Machine Learning Model to Predict Graft Rejection After Kidney Transplantation.

BACKGROUND: There are few predictive studies about early posttransplant outcomes taking into account baseline and posttransplant variables. The objective of this study was to create a predictive model for 30-day graft rejection using machine learning techniques. METHODS: Retrospective study with 1255 patients undergoing transplant from living and deceased donors at a tertiary health service in Brazil. Recipient, donor, transplantation, and postoperative period data were collected from physical and electronic records. We split the data into derivation (training) and validation (test) datasets. Five supervised machine learning algorithms were developed with this subset of variables in the training set: Simple Logistic Regression, Lasso, Multilayer Perceptron, XGBoost, and Light GBM. RESULTS: There were 147 (12.48%) cases of graft rejection within 30 days of transplantation. The best model was XGBoost (accuracy, 0.839; receiver operating characteristic area under the curve, 0.715; precision, 0.900). The model showed that deceased donor transplantation, glomerulopathy as an underlying disease, and donor's use of vasoactive drugs had more than 20% importance as rejection risk factors. The variables with the greatest predictive values were thymoglobulin induction and delayed graft function. CONCLUSIONS: We fitted a machine learning model to predict 30-day graft rejection after kidney transplantation that reaches a higher accuracy and precision. Machine learning models could contribute to predicting kidney survival using nontraditional approaches.

The future is coming: promising perspectives regarding the use of machine learning in renal transplantation / O futuro está chegando: perspectivas promissoras sobre o uso de machine learning no transplante renal

Abstract Introduction: The prediction of post transplantation outcomes is clinically important and involves several problems. The current prediction models based on standard statistics are very complex, difficult to validate and do not provide accurate prediction. Machine learning, a statistical technique that allows the computer to make future predictions using previous experiences, is beginning to be used in order to solve these issues. In the field of kidney transplantation, computational forecasting use has been reported in prediction of chronic allograft rejection, delayed graft function, and graft survival. This paper describes machine learning principles and steps to make a prediction and performs a brief analysis of the most recent applications of its application in literature. Discussion: There is compelling evidence that machine learning approaches based on donor and recipient data are better in providing improved prognosis of graft outcomes than traditional analysis. The immediate expectations that emerge from this new prediction modelling technique are that it will generate better clinical decisions based on dynamic and local practice data and optimize organ allocation as well as post transplantation care management. Despite the promising results, there is no substantial number of studies yet to determine feasibility of its application in a clinical setting. Conclusion: The way we deal with storage data in electronic health records will radically change in the coming years and machine learning will be part of clinical daily routine, whether to predict clinical outcomes or suggest diagnosis based on institutional experience.

Resumo Introdução: A predição de resultados pós-transplante é clinicamente importante e envolve vários problemas. Os atuais modelos de previsão baseados em padrões estatísticos são muito complexos, difíceis de validar e não fornecem previsões precisas. Machine Learning, é uma técnica estatística que permite que o computador faça previsões futuras usando experiências anteriores, está começando a ser usada para resolver essas questões. No campo do transplante renal, o uso da previsão computacional foi relatado na predição de rejeição crônica de aloenxerto, função tardia do enxerto e sobrevida do enxerto. Este artigo descreve os princípios e etapas de machine learning para fazer uma previsão e realiza uma breve análise das aplicações mais recentes de seu uso na literatura. Discussão: Existem evidências convincentes de que as abordagens de machine learning baseadas nos dados do doador e do receptor são melhores para proporcionar melhor prognóstico dos resultados do enxerto do que a análise tradicional. As expectativas imediatas que emergem dessa nova técnica de modelagem de previsão são que ela gerará melhores decisões clínicas baseadas em dados de práticas dinâmicas e locais e aperfeiçoará a alocação de órgãos, bem como o gerenciamento de cuidados pós-transplante. Apesar dos resultados promissores, ainda não há um número substancial de estudos para determinar a viabilidade de sua aplicação em um cenário clínico. Conclusão: A forma como lidamos com dados de armazenamento em prontuários eletrônicos de saúde mudará radicalmente nos próximos anos e a machine learning fará parte da rotina clínica diária, seja para prever resultados clínicos ou sugerir um diagnóstico baseado na experiência institucional.

The future is coming: promising perspectives regarding the use of machine learning in renal transplantation.

