Techniques for predicting dialysis access site outcomes

Abstract

Techniques for determining a vascular access (VA) prediction of a patient based on evaluating biomarkers with a VA profile library are described. The biomarkers may include targeted metabolomics, untargeted metabolomics, micro-RNA (miRNA), proteomic markers, and / or vasculature remodeling markers (e.g., following surgical interventions). The VA profile library may be determined based on real-world clinical data of a population of patients with known VA outcomes. In one example, a VA is an arteriovenous fistulas (AVF). In one example, a method of generating a VA prediction for a dialysis patient includes determining biomarker information of a patient, and determining a VA prediction based on evaluating the biomarker information with a VA profile library, the VA prediction indicating a likelihood of success of a VA of the patient for dialysis. The metabolites may include one or more of decanoyl-L-carnitine, acetylcarnitine, 2-methylbutyroylcarnitine, L-valine, L-hexanoylcarnitine, N-methylanthranilic acid, 1-palmitoyllysophosphatidylcholine (LysoPC 16:0), L-(+)-ergothioneine, or D-maltose.

Description

TECHNIQUES FOR PREDICTING DIALYSIS ACCESS SITE OUTCOMESCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 639,174, filed April 26, 2024, and entitled “Techniques For Predicting Dialysis Access Site Outcomes,” and U.S. Provisional Patent Application No. 63 / 676,974, filed July 30, 2024, and entitled “Techniques For Predicting Dialysis Access Site Outcomes,” the entirety of each of which is incorporated by reference herein.FIELD OF THE DISCLOSURE

[0002] The present disclosure relates to techniques for dialysis treatments. More particularly, the present disclosure relates to techniques for predicting dialysis access site outcomes.BACKGROUND

[0003] Dialysis treatment requires access to the patient circulatory system via a dialysis access site in order to process patient blood using a dialysis treatment unit. For peritoneal dialysis (PD), the dialysis access site may be accessed via a catheter. Hemodialysis (HD) treatment requires access to blood circulation in an extracorporeal circuit connected to the main cardiovascular circuit of the patient through a vascular or arteriovenous (AV) access. Arteriovenous fistulas (AVFs) are regarded as the preferred form of vascular access for HD because of high delivered blood flows, low infection risk, and access longevity. During an HD treatment, blood is removed from the vascular access by an arterial needle fluidly connected to the extracorporeal circuit and provided to an HD treatment unit. After processing via the HD treatment unit, the blood is sent back to the vascular access through a venous needle and back into the patient cardiovascular circuit.

[0004] The health of the vascular access site of a patient is of primary importance to the efficacy of the dialysis treatment. For example, an access site should be capable of providing adequate blood flow for HD treatment and should be free of serious complications, such as pain and / or swelling, aneurysms, and / or the like.

[0005] After surgical (or endovascular) creation, AVFs need to undergo maturation before they can be cannulated safely and repeatedly. AVF maturation entails complex physiological and morphological changes and an intricate interplay between various biological mechanisms andprocesses. Currently, more than half of surgically created AVFs fail to develop into a usable VA. Unsuccessful AVF maturation is a major factor contributing to low AVF prevalence and, for patients having an AVF, HD complications. However, the biology of AVF maturation is not fully understood. As a result, with conventional medical techniques, healthcare professionals are limited in their ability to increase the success rate of AVF maturation.

[0006] Accordingly, patients with impaired kidney function and healthcare providers would benefit from processes capable of efficiently and effectively determining patient characteristics that may affect AVF maturation and success rates.SUMMARY

[0007] This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to necessarily identify key features or essential features of the claimed subject matter, nor is it intended as an aid in determining the scope of the claimed subject matter.

[0008] The present disclosure is directed to systems and methods for planning, treating, and / or implanting vascular accesses (VAs), including arteriovenous fistulas (AVFs), for patients, for instance, individuals with impaired kidney function. In some embodiments, patient biological information is determined and used to provide a targeted, personalized treatment, treatment recommendation, AVF implantation process, drug regimen, and / or the like to facilitate maturation of an AVF for a specific patient. In various embodiments, an AVF analysis process may support clinical decision making for VA creation (AVF, AVG, CVC, bioengineered vessel).

[0009] In various embodiments, the patient’s biological information is or includes biomarkers of the patient. In some embodiments, an AVF analysis process may operate to identify biomarkers to reliably predict AVF maturation success or failure that would be applicable to newly created AVFs or during assessment for suitable access placement. In various embodiments, the biomarkers may include targeted metabolomics, untargeted metabolomics, micro-RNA (miRNA), proteomic markers, and / or vasculature remodeling markers (e.g., following surgical interventions).

[0010] In one embodiment, historical clinical data may be used to identify biomarkers that can characterize and predict AVF maturation and failure prior to access creation. In various embodiments, biomarker profiles generated from the historical clinical data can be expanded to distinguish between cardiovascular and vascular access complications.

[0011] In some embodiments, identified biomarkers are provided for predicting AVF outcomes (“maturation biomarkers” or “maturation factors”). In various embodiments, the maturation biomarkers include metabolites, for example, plasma metabolites found in patient plasma. In exemplary embodiments, the maturation biomarkers are pre-surgical metabolites measured, determined, or otherwise detected prior to an AVF creation surgery. In some embodiments, the maturation biomarkers are identified based on being differentially regulated between patients with successful or failed AVF maturation.

[0012] In some embodiments, maturation biomarkers may include metabolites linked to the metabolism of lipids, amino acids, and / or starches. In various embodiments, a maturation biomarker includes an acylcarnitine. In various embodiments, a maturation biomarker may include at least one of decanoyl-L-carnitine, acetylcarnitine, 2-methylbutyroylcarnitine, L-valine, L- hexanoylcarnitine, N-methylanthranilic acid, 1-palmitoyllysophosphatidyl choline, L-(+)- ergothioneine, or D-maltose.

[0013] In various embodiments, maturation biomarker information may be determined for one or more maturation biomarkers. Maturation biomarker information may include the presence / absence of the maturation biomarker, for instance, in patient fluids such as plasma. Maturation biomarker information may include the concentration of the maturation biomarker, regulation information (e g., up-regulation, down-regulation, concentration increase, concentration decrease, etc.), and / or the like. In some embodiments, vascular access maturation predictions may be based on maturation biomarker information.BRIEF DESCRIPTION OF THE DRAWINGS

[0014] By way of example, specific embodiments of the disclosed machine will now be described, with reference to the accompanying drawings, in which:

[0015] FIG. 1 illustrates an aspect of the present disclosure in accordance with one embodiment.

[0016] FIG. 2A illustrates a metabolite feature cluster in accordance with one embodiment.

[0017] FIG. 2B illustrates an aspect of the present disclosure in accordance with one embodiment.

[0018] FIG. 3 includes Table 1 that illustrates a demographic baseline by AVF maturation in accordance with one embodiment.

[0019] FIG. 4 includes Table 2 that illustrates a demographic baseline by metabolomic cluster in accordance with one embodiment.

[0020] FIG. 5 includes Table 3 that illustrates univariate and multivariate logistic regression models in accordance with one embodiment.

[0021] FIG. 6 illustrates an aspect of the present disclosure in accordance with one embodiment.

[0022] FIG. 7A illustrates an aspect of the present disclosure in accordance with one embodiment.

[0023] FIG. 7B illustrates an aspect of the present disclosure in accordance with one embodiment.

[0024] FIG. 8 is a data table illustrating an aspect of the present disclosure in accordance with one embodiment.

[0025] FIG. 9 illustrates an aspect of the present disclosure in accordance with one embodiment.

[0026] FIG. 10 illustrates an aspect of the present disclosure in accordance with one embodiment.

[0027] FIG. 11 illustrates changes of metabolite levels after the AVF surgery for cohort patients for the maturation biomarkers in accordance with one embodiment.

[0028] FIG. 12A provides additional details regarding biomarkers in accordance with one embodiment.

[0029] FIG. 12B provides additional details regarding biomarkers in accordance with one embodiment.

[0030] FIG. 12C provides additional details regarding biomarkers in accordance with one embodiment.

[0031] FIG. 12D illustrates a process of oxidative stress in accordance with one embodiment.

[0032] FIG. 13 illustrates unsupervised hierarchical clustering of pre-surgery patient plasma samples based on their metabolomic features in accordance with one embodiment.

[0033] FIG. 14 includes Table 4 that illustrates a demographic baseline by metabolomic clusters in accordance with one embodiment.

[0034] FIG. 15 illustrates differential pre-surgery plasma metabolites between AVF maturation success and failure groups in accordance with one embodiment.

[0035] FIG. 16 includes Table 5 and demonstrates differentially abundant pre-surgery plasma metabolites in AVF maturation success versus failure group in accordance with one embodiment.

[0036] FIG. 17A illustrates unsupervised cluster analysis of patient plasma samples in accordance with one embodiment.

[0037] FIG. 17B illustrates variable importance represented by coefficients of the predictor metabolites in the Lasso logistic regression model in accordance with one embodiment.

[0038] FIG. 17C illustrates receiver operating characteristic analysis in accordance with one embodiment.

[0039] FIG. 18A illustrates Pearson correlation analysis between the listed clinical and demographic variables and the predictor metabolites in accordance with one embodiment.

[0040] FIG. 18B illustrates variable importance represented by coefficients for the retained variables in the Lasso logistic regression model in accordance with one embodiment.

[0041] FIG. 18C illustrates receiver operating characteristic analysis in accordance with one embodiment.

[0042] FIG. 19A is a block diagram of a processing flow in accordance with one embodiment.

[0043] FIG. 19B is a block diagram of a processing flow in accordance with one embodiment.

[0044] FIG. 20 is a block diagram of a computing architecture in accordance with one embodiment.DETAILED DESCRIPTION

[0045] Hemodialysis (HD) is a life-sustaining therapy for patients with advanced kidney failure. Delivery of high-quality HD is the key to the well-being of patients experiencing renal complications, such as end-stage kidney disease (ESKD) patients. HD requires at least one vascular access (VA) site for accessing the circulatory system of a patient. The success of HD is highly dependent on the quality of the VA.

[0046] AVFs are regarded as the preferred VA for most HD patients. In general, an AVF is a vein surgically connected to an artery. AVFs are preferred because of higher delivered-blood flows, low infection risk, and access longevity compared with other methods, such as using a catheter. In addition, HD patients with AVFs as VA have lower morbidity and mortality and may incur lower long-term costs. Notwithstanding the knowledge that AVFs are a superior form of VA, emphasis on AVF usage by all major clinical practice guidelines, and significant policy interventions such as the “fistula first” policy, AVFs rates have not risen significantly. In fact, there has been a significant rise in HD catheter usage, especially in incident HD patients, leading to worse patient outcomes.

[0047] Although AVFs are used as examples in the present disclosure, embodiments are not so limited as this is for illustrative purposes only. For example, outcomes for other types of VAsand / or blood vessel procedures may be determined according to some embodiments, including, without limitation, AVGs, CVCs, bioengineered vessels, and / or the like.

[0048] Practice patterns and poor maturation rates are the most important causes of low AVF usage in HD programs. VA planning for an individual is mostly reliant on a traditional rather than a personalized approach. Conventional AVF analysis techniques include vein mapping and ultrasounds, which are highly subjective and error prone, and, as a result, have been unsuccessful for monitoring and improving primary AVF patency.

[0049] Lack of availability of clinical tools or biomarkers that can reliably predict AVF maturation outcomes, especially prior to AVF creation, is a significant reason behind the lack of personalized VA planning. Predicting AVF outcomes has become more important in light of recent advances, which have expanded the available therapeutic options for arteriovenous access creation. However, without an accurate and reliable process for predicting the outcome of an AVF, healthcare professionals are not interested in increasing AVF adoption due to the risk of unknown outcomes.

[0050] After surgical (or endovascular) creation, AVFs need to undergo maturation before they can be cannulated safely and repeatedly. AVF maturation entails complex physiological and morphological changes and an intricate interplay between various biological mechanisms and processes. Within a week of surgical anastomosis creation, in a successful AVF, the arterial flow increases, for example, about 20-fold. Increase in flow is coupled with arterial and venous dilatation and remodeling. These changes, in the early period, are largely driven by increases in blood flow velocity and wall shear stress which leads to release of vasodilators such as endothelial nitrous oxide (eNO). Further dilatation and remodeling of the feeding artery requires breakdown of elastic lamina which is mediated by reactive oxygen species and matrix metalloproteinases, which is also eNO dependent. Vein dilatation and wall thickening are the clinically more apparent and desirable changes in an AVF, but have been studied to a much lesser extent. Wall thickening in veins is characterized by neo-intimal hyperplasia (NIH). Successful maturation requires an optimal balance between NIH and dilatation.

[0051] Successful AVF maturation may be determined based on various maturation factors. One non-limiting example of a maturation factor includes use of the AVF as vascular access for HD with delivered blood flows above a threshold, such as greater than 250 ml / min, at or above 300 ml / min, or other blood flow thresholds. Another non-limiting example of a maturation factor includes use of the AVF for the delivery of a prescribed blood flow rate on >75% of dialysissessions during a 4-week period. An additional non-limiting example of a maturation factor includes an AVF diameter of greater than 4 mm, a blood flow greater than or equal to 500 ml / min (for instance, quantified by medical analysis, such as via Doppler ultrasound). A further nonlimiting example of a maturation factor includes an AVF being deemed clinically ‘mature’ or ‘usable for dialysis’ via inspection by a healthcare professional.

[0052] However, other, unknown biological and physiological processes and patient characteristics operate to determine the success of AVF maturation, leading to a complex interplay of factors that decides the eventual AVF outcome of an individual patient. Existing AVF treatment, planning, and implantation procedures are not able to take advantage of patient characteristics to improve patient outcomes. Accordingly, more than half of surgically created AVFs fail to develop into mature, usable VAs. Unsuccessful AVF maturation is a major factor contributing to complications in HD patients with an AFV and low AVF prevalence among the population of patients, despite the many advantages of AVF adoption.

[0053] The biology of AVF maturation is ill-defined and not well understood. Accordingly, existing processes are not able to provide patient care targeted to improve AVF maturation outcomes, particularly personalized to a specific patient and their unique biology. Therefore, techniques to improve the understanding of AVF maturation biology could provide evidence-based interventions to improve maturation rates, thereby increasing AVF use and improving patient outcomes for patients with an AVF.

[0054] Accordingly, some embodiments provide an AVF analysis process configured to access patient biological information to provide a targeted, personalized treatment, treatment recommendations or plans, AVF implantation processes, drug regimens, and / or the like to facilitate maturation of an AVF for a specific patient and / or to determine AVF candidates.

[0055] In various embodiments, the patient biological information is or includes metabolic information of the patient. Metabolomics refers to a high throughput approach to measure small molecular metabolites in biological samples. These metabolites are intermediates of biochemical pathways and processes and they play a fundamental role in determining the phenotype of an individual. Metabolomics has not been effectively applied in the field of human AVF maturation.

[0056] In some embodiments, the AVF analysis process may use data models to determine metabolomic profiles of patients associated with AVF outcomes. For example, training data may be determined from a clinical study of patients with known AVF outcomes. A non-limitingexample of a clinical study is the Manchester Vascular Access study (MANVAS) performed as a prospective multi-center observational study designed to investigate the natural history and maturation of newly created AVFs (Nikam M., “Clinical Investigation of the Arteriovenous Access for Haemodialysis,” Doctoral thesis, The University of Manchester (2014), the contents of which are incorporated by reference in the present disclosure). In some embodiments, the training data may be or may include supervised data and / or unsupervised data. The training data may be the result of a metabolomics analysis of plasma samples collected prior to AVF creation of patients in the clinical study to identify metabolomic patterns associated with AVF outcomes. The training data may be used to learn, determine, or otherwise analyze metabolomic patterns associated with certain AVF outcomes.

[0057] A metabolomic analysis of a new patient may be performed and analyzed via the data models trained or otherwise configured based on the training data. The data model may receive the patient’s metabolomic profile as input and generate an AVF prediction as output. In some embodiments, the AVF prediction may provide a prediction of the AVF success of the patient, such as a numerical score, risk of failure, and / or the like. In various embodiments, the AVF prediction may be used to determine candidates for an AVF, for instance, patients with a score over a threshold value may be recommended for an AVF. In some embodiments, the AVF prediction may include a treatment recommendation or treatment plan, for example, procedure recommendations, drug regimens, and / or the like that may be used to treat the patient to facilitate a successful AVF outcome. In various embodiments, a medical method includes treating a patient according to one or more treatment recommendations, treatment plans, drug regimens, AVF implantation procedures, and / or the like that may be used to treat the patient to facilitate a successful AVF outcome.

[0058] In various embodiments, the AVF analysis process may use a profile library, heatmaps, fingerprints, and / or the like of biomarker information of patients with known AVF outcomes. The biomarker information of a new patient (e.g., metabolomic information from patient plasma) may be compared with the profile library to determine which AVF category or cluster the new patient belongs to. For example, a patient may have a biomarker or biomarker pattern that matches a first cluster of patients with known AVF outcomes. A prediction may be determined that the patient may likely have a similar AVF outcome to the outcome of the first cluster.

[0059] In some embodiments, the AVF analysis process may include or may be combined with machine learning (ML) techniques, including, without limitation, artificial intelligence (Al) processes, neural networks (NN), and / or the like. For example, patient metabolomic and patient characteristics (e.g., physical characteristics, demographic information, medical history, health conditions, and / or the like) may be used in ML / Al applications to analyze, predict, or otherwise determine AVF predictions and / or to determine a recommended treatment or other course of action based on a patient profile. In various embodiments, data models may include patient profile computational models (e.g., regression models, ML processes, Al processes, neural networks (NNs), convoluted neural networks (CNNs), and / or the like). In some embodiments, for example. ML / AI processes may correlate specific metabolomic patterns with specific AVF maturation success or failure outcomes.

[0060] For example, in some embodiments, ML / AI algorithms, processes, and / or the like may be used to learn the optimal parameters of the predictive model by investigating past examples with known inputs and known outputs. After training, the predictive model can be used to make predictions on unseen inputs (i.e., generalization). For example, AVF analysis processes may involve a classification supervised learning problem in which the output belongs to a set of distinct classes (e g., AVF maturation outcomes). Non-limiting types of ML algorithms for building predictive models according to some embodiments may include, without limitation, logistic regression, univariate and / or multivariate logistic regression, tree-based methods, Random Forest methods, Gradient Boosting methods, deep learning (DL) algorithms such as Recurrent Neural Networks (RNNs), which process sequence of input, and / or the like. Embodiments are not limited in this context.EXAMPLE CASE STUDY 1

[0061] In Example Case Study 1, a population of patients implanted with AVFs were used to determine training data. The population of patients were from MANVAS, which included a total of 170 patients with planned upper limb AVF for HD and who were able to attend appointments arranged by the study team.

[0062] Successful AVF maturation was defined as use as vascular access for HD with delivered blood flows >300 ml / min or the delivery of the prescribed blood flow rate on >75% of dialysis sessions during a 4-weeks period, or an AVF diameter of >4 mm and a blood flow > 500 ml / min.

