscle at L3(cm\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e98.5 (85.8, 126)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e100.5 (87.0, 126)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e92.1 (82.5, 120)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.139\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea visceral fat(cm\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e86.4 (56.0, 125)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e82.1 (58.2, 126)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e90.2 (50.9, 115.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.528\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea visceral Fat at L3 (cm\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77.2 (38.9, 127)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e77.2 (46.1, 135)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81.1 (38.2, 118)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.516\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eContinuous data are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD or median (IQR): CVD, cardiovascular disease; WBC, white blood cell; Hb, hemoglobin; Cre, creatinine; eGFR, estimated glomerular filtration rate; BUN, blood urea nitrogen; Na, sodium; K, potassium; AST, aspartate aminotransferase; ALT, alanine aminotransferase; TP, total protein; CRP, C-reactive protein; T.chol, total cholesterol; UA, uric acid; HbA1C, hemoglobinA1C\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCT volumetry data\u003c/h3\u003e\n\u003cp\u003eThere were no significant differences in CT volumetry data between training cohort and validation cohort (Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The correlation between CT volumetry data and preoperative Cre level was investigated using Pearson\u0026rsquo;s correlation coefficient (r). The analysis revealed moderate positive correlations between area of psoas muscle/skeletal muscle/visceral fat and pre-operative Cre level (Supplementary table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Significant positive correlations were observed between CT volumetry data and pre-operative body weight (Supplementary table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e\n\u003ch3\u003eCorrelation between explanatory variables and post-donation Cre level\u003c/h3\u003e\n\u003cp\u003eAll correlation coefficients between preoperative data and Cre level at 1 year post-donation are presented in Supplemental Fig.\u0026nbsp;1. Additionally, the point-biserial correlation coefficients (r values) for the top 10 variables with the strongest positive correlations are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. As with the pre-operative Cre level, the Cre level at 1 year post-donation demonstrated a positive correlation with preoperative skeletal muscle volume (r\u0026thinsp;=\u0026thinsp;0.62, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), along with weak positive correlations with the psoas muscle and visceral fat volumes.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCorrelation between pre-operative variables and creatinine level at 1 year post-donation.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003er (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCre\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.84 (0.80\u0026ndash;0.88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.65 (0.56\u0026ndash;0.72)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea skeletal muscle at L3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.62 (0.53\u0026ndash;0.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBody weight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.57 (0.47\u0026ndash;0.65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.51(0.40\u0026ndash;0.60)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeight\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.47 (0.35\u0026ndash;0.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea visceral Fat at L3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.46 (0.34\u0026ndash;0.56)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHb\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.40 (0.27\u0026ndash;0.51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea visceral fat\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.37 (0.24\u0026ndash;0.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArea psoas muscle at L3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.25 (0.12\u0026ndash;0.37)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eCre, creatinine; UA, uric acid; Hb, hemoglobin\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eAnalysis of predictive variables for Cre level post-donation using ML\u003c/h3\u003e\n\u003cp\u003eBy using 34 explanatory variables, Data Modeler automatically generated 1123 models that calculate Cre level at 1-year post-donation. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e show their distribution in the function space. We selected 98 models with lower complexities and lower 1-R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e in each epoch to limit the number of models to 9% of all generated models. We ranked the frequencies of all explanatory variables that were used in the selected models (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The selected factors for generating models included age, male, BW, the history of CVD, BUN, Cre, HbA1C, and the volume of the non-excised kidney.