INTRODUCTION: The prediction of post transplantation outcomes is clinically important and involves several problems. The current prediction models based on standard statistics are very complex, difficult to validate and do not provide accurate prediction. Machine learning, a statistical technique that allows the computer to make future predictions using previous experiences, is beginning to be used in order to solve these issues. In the field of kidney transplantation, computational forecasting use has been reported in prediction of chronic allograft rejection, delayed graft function, and graft survival. This paper describes machine learning principles and steps to make a prediction and performs a brief analysis of the most recent applications of its application in literature. DISCUSSION: There is compelling evidence that machine learning approaches based on donor and recipient data are better in providing improved prognosis of graft outcomes than traditional analysis. The immediate expectations that emerge from this new prediction modelling technique are that it will generate better clinical decisions based on dynamic and local practice data and optimize organ allocation as well as post transplantation care management. Despite the promising results, there is no substantial number of studies yet to determine feasibility of its application in a clinical setting. CONCLUSION: The way we deal with storage data in electronic health records will radically change in the coming years and machine learning will be part of clinical daily routine, whether to predict clinical outcomes or suggest diagnosis based on institutional experience.

Wound Healing Complications in Kidney Transplant Recipients Receiving Everolimus.

BACKGROUND: De novo use of mammalian target of rapamycin inhibitors after kidney transplantation is associated with a concentration-dependent incidence of wound healing adverse events (WHAE). The objective of this analysis was to compare the incidence of WHAE in patients receiving everolimus (EVR) or mycophenolate sodium (MPS). METHODS: This was a predefined subanalysis of a single-center prospective randomized study in which 288 kidney transplant recipients receiving tacrolimus and prednisone were randomized for 3 different regimens: rabbit antithymocyte globulin (r-ATG)/EVR (N = 85); basiliximab (BAS)/EVR (N = 102); BAS/MPS (N = 101). Clinical WHAE were prospectively collected using a prespecified case report form in all study visits. Abdominal ultrasound was performed at 30 days posttransplant to capture subclinical abnormalities. Surgeons were blinded to randomized treatment and no specific surgical procedures were implemented. RESULTS: A higher proportion of patients in BAS/EVR showed at least 1 clinical WHAE (22.3% vs 35.3% vs 22.0%, P = 0.03) and total clinical and subclinical WHAE (35% vs 42% vs 26%, P = 0.014) compared with BAS/MPS, respectively. A higher proportion of patients in r-ATG/EVR showed subclinical WHAE (13% vs 7% vs 4%, P = 0.025) compared with BAS/MPS, respectively. Patients receiving EVR showed a higher risk of developing clinical or subclinical WHAE (r-ATG/EVR vs BAS/MPS hazard ratio 1.30; BAS/EVR vs BAS/MPS hazard ratio 1.73, P = 0.028). CONCLUSIONS: In this cohort of de novo kidney transplant recipients receiving tacrolimus and prednisone, the use of EVR was associated with higher incidence of combined clinical and subclinical WHAE compared with MPS.

Prospective Randomized Trial Investigating the Influence of Pharmaceutical Care on the Intra-Individual Variability of Tacrolimus Concentrations Early After Kidney Transplant.

BACKGROUND: This study evaluated the influence of pharmaceutical care (PhC) in the intra-individual variability of dose-corrected whole blood tacrolimus (TAC) trough concentrations, adherence to immunosuppressive therapy and clinical outcomes. METHODS: We randomized 128 kidney transplant recipients to receive PhC consisted of predefined instructions provided by a pharmacist (PhC group, n = 64) or standard nurse staff instructions (control group, n = 64) from day 3 to day 90 after kidney transplantation. The study was powered to detect at least 50% reduction in the coefficient of variation (%CV), calculated from 6 dose-corrected whole blood TAC trough concentrations, in the PhC group. Patient adherence was evaluated using Basel Assessment of Adherence to Immunosuppressive Medication Scale (BAASIS) questionnaire. RESULTS: There was no difference in the %CV comparing PhC and control group (31.4% ± 12.3% versus 32.5% ± 16.1%, P = 0.673). There were no differences in the proportion of patients showing TAC concentrations within predefined target concentrations in each study visit. There was no difference in the proportion of nonadherent patients at day 28 (17% versus 26%, P = 0.135) and day 90 (27% versus 25%, P = 0.457) based on BAASIS questionnaire answers, respectively. There were no differences in clinical outcomes. CONCLUSIONS: Universal PhC in addition to standard nurse staff instruction was not associated with reduced intra-individual variability of dose-corrected whole blood TAC trough concentrations or improved adherence.

De novo everolimus for recipients of kidney transplants from HLA identical donors. / Uso de everolimo de novo em receptores de transplante renal com doador vivo HLA idêntico.

