Survival Prediction in Kidney Transplantation Using Deeplearning4j Model

The variations in outcome and frequent occurrence of kidney allograft failure continue to pose important clinical and research challenges despite recent advances in kidney transplantation. The aim of this systematic review was to examine the current application of machine learning models in kidney transplantation and perform a meta-analysis of these models in the prediction of graft survival. This review was registered with the PROSPERO database (CRD42021247469) and all peer-reviewed original articles that reported machine learning model-based prediction of graft survival were included. Quality assessment was performed by the criteria defined by Qiao and risk-of-bias assessment was performed using the PROBAST tool. The diagnostic performance of the meta-analysis was assessed by a meta-analysis of the area under the receiver operating characteristic curve and a hierarchical summary receiver operating characteristic plot. A total of 31 studies met the inclusion criteria for the review and 27 studies were included in the meta-analysis. Twenty-nine different machine learning models were used to predict graft survival in the included studies. Nine studies compared the predictive performance of machine learning models with traditional regression methods. Five studies had a high risk of bias and three studies had an unclear risk of bias. The area under the hierarchical summary receiver operating characteristic curve was 0.82 and the summary sensitivity and specificity of machine learning-based models were 0.81 (95 per cent c.i. 0.76 to 0.86) and 0.81 (95 per cent c.i. 0.74 to 0.86) respectively for the overall model. The diagnostic odds ratio for the overall model was 18.24 (95 per cent c.i. 11.00 to 30.16) and 29.27 (95 per cent c.i. 13.22 to 44.46) based on the sensitivity analyses. Prediction models using machine learning methods may improve the prediction of outcomes after kidney transplantation by the integration of the vast amounts of non-linear data.

Clinical Significance of Postoperative Nutritional Status as a Prognostic Factor in Kidney Transplant Recipients

BackgroundIn an environment of limited kidney donation resources, patient recovery and survival after kidney transplantation (KT) are highly important. We used pre-operative data of kidney recipients to build a statistical model for predicting survivability after kidney transplantation.Material/MethodsA dataset was constructed from a pool of patients who received a first KT in our hospital. For allogeneic transplantation, all donated kidneys were collected from deceased donors. Logistic regression analysis was used to change continuous variables into dichotomous ones through the creation of appropriate cut-off values. A regression model based on the least absolute shrinkage and selection operator (LASSO) algorithm was used for dimensionality reduction, feature selection, and survivability prediction. We used receiver operating characteristic (ROC) analysis, calibration, and decision curve analysis (DCA) to evaluate the performance and clinical impact of the proposed model. Finally, a 10-fold cross-validation scheme was implemented to verify the model robustness.ResultsWe identified 22 potential variables from which 30 features were selected as survivability predictors. The model established based on the LASSO regression algorithm had shown discrimination with an area under curve (AUC) value of 0.690 (95% confidence interval: 0.557–0.823) and good calibration result. DCA demonstrated clinical applicability of the prognostic model when the intervention progressed to the possibility threshold of 2%. An average AUC value of 0.691 was obtained on the validation data.ConclusionsOur results suggest that the proposed model can predict the mortality risk for patients after kidney transplants and could help kidney specialists choose kidney recipients with better prognosis.

BACKGROUND Kidney transplant recipients have higher life expectancy but may require subsequent transplantations, raising ethical concerns regarding organ allocation. We assessed the safety of multiple kidney transplants through long-term follow-up. MATERIAL AND METHODS A retrospective cohort study was conducted at a single center, categorizing patients based on the number of kidney transplantations received. The primary outcome was the composite of death-censored graft failure and overall mortality. The secondary outcome was death-censored graft failure. RESULTS Between 2000 and 2019, our center performed 2152 kidney transplantations. Patients were divided into 3 groups: A (1 transplant; n=1850), B (2 transplants; n=285), and C (3 or more transplants; n=75). Group C patients were younger, had fewer comorbidities, and received more aggressive induction therapy. The primary outcomes, including death-censored graft loss and overall mortality, showed similar rates across groups (A: 21.3%, B: 25.2%, C: 21.7%, p=0.068). However, the secondary outcome of death-censored graft failure alone was significantly lower in group A compared to the other groups. No significant difference was observed between groups B and C (8% vs 16% and 13%, respectively, p=0.001, p=0.845). Multivariate analysis identified having a living donor as the strongest predictor of patient and graft survival in all study groups. CONCLUSIONS Graft and patient survival rates were similar between first and multiple transplant recipients. Multiple transplant recipients had lower death-censored graft failure risk compared to first transplant recipients. However, the risk did not differ among second and subsequent transplant recipients. Younger patients, especially those with a living donor, should be considered for repeat kidney transplantation.

Computerized tomography (CT) is the most accurate method for evaluating pelvic calcifications, which are of utmost importance for planning kidney transplantation (KT). The aim of our study was to evaluate the incidence and distribution of iliac artery calcifications and correlate the novel pelvic calcification score (PCS) with cardiovascular risk factors and graft and overall survival in KT patients. We retrospectively included 118 KT patients operated at our institution with pretransplant pelvic CT. Calcification morphology, circumference and length of both common and external iliac arteries were independently scored by two uroradiologists. PCS was calculated as the total score sum of all three calcification features in all vessels. PCS correlation with graft and patient survival was performed. Calcification in at least one vascular segment was found in 79% of patients. PCS was significantly higher in male patients (p = 0.006), patients over 55 years (p < 0.001), and patients on haemodialysis (p = 0.016). Patients with a PCS >3 had significantly shorter graft and overall survival rates (p = 0.041 and p = 0.039, respectively). The extent of iliac artery calcification in KT recipients quantified by PCS on pretransplant CT correlates with graft and overall patient survival. A PCS over three was associated with worse clinical outcomes and could become a possible prognostic factor. Our novel PCS is a robust method for quantifying iliac artery calcification burden. Since higher a PCS correlates with worse patient and graft survival, PCS has the potential to become a prognostic factor in kidney transplant patients.

Comprehensive Analysis of Donor Factors for Allograft Survival in Living Kidney Transplantation: A Single-Center Study in Japan

Objective To determine the relationship between Notch receptor-1 (Notch1) and ongoing immune processes within the graft. Methods Flow cytometry (FCM) was performed to detect Notch1 expression in peripheral blood mononuclear cells ( PBMCs) of 40 recipients post-transplantatioa Results The Notch1 expression level was higher in PBMCs of patients with ongoing acute rejection episodes proven by histologically thanCNI toxicosis group (1. 66±0. 64),control group (1. 36 ± 0. 50) and stable graft function group (1. 67 ± 0. 45) (P<0. 05). Most importantly,elevated levels of Notchl levels were correlated with the serum creatinine levels. Conclusion These data reveal a correlation between elevated Notch1 expression in PBMCs post-transplantation and intragraft immune activation that leads to acute rejection. Notch1 may be a good predictor of acute renal allograft rejection and long-term survival. Key words: Notch1 ; Graft rejection; Kidney transplantation; Flow cytometry

New-onset Diabetes Mellitus After Kidney Transplantation—A Paired Kidney Analysis

The use of antibody induction after kidney transplantation has increased from 25% to 63% in the past decade and roughly one half of the induction agent used is anti-interleukin-2 receptor antibody (IL-2RA, ie, basiliximab or daclizumab). When combined with calcineurin inhibitor (CNI)-based immunosuppression, IL-2RAs have been shown to reduce the incidence of acute rejection, one of the predictors of poor graft survival, without increasing risks of infections and malignancies in kidney transplantation. For low-immunological-risk patients, IL-2RAs, as compared with lymphocyte-depleting antibodies, are equally efficacious and have better safety profiles. For high-risk patients, however, IL-2RAs may be inferior to lymphocyte-depleting antibodies for the prophylaxis of acute rejection. In an effort to reduce toxicities of other immunosuppressive medications without increasing the risk of acute rejection and chronic graft loss, IL-2RAs have often been combined with steroid- and CNI-sparing immunosuppression protocols. More data support the benefits of early steroid withdrawal with IL-2RA in low-risk patients, but preferred induction therapy for high-risk patients has yet to be determined. Although CNI-sparing protocols with IL-2RA may preserve renal function and improve long-term survival in selected patients, further studies are needed to identify those who benefit most from this strategy.

BackgroundBrain death initiates hemodynamic, immunological, and hormonal changes that potentially compromise organ quality for transplantation. Therefore, it is generally believed that organs should be procured as soon as possible after the declaration of brain death. However, conflicting data exist regarding the impact of brain death duration on long-term graft function and survival.Material/MethodsThe effect of duration of brain death on graft survival and function of 1869 adult transplant recipients receiving kidneys from deceased donors after brain death was analyzed, using relevant donor and recipient characteristics and allograft related factors.ResultsDuration of brain death was a significant predictor for long-term graft survival, whilst there was no significant effect of duration of brain death on the incidence of delayed graft function or acute graft rejection after kidney transplantation. After dividing the study population into a “short durBD” (<10.6 hours) group and a “long durBD” (>10.6 hours) group, the 15-year graft survival estimates were significantly higher and the serum creatinine at 3 months after transplantation was significantly lower in the “long durBD” group.ConclusionsDuration of brain death does not affect the incidence of delayed graft function or acute rejection after kidney transplantation. However, longer duration of brain death is associated with better kidney allograft function and survival.

Should Immunosuppression After Kidney Transplant Be Adjusted Based on Renal Resistance During Pretransplant Hypothermic Machine Perfusion?