\u003c/p\u003e\n\u003ch3\u003eCreating a predictive model for Cre level at 1-year\u003c/h3\u003e\n \u003cp\u003eTo establish an optimized model for predicting Cre levels 1 year post-donation, ensemble learning was performed using the bagging method, which calculates the trimmed means of the selected models. The formula for the optimized DKF model, incorporating key variables extracted from the developed models, is provided in Supplemental Document 1. The R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, RMSE, and MAE values between the predicted Cre levels generated by the optimized DKF model and the measured values were 0.72, 0.1, 0.073mg/dl mg/dL in the training cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA) and 0.69, 0.1, and 0.079 mg/dl in the validation cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), respectively.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eThe impact of each variable in the generated formula\u003c/h2\u003e \u003cp\u003eWe investigated factors that were used in the generated formula to determine which factors had higher impact on the target. A simulation was performed by changing each variable within its range while keeping other variables fixed at their median values. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e lists the 8 most significant driver variables that influenced the result. Pre-Cre level and the volume of the non-excised kidney were the most influential factors for Cre levels at 1 year, with high ΔTarget values. Supplementary Fig.\u0026nbsp;2 illustrates the extent to which the target Cre value changes as the top 3 driver variables are varied.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe changes in the Cre level as each explanatory variable changes from its minimum to its maximum with other variables fixed at their median (ΔTarget).\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eΔTarget\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.685\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003epre Cre level\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.450\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003evolume of non-excised kidney\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.281\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ebody weight\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.102\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.062\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBUN\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.047\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePre HbA1C\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.047\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHistory of CVD\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAge at donation\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eCre, creatinine; BUN, blood urea nitrogen; HbA1C, hemoglobinA1C; CVD, cardiovascular disease\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eThe development of a sparse model\u003c/h3\u003e\n\u003cp\u003eTo enhance clinical applicability, we developed a simplified prediction formula (referred to as the sparse DKF model), where variables with smaller ΔTarget were fixed at their median values. Specifically, BUN and HbA1C were fixed at their median values, while the history of CVD was set to its mode, which was 0. The key driving factors for this simplified model included age, sex (male), body weight (BW), Pre-Cre, and the volume of the non-excised kidney. The developed model has been uploaded to GitHub and implemented as a web application for convenient calculations, accessible at the following address: [\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://donorcrcalculatoren-89wgze554kd9n2yjkcyubb.streamlit.app/\u003c/span\u003e\u003cspan address=\"https://donorcrcalculatoren-89wgze554kd9n2yjkcyubb.streamlit.app/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e].\u003c/p\u003e\n\u003ch3\u003eVerification of accuracy of developed models\u003c/h3\u003e\n\u003cp\u003eTo assess the accuracy of the developed models, we calculated the R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, RMSE, and MAE for the values predicted by the optimized DKF model, the sparse DKF model, and the conventional DKF model, comparing them to the measured eGFR values in the validation cohort (N\u0026thinsp;=\u0026thinsp;61). (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e) The optimized DKF model achieved the highest R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e and the lowest RMSE and MAE, indicating minimal prediction error and high predictive accuracy. The sparse DKF model, designed with clinical practicality in mind, also demonstrated significantly better predictive accuracy compared to the conventional DKF model. Consequently, both the optimized DKF model and the sparse DKF model showed superior data fit and can be considered more reliable than the conventional DKF model.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eR-squared, RMSE and MAE in comparison of the predicted values by the optimized DKF model/the sparse DKF model/ the conventional DKF model and the observed eGFR values.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRoot Mean Squared Error\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMean Absolute Error\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe optimized DKF model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.92\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe sparse DKF model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.17\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eThe conventional DKF model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5.80\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eDKF, donor\u0026rsquo;s kidney function\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eNext, we analyzed the correlation between the predicted values from the optimized DKF model and the measured eGFR values (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA), as well as the predicted values from the conventional DKF model and the measured eGFR values (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). Three cases with |Z| \u0026ge; 2 were identified, and these cases completely overlapped between the optimized DKF model and the conventional DKF model. Supplementary Table\u0026nbsp;2 compares the patient characteristics among inliers (-2\u0026thinsp;\u0026lt;\u0026thinsp;Z\u0026thinsp;\u0026lt;\u0026thinsp;2), high outliers (Z\u0026thinsp;\u0026gt;\u0026thinsp;2), and low outliers (Z \u0026lt; -2) in the prediction of Cre at 1 year post-donation. Although the small sample size limits statistical interpretation, cases with higher preoperative Cre levels appeared more likely to be classified as outliers in both the optimized DKF model and the conventional DKF model.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn the present study, we generated a predictive model for Cre level at 1-year post-donation, using the SR via GP technique, which is one of the ML techniques. The developed model, referred to \u0026ldquo;the optimized DKF model\u0026rdquo; by ML was found to have higher accuracy, demonstrating higher R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e values and lower RMSE and MAE, compared to the conventional DKF model which assumed eGFR post-donation was 70% of the preoperative value. Preoperative Cre level and the volume of the non-excised kidney were identified as the most influential predictive factors for Cre levels at 1 year post-donation. Predicting the kidney function post-donation will facilitate more rigorous management and help achieve optimal post-transplant outcomes.\u003c/p\u003e \u003cp\u003eStatistical analysis revealed that there was a moderate correlation between the skeletal muscle volume and Cre levels at 1 year, as well as between the visceral fat volume and Cre levels at 1 year. The volume of non-excised kidney was not corelated to Cre level post-donation. Approximately 95% of the human body\u0026rsquo;s total creatine is located in skeletal muscle and serum Cre can be served as a surrogate marker of skeletal muscle mass even in CKD patients.\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e The previous study revealed that visceral adipose tissue detected by CT scan was associated with CKD when defined using cystatin C estimating equations but not when using a Cre-based estimating equation.\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e On the other hand, area of skeletal muscle and area of visceral fat were not identified as influential factors in the ML analysis. Volume of non-excised kidney was found to be a factor affecting Cre level at 1-year. The larger the volume of the non-excised kidney, the lower the Cre levels at 1 year tended to be. (Supplementary Fig.\u0026nbsp;2A) Shimada et al demonstrated that body surface area-adjusted preserved kidney volume calculated by the 3D reconstructed image was an independent risk factor for \u0026gt;\u0026thinsp;30% reduction of eGFR at 1-year post-donation (odds ratio, 0.93)\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, which aligned with our findings. It is well known that adaptive hyperfiltration after donor nephrectomy is attributable to hyper perfusion and hypertrophy of the remaining glomeruli.\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e The increases in single-kidney renal plasma flow, cortical volume, and GFR continue through to the late post-donation period.\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e From the perspective of donor outcomes, it is suggested to determine which side of kidney to be excised based on the volume of the remaining kidney.\u003c/p\u003e \u003cp\u003eAs previously mentioned, the post-donation eGFRs are approximately 60\u0026ndash;70% of the pre-donation values based on past reports.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e To assess the generated model\u0026rsquo;s goodness-of-fit, we compared its R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, RMSE, and MAE values to those of the conventional prediction method. The optimized DKF model showed higher explanatory power and accuracy for predicting Cre level at 1 year compared to the conventional method. The eGFR was reported to increase by +\u0026thinsp;0.35 ml/min/ 1.73 m\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e per year from 6 weeks post-donation onward, with this increase leveling off by the fifth year.