High‐Throughput Screening of Metabolic Biomarkers and Wearable Biosensors for the Quantification of Metabolites

Metabolic biomarkers facilitate pathological/physiological investigations and clinic decisions. Noninvasive quantification of metabolic biomarkers allows patients to monitor their health without time, location, and environmental constraints. Recently, the synergistic integration between high‐throughput screening strategy via mass spectrometry and noninvasive quantification techniques via wearable biosensors may provide fundamental insights into the application of metabolic biomarkers in clinical diagnostics. Here, mass spectrometry for the screening of metabolic biomarkers in several major clinical diseases (retinoblastoma, coronary heart disease, lung adenocarcinoma) is introduced. Then, three types of wearable biosensors based on electrochemistry, organic electrochemical transistors, and field effect transistors for the noninvasive quantification of metabolic biomarkers are summarized. The review may serve as a handbook for high‐throughput screening of metabolic biomarkers and wearable biosensors for the quantification of metabolites. It is anticipated that biomarker discovery and quantification will facilitate the next generation of wearable biosensors toward personalized, intelligent, and precise home–healthcare on a large‐scale.


Introduction
As the end products of biopathway, metabolic biomarkers reflected the biological/pathogenic processes in a more terminal pattern than genomic and proteomic strategies for accurate diagnostics. Meanwhile, screening/quantification of metabolic biomarkers holds the promises to precisely monitor even regulate biological/pathogenic processes. [1,2] Coincidentally, partial metabolic biomarkers could be noninvasively induced from human biospecimens (e.g., sweat and tears, etc.), which may afford a noninvasive biosample for wearable biosensors. [3,4] DOI: 10.1002/adsr.202200052 Noninvasive quantification of metabolic biomarkers allows patients to monitor their health without time, location, and environment constraints. [5][6][7] For the economic perspective, the wearable device is expected to grow exponentially in the following years, with more than 20% growth rate annually and the total market is anticipated to be EUR 40 billion per year. [8] We summarized typical applications ( Table 1) of wearable devices in the type of mask, smartphone, and the devices would be assembled as a bracelet, ring, armband around the human body (e.g., face, arm, ear, and waist). Therefore, a comprehensive summary of current achievements and challenges is necessary to understand the principles of designing advanced platform for highthroughput screening and noninvasive quantification of metabolic biomarkers.
For high-throughput screening of metabolic biomarkers, mass spectrometry (MS) has been regarded as the gold standard for quantification of metabolites, due to the superiorities in sensitivity, accuracy, high-throughput, and broad biosample category. [17,18] The mechanism of MS is ionization and fragmentation of sample molecules with gas phase, whose molecular fragmentation with unique structure and weight can be used to quantify targeted molecules. [19,20] Moreover, the matrix assisted laser desorption/ionization mass spectrometry (MALDI) is widely adopted in metabolite detections with enhanced sensitivity via the assistance of nanomaterials as matrices, which demonstrated huge potential for efficient screening of metabolic biomarkers toward practical clinical applications. [21][22][23][24] In this regard, the advanced screening technique for metabolic biomarkers awaits further exploration of selection and optimization of tailored matrix. [25][26][27] For noninvasive quantification of metabolic biomarkers in a wearable manner, biosensors based on electrochemistry (EC), [28,29] organic electrochemical transistors (OECTs), [30,31] and field effect transistors (FETs) [32,33] have been widely applied with the advantages of miniaturization, favorable sensitivity, selectivity, stability, reproducibility. Briefly, a rational design of wearable biosensors opens a patient-oriented strategy for noninvasive quantification of metabolic biomarkers. [34] The principle of electrochemistry relies on the relationship between the electrical configurations and the identical amperometric signal. [35,36] The OECTs and FETs were also branches of the Bracelet device AVA bracelet Measurement of skin temperature and data transmission to an App [9] Ring device Ora ring Measurement of finger temperature and send data to an App [10] Mask Lumos smart sleep mask A device that assists the travellers fight jet lag [11] Armband device BodyMedia sense wear Measurement of the arm temperature and App display [12] Ear device In-ear wearable thermometer Measurement of temperature of ear [13] Smart-watch Garmin approach S20 golf watch Provides the player with useful GPS information [14] Waist device Ran's night self Measurement of temperature of clothes [15] Smart-watch Parkinson smartwatch Monitoring of Parkinson disease [16] electroanalytical method to qualitatively or quantitatively analyze target molecules. [37] The FETs is a sensitive biosensor type whose gate piezoelectric field would regulate the conductivity and charge transport properties of the semiconductor, resulting in various migration ability of holes or electrons in the semiconductor. [38] Meanwhile, the OECTs uses the ion injection from an electrolyte to modulate the conductivity and mobility of organic semiconducting channels, which endows OECTs with higher transconductance compared with that of FETs, but causes the increase of the response time. [39] Briefly, for the noninvasive quantification of metabolites, the main challenges are the low detection limit, specificity, and bio-interfacial adaptation (skinaffinity, biosafety, mechanical compliance, and signal fidelity). [3] Remarkably, integrating micro/nanomaterials with detection techniques has been proved beneficial for clinical diagnosis in the following aspects: i) tunable dimensional effect of micro/nanomaterials allows them to combine with detection platform easily; [1,[40][41][42] ii) attractive characteristics, such as electrochemical properties, large surface ratio, and easy functionalization; [43][44][45] iii) They can sever as capturing/recognizing components, signal reporter/amplifier, electron-transfer mediator, which facilitate the bioanalytical reaction, resulting in enhanced sensitivity. [46,47] In short, the rational design of micro/nanomaterials may open a new avenue for screening and quantification of metabolic biomarkers, by facilitating metabolic analysis platform with improved performances.
Herein, we summarized and described the recent progress in micro/nanomaterials-assisted high-throughput screening and noninvasive quantification of metabolic biomarkers in several major clinical diseases (retinoblastoma, coronary heart disease, and early-stage lung adenocarcinoma, Cushing's disease, Figure  1). The roles of micro/nanomaterials played significant functionality in screening and quantification of metabolic biomarkers when coupled with the techniques of mass spectrometry (MS), electrochemistry (EC), organic electrochemical transistors (OECTs), and field effect transistors (FETs). Accordingly, tremendous strategies including MS, electroanalytical method, optical detection, and paper-based platform have been summarized for the metabolic analysis with their own merits and shortcomings ( Table 2). The integration of MS and wearable biosensor techniques may open a novel avenue for screening and quantification of metabolic biomarkers, embracing the future development of wearable device platforms toward personalized, intelligent, and precise home-healthcare in large-scale. Schematic of micro/nanomaterials, techniques for screening and quantification of metabolic biomarkers. a) Inorganic matrix, Reproduced with permission. [1] Copyright 2022, Wiley-VCH GmbH; b) MXene, Reproduced with permission. [48] Copyright 2021, Wiley-VCH GmbH; c) Hydrogel, Reproduced with permission. [49] Copyright 2021, Elsevier; d) PEDOT:PSS. Reproduced with permission. [50] Copyright 2022, Elsevier. We summarized the roles of micro/nanomaterials in the metabolic analysis including the techniques of mass spectrometry (MS), electrochemistry (EC), organic electrochemical transistors (OECTs), and field effect transistors (FETs). We compared the differences of techniques to bridge the technology gap toward practical high-throughput metabonomics-based wearable platform toward noninvasive healthcare management.