INTRODUCTION: Kidney transplant recipients from HLA-identical living donor have lower risk of acute rejection and greater graft survival compared to other types of kidney transplantation. Immunosuppressive regimens without calcineurin inhibitors (CNI) can further improve these results by reducing cardiovascular, metabolic and toxic events related to this drug class. OBJECTIVE: This study aimed to evaluate efficacy and safety of a new immunosuppressive regimen with planned suspension of CNI. METHODS: This was a prospective, single center and single treatment arm study to evaluate HLA-identical kidney transplant recipients receiving everolimus (EVR), tacrolimus (TAC) and corticosteroids, followed by TAC discontinuation 30 days after transplantation. TAC discontinuation was later postponed to the third month after an interim efficacy analysis. RESULTS: Thirty-nine patients were included. Although mean TAC and EVR blood concentrations have remained within the proposed therapeutic ranges, five patients had biopsy-proven acute rejection and one patient had an episode of C4D-positive glomerulitis. This result led to the end of the inclusions. Interestingly, the proportion of patients with proteinuria greater than 0.5 g/L has not reached more than 22% of patients in any visit. Adverse events related to EVR use were the most incident in this population: oral ulcers, dyslipidemia and peripheral edema. CONCLUSION: The proposed scheme was not effective for this population, particularly due to a high incidence of acute rejection. Safety profile showed that prolonged exposure to a high concentration of blood EVR increases the incidence of adverse events related to this drug.

Uso de everolimo de novo em receptores de transplante renal com doador vivo HLA idêntico / De novo everolimus for recipients of kidney transplants from HLA identical donors

Resumo Introdução: Receptores de rim de doadores vivos HLA-idêntico apresentam menor risco para rejeição aguda e maior sobrevida do enxerto, quando comparado a outros tipos de transplante. Um regime imunossupressor sem inibidor de calcineurina (ICN) pode melhorar ainda mais esses resultados, através da redução de eventos cardiovasculares, metabólicos e tóxicos secundários a esse fármaco. Objetivo: Avaliar eficácia e segurança do novo tratamento imunossupressor com suspensão planejada do ICN. Métodos: Estudo prospectivo, aberto, braço único de tratamento em único centro para avaliar resultados do transplante renal HLA-idêntico em pacientes que recebem everolimo (EVR), tacrolimo (TAC) e corticoide, seguido da descontinuação do TAC 30 dias pós-transplante. Após análise interina de eficácia, a descontinuação do TAC foi postergada para o terceiro mês pós-transplante, através de emenda ao protocolo. Resultados: Trinta e nove pacientes foram incluídos. Apesar de as médias das concentrações de TAC e EVR terem respeitado os intervalos propostos, cinco pacientes tiveram rejeição aguda comprovada por biópsia e um paciente apresentou um episódio de glomerulite com depósitos de C4D. Esse resultado demandou o fim das inclusões. A proporção de pacientes com proteinúria > 0.5g/L não atingiu mais que 22% dos pacientes em nenhuma visita. Os eventos adversos mais frequentes foram relacionados ao uso de EVR: úlceras orais, dislipidemia e edema periférico. Conclusão: O regime proposto não foi eficaz para essa população, principalmente pela alta incidência de rejeição aguda. O perfil de segurança mostrou que a exposição prolongada a altas concentrações sanguíneas de EVR aumenta a incidência dos eventos adversos relacionados ao fármaco.

Abstract Introduction: Kidney transplant recipients from HLA-identical living donor have lower risk of acute rejection and greater graft survival compared to other types of kidney transplantation. Immunosuppressive regimens without calcineurin inhibitors (CNI) can further improve these results by reducing cardiovascular, metabolic and toxic events related to this drug class. Objective: This study aimed to evaluate efficacy and safety of a new immunosuppressive regimen with planned suspension of CNI. Methods: This was a prospective, single center and single treatment arm study to evaluate HLA-identical kidney transplant recipients receiving everolimus (EVR), tacrolimus (TAC) and corticosteroids, followed by TAC discontinuation 30 days after transplantation. TAC discontinuation was later postponed to the third month after an interim efficacy analysis. Results: Thirty-nine patients were included. Although mean TAC and EVR blood concentrations have remained within the proposed therapeutic ranges, five patients had biopsy-proven acute rejection and one patient had an episode of C4D-positive glomerulitis. This result led to the end of the inclusions. Interestingly, the proportion of patients with proteinuria greater than 0.5 g/L has not reached more than 22% of patients in any visit. Adverse events related to EVR use were the most incident in this population: oral ulcers, dyslipidemia and peripheral edema. Conclusion: The proposed scheme was not effective for this population, particularly due to a high incidence of acute rejection. Safety profile showed that prolonged exposure to a high concentration of blood EVR increases the incidence of adverse events related to this drug.

Efficacy and Safety of Elective Conversion From Sotrastaurin (STN) to Tacrolimus (TAC) or Mycophenolate (MPS) in Stable Kidney Transplant Recipients.