\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e It is reported that increase in GFR can begin as early as 8 hours post-donation, however, this acute compensation is less efficient in older donors.\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e Thus, predicting Cre levels at 1-year by ML may be more beneficial than the conventional prediction method in clinical practice. However, outliers also exist in the optimized DKF model. Since the cases with pre-operative Cre values deviating from the median were likely to become outliers, the prediction may not be applied in the cases with extreme Cre level before the donation.\u003c/p\u003e \u003cp\u003eThis model was developed by analyzing the characteristics of donors who were ultimately selected for kidney donation, excluding those who did not meet the criteria for a living kidney donor. Therefore, it is essential to note that even if a good Cre level was expected by the generated model, it does not guarantee the safety of prognosis for cases which don\u0026rsquo;t meet the criteria for a living kidney donor. It is important to adhere to the guidelines for evaluating eligibility for a living kidney donor. How to approach to donor selection has been discussed nowadays. While the Kidney Disease Improving Global Outcomes Guideline recommends the use of fixed cutoffs\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, age and sex- based GFR cutoffs are commonly used in the British Transplantation Society, the European Renal Best Practice, and the Canadian Society of Transplantation, considering kidney function decline with healthy aging.\u003csup\u003e\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e It is challenging to judge whether older candidates' kidney function is appropriate for their age because they tend to be more medically complex. In such cases, which include so-called marginal donors, the prediction of kidney function post-donation by the generated model might be useful in planning follow-up. In cases where a donor is acceptable for kidney donation but is predicted to have poor post-donation kidney function, careful follow-up should be warranted after donation. Additionally, as non-excised kidney volume was an influential factor for predicting Cre level at 1-year, it is possible to simulate the impact on kidney function based on CT imaging data. It can be used as a reference to consider the impact on kidney function depending on which kidney will be donated. Overall, although the optimized DKF model/the sparse DKF model cannot be used directly for donor selection, it might be helpful as a reference for donor selection, for simulating kidney function post-donation, and for planning follow-up.\u003c/p\u003e \u003cp\u003eWe acknowledge there are limitations in our study. This model was generated by data set from a single institution with a small sample size. There might be other factors affecting kidney function after donation beyond our dataset. Since the validation utilized a validation cohort included within the dataset, external data is required for further validation. Larger studies with more factors are warranted to verify our findings. However, this study indicated the strong potential of ML system for identifying unknown risk factors related to post-donated kidney function.\u003c/p\u003e \u003cp\u003eIn conclusion, the ML model achieved an effective predictive performance for predicting Cre level post-donation. It is meaningful for transplant physicians to pay sufficient attention to the care for donors after donation. Applying ML techniques to the clinical field has the potential to lead to better healthcare for patients.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003ePatients and data collection\u003c/h2\u003e \u003cp\u003eWe retrospectively enrolled 204 patients who underwent enhanced CT prior to donor nephrectomy for living kidney transplantation at Tokyo Women\u0026rsquo;s Medical University Hospital between January 2012 and December 2016. Patients who were lost to follow-up after discharge or whose follow-up period after nephrectomy was less than one year were excluded. This study was approved by the Institutional Review Board of the Tokyo Women\u0026rsquo;s Medical University Hospital (#2021\u0026thinsp;\u0026minus;\u0026thinsp;0122). Since this is a retrospective study, informed consent was waived by the Institutional Review Board of the Tokyo Women\u0026rsquo;s Medical University Hospital. All methods of research procedures were performed in accordance with the Declaration of Helsinki.\u003c/p\u003e \u003cp\u003eBasic information about the patients was obtained from donor medical records within 3 months before donation, including age, sex, medications for hypertension, hyperlipidemia, and hyperuricemia, smoking habits, systolic/diastolic blood pressure, body mass index, and laboratory data. The laboratory data comprised white blood cell (WBC) count, hemoglobin (Hb), platelets, AST (aspartate aminotransferase), ALT (alanine aminotransferase), blood urea nitrogen (BUN), serum creatinine (Cre), estimated glomerular filtration rate (eGFR), sodium (Na), potassium (K), total protein (TP), C-reactive protein (CRP), uric acid (UA), hemoglobin A1c (HbA1c), total cholesterol (T. Chol), urinary protein, and urinary occult blood as qualitative tests. Blood pressure was measured with a brachial sphygmomanometer at the first day of hospitalization prior to donation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eCT volumetry\u003c/h2\u003e \u003cp\u003ePreoperative dynamic CT was performed using 64-,80- or 320-multidetector CT scanners (Aquilion 64, Aquilion Prime SP, Aquilion ONE; Canon Medical Systems Corporation, Otawara, Tochigi, Japan). Iodinated contrast media (Iohexol, Omnipaque 300; GE Healthcare Pharmaceutical Diagnostics, Tokyo, Japan) was injected using a power injector following an unenhanced CT scan. The nephrographic phase images were transferred and analyzed using dedicated software (SYNAPSE VINCENT;FUJIFILM Corporation Tokyo, Japan). The software semi-automatically generated a 3D model of the kidney by stacking 1-mm axial slices and calculated the kidney's volume (Supplementary Fig.