Mass Spectrometry
The metabolites in biofluids may be abnormally secreted, over/under-expressed, modified/degraded by unknown activation of the metabolic pathways, which could be detected via metabonomics. [25,66] The mass spectrometry (MS) profiling combined with advantages of high-throughput is gaining popularity in clinical metabonomics research. [20,[67][68][69] A general workflow includes biosample preparations followed by MS acquisition, processing procedures, and statistical analysis. Then, pronounced www.advancedsciencenews.com www.advsensorres.com  m/z features with obvious discriminations are selected for the identification of biomarkers. In particular, matrix-assisted laser desorption ionization (MALDI) MS technique is commonly used in clinical investigations to screen new potential biomarkers associated with pathological/psychological status. [26,70] The MALDI not only offers various advantages in profiling studies such as simplicity and low consumption of the biosamples, but also enhances detection sensitivity and accuracy due following two aspects: 1) surface plasmon resonance improves laser energy transfer and 2) electronic conductivity facilitates molecular ionization process. [22,71] Herein, we summarized several examples of biomarker screening via MALDI MS toward practical diagnosis of major diseases. The nanomaterials have been widely used as the matrices for the LDI MS detection of metabolites, and comparisons of functional nanomaterials were summarized with unique sizes, elemental compositions, and architectures ( Table 3). [72][73][74][75] As a common intraocular pediatric malignancy, retinoblastoma (RB) that arises in the retina, accounts for almost 10% of cancer in children. [1] Humor biosamples were harvested from retinoblastoma patients and microarrayed with matrix for enhanced MALDI metabolic fingerprinting, and subsequent metabolic biomarkers could be identified via MS values ( Figure  2a). [1] The MALDI metal fingerprinting were performed on inorganic nanoparticles as the matrix with SEM characterizations ( Figure 2b) and high-angle annular dark-field (HAADF) photograph and elemental mapping with O in purple and Fe in green ( Figure 2c) validated the morphology and composition of inorganic nanoparticles. Via the metabolic analysis, normalized intensity of seven metabolites (Nor, Ace, Asp, Lev, Lac, Val, and Prc) were screened and demonstrated pronounced differences between the early and advanced retinoblastoma, which could sever as the potