ABSTRACT
Sepsis is a life-threatening condition and understanding the disease pathophysiology through the use of host immune response biomarkers is critical for patient stratification. Lack of accurate sepsis endotyping impedes clinicians from making timely decisions alongside insufficiencies in appropriate sepsis management. This work aims to demonstrate the potential feasibility of a data-driven validation model for supporting clinical decisions to predict sepsis host-immune response. Herein, we used a machine learning approach to determine the predictive potential of identifying sepsis host immune response for patient stratification by combining multiple biomarker measurements from a single plasma sample. Results were obtained using the following cytokines and chemokines IL-6, IL-8, IL-10, IP-10 and TRAIL where the test dataset was 70%. Supervised machine learning algorithm naïve Bayes and decision tree algorithm showed good accuracy of 96.64% and 94.64%. These promising findings indicate the proposed AI approach could be a valuable testing resource for promoting clinical decision making.
Subject(s)
Algorithms , Biomarkers/analysis , Machine Learning , Sepsis/diagnosis , Bayes Theorem , Case-Control Studies , Clinical Decision-Making , Humans , Reproducibility of ResultsABSTRACT
In this work, we demonstrate a robust, dual marker, biosensing strategy for specific and sensitive electrochemical response of Procalcitonin and C-reactive protein in complex body fluids such as human serum and whole blood for the detection of sepsis. Enhanced sensitivity is achieved by leveraging the physicochemical properties of zinc oxide at the electrode-solution interface. Characterization techniques such as SEM, EDAX, AFM, FTIR and fluorescence microscopy were performed to ensure a suitable biosensing surface. The characteristic biomolecular interactions between the target analyte and specific capture probe is quantified through unique frequency signatures using non-faradaic electrochemical impedance spectroscopy (EIS). The developed biosensor demonstrated a detection limit of 0.10â¯ngâ¯mL-1 for PCT in human serum and whole blood with an R2 of 0.99 and 0.98 respectively. CRP demonstrated a detection limit of 0.10⯵gâ¯mL-1 in human serum and whole blood with an R2 of 0.90 and 0.98 respectively. Cross-reactivity analysis demonstrated robust selectivity to PCT and CRP with negligible interaction to non-specific biomolecules. The novel aspect of this technology is the ability to fine-tune individual biomarkers response owing to the optimal frequency tuning capability. The developed biosensor requires an ultra-low sample volume of 10⯵L without the need for sample dilution for rapid analysis. We envision the developed dual marker biosensor to be useful as a sepsis-screening device for prognostic monitoring.
ABSTRACT
AIM: Development of a label-free multiplexed point-of-care diagnostic device for a panel of cardiac biomarkers - cardiac troponin-T (cTnT), troponin-I (cTnI) and B-type natriuretic peptide (BNP). METHODS: A nonfaradaic electrochemical immunoassay designed with anisotropic high surface area ZnO nanostructures grown using low-temperature hydrothermal methods was selectively immobilized with capture antibodies. Multiplexed detection in human serum using ZnO nanostructures based on complementary electrochemical measurement techniques - electrochemical impedance spectroscopy and Mott-Schottky. RESULTS: Linear signal response for detection of three biomarkers in human serum with dynamic range of 1 pg/ml-100 ng/ml and limit of detection at 1 pg/ml and low signal response to background interferences was achieved. CONCLUSION: First demonstration of simultaneous detection of three cardiac biomarkers in clinically relevant range with sensor's analytical performance and linear response of detection showed potential utility in screening clinical samples for early diagnosis of acute myocardial infarction and chronic heart failure.