\u0026nbsp;3A, B). Whole-body skeletal muscle mass was estimated from the trunk muscle area at the L3-4 level which is known to be highly correlated with the total body skeletal muscle volume (Supplementary Fig.\u0026nbsp;3C).\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e As for the visceral fat area at the naval level, a single axial slice at the L3-L4 intervertebral space was chosen for analysis, as it is frequently used as a surrogate for abdominal adiposity (Supplementary Fig.\u0026nbsp;3C).\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eEvaluation of renal function\u003c/h2\u003e \u003cp\u003eRenal function, measured by serum Cre levels, was evaluated pre-donation and at 1-year post-donation. The pre-donation serum Cre level was defined as the most recent result obtained within 3 months before donation. Serum Cre level at 1-year post-donation was employed as the primary outcome for this study. The eGFR was calculated using a formula specific to Japanese patients with CKD (eGFR [mL/min/1.73 m\u003csup\u003e2\u003c/sup\u003e]\u0026thinsp;=\u0026thinsp;194 x Serum creatinine(-1.094) \u0026times; Age(-0.287) \u0026times; 0.739 [if female]).\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eContinuous data were described as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation or median (interquartile range). Unpaired t-test or Mann\u0026ndash;Whitney U test was used to compare continuous variables. The chi-square test or Fisher\u0026rsquo;s exact test was used to compare the categorical variables. A Pearson\u0026rsquo;s correlation test was performed on correlation between clinical factors and Cre level at 1-year post donation. A Pearson\u0026rsquo;s correlation test was also performed to evaluate the correlation between the eGFR predicted by the generated/conventional model and the observed eGFR measured in practice. Values for which \u003cem\u003ep\u003c/em\u003e was less than 0.05 were inferred as significant.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eModel generation\u003c/h2\u003e \u003cp\u003eThe data were randomly divided into a training and a validation data in a 7:3 ratio. By using training data, we conducted SR via GP to construct machine learning models for predicting Cre levels at 1-year post-donation using Data Modeler version 9.3 (Evolved Analytics LLC, Rancho Santa Fe, CA, USA) which runs on Mathematica version 12.1 (Wolfram Research Incorporated, Champaign, IL, USA). The explanatory variables included 34 pre-operative variables (Supplementary Table\u0026nbsp;3) such as gender, age, height, body weight, blood pressure, smoking history, disease history, laboratory data, and CT volumetry data. DataModeler was executed with 8 independent evolutions and 6 minutes modeling time. During the evolution, SR via GP automatically discards less important variables and selects more important ones, effectively performing dimensionality reduction. Then it generates hundreds of predictive models, selects a few to a few dozen of simpler and less erroneous models, and uses their trimmed mean as the optimized predictive model, just like a bagging method used in random forest. The optimized model created through the process is referred to as \u0026ldquo;optimized donor\u0026rsquo;s kidney function (DKF) model\u0026rdquo; in this paper.\u003c/p\u003e \u003cp\u003eWe also explored the driver variables in the optimized DKF model and their impact on the overall model, illustrating their effects on the target using an explore plot. In this plot, each variable was varied within its range while keeping all other variables fixed at their median values. Finally, we developed a simplified prediction formula from the optimized DKF model, in which variables with smaller effects on the target were fixed at their median values. This model is referred to as the \u0026ldquo;sparse DKF model.\u0026rdquo; As a conventional method for estimating renal function post-donation, a formula 0.7 \u0026times; [eGFR at 1-year post-donation] was used and is referred to as the \u0026ldquo;conventional DKF model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eVerification of accuracy of developed models\u003c/h2\u003e \u003cp\u003eTo compare the accuracy among the optimized DKF model, the sparse DKF model, and the conventional DKF model, metrics such as R\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, root mean squared error (RMSE), mean absolute error (MAE), and Pearson\u0026rsquo;s correlation coefficient between each predicted value and measured eGFR were utilized.