BACKGROUND: This study aimed to evaluate the efficacy and safety outcomes of conversion strategies in stable kidney transplant recipients after premature termination of the sotrastaurin (STN) development program. METHODS: This is an exploratory and prospective study, including 38 stable renal transplant recipients. Tacrolimus (TAC) group [STN → mycophenolate sodium (MPS)] consisted of 9 patients receiving TAC, STN, and prednisone that were converted from STN to MPS. Everolimus (EVR) group (STN → TAC) consisted of 29 patients receiving EVR, STN, and prednisone that were converted from STN to TAC. RESULTS: In TAC (STN → MPS) group, dose-adjusted TAC concentrations decreased from baseline to first week (2.3 ± 1.1 versus 1.5 ± 1.0 ng·mL·mg, P < 0.05). Two patients experienced a first acute rejection episode. Conversion to MPS was associated with a higher incidence of adverse events. In EVR (STN → TAC) group, dose-adjusted EVR concentrations decreased from baseline to first week (3.6 ± 2.3 ng·mL·mg versus 1.9 ± 0.8 ng·mL·mg, P < 0.01). The proportion of patients with donor-specific antibodies was lower in TAC (STN → MPS) (11%) compared to EVR (STN → TAC) (31%) before conversion. Conversion from STN to TAC was associated with a reduction in estimated glomerular filtration rate (69.6 ± 16.9 versus 61.0 ± 18.8 mL·min·1.73 m, P < 0.01) and a decreased proportion of patients with donor-specific antibodies (31% versus 14%) at 12 months. CONCLUSIONS: Conversion from TAC/STN to TAC/MPS or from EVR/STN to TAC/EVR was associated with significant pharmacokinetic changes in both TAC and EVR whole-blood trough concentrations due to known drug-to-drug interaction, which were associated with changes in efficacy and safety.

Pharmacokinetics and Long-Term Safety and Tolerability of Everolimus in Renal Transplant Recipients Converted From Cyclosporine.

BACKGROUND: Conversion from cyclosporine (CsA) to everolimus (EVR) in kidney transplant recipients receiving mycophenolate sodium (MPS) and corticosteroids has been used to reduce CsA associated toxicities. Nevertheless, exposures produced by the initial EVR dose, the steady state pharmacokinetic and long-term safety and tolerability have not been explored in detail. METHODS: Twenty-four stable kidney transplant recipients receiving CSA, MPS, and corticosteroids were converted from CSA to EVR. The initial EVR dose was 3 mg BID. Weekly monitoring of EVR blood concentrations was followed by a full 12 hour pharmacokinetic profile 28 days after conversion. Therapeutic drug monitoring, safety, and tolerability were analyzed during 5 years of follow-up. RESULTS: The study population was relatively young (mean of 42 years) with a predominance of males (62%) and White (67%) recipients of kidneys from living (54%) or deceased (46%) donors. Mean time of the conversion was 61 months after transplantation. In the first 7 patients, the initial EVR dose of 3 mg BID resulted in mean EVR trough blood concentration of 14.7 ± 3.7 ng/mL at day 7. The initial EVR dose was then reduced to 2 mg BID for the following 17 patients. Four weeks after conversion, mean EVR dose was 1.7 ± 0.5 mg BID (7 patients were receiving 1 mg BID and 17 were receiving 2 mg BID) resulting in mean EVR trough blood concentration of 4.0 ± 1.4 ng/mL. Whereas mean maximum concentration (13.4 ± 2.8 versus 22.9 ± 7.4 ng/mL, P = 0.003) and mean apparent clearance (232 ± 79 versus 366 ± 173 mL/min, P = 0.016) were higher, mean area under the curve (78.2 ± 22.1 versus 102.5 ± 38.5 ng.h/mL, P = 0.067) and mean C0 (3.7 ± 1.3 versus 4.1 ± 1.5 ng/mL, P = 0.852) were no different comparing patients receiving 1 mg and 2 mg EVR BID. Mean inter-subject variability of area under the curve, trough concentration, and maximum concentration was 38%, 36%, and 38%. EVR treatment was discontinued in 29% of patients due to proteinuria (N = 2), pneumonia (N = 2), dyslipidemia (N = 2), and anemia (N = 1) and MPS dose was reduced in 58% of patients. CONCLUSIONS: The initial 3 mg BID dose produced high EVR trough blood concentrations. The 2 mg BID dose appears to be the appropriate initial dose to provide therapeutic concentrations but still requires initial intensive therapeutic monitoring to achieve and maintain blood concentrations within the therapeutic target concentration. The combination of EVR and full dose MPS has limited long-term tolerability and safety.