[0063] Pre-surgery information and biological samples were collected, including patient baseline demographic information, comorbidity burden, access, dialysis history, medication history, and social history were documented, and plasma samples were collected. Participants would typically undergo four post-surgery visits at 2, 6, 12, and 24 weeks. During post-surgery visits, the AVF was reviewed clinically and blood samples were collected, the anatomy of the AVF and corresponding vessels were analyzed and arterial and AVF blood flows were measured.

[0064] A subset of these plasma samples was used for metabolomic analysis (deceased patients) (the “Cohort” or “Sample Cohort”). FIGS. 2A and 2B depict the results of the metabolomic analysis.Metabolomic Analysis

[0065] Sample preparation for liquid chromatography mass spectrometry (LC-MS) analysis: Cold methanol extraction of metabolites from plasma samples was performed using a Biomek 4000 automated liquid handler (Beckman Coulter, Brea, CA, USA). In general, plasma samples, stored at -80°C, were thawed and aliquoted to 2-D barcoded biobank tubes on a 96-tube rack (Greiner Bio-One, Germany). 25 pL of internal standard solution (20 mg / L [13Cn]-L-tryptophan, 20 mg / L creatinine (N-methyl-D3), 0.2 mg / L thymine (1,3-15N ) and 0.2 mg / L, [13Cs]-indoxyl sulfate) was added to 75 pL plasma sample and mixed well. 150 pL of ice-cold methanol was then added to the solution. The mixture was vortexed at 2,000 rpm for 2 minutes and then centrifuged at 4,300 rpm for 15 minutes at 4°C. 75 pL of the supernatants were transferred to a 96-well plate (Agilent Technologies, Santa Clara, CA) and evaporated to dryness using cold-trap vacuum drier (Labconco, Kansas City, MO, USA). The dry extracts were stored at -80°C and reconstituted with 50% methanol before analysis.

[0066] Chromatographic conditions: Liquid chromatography separation was performed using a reverse phase liquid chromatography (RPLC) column (Phenomenex Luna Cl 8, 4.6x150mm). The column temperature was set to 37 °C and the injection volume was 5 pL. Mobile phases for RPLC are 0.1% formic acid and 10 mM ammonium formate in water (A) and in methanol (B). Metabolites were eluted at a flow rate of 0.75 mL / min. The initial condition was set at 1% phase B. The gradient was as follows: 1% B for 0.5 min followed by a linear gradient to 40% B for 6.5 min and then to 90% B for 1.5 min and to 100% B for 3.5 min. After a stay of 0.5 min at 100% B,the mobile phase was returned to initial condition in 0.5 min and maintained for 3 min until the end of run.

[0067] Mass spectrometry (MS) instrumentation and data acquisition: Samples were analyzed using an Agilent 1290 Infinity II HPLC system coupled to an Agilent 6546 Q-TOF Mass Spectrometer. The Q-TOF is equipped with Jet Stream technology and operated in negative mode. The MS data were acquired between 50 and 1 , 100 m / z at a scan rate of 2 spectra / s in profile mode. Ion source conditions were as follows: gas temperature 350 °C, drying gas 12 L / min, nebulizer 35 psi, fragmentor 51 V, skimmer 65 V and capillary voltage -2000 V in the negative modes. Reference masses 112.9856 (TFA anion) and 1033.9881 (HP-0921 (hexakis(lH, 1H, 3H-tetra- fluoropropoxy)phosphazine) + TFA) were used for internal mass correction during the runs.

[0068] Metabolomic data processing and analysis: For LC-MS data, features (characterized by a unique combination of mass and retention time) were extracted using the MassHunter Profinder Software B.10.0 SP1 (Agilent Technologies, Santa Clara, CA, USA) with the absolute height filter set to 5,000 counts. The features were then selected for further statistical analysis based on the following criteria: 1) coefficients of variation of their abundance in pooled quality control samples were less than 30%; 2) their abundance negatively correlated with dilution fold; 3) ratios of their abundance in solvent blanks over that in pooled quality control samples were less than 0.2; 4) the percentage of missing values was less than 75% across all individual patient samples. The dataset with selected metabolite features was further pre-processed before analysis. Missing values were imputed with half of the minimal value in the data set. The dataset was then subject to log transformation and pareto normalization before hierarchical clustering and principal component analysis (PCA).Statistical Analysis

[0069] Descriptive statistics are presented as mean and standard deviation, or number and percentage, as appropriate. Binary logistic regression was used to compare variables between AVF maturation outcome groups and metabolomics clusters, respectively. Unadjusted and incrementally adjusted logistic regression models were constructed with either AVF maturation success (yes / no) or metabolome clusters (model 1) as binary outcomes. Adjustments were made for age at AVF creation, gender, diabetes, cardiovascular disease, the patient's dialytic status (e.g., CKD stage 5D, yes / no), and use of aspirin. For pre-processing metabolomic data a custom R scriptwas used. Metabolomics data were then analyzed using hierarchical clustering, PCA, and K-means clustering through the R package “tidymodels” (R version 4.2.2, R Foundation for Scientific Computing, Vienna, Austria).Description of the Study Cohort

[0070] FIG. 1 depicts a process flow for selecting patients of the Study Cohort. The original MANVAS cohort comprised 170 patients. Of the deceased patients (n=70), 63 patients had completed the study. Among them, pre-surgery plasma samples were available for metabolomic analysis in 44 patients. This analytic cohort comprised 26 men and 18 women, with an average age of 68 years. Twenty -three patients underwent a brachial AVF creation and 21 had a radial AVF. Successful AVF maturation was observed in 28 patients (63.6%). The AVF maturation success rate did not differ by anatomical location (p=0.39). No difference was observed between the maturation outcome groups in their baseline clinical characteristics, such as demographics, comorbidities, dialysis status at the time of AVF creation, and pre-surgery laboratory tests (see Table 1 of FIG. 3). Aspirin therapy was more common in patients with successful AVF maturation (p=0.04), while outcomes were independent of use of warfarin and clopidogrel, respectively.Cluster Analysis of Metabolomic Features

[0071] Following data pre-processing, a total of 819 metabolite features were identified for further analysis. FIGS. 2A and 2B depict analysis results, for example, generated based on unsupervised hierarchical clustering and principal component analysis of patient plasma samples based on their metabolite features.

[0072] FIG. 2A depicts a heatmap of metabolite feature levels in two clusters (Cluster 1 and Cluster 2) of patient plasma samples as revealed by hierarchical clustering analysis (top dendrogram). The color scale represents metabolite feature levels after the raw data were pre- processed. Example processing includes metabolomics data analysis using unsupervised cluster analysis. The association between metabolomic clusters and AVF maturation outcomes was assessed using unadjusted and incrementally stepwise adjusted logistic regression models. Patient clustering was performed using correlation distance with complete linkage. Metabolite feature clustering used Euclidean distance was performed with complete linkage.

[0073] FIG. 2B depicts a two-dimensional principal component analysis plot. Each dot represents the plasma sample from an HD patient in the Study Cohort and dot colors represent different clusters as shown in FIG. 2A (red for Cluster 1; magenta for Cluster 2). The dots are then separated into three groups based on K-means clustering of their top two principal components. Each group is represented by an ellipse.

[0074] Hierarchical clustering analysis identified two clusters (FIG. 2A). Clusters 1 and 2 comprised 21 and 23 patients, respectively. Further PCA showed that patients were separated into two main groups that mostly overlapped with Clusters 1 and 2, respectively. Notable exceptions were patients P188, P089, and P155, who formed a third, more distant, group (FIG. 2B). While all three patients were members of Cluster 2, they were most distant from other patients in Cluster 2, evidently due to very low levels of several metabolites (FIG. 2A). These three patients were elderly males (mean age 77 years), they were taking aspirin, and had brachial AVFs created. Their primary renal diseases were hypertensive nephropathy (P089, Pl 15) and diabetic nephropathy (P188), respectively. At the time of AVF creation, patients P089 and P188 were already on hemodialysis while patient (P155) was pre-dialysis.Clinical presentation of Two Clusters

[0075] The AVF maturation success rate was significantly higher in Cluster 1 (85.7% vs. 43.5% in Cluster 2; P=0.006). Other clinical variables (e.g., age, gender, diabetes, cardiovascular disease) and drug use (including aspirin) did not differ between the clusters tests (see Table 2 of FIG. 4). A non-significant (p=0.054) trend towards higher reported comorbidity of peripheral vascular disease was found in Cluster 1 (28.6%) compared to Cluster 2 (4.3%) (see Table 2 of FIG. 4).Predictors of AVF Maturation Outcome

[0076] To evaluate the significance of metabolome-associated clusters as a predictor of AVF maturation, data analysis was performed to complement a univariate analysis with three multivariate regression models. Metabolomic cluster assignment was used as an independent predictor in all models tests (see Table 3 of FIG. 5). In multivariate Model 1, in addition to cluster assignment, aspirin use was added as an independent variable, as it was significantly associated with AVF outcomes in univariate analysis tests (see Table 1 of FIG. 3). Key demographics (age and gender) and clinical variables (e.g., diabetes, cardiovascular disease, etc.) were added in Model2. Additionally, CKD stage 5D was incorporated in Model 3 to assess potential impact of ESKD on AVF maturation outcomes. The odds ratios of cluster assignment showed only minor variations as more independent variables were included in the multivariate models. Patients in Cluster 2 consistently had reduced odds of successful AVF maturation compared to those in Cluster 1, irrespective of other independent predictors. Notably, in all models, the odds ratios of successful AVF maturation were increased in patients who used aspirin (see Table 3 of FIG. 5).

[0077] In the Case Study, metabolomic profiling of plasma samples was performed on samples collected prior to AVF surgery in a cohort of 44 patients from the MANVAS study. The Case Study investigated the feasibility of pre-surgeiy serum metabolomes to predict AVF maturation outcomes. The Case Study demonstrated that using plasma metabolomic signatures, patients could be separated into two clusters, and these clusters were statistically significantly associated with successful and unsuccessful AVF maturation, respectively.

[0078] Non-limiting technological advantages of some embodiments as evidenced by the Case study, include providing prospective designs of clinical case studies, use of a well-documented cohort, clinically and radiologically validated endpoints, and a novel metabolomics approach to improve the knowledge of AVF maturation biology. Processes according to some embodiments could increase the ability of healthcare professionals to predict outcomes prior to AVF creation.

[0079] Processes according to some embodiments may operate to identify biomarkers capable of predicting AVF maturation outcomes even before AVF creation, thus allowing personalized vascular access planning. Targeted metabolomic analysis in combination with bioinformatics may provide insight into upstream biological pathways influencing AVF maturation.EXAMPLE CASE STUDY 2

[0080] In Example Case Study 2, specific metabolites associated with AVF maturation were identified. More specifically, metabolites and AVF maturation of a study cohort was analyzed. The study cohort included 28 patients with successful AVF maturation and 16 with AVF failure. 179 metabolites were annotated in pre-surgery plasma samples. FIG. 6 depicts metabolite annotation according to the present disclosure. FIGS. 7A and 7B depict volcano plots of identified metabolites excluding or including drugs, respectively.

[0081] Successful AVF maturation was defined as either adequate HD or a combination of ultrasound features (vein size > 4 mm with AVF flow > 500 ml / min) and clinical assessment. Pre-surgery plasma samples from the day of AVF creation surgery were analyzed by liquid chromatography -mass spectrometry. Metabolites were identified by matching to in-house and METLIN libraries.

[0082] In patients with successful AVF maturation, five metabolites were significantly up- regulated and four metabolites were down-regulated. The metabolites are linked to the metabolism of lipids, amino acids, or starches. Four of the nine metabolites are acylcarnitines, which are responsible for transporting fatty acids into the mitochondria for P-oxidation (suggesting a role of energy metabolism in AVF maturation. AVF maturation outcomes did not correlate with age, gender, presence of diabetes, or presence of cardiovascular disease).

[0083] Nine pre-surgery plasma metabolites, listed in FIG. 8, were identified as maturation biomarkers that were differentially regulated between patients with successful or failed AVF maturation. FIG. 9 depicts graphs of intensity vs. outcome for the maturation biomarkers. FIG. 10 depicts the results of a metabolomic analysis for patients with AVF maturation success / failure for maturation biomarkers. FIG. 11 depicts changes of metabolite levels after the AVF surgery for cohort patients for the maturation biomarkers.

[0084] The maturation biomarkers may be associated with certain physiological functions, such as metabolism of certain compounds. For example, lipid metabolism maturation biomarkers may include decanoyl-L-carnitine, acetylcarnitine, 2-methylbutyroylcarnitine, L-hexanoylcarnitine, and 1-palmitoyllysophosphatidylcholine; amino acid metabolism maturation biomarkers may include L-valine, N-methylanthranilic acid, and L-(+)-ergothioneine; and starch and sucrose metabolism maturation biomarkers may include D-maltose.

[0085] Acylcarnitines are fatty acid metabolites that play important roles in many cellular energy metabolism pathways. They have historically been used as important diagnostic markers for inborn errors of fatty acid oxidation and are being intensively studied as markers of energy metabolism, deficits in mitochondrial and peroxisomal P-oxidation activity, insulin resistance, and physical activity.

[0086] Accordingly, the maturation biomarkers may be different metabolites linked to metabolism of lipid, amino acid and starch, and functionally to energy metabolism and inflammation.

[0087] The maturation biomarkers may be determined as biomarkers (e.g., established as having AV outcome predictive value) or evaluated (e.g., against patient biomarker information to make a prediction or treatment plan for the patient) based on one or more biomarker factors. In oneexample, maturation biomarkers may be determined or evaluated based on their presence / absence in the plasma samples of patients with different AV outcomes. For instance, Biomarker A may be present in patients (e.g., in patient fluid samples) with successful AV maturation and may be absent in patients with unsuccessful AV maturation. In one example, the maturation biomarkers may be determined or evaluated as biomarkers due to their concentration in patients with different AV outcomes. For instance, Biomarker B may have a concentration above a threshold concentration in patients with successful AV maturation and may have a concentration below the threshold concentration in patients with unsuccessful AV maturation. In one example, the maturation biomarkers may be determined as biomarkers due to differential regulation in patients with different AV outcomes. For instance, Biomarker C may be up-regulated (e.g., an increase in concentration after an AV procedure) in patients with successful AV maturation and may not be up-regulated (or may be down-regulated) in patients with unsuccessful AV maturation.

[0088] A VA library, heatmap, fingerprint, data clusters, AI / ML model training data and / or any other information used to evaluate or predict AV maturation outcomes may be based on the one or more various biomarker factors. For example, up-regulation of Biomarker D after AV implantation may have predictive value. Accordingly, the regulation of Biomarker D of a patient following an AV implantation procedure may be compared to a VA library that includes Biomarker D regulation information of a population of patients with known AV maturation outcomes to form a VA prediction for the patient.

[0089] FIGS. 12A-12D provide additional details on the above biomarkers as well as a diagram on oxidative stress. FIG. 12A provides additional details regarding biomarkers Decanoylcarnitine (C 10:0), and 2-Methylbutyrylcamitine (C5:0 M). FIG. 12B provides additional details regarding biomarkers Acetyl carnitine (C2:0) and Caproylcarnitine (C6:0). For FIG. 12A and FIG. 12 B, see also Dambrova M, Makrecka-Kuka M, Kuka J, et al. Acylcarnitines: Nomenclature, Biomarkers, Therapeutic Potential, Drug Targets, and Clinical Trials. Hakkola J, ed. Pharmacol Rev. 2022;74(3):506-551. doi: 10.1124 / pharmrev.121.000408. FIG. 12C provides additional details regarding biomarker LPC 16:0. See Liu P, Zhu W, Chen C, et al. The mechanisms of lysophosphatidylcholine in the development of diseases. Life Sciences. 2020;247: 117443. doi: 10.1016 / j.lfs.2020.117443.EXAMPLE CASE STUDY 3

[0090] In Example Case Study 3, 44 patients were studied, of whom 28 experienced successful fistula maturation, as described in Example Case Study 1. The description of Example Case Study 3 shares some similarities with the prior case studies and thus refers to some of the figures described above and included further chemical and statistical analysis. Metabolomic profiles with 2,768 features were correlated with maturation outcomes. Lasso logistic regression identified six metabolites predictive of maturation outcomes, with an area under the receiver operating characteristics curve of 0.917. The six metabolites are associated with cellular bioenergetics and inflammation. The results remained consistent after adjusting for clinical and demographic variables. Untargeted metabolomics analysis was performed on plasma samples collected before arteriovenous fistula creation in patients from MANVAS. Successful fistula maturation was defined as either adequate hemodialysis using the newly created fistula, or a combination of ultrasound criteria (fistula diameter > 4 mm with a blood flow > 500 ml / min) and clinical assessment. Metabolomics data were analyzed via unsupervised cluster analysis, and Lasso logistic regression was employed to assess associations between metabolites and fistula maturation outcomes.

[0091] Lasso logistic regression is a type of regression analysis that combines logistic regression with LI regularization, also known as the least absolute shrinkage and selection operator (LASSO). This method is used to enhance the prediction accuracy and interpretability of statistical models by performing both variable selection and regularization. The Lasso technique works by adding a penalty equal to the absolute value of the magnitude of coefficients to the loss function, which forces some of the coefficients to be exactly zero. This results in a sparse model where only the most significant variables are retained, making the model simpler and more interpretable.

[0092] In practice, lasso logistic regression is particularly useful when dealing with highdimensional data where the number of predictors exceeds the number of observations. By shrinking less important coefficients to zero, it effectively reduces overfitting and improves the model’s generalization to new data.

[0093] As described previously, in this example case study, a population of patients implanted with AVFs were used to determine training data. The population of patients were from MANVAS, which included a total of 170 patients with planned upper limb AVF for HD and who were able to attend appointments arranged by the study team.

[0094] Successful AVF maturation was defined as (1) use as vascular access for HD with delivered blood flows >300 ml / min, or (2) the delivery of the prescribed blood flow rate on >75% of dialysis sessions during a 4-weeks period, or (3) an AVF diameter of >4 mm, a blood flow > 500 ml / min, both quantitated by Doppler ultrasound and AVF deemed clinically ‘mature’ or ‘usable for dialysis’ by an experienced dialysis nurse.

[0095] On the first pre-surgery visit, patient baseline demographic information, comorbidity burden, access, dialysis history, medication history, and social history were documented. The most recent routine laboratory test results, typically conducted within a one-month period prior to the AVF creation surgery, were also recorded. On the day of surgery, a plasma sample for the study was collected prior to the procedure. Participants would typically undergo four post-surgery visits after 2, 6, 12, and 24 weeks. During post-surgery visits, the AVF was reviewed clinically, and plasma samples were collected and stored in a renal biobank. Per MANVAS protocol, post-surgery doppler ultrasound scans were performed, including measurement of arterial and fistula volume flow.

[0096] In Example Case Study 3, plasma samples collected prior to AVF creation underwent metabolomics analysis as shown in FIG. 1. Only plasma samples from deceased patients were included in this analysis to mitigate potential concerns due to institutional and national privacy regulations.Metabolomics analysis

[0097] Sample preparation for liquid chromatography mass spectrometry (LC-MS) analysis. Cold methanol extraction of metabolites from plasma samples was performed using a Biomek 4000 automated liquid handler (Beckman Coulter, Brea, CA, USA). In brief, plasma samples, stored at -80°C, were thawed and aliquoted to 2-D barcoded biobank tubes on a 96-tube rack (Greiner Bio- One, Germany). Twenty-five pL of internal standard solution (20 mg / L [13Cn]-L-tryptophan, 20 mg / L creatinine (N-methyl-Da), 0.2 mg / L thymine (1,3-15N2) and 0.2 mg / L, [13Ce]-indoxyl sulfate) was added to 50 pL plasma sample and then mixed with 150 pL of ice-cold methanol. The mixture was vortexed at 2,000 rpm for 2 min and then centrifuged at 4,300 rpm for 15 min at 4°C. 75 pL supernatant was transferred to a 96-well plate (Agilent Technologies, Santa Clara, CA) and evaporated to dryness using cold-trap vacuum drier (Labconco, Kansas City, MO, USA). The dry extracts were stored at -80°C until analysis.

[0098] Chromatographic conditions. Liquid chromatography separation was performed using a reverse phase liquid chromatography (RPLC) column (PhenomenexLuna C18, 4.6x 150 mm). The column temperature was set to 37 °C and the injection volume was 5 pL. Mobile phases for RPLC were 0.01% formic acid and 1 mM ammonium formate in water (A) and in methanol (B). Metabolites were eluted at a flow rate of 0.75 mL / min. The initial condition was set at 1% phase B. The gradient was as follows: 1% B for 0.5 min followed by a linear gradient to 40% B for 6.5 min and then to 90% B for 1.5 min and to 100% B for 3.5 min. After 0.5 min at 100% B, the mobile phase was returned to initial condition within 0.5 min and maintained for 3 min until the end of run.

[0099] Mass spectrometry (MS) instrumentation and data acquisition. Samples were analyzed using an Agilent 1290 Infinity II HPLC system coupled to an Agilent 6546 Q-TOF mass spectrometer. The Q-TOF was equipped with Jet Stream technology and data were acquired in both positive and negative ionization modes between 50 and 1,100 m / z at a scan rate of 2 spectra / s in profile mode. Ion source conditions for positive mode were as follows: gas temperature 350 °C, drying gas flow 12 L / min, nebulizer 35 psi, fragmentor 125 V, skimmer 65 V, and capillary voltage 4000 V. For negative mode, ion source conditions were the same except fragmentor 51 V and capillary voltage -2000 V. Iterative MS / MS spectra were acquired at a scan rate of 4 spectra / s and under collision energies of 10 V, 20 V, and 40 V.

[0100] The samples were randomized prior to analysis. Additionally, a pooled quality control sample - prepared by combining small aliquots from the study samples - was injected after every eighth sample to monitor signal stability of the instrument during the analysis.

[0101] Metabolomics data processing and analysis. For LC-MS data, features (characterized by a unique combination of accurate mass and retention time) were extracted using the MassHunter Profinder Software B.10.0 SP1 (Agilent Technologies, Santa Clara, CA, USA) with the absolute height filter set to 5,000 counts. The features were then selected for further statistical analysis based on the following criteria: 1) coefficients of variation of their abundance in pooled quality control samples were less than 30%; 2) their abundance in diluted pooled quality control samples negatively correlated with the dilution fold; 3) their mean abundance in solvent blanks were less than 20% of that of pooled quality control samples; 4) the percentage of missing values was less than 75% across all individual patient samples.

[0102] The dataset with selected metabolomic features was further pre-processed as follows before analysis. Missing values were imputed with one fifth of the minimal value for the corresponding features in the dataset. The dataset was then subjected to base 2 log transformation and pareto normalization before hierarchical clustering.

[0103] Metabolite identification. Metabolites were identified based on acquired LC-MS data, including accurate mass, retention time, isotopic pattern, and MS / MS spectra. MassHunter Qualitative Analysis software was employed to match this data against our in-house metabolite library and the MassHunter METLIN library (Agilent Technologies, Santa Clara, CA). Our inhouse library contains data on 703 chemical standards (MetaSci, Toronto, Canada), which were obtained using the same LC-MS method and instruments as those used for the study samples. Metabolites identified by matching accurate mass (±5 ppm), retention time (±15 s), and MS / MS spectra using our in-house library, and those identified using the METLIN library without retention time, are considered as Metabolomics Standards Initiatives (MSI) level 1 and level 2 identifications, respectively.Statistical analysis

[0104] Descriptive statistics were calculated as mean and standard deviation, or number and percentage, as appropriate. The Wilcoxon rank sum test was used to compare the mean intensity differences of metabolites between AVF maturation success and failure groups. Metabolites were considered as significantly differential if p < 0.05 and fold change > 1.3 or fold change < 0.7. The hierarchical clustering analysis was performed and the resulting heatmap plotted using the R ComplexHeatmap package. Lasso logistic regression classifier was trained with leave- one-out cross validation using the R glmnet package. The Lasso model-based receiver operating characteristics (ROC) analysis was performed for selected variables and ROC curves were plotted using the R pROC package. Performance difference between two models were assessed by comparing their area under the receiver operating characteristic curves using DeLong’s test. All analyses were performed in R version 4.2.2 (R Foundation for Scientific Computing, Vienna Austria).Description of the study cohort

[0105] The MANVAS cohort as described above and in FIG. 1 comprised 170 patients. Of the deceased patients (n=70), 63 patients had completed the study. Among them, pre-surgery plasma samples were available for metabolomics analysis in 44 patients. This final cohort comprised 26 men and 18 women, with an age of 68.0 ± 13.4 years, 34 were Caucasians, and 21 had diabetes. Twenty-three patients underwent a brachial AVF creation and 21 had a radial AVF. Fourteen patients were on HD prior to AVF creation. Thirteen patients underwent dialysis via a catheter, while one patient utilized a brachial AVF. Successful AVF maturation was observed in 28 patients (63.6%), and three of them had secondary maturation procedures: one required a surgical refashioning of the anastomotic segment; one had one percutaneous angioplasty; and one had two percutaneous angioplasties. The AVF maturation success rate did not differ by AVF anatomical location (p = 0.39). Maturation outcomes did not differ by baseline clinical characteristics, demographics, comorbidities, dialysis status at the time of AVF creation, and presurgery routine laboratory results. See Table 1 from FIG. 3 above. Aspirin therapy was more common in patients with successful AVF maturation (p = 0.04), while outcomes were independent of the use of warfarin and clopidogrel.Cluster analysis of metabolomic features

[0106] Following LC-MS analysis and data processing, a total of 2,768 metabolomic features were observed, including 830 in positive mode and 1,938 in negative mode. As shown in FIG. 13, hierarchical clustering analysis of metabolomic profdes identified two clusters. Clusters 1 and 2 comprised 21 and 23 patients, respectively. The AVF maturation success rate was significantly higher in Cluster 1 (81.0% vs. 47.8% in Cluster 2; p = 0.03). As shown in FIG. 14 - Table 4, clinical variables (including use of aspirin) did not differ between the two clusters. The cluster analysis results indicate that pre-surgery plasma metabolomes are associated with AVF maturation outcomes.Differentially regulated metabolites in pre-surgery plasmas in patients with matured AVF or failed AVF

[0107] In the next step, identification of metabolites that are associated with AVF maturation outcomes were sought. 156 metabolites were annotated with MSI level 1 or level 2identification (see Methods). In patients with successful AVF maturation, five annotated metabolites were significantly up-regulated, and four were down-regulated, as shown at FIG. 15 and FIG. 16 - Table 5. As expected, levels of the common uremic solutes (e.g., creatinine, indoxyl sulfate, p-cresol, hippuric acid, trimethyl N-oxide) were unrelated to maturation outcomes.Predictive modeling analysis of AVF maturation outcomes

[0108] As shown at FIG. 17A, unsupervised cluster analysis revealed that nine plasma metabolites significantly distinguished AVF maturation outcomes (FIG. 17A, p = 0.002). To further assess the predictive value of these metabolites for AVF maturation outcomes, a Lasso logistic regression classifier was developed using leave-one-out cross-validation to classify AVF maturation outcomes in all 44 patients. In the final model, feature selection reduced the number of predictive metabolites from nine to six (decanoyl -L-carnitine, 2-methylbutyroylcarnitine, L- valine, 1-palmitoyllysophospgatidylcholine, N-methylanthranilic acid, and D-maltose; FIG. 17B). The area under the receiver operating characteristic curve (AUROC) was 0.917 (95% CI: 0.833 - 1.000; See FIG. 17C).

[0109] Then it was investigated whether additional key demographic and clinical variables could improve the performance of the metabolite-based model. These variables include patient age, sex, ethnicity, diabetes status, cardiovascular disease status, chronic kidney disease (CKD) Stage 5D, and aspirin prescription. Decanoyl-L-camitine and age were found to be moderately correlated (r = 0.52; FIG. 18A). Subsequently the six metabolites with these key demographic and clinical variables were combined to train a Lasso logistic regression classifier. In the final model, aspirin use and age were retained together with the six metabolites (FIG. 18B), resulting in non- significantly (p = 0.465) higher AUROC (0.955; 95% CI: 0.901 - 1.000; See FIG. 18C).Discussion

[0110] Through metabolomic profiling of plasma samples collected prior to AVF creation, six plasma metabolites predictive of AVF maturation outcomes were identified. These metabolites are linked to bioenergetics and inflammation. The results remained consistent after adjusting for clinical and demographic variables, highlighting the potential of these metabolites as biomarkers for AVF outcomes.

[0111] Two of these six metabolites, decanoyl-L-carnitine and 2-methylbutyroylcarnitine, are acylcarnitines that are essential for energy metabolism. Acylcarnitines facilitate the transport of acyl groups to the mitochondria for 0-oxidation, a critical energy-producing pathway, particularly in muscle tissue. The prominence of acylcarnitine among these metabolites suggests a potential link between cellular energy metabolism and AVF maturation. Prior studies have associated acylcarnitine-related energy metabolism with various inherited metabolic disorders, diabetes, and cardiovascular diseases. Elevated serum acylcarnitine levels are common in CKD patients, likely due to impaired renal excretion. In our study, lipid-derived acylcarnitines, such as decanoyl-L-carnitine, were elevated in patients with successful AVF maturation. In contrast, 2- methylbutyroylcarnitine, an amino acid-derived acylcarnitine, was higher in patients with failed AVF maturation. It is interesting to note that high amino acid-derived acylcarnitines have been linked to progression to end stage kidney disease (ESKD) in patients with type 2 diabetes nephropathy. How these different classes of acylcamitine connect pathophysiologically to AVF maturation outcomes remains to be investigated.

[0112] Another predictive metabolite, the branched-chain amino acid (BCAA) valine, is also linked to energy metabolism through its roles in tricarboxylic acid cycle and gluconeogenesis. Plasma levels of BCAAs are generally lower in CKD patients compared to healthy individuals possibly due to metabolic acidosis, a common feature of CKD that promotes protein breakdown and increases BCAA catabolism in muscle and liver tissues. In our study, pre-surgery plasma valine levels were significantly higher in the AVF success group, while leucine and isoleucine, the other two BCAAs, showed no significant differences. Notably, low plasma valine levels in uremic patients cannot be effectively corrected by dietary valine supplementation, unlike leucine and isoleucine. This discrepancy may reflect distinct metabolic or regulatory pathways among BCAAs in kidney patients. Elevated valine levels in the success group suggest a possible role in better supporting cellular protein synthesis, endothelial function, and vascular remodeling during AVF maturation.

[0113] Elevated levels of 1-Palmitoyllysophosphatidylcholine (LPC 16:0) in pre-surgery plasma were associated with AVF failure. This metabolite belongs to lysophosphatidylcholine (LPC), a primary component of oxidatively modified low-density lipoprotein. LPC exerts diverse biological effects on various cell types such as endothelial cells, vascular smooth muscle cells, and immune cells. Studies have shown that LPC induces the production of reactive oxygen species inboth endothelial cells and vascular smooth muscle cells, contributing to oxidative stress. It also promotes the release of pro-inflammatory cytokines such as IL-6 and IL-8 in endothelial cells and triggers protein kinase A-mediated vascular calcification, reducing the elasticity of smooth muscle cells. Moreover, LPC inhibits endothelial nitric oxide synthase, leading to reduced nitric oxide levels and impaired endothelial function. High plasma LPC levels have been shown to correlate with vascular damage in various pathological conditions, including atherosclerosis and hypertension. Therefore, elevated LPC 16:0 levels could contribute to vascular dysfunction and affect AVF maturation by exacerbating inflammation, oxidative stress, and vascular stiffness.

[0114] N-Methylanthranilic acid, a derivative of anthranilic acid and commonly found in plants, was associated with AVF failure in our study, whereas anthranilic acid itself did not show significant differences between the two groups. Anthranilic acid is a key intermediate in the kynurenine pathway of tryptophan metabolism. This pathway is crucial for immune regulation and inflammation and has been implicated in numerous conditions, including autoimmune diseases, metabolic disorders and cardiovascular diseases. Elevated plasma levels of kynurenine pathway metabolites have been linked to systemic inflammation. Whether N-methylanthranilic acid plays a role comparable to anthranilic acid remains to be determined.

[0115] Higher plasma maltose levels were observed in the AVF failure group. Maltose, a disaccharide composed of two glucose molecules, serves as an intermediate product of starch and glycogen digestion in humans. It is found in foods such as wheat, com, barley, and rye. Under normal physiological conditions, maltose is barely detectable in the bloodstream because it is rapidly hydrolyzed into glucose by maltase enzymes in the intestinal mucosa. Furthermore, maltose cannot be absorbed by the intact intestinal mucosa. However, in patients with ESKD, a condition often associated with increased gut permeability or a "leaky gut," this barrier function is compromised. This phenomenon is primarily driven by chronic inflammation. As a result, maltose may translocate from the intestinal lumen into the bloodstream through the now-permeable intestinal barrier. The elevated maltose levels in ESKD patients is interpreted as an epiphenomenon of increased gut permeability rather than a direct causal factor in AVF failure. The higher plasma maltose levels observed in the AVF failure group compared to the AVF success group may indicate greater gut permeability, potentially reflecting increased systemic inflammation, which could adversely affect AVF maturation.

[0116] Among clinical variables, only patient age and aspirin use were retained in the final prediction model. However, their contributions were minimal compared to those of the metabolites. This finding corroborates current literature, which indicates no consistent associations between clinical characteristics and AVF maturation outcomes. In addition, the evidence regarding the impact of antiplatelet drugs, such as aspirin and clopidogrel, on AVF maturation outcomes remains inconclusive.

[0117] Overall, the results suggest a prominent role of bioenergetics and inflammation for AVF maturation. Consistent with this notion, two additional lipid-derived acylcarnitines (acetylcamitine and L-hexanoyl carnitine) and the antioxidant L-(+)-ergothioneine in pre-surgery plasma showed differences between the AVF maturation success and failure groups in our study, although they were not included in the final model. Furthermore, a previous report has linked increased expression of pro-inflammatory genes in pre-access veins to higher risk of AVF maturation failure.

[0118] This example case study has several notable strengths. These include its prospective design, a well-characterized cohort, rigorously validated clinical and radiological endpoints, and the application of an innovative metabolomics approach that utilizes plasma samples collected prior to AVF creation. Notably, this study is the first to demonstrate the potential of plasma metabolites as biomarkers for predicting AVF maturation outcomes, even before AVF creation.

[0119] FIG. 19A illustrates an example of a processing flow in accordance with embodiments described in the present disclosure. More specifically, FIG. 19A illustrates an example of a processing flow for training an AVF prediction model 650. As shown in FIG. 19A, a training module 630 may be configured to access training data 610. In various embodiments, the training data 610 includes real -world clinical data 612 associated with patients receiving a vascular access, such as an AVF. In some embodiments, clinical data 612 includes known AVF outcomes. A non-limiting example of the clinical data 612 includes data the same as or similar to MANVAS data. In some embodiments, the clinical data 612 includes patient information, including demographics, age, gender, health condition(s), physical characteristics, medications, dialysis treatment, and / or the like.

[0120] In some embodiments, the training data 610 includes biomarker data (which may be or may include maturation biomarker information) 614 of patients, for example, associated with the clinical data 612. The biomarker data 614 include targeted metabolomics, untargetedmetabolomics, micro-RNA (miRNA), proteomic markers, and / or vasculature remodeling markers (e.g., following surgical interventions), concentrations thereof, intensities thereof, regulation information thereof, and / or the like.

[0121] The training data 610 may include biomarkers (via the biomarker data 614) associated with known AVF outcomes (via the clinical data 612) such as the data described above in the various case studies.

[0122] The training data 610 is fed into the training module 630 configured to train AVF prediction model 650. The AVF prediction model 650 may include various types of data models, mathematical models, regression models, algorithms, Al models, ML models, and / or the like. The AVF prediction model 650 may be trained to predict the outcome of an AVF (e.g., successful or not successful based on various success criteria; numerical score indicating likelihood of success; success probability; and / or the like) based on patient biomarkers (e.g., MS profile of patient plasma; distribution of metabolomics; biomarker regulation, and / or the like) alone or in combination with patient information (e.g., demographics, age, gender, health condition(s), and / or the like).

[0123] FIG. 19B illustrates an example of a processing flow in accordance with embodiments described in the present disclosure. More specifically, FIG. 19B illustrates an example of a processing flow for determining an AVF prediction. As shown in FIG. 19B, patient information 660 may be fed into a trained AVF prediction model 650. The patient information 660 may include patient information (e.g., demographics, age, gender, health condition(s), and / or the like) and patient biomarkers (e.g., MS analysis of patient plasma, biomarker regulation, and / or the like). The AVF prediction model 650 may operate to generate an AVF prediction 670 configured to indicate an AVF outcome. In some embodiments, the AVF prediction 670 may indicate a probability of an AVF outcome, such as AVF maturation success.

[0124] In various embodiments, the AVF prediction 670 may be configured to provide treatment recommendations for the patient for achieving a successful AVF outcome. For example, the AVF prediction 670 may recommend treatment with drug X with a specified regimen.

[0125] In some embodiments, a healthcare provider 680 may access the AVF prediction 670 for use in treating a patient 685. For example, a doctor may recommend a particular type or location of an AVF based on the AVF prediction. A healthcare professional may treat a patientaccording to the treatment recommendation to implant an AVF and / or to facilitate the successful maturation of an implanted AVF.[00126J FIG. 20 illustrates an embodiment of an exemplary computing architecture 700 suitable for implementing various embodiments as previously described. In various embodiments, the computing architecture 700 may comprise or be implemented as part of an electronic device. In some embodiments, the computing architecture 700 may be representative, for example, of computing device 702 and / or components thereof. The embodiments are not limited in this context.