\u003c/p\u003e \u003cp\u003eAdditionally, outliers in the optimized DKF model were identified using Z-scores, which represent the distance of a data point from the mean in terms of standard deviations. The standard cutoff values for defining outliers were Z-scores of \u0026plusmn;\u0026thinsp;2 or more extreme. Based on the Z-scores, the validation cohort was divided into three groups: inliers (-2\u0026thinsp;\u0026lt;\u0026thinsp;Z\u0026thinsp;\u0026lt;\u0026thinsp;2), high outliers (Z\u0026thinsp;\u0026gt;\u0026thinsp;2), and low outliers (Z \u0026lt; -2). This classification was used to investigate preoperative factors that are likely to contribute to outliers.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthorship\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRO, TH, and KI conceived the idea of the study. RO, TH, KI and YK developed and conducted the statistical analysis and machine learning. TB, KU, KO, TS, JH, TT, and HI contributed to the interpretation of the results. RO drafted the original manuscript and TH revised it. TH, KI and YK supervised the conduct of this study. All authors reviewed the manuscript draft and revised it critically on intellectual content. All authors approved the final version of the manuscript to be published.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data which support the findings of this study are available from the corresponding author, [T.H.], upon reasonable request. The sparse DKF model was published in GitHub (https://github.com/thirai-0813/Donor_Cr_Calculator_EN.git).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by Funds for the Development of Human Resources in Science and Technology, Initiative for Realizing Diversity in the Research Environment, Tokyo Women\u0026rsquo;s Medical University.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing financial or other interest or personal relationship that could have influenced this paper or the study it describes.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eIbrahim, H. N. et al. Long-term consequences of kidney donation. \u003cem\u003eN Engl. J. Med. Jan\u003c/em\u003e. \u003cb\u003e29\u003c/b\u003e (5), 459\u0026ndash;469. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa0804883\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa0804883\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMuzaale, A. D. et al. Risk of end-stage renal disease following live kidney donation. \u003cem\u003eJama\u003c/em\u003e. Feb 12. ;311(6):579\u0026thinsp;\u0026ndash;\u0026thinsp;86. (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jama.2013.285141\u003c/span\u003e\u003cspan address=\"10.1001/jama.2013.285141\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTextor, S. C. Medically Complex Living Kidney Donors: Where Are We Now? \u003cem\u003eKidney Int. Rep.\u003c/em\u003e \u003cb\u003e1\u003c/b\u003e, 4\u0026ndash;6 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMueller, T. F. \u0026amp; Luyckx, V. A. The natural history of residual renal function in transplant donors. \u003cem\u003eJ. Am. Soc. Nephrol. Sep.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e (9), 1462\u0026ndash;1466. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1681/asn.2011111080\u003c/span\u003e\u003cspan address=\"10.1681/asn.2011111080\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrams, M. E. et al. Kidney-Failure Risk Projection for the Living Kidney-Donor Candidate. \u003cem\u003eN Engl. J. Med. Feb\u003c/em\u003e. \u003cb\u003e4\u003c/b\u003e (5), 411\u0026ndash;421. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJMoa1510491\u003c/span\u003e\u003cspan address=\"10.1056/NEJMoa1510491\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAugustine, J. J., Arrigain, S., Mandelbrot, D. A., Schold, J. D. \u0026amp; Poggio, E. D. Factors Associated With Residual Kidney Function and Proteinuria After Living Kidney Donation in the United States. \u003cem\u003eTransplantation Feb\u003c/em\u003e. \u003cb\u003e1\u003c/b\u003e (2), 372\u0026ndash;381. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/tp.0000000000003210\u003c/span\u003e\u003cspan address=\"10.1097/tp.0000000000003210\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLocke, J. E. et al. Obesity increases the risk of end-stage renal disease among living kidney donors. \u003cem\u003eKidney Int. Mar.\u003c/em\u003e \u003cb\u003e91\u003c/b\u003e (3), 699\u0026ndash;703. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.kint.2016.10.014\u003c/span\u003e\u003cspan address=\"10.1016/j.kint.2016.10.014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLentine, K. L. \u0026amp; Patel, A. Risks and outcomes of living donation. \u003cem\u003eAdv. Chronic Kidney Dis. Jul\u003c/em\u003e. \u003cb\u003e19\u003c/b\u003e (4), 220\u0026ndash;228. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.ackd.2011.09.005\u003c/span\u003e\u003cspan address=\"10.1053/j.ackd.2011.09.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShinoda, K. et al. Pre-donation BMI and preserved kidney volume can predict the cohort with unfavorable renal functional compensation at 1-year after kidney donation. \u003cem\u003eBMC Nephrol. Feb\u003c/em\u003e. \u003cb\u003e8\u003c/b\u003e (1), 46. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12882-019-1242-0\u003c/span\u003e\u003cspan address=\"10.1186/s12882-019-1242-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBzdok, D., Altman, N. \u0026amp; Krzywinski, M. Statistics versus machine learning. \u003cem\u003eNat. Methods Apr\u003c/em\u003e. \u003cb\u003e15\u003c/b\u003e (4), 233\u0026ndash;234. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nmeth.4642\u003c/span\u003e\u003cspan address=\"10.1038/nmeth.4642\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmidt, M. \u0026amp; Lipson, H. Distilling free-form natural laws from experimental data. \u003cem\u003eScience\u003c/em\u003e. Apr 3. ;324(5923):81\u0026thinsp;\u0026ndash;\u0026thinsp;5. (2009). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.1165893\u003c/span\u003e\u003cspan address=\"10.1126/science.1165893\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUshio, Y. et al. Machine learning for morbid glomerular hypertrophy. \u003cem\u003eSci. Rep. Nov\u003c/em\u003e. \u003cb\u003e9\u003c/b\u003e (1), 19155. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-022-23882-7\u003c/span\u003e\u003cspan address=\"10.1038/s41598-022-23882-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMadero, M. et al. Comparison between Different Measures of Body Fat with Kidney Function Decline and Incident CKD. \u003cem\u003eClin. J. Am. Soc. Nephrol. Jun\u003c/em\u003e. \u003cb\u003e7\u003c/b\u003e (6), 893\u0026ndash;903. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2215/cjn.07010716\u003c/span\u003e\u003cspan address=\"10.2215/cjn.07010716\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTagliafico, A. S., Bignotti, B., Torri, L. \u0026amp; Rossi, F. Sarcopenia: how to measure, when and why. \u003cem\u003eRadiol. Med.\u003c/em\u003e \u003cb\u003e127\u003c/b\u003e (3), 228\u0026ndash;237. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s11547-022-01450-3\u003c/span\u003e\u003cspan address=\"10.1007/s11547-022-01450-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (Mar 2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilkinson, T. J. et al. Association of sarcopenia with mortality and end-stage renal disease in those with chronic kidney disease: a UK Biobank study. \u003cem\u003eJ. Cachexia Sarcopenia Muscle Jun\u003c/em\u003e. \u003cb\u003e12\u003c/b\u003e (3), 586\u0026ndash;598. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jcsm.12705\u003c/span\u003e\u003cspan address=\"10.1002/jcsm.12705\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatel, S. S. et al. Serum creatinine as a marker of muscle mass in chronic kidney disease: results of a cross-sectional study and review of literature. \u003cem\u003eJ. Cachexia Sarcopenia Muscle Mar.\u003c/em\u003e \u003cb\u003e4\u003c/b\u003e (1), 19\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13539-012-0079-1\u003c/span\u003e\u003cspan address=\"10.1007/s13539-012-0079-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoung, J. A. et al. Association of visceral and subcutaneous adiposity with kidney function. \u003cem\u003eClin. J. Am. Soc. Nephrol. Nov\u003c/em\u003e. \u003cb\u003e3\u003c/b\u003e (6), 1786\u0026ndash;1791. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2215/cjn.02490508\u003c/span\u003e\u003cspan address=\"10.2215/cjn.02490508\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLenihan, C. R. et al. Longitudinal study of living kidney donor glomerular dynamics after nephrectomy. \u003cem\u003eJ. Clin. Invest. Mar.\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e (3), 1311\u0026ndash;1318. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1172/jci78885\u003c/span\u003e\u003cspan address=\"10.1172/jci78885\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLam, N. N. et al. Changes in kidney function follow living donor nephrectomy. \u003cem\u003eKidney Int. Jul\u003c/em\u003e. \u003cb\u003e98\u003c/b\u003e (1), 176\u0026ndash;186. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.kint.2020.03.034\u003c/span\u003e\u003cspan address=\"10.1016/j.kint.2020.03.034\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDelanaye, P. et al. Outcome of the living kidney donor. \u003cem\u003eNephrol. Dial Transpl. Jan\u003c/em\u003e. \u003cb\u003e27\u003c/b\u003e (1), 41\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/ndt/gfr669\u003c/span\u003e\u003cspan address=\"10.1093/ndt/gfr669\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLentine, K. L. et al. KDIGO Clinical Practice Guideline on the Evaluation and Care of Living Kidney Donors. \u003cem\u003eTransplantation Aug\u003c/em\u003e. \u003cb\u003e101\u003c/b\u003e (8S Suppl 1), S1\u0026ndash;s109. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/tp.0000000000001769\u003c/span\u003e\u003cspan address=\"10.1097/tp.0000000000001769\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndrews, P. A., Burnapp, L. \u0026amp; British Transplantation Society / Renal Association UK Guidelines for Living Donor Kidney Transplantation 2018. Summary of Updated Guidance. \u003cem\u003eTransplantation Jul\u003c/em\u003e. \u003cb\u003e102\u003c/b\u003e (7), e307. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/tp.0000000000002253\u003c/span\u003e\u003cspan address=\"10.1097/tp.0000000000002253\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eERBP Guideline on the Management and Evaluation of the Kidney Donor and Recipient. \u003cem\u003eNephrol. Dial Transpl. Aug\u003c/em\u003e ;\u003cb\u003e28\u003c/b\u003e Suppl 2:ii1\u0026ndash;71. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/ndt/gft218\u003c/span\u003e\u003cspan address=\"10.1093/ndt/gft218\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRichardson, R. et al. Kidney Paired Donation Protocol for Participating Donors 2014. \u003cem\u003eTransplantation Oct.\u003c/em\u003e \u003cb\u003e99\u003c/b\u003e (10 Suppl 1), S1\u0026ndash;s88. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/tp.0000000000000918\u003c/span\u003e\u003cspan address=\"10.1097/tp.0000000000000918\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShen, W. et al. Total body skeletal muscle and adipose tissue volumes: estimation from a single abdominal cross-sectional image. \u003cem\u003eJ Appl Physiol ()\u003c/em\u003e. Dec 2004;97(6):2333-8. (1985). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1152/japplphysiol.00744.2004\u003c/span\u003e\u003cspan address=\"10.1152/japplphysiol.00744.2004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSrikumar, T. et al. Semiautomated Measure of Abdominal Adiposity Using Computed Tomography Scan Analysis. \u003cem\u003eJ. Surg. Res. May\u003c/em\u003e. \u003cb\u003e237\u003c/b\u003e, 12\u0026ndash;21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jss.2018.11.027\u003c/span\u003e\u003cspan address=\"10.1016/j.jss.2018.11.027\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatsuo, S. et al. Revised equations for estimated GFR from serum creatinine in Japan. \u003cem\u003eAm. J. Kidney Dis. Jun\u003c/em\u003e. \u003cb\u003e53\u003c/b\u003e (6), 982\u0026ndash;992. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.ajkd.2008.12.034\u003c/span\u003e\u003cspan address=\"10.1053/j.ajkd.2008.12.034\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2009).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"kidney transplantation, living donor, machine learning, kidney function post-donation, prediction","lastPublishedDoi":"10.21203/rs.3.rs-5862042/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5862042/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eLiving kidney donors typically experience approximately a 30% reduction in kidney function after donation, although the degree of reduction varies among individuals. This study aimed to develop a machine learning (ML) model to predict serum creatinine (Cre) levels at one year post-donation using preoperative clinical data, including kidney-, fat-, and muscle-volumetry values from computed tomography. A total of 204 living kidney donors were included. Symbolic regression via genetic programming was employed to create an ML-based Cre prediction model using preoperative clinical variables. Validation was conducted using a 7:3 training-to-test data split. The ML model demonstrated a median absolute error of 0.079 mg/dL for predicting Cre. In the validation cohort, it outperformed conventional methods (which assume post-donation eGFR to be 70% of the preoperative value) with higher R\u0026sup2; (0.58 vs. 0.27), lower root mean squared error (5.27 vs. 6.89), and lower mean absolute error (3.92 vs. 5.8). Key predictive variables included preoperative Cre and remnant kidney volume. The model was deployed as a web application for clinical use. The ML model offers accurate predictions of post-donation kidney function and may assist in monitoring donor outcomes, enhancing personalized care after kidney donation.\u003c/p\u003e","manuscriptTitle":"Personalized Prediction Model Generated with Machine Learning for Kidney Function One Year After Living Kidney Donation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-01-23 16:43:07","doi":"10.21203/rs.3.rs-5862042/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-03-04T09:18:44+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-03-03T22:13:30+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-02-22T08:36:50+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"257275007403858373947563430315802620908","date":"2025-02-19T18:49:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"71183972631217017547374830869002692540","date":"2025-02-01T10:00:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-01-30T09:49:54+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-01-30T09:46:26+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-01-22T15:13:45+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-01-21T14:26:56+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-01-20T02:47:59+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8b4013b1-adc8-4e68-9c2a-a4958ad1c819","owner":[],"postedDate":"January 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":43258979,"name":"Health sciences/Nephrology/Kidney"},{"id":43258980,"name":"Health sciences/Health care"}],"tags":[],"updatedAt":"2025-07-07T16:14:31+00:00","versionOfRecord":{"articleIdentity":"rs-5862042","link":"https://doi.org/10.1038/s41598-025-02879-y","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-07-01 15:57:34","publishedOnDateReadable":"July 1st, 2025"},"versionCreatedAt":"2025-01-23 16:43:07","video":"","vorDoi":"10.1038/s41598-025-02879-y","vorDoiUrl":"https://doi.org/10.1038/s41598-025-02879-y","workflowStages":[]},"version":"v1","identity":"rs-5862042","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5862042","identity":"rs-5862042","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}