Machine Learning Method for the Management of Acute Kidney Injury: More Than Just Treating Biomarkers Individually

Wei

et al. 2020

Preprint

Background Acute kidney injury (AKI) is a serve and harmful syndrome in the intensive care unit. Our previous study has shown that patients who will progress to AKI 3 stage are considered to receive RRT. This study aimed to develop a prediction model that can predict whether progression to AKI stage 3. Methods Patients with AKI stage 1/2, when they were first diagnosed with AKI in the Medical Information Mart for Intensive Care (MIMIC-III), were included. Patients who receive RRT or progress to AKI 3 stage within 72 hours of first AKI diagnosis were excluded. We build two predictive models, respectively using machine learning extreme gradient boosting (XGBoost) and logistic regression, to predict patients who will progress to AKI stage 3. Established models were evaluated by cross-validation and area under receiver operating characteristic curve (AU-ROC). Results Of the 29238 patients included in the analysis, 3237 (11.1%) patients progressed to AKI stage 3. Creatinine, blood urea nitrogen (BUN), sepsis and respiratory failure were the important predictors of AKI progression. The machine learning XGBoost model has a better performance than the Cox regression model on predicting AKI stage 3 progression (AU-ROC, 0.860 vs. 0.728, respectively). Conclusions The XGBoost model was able to identify patients with AKI progression better than the Cox regression model. Machine learning techniques may improve predictive modeling in medical research.

Section: Discussionmentioning

confidence: 99%

WITHDRAWN: Machine Learning for the Prediction of Progression in Patients With Acute Kidney Injury in Critical Care

Wei

et al. 2020

Preprint

“…In supervised learning, the algorithm makes a functional map from the variables to the outcomes. [ 8 ] One study about biomarker finding used an unsupervised ANN for clustering. The reported prediction accuracies differed among the studies, owing to the study subjects, predictors, validation types, and algorithms.…”

Section: Discussionmentioning

confidence: 99%

Artificial intelligence with kidney disease

Park¹,

Park²,

Lee³

et al. 2021

Medicine

Background: Artificial intelligence (AI) has had a significant impact on our lives and plays many roles in various fields. By analyzing the past 30 years of AI trends in the field of nephrology, using a bibliography, we wanted to know the areas of interest and future direction of AI in research related to the kidney. Methods: Using the Institute for Scientific Information Web of Knowledge database, we searched for articles published from 1990 to 2019 in January 2020 using the keywords AI; deep learning; machine learning; and kidney (or renal). The selected articles were reviewed manually at the points of citation analysis. Results: From 218 related articles, we selected the top fifty with 1188 citations in total. The most-cited article was cited 84 times and the least-cited one was cited 12 times. These articles were published in 40 journals. Expert Systems with Applications (three articles) and Kidney International (three articles) were the most cited journals. Forty articles were published in the 2010s, and seven articles were published in the 2000s. The top-fifty most cited articles originated from 17 countries; the USA contributed 16 articles, followed by Turkey with four articles. The main topics in the top fifty consisted of tumors (11), acute kidney injury (10), dialysis-related (5), kidney-transplant related (4), nephrotoxicity (4), glomerular disease (4), chronic kidney disease (3), polycystic kidney disease (2), kidney stone (2), kidney image (2), renal pathology (2), and glomerular filtration rate measure (1). Conclusions: After 2010, the interest in AI and its achievements increased enormously. To date, AIs have been investigated using data that are relatively easy to access, for example, radiologic images and laboratory results in the fields of tumor and acute kidney injury. In the near future, a deeper and wider range of information, such as genetic and personalized database, will help enrich nephrology fields with AI technology.

“…Zacharias et al [ 10 ] predicted AKI following cardiac surgery based on urinary metabolome profiles (NMR) of patients and identified that the AUC at 24 h after the surgery was 83%, which is comparable to the AUC at 24 h after ICU admission in our study (89%) using the 1st group of patients. Various machine learning techniques have been applied for AKI prediction [ 33 ], although the AUCs were usually below 80% when using biomarker candidates as inputs [ 31 , 34 ]. AKI prediction based on clinical variables may be less laborious than that based on body fluid biomarkers, but in our study, clinical parameters were not selected as efficient predictors by LASSO.…”

Section: Discussionmentioning

confidence: 99%

Urinary Metabolome Analyses of Patients with Acute Kidney Injury Using Capillary Electrophoresis-Mass Spectrometry

et al. 2021

Acute kidney injury (AKI) is defined as a rapid decline in kidney function. The associated syndromes may lead to increased morbidity and mortality, but its early detection remains difficult. Using capillary electrophoresis time-of-flight mass spectrometry (CE-TOFMS), we analyzed the urinary metabolomic profile of patients admitted to the intensive care unit (ICU) after invasive surgery. Urine samples were collected at six time points: before surgery, at ICU admission and 6, 12, 24 and 48 h after. First, urine samples from 61 initial patients (non-AKI: 23, mild AKI: 24, severe AKI: 14) were measured, followed by the measurement of urine samples from 60 additional patients (non-AKI: 40, mild AKI: 20). Glycine and ethanolamine were decreased in patients with AKI compared with non-AKI patients at 6–24 h in the two groups. The linear statistical model constructed at each time point by machine learning achieved the best performance at 24 h (median AUC: 89%, cross-validated) for the 1st group. When cross-validated between the two groups, the AUC showed the best value of 70% at 12 h. These results identified metabolites and time points that show patterns specific to subjects who develop AKI, paving the way for the development of better biomarkers.