[0127] As used in this application, the terms “system” and “component” and “module” are intended to refer to a computer-related entity, either hardware, a combination of hardware and software, software, or software in execution, examples of which are provided by the exemplary computing architecture 700. For example, a component can be, but is not limited to being, a process running on a processor, a processor, a hard disk drive, multiple storage drives (of optical and / or magnetic storage medium), an object, an executable, a thread of execution, a program, and / or a computer. By way of illustration, both an application running on a server and the server can be a component. One or more components can reside within a process and / or thread of execution, and a component can be localized on one computer and / or distributed between two or more computers. Further, components may be communicatively coupled to each other by various types of communications media to coordinate operations. The coordination may involve the uni-directional or bi-directional exchange of information. For instance, the components may communicate information in the form of signals communicated over the communications media. The information can be implemented as signals allocated to various signal lines. In such allocations, each message is a signal. Further embodiments, however, may alternatively employ data messages. Such data messages may be sent across various connections. Exemplary connections include parallel interfaces, serial interfaces, and bus interfaces.

[0128] The computing architecture 700 includes various common computing elements, such as one or more processors, multi-core processors, co-processors, memory units, chipsets, controllers, peripherals, interfaces, oscillators, timing devices, video cards, audio cards, multimedia input / output (I / O) components, power supplies, and so forth. The embodiments, however, are not limited to implementation by the computing architecture 700.

[0129] As shown in FIG. 20, the computing architecture 700 comprises a processing unit 704, a system memory 706 and a system bus 708. The processing unit 704 can be any of variouscommercially available processors, including without limitation an AMD® Athlon®, Duron® and Opteron® processors; ARM® application, embedded and secure processors; IBM® and Motorola® DragonBall® and PowerPC® processors; IBM and Sony® Cell processors; Intel® Celeron®, Core (2) Duo®, Itanium®, Pentium®, Xeon®, and XScale® processors; and similar processors. Dual microprocessors, multi-core processors, and other multi-processor architectures may also be employed as the processing unit 704.

[0130] The system bus 708 provides an interface for system components including, but not limited to, the system memory 706 to the processing unit 704. The system bus 708 can be any of several types of bus structure that may further interconnect to a memory bus (with or without a memory controller), a peripheral bus, and a local bus using any of a variety of commercially available bus architectures. Interface adapters may connect to the system bus 708 via a slot architecture. Example slot architectures may include without limitation Accelerated Graphics Port (AGP), Card Bus, (Extended) Industry Standard Architecture ((E)ISA), Micro Channel Architecture (MCA), NuBus, Peripheral Component Interconnect (Extended) (PCI(X)), PCI Express, Personal Computer Memory Card International Association (PCMCIA), and the like.

[0131] The system memory 706 may include various types of computer-readable storage media in the form of one or more higher speed memory units, such as read-only memory (ROM), random-access memory (RAM), dynamic RAM (DRAM), Double-Data-Rate DRAM (DDRAM), synchronous DRAM (SDRAM), static RAM (SRAM), programmable ROM (PROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, polymer memory such as ferroelectric polymer memory, ovonic memory, phase change or ferroelectric memory, silicon-oxide-nitride-oxide-silicon (SONOS) memory, magnetic or optical cards, an array of devices such as Redundant Array of Independent Disks (RAID) drives, solid state memory devices (e.g., USB memory, solid state drives (SSD) and any other type of storage media suitable for storing information. In the illustrated embodiment shown in FIG. 20, the system memory 706 can include non-volatile memory 710 and / or volatile memory 712. A basic input / output system (BIOS) can be stored in the non-volatile memory 710.

[0132] The computer 702 may include various types of computer-readable storage media in the form of one or more lower speed memory units, including an internal (or external) hard disk drive (HDD) 714, a magnetic floppy disk drive (FDD) 716 to read from or write to a removable magnetic disk 718, and an optical disk drive 720 to read from or write to a removable optical disk722 (e.g., a CD-ROM or DVD). The HDD 714, FDD 716 and optical disk drive 720 can be connected to the system bus 708 by a HDD interface 724, an FDD interface 726 and an optical drive interface 729, respectively. The HDD interface 724 for external drive implementations can include at least one or both of Universal Serial Bus (USB) and IEEE 1384 interface technologies.

[0133] The drives and associated computer-readable media provide volatile and / or nonvolatile storage of data, data structures, computer-executable instructions, and so forth. For example, a number of program modules can be stored in the drives and memory units 710, 712, including an operating system 730, one or more application programs 732, other program modules 734, and program data 736. In one embodiment, the one or more application programs 732, other program modules 734, and program data 736 can include, for example, the various applications and / or components of computing device 110.

[0134] A user can enter commands and information into the computer 702 through one or more wire / wireless input devices, for example, a keyboard 738 and a pointing device, such as a mouse 740. Other input devices may include microphones, infra-red (IR) remote controls, radiofrequency (RF) remote controls, game pads, stylus pens, card readers, dongles, finger print readers, gloves, graphics tablets, joysticks, keyboards, retina readers, touch screens (e.g., capacitive, resistive, etc.), trackballs, trackpads, sensors, styluses, and the like. These and other input devices are often connected to the processing unit 704 through an input device interface 742 that is coupled to the system bus 708, but can be connected by other interfaces such as a parallel port, IEEE 994 serial port, a game port, a USB port, an IR interface, and so forth.

[0135] A monitor 744 or other type of display device is also connected to the system bus 708 via an interface, such as a video adaptor 746. The monitor 744 may be internal or external to the computer 702. In addition to the monitor 744, a computer typically includes other peripheral output devices, such as speakers, printers, and so forth.

[0136] The computer 702 may operate in a networked environment using logical connections via wire and / or wireless communications to one or more remote computers, such as a remote computer 749. The remote computer 749 can be a workstation, a server computer, a router, a personal computer, portable computer, microprocessor-based entertainment appliance, a peer device or other common network node, and typically includes many or all of the elements described relative to the computer 702, although, for purposes of brevity, only a memory / storage device 750 is illustrated. The logical connections depicted include wire / wireless connectivity to alocal area network (LAN) 752 and / or larger networks, for example, a wide area network (WAN) 754. Such LAN and WAN networking environments are commonplace in offices and companies, and facilitate enterprise- wide computer networks, such as intranets, all of which may connect to a global communications network, for example, the Internet.

[0137] When used in a LAN networking environment, the computer 702 is connected to the LAN 752 through a wire and / or wireless communication network interface or adaptor 756. The adaptor 756 can facilitate wire and / or wireless communications to the LAN 752, which may also include a wireless access point disposed thereon for communicating with the wireless functionality of the adaptor 756.

[0138] When used in a WAN networking environment, the computer 702 can include a modem 758, or is connected to a communications server on the WAN 754, or has other means for establishing communications over the WAN 754, such as by way of the Internet. The modem 759, which can be internal or external and a wire and / or wireless device, connects to the system bus 708 via the input device interface 742. In a networked environment, program modules depicted relative to the computer 702, or portions thereof, can be stored in the remote memory / storage device 750. It will be appreciated that the network connections shown are exemplary and other means of establishing a communications link between the computers can be used.

[0139] The computer 702 is operable to communicate with wire and wireless devices or entities using the IEEE 802 family of standards, such as wireless devices operatively disposed in wireless communication (e.g., IEEE 802.16 over-the-air modulation techniques). This includes at least Wi-Fi (or Wireless Fidelity), WiMax, and Bluetooth™ wireless technologies, among others. Thus, the communication can be a predefined structure as with a conventional network or simply an ad hoc communication between at least two devices. Wi-Fi networks use radio technologies called IEEE 802.1 lx (a, b, g, n, etc.) to provide secure, reliable, fast wireless connectivity. A WiFi network can be used to connect computers to each other, to the Internet, and to wire networks (which use IEEE 802.3-related media and functions).

[0140] Numerous specific details have been set forth herein to provide a thorough understanding of the embodiments. It will be understood by those skilled in the art, however, that the embodiments may be practiced without these specific details. In other instances, well-known operations, components, and circuits have not been described in detail so as not to obscure theembodiments. It can be appreciated that the specific structural and functional details disclosed herein may be representative and do not necessarily limit the scope of the embodiments.

[0141] Some embodiments may be described using the expression "coupled" and "connected" along with their derivatives. These terms are not intended as synonyms for each other. For example, some embodiments may be described using the terms “connected” and / or “coupled” to indicate that two or more elements are in direct physical or electrical contact with each other. The term "coupled,” however, may also mean that two or more elements are not in direct contact with each other, but yet still co-operate or interact with each other.

[0142] Unless specifically stated otherwise, it may be appreciated that terms such as “processing,” “computing,” “calculating,” “determining,” or the like, refer to the action and / or processes of a computer or computing system, or similar electronic computing device, that manipulates and / or transforms data represented as physical quantities (e.g., electronic) within the computing system’s registers and / or memories into other data similarly represented as physical quantities within the computing system’s memories, registers or other such information storage, transmission or display devices. The embodiments are not limited in this context.

[0143] It should be noted that the methods described herein do not have to be executed in the order described, or in any particular order. Moreover, various activities described with respect to the methods identified herein can be executed in serial or parallel fashion.

[0144] Although specific embodiments have been illustrated and described herein, it should be appreciated that any arrangement calculated to achieve the same purpose may be substituted for the specific embodiments shown. This disclosure is intended to cover any and all adaptations or variations of various embodiments. It is to be understood that the above description has been made in an illustrative fashion, and not a restrictive one. Combinations of the above embodiments, and other embodiments not specifically described herein will be apparent to those of skill in the art upon reviewing the above description. Thus, the scope of various embodiments includes any other applications in which the above compositions, structures, and methods are used.