BackgroundKidney transplantation is the preferred treatment option for patients with end-stage renal disease. To maximize patient and graft survival, the allocation of donor organs to potential recipients requires careful consideration.ObjectiveThis study aimed to develop an innovative technological solution to enable better prediction of kidney transplant survival for each potential donor-recipient pair.MethodsWe used deidentified data on past organ donors, recipients, and transplant outcomes in the United States from the Scientific Registry of Transplant Recipients. To predict transplant outcomes for potential donor-recipient pairs, we used several survival analysis models, including regression analysis (Cox proportional hazards), random survival forests, and several artificial neural networks (DeepSurv, DeepHit, and recurrent neural network [RNN]). We evaluated the performance of each model in terms of its ability to predict the probability of graft survival after kidney transplantation from deceased donors. Three metrics were used: the C-index, integrated Brier score, and integrated calibration index, along with calibration plots.ResultsOn the basis of the C-index metrics, the neural network–based models (DeepSurv, DeepHit, and RNN) had better discriminative ability than the Cox model and random survival forest model (0.650, 0.661, and 0.659 vs 0.646 and 0.644, respectively). The proposed RNN model offered a compromise between the good discriminative ability and calibration and was implemented in a technological solution of technology readiness level 4.ConclusionsOur technological solution based on the RNN model can effectively predict kidney transplant survival and provide support for medical professionals and candidate recipients in determining the most optimal donor-recipient pair.

BackgroundPredicting allograft survival is vital for efficient transplant success. With dynamic changes in patient conditions, clinical indicators may change longitudinally, and doctors’ judgments may be highly variable. It is necessary to establish a dynamic model to precisely predict the individual risk/survival of new allografts.MethodsThe follow-up data of 407 patients were obtained from a renal allograft failure study. We introduced a landmarking-based dynamic Cox model that incorporated baseline values (age at transplantation, sex, weight) and longitudinal changes (glomerular filtration rate, proteinuria, hematocrit). Model performance was evaluated using Harrell’s C-index and the Brier score.ResultsSix predictors were included in our analysis. The Kaplan–Meier estimates of survival at baseline showed an overall 5-year survival rate of 87.2%. The dynamic Cox model showed the individual survival prediction with more accuracy at different time points (for the 5-year survival prediction, the C-index = 0.789 and Brier score = 0.065 for the average of all time points) than the static Cox model at baseline (C-index = 0.558, Brier score = 0.095). Longitudinal covariate prognostic analysis (with time-varying effects) was performed.ConclusionsThe dynamic Cox model can utilize clinical follow-up data, including longitudinal patient information. Dynamic prediction and prognostic analysis can be used to provide evidence and a reference to better guide clinical decision-making for applying early treatment to patients at high risk.

Limited information exists about acute renal failure (ARF) early after heart transplantation (HTx). We correlated pre-, per-, and post-operative patient and donor parameters to the risk of developing ARF. We also analyzed the consequences of ARF on kidney function after HTx, risk of later need for chronic dialysis or kidney transplantation, and mortality. In a retrospective study from 1983 to 2007, 145 (25%) of 585 HTx recipients developed ARF, defined as ≥ 26.4 micromol/L or ≥ 50% increase in serum creatinine from pre-operatively to the seventh day post-HTx and/or the need of early post-operative dialysis. Independent risk factors for ARF were intravenous cyclosporine immediately post-operatively (odds ratio [OR] 2.16, 95% CI 1.34-3.50, p = 0.02), donor age (OR 1.02, 95% CI 1.00-1.04, p = 0.02), and pre-operative cardiac output (OR 1.38, 95% CI 1.12-1.71, p = 0.003). The development of ARF was a predictor for short-term survival (≤ 3 months) ranging from 98% for patients who improved their creatinine after HTx vs. 79% for those in need of dialysis (p < 0.001). However, ARF did not predict subsequent end stage renal disease in need of dialysis or renal transplantation. ARF is a common complication post-HTx. As ARF is associated with short-term survival, post-operative strategies of preserving renal function have the potential of reducing mortality. Of avoidable risk factors, the use of intravenous CsA should be discouraged.

The United Network for Organ Sharing (UNOS) Expanded Criteria Donor (ECD) system utilizes pre-transplant variables to identify deceased donor kidneys with an increased risk of graft loss. The aim of this study was to compare the ECD system with a quantitative approach, the deceased donor score (DDS), in predicting outcome after kidney transplantation. We retrospectively reviewed 49 111 deceased donor renal transplants from the UNOS database between 1984 and 2002. DDS: 0-39 points; >or=20 points defined as marginal. Recipient outcome variables were analyzed by ANOVA or Kaplan-Meier method. There was a 90% agreement between the DDS and ECD systems as predictors of renal function and graft survival. However, DDS identified ECD- kidneys (10.7%) with a significantly poorer outcome than expected (DDS 20-29 points, n = 5,252). Stratification of ECD+ kidneys identified a group with the poorest outcome (DDS >or=30 points). Predictability of early post-transplant events (i.e. need for hemodialysis, decline of serum creatinine and length of hospital stay) was also improved by DDS. DDS predicted outcome of deceased donor renal transplantation better than the ECD system. Knowledge obtained by stratification of deceased donor kidneys can allow for improved utilization of marginal kidneys which is not achieved by the UNOS ECD definition alone.