[0145] Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

[0146] As used herein, an element or operation recited in the singular and proceeded with the word “a” or “an” should be understood as not excluding plural elements or operations, unless such exclusion is explicitly recited. Furthermore, references to “one embodiment” of the present disclosure are not intended to be interpreted as excluding the existence of additional embodiments that also incorporate the recited features.

[0147] The present disclosure is not to be limited in scope by the specific embodiments described herein. Indeed, other various embodiments of and modifications to the present disclosure, in addition to those described herein, will be apparent to those of ordinary skill in the art from the foregoing description and accompanying drawings. Thus, such other embodiments and modifications are intended to fall within the scope of the present disclosure. Furthermore, although the present disclosure has been described herein in the context of a particular implementation in a particular environment for a particular purpose, those of ordinary skill in the art will recognize that its usefulness is not limited thereto and that the present disclosure may be beneficially implemented in any number of environments for any number of purposes.

[0148] Listed herein below are several example embodiments of the present disclosure:

[0149] In some embodiments, the present disclosure describes a method of generating a vascular access (VA) prediction for a patient. In some examples, the method includes determining biomarker information of a patient; and determining a VA prediction based on evaluating the biomarker information with a VA profile library, the VA prediction indicating a likelihood of success of a VA of the patient for dialysis.

[0150] In some embodiments, the VA comprises an arteriovenous fistulas (AVF). In some embodiments, the biomarker information comprises metabolomic data. In some embodiments, the metabolomic data is determined based on mass analysis of plasma of the patient. In some embodiments, the method further comprises determining a treatment recommendation for the AV based on the AV prediction.

[0151] In some embodiments, the method further comprises treating the patient according to the treatment recommendation for the AV. In some embodiments, the method further comprises implanting an AV in the patient based on the treatment recommendation. In some embodiments, the VA profile library comprising biomarker information of a plurality of unknown biomarkers associated with AV maturation success. In some embodiments, the VA profile library comprises clusters of patients with known AVF maturation outcomes clustered based on metabolomic data.

[0152] In some embodiments, determining the biomarker information includes applying a Lasso logistic regression model to metabolomic data of patients.[00153J In some embodiments, the present disclosure describes an apparatus comprising at least one processor and a memory coupled to the at least one processor, the memory comprising instructions that, when executed by the at least one processor, cause or configure the at least one processor to perform various operations. In some embodiments, the at least one processor is caused to or configured to access a computational model trained using training data to determine a vascular access (VA) prediction indicating a probability of success of the VA based on biomarker information. In some embodiments, the at least one processor is further caused or configured to receive biomarker information of a patient. In some further embodiments, the at least one processor is caused or configured to determine the VA prediction for the patient via providing the biomarker information to the computational model.

[0154] In some embodiments, the training data comprises biomarker information of a population of patients with known VA maturation outcomes. In some embodiments, the biomarker information for the population of patients comprises information for at least one of the following biomarkers: decanoyl-L-camitine, acetylcarnitine, 2-methylbutyroylcarnitine, L-valine, L- hexanoyl carnitine, N-methylanthranilic acid, 1-palmitoyllysophosphatidylcholine (LysoPC 16:0), L-(+)-ergothioneine, or D-maltose. In some embodiments, the biomarker information for the population of patients comprises information for at least one of the following biomarkers: decanoyl-L-carnitine, 2-methylbutyroylcarnitine, L-valine, 1-palmitoyllysophospgatidyl choline, N-methylanthranilic acid, and D-maltose. In some embodiments, the biomarker information comprises regulation information of the biomarkers indicating one of up-regulation or downregulation of at least a portion of the biomarkers.

[0155] In some embodiments of the present disclosure, a method of predicting vascular access (VA) maturation for a patient is described. In some embodiments, the method includes determining patient biomarker information of the patient. In some embodiments, the method includes determining a VA prediction for the patient based on evaluating the patient biomarker information with a VA profile library of biomarker information of a population of patients with known VA maturation outcomes, the VA prediction indicating a likelihood of success of VA maturation of the patient, the biomarker information for the population of patients comprising information for at least one of the following biomarkers: decanoyl-L-carnitine, acetylcamitine, 2-methylbutyroylcamitine, L-valine, L-hexanoylcarnitine, N-methylanthranilic acid, 1- palmitoyllysophosphatidylcholine (LysoPC 16:0), L-(+)-ergothioneine, or D-maltose.[00156J In some embodiments, the biomarker information comprising regulation information of the biomarkers indicating one of up-regulation or down-regulation of at least a portion of the biomarkers. In some embodiments, the patient biomarker information comprising patient biomarker regulation, wherein the VA prediction is determined based on evaluating the patient biomarker regulation with the VA profile library. In some embodiments, the AV is an AV fistula (AVF). In some embodiments, the method further comprises determining a treatment recommendation for the patient based on the VA prediction. In some embodiments, the method further comprises treating the patient according to the treatment recommendation for the AV.

[0157] In some embodiments, the method further comprises implanting an AV in the patient based on the treatment recommendation. In some embodiments, the biomarker information for the population of patients comprises information for at least one of the following biomarkers: decan oyl -L-carnitine, 2-methylbutyroyl carnitine, L-valine, 1-palmitoyllysophospgatidyl choline, N-methylanthranilic acid, and D-maltose. In some embodiments, determining the biomarker information includes applying a Lasso logistic regression model to metabolomic data of patients.

Claims

CLAIMSWhat is claimed is:

1. A method of generating a vascular access (VA) prediction for a patient, the method comprising: determining biomarker information of a patient; and determining a VA prediction based on evaluating the biomarker information with a VA profile library, the VA prediction indicating a likelihood of success of a VA of the patient for dialysis.

2. The method of claim 1, the VA comprising an arteriovenous fistulas (AVF).

3. The method of claim 1, the biomarker information comprising metabolomic data.

4. The method of claim 3, the metabolomic data determined based on mass analysis of plasma of the patient.

5. The method of claim 1, further comprising determining a treatment recommendation for the AV based on the AV prediction.

6. The method of claim 5, further comprising treating the patient according to the treatment recommendation for the AV.

7. The method of claim 5, further comprising implanting an AV in the patient based on the treatment recommendation.

8. The method of claim 1, the VA profile library comprising biomarker information of a plurality of unknown biomarkers associated with AV maturation success.

9. The method of claim 1, the VA profile library comprising clusters of patients with known AVF maturation outcomes clustered based on metabolomic data.

10. The method of claim 1, wherein determining the biomarker information includes applying a Lasso logistic regression model to metabolomic data of patients.

11. An apparatus, comprising: at least one processor; a memory coupled to the at least one processor, the memory comprising instructions that, when executed by the at least one processor, cause the at least one processor to: access a computational model trained using training data to determine a vascular access (VA) prediction indicating a probability of success of the VA based on biomarker information; receive biomarker information of a patient; and determine the VA prediction for the patient via providing the biomarker information to the computational model.

12. The apparatus of claim 11, the training data comprising biomarker information of a population of patients with known VA maturation outcomes.

13. The apparatus of claim 12, wherein the biomarker information for the population of patients comprises information for at least one of the following biomarkers: decanoyl-L- carnitine, acetylcarnitine, 2-methylbutyroylcarnitine, L-valine, L-hexanoylcarnitine, N- methylanthranilic acid, 1-palmitoyllysophosphatidylcholine (LysoPC 16:0), L-(+)-ergothioneine, or D-maltose.

14. The apparatus of claim 13, wherein the biomarker information for the population of patients comprises information for at least one of the following biomarkers: decanoyl-L- carnitine, 2-methylbutyroylcarnitine, L-valine, 1-palmitoyllysophospgatidylcholine, N- methylanthranilic acid, and D-maltose.

15. The apparatus of claim 13, the biomarker information comprising regulation information of the biomarkers indicating one of up-regulation or down-regulation of at least a portion of the biomarkers.

16. A method of predicting vascular access (VA) maturation for a patient, the method comprising: determining patient biomarker information of the patient; and determining a VA prediction for the patient based on evaluating the patient biomarker information with a VA profde library of biomarker information of a population of patients with known VA maturation outcomes, the VA prediction indicating a likelihood of success of VA maturation of the patient, the biomarker information for the population of patients comprising information for at least one of the following biomarkers: decanoyl-L-camitine, acetylcamitine, 2- methylbutyroylcarnitine, L-valine, L-hexanoylcarnitine, N-methylanthranilic acid, 1- palmitoyllysophosphatidylcholine (LysoPC 16:0), L-(+)-ergothioneine, or D-maltose.

17. The method of claim 16, the biomarker information comprising regulation information of the biomarkers indicating one of up-regulation or down-regulation of at least a portion of the biomarkers.

18. The method of claim 17, the patient biomarker information comprising patient biomarker regulation, wherein the VA prediction is determined based on evaluating the patient biomarker regulation with the VA profile library.

19. The method of claim 16, wherein the AV is an AV fistula (AVF).

20. The method of claim 16, further comprising determining a treatment recommendation for the patient based on the VA prediction.

21. The method of claim 20, further comprising treating the patient according to the treatment recommendation for the AV.

22. The method of claim 20, further comprising implanting an AV in the patient based on the treatment recommendation.

23. The method of claim 16, wherein the biomarker information for the population of patients comprises information for at least one of the following biomarkers: decanoyl-L-carnitine, 2- methylbutyroylcarnitine, L-valine, 1-palmitoyllysophospgatidyl choline, N-methylanthranilic acid, and D-maltose.

24. The method of claim 16, wherein determining the biomarker information includes applying a Lasso logistic regression model to metabolomic data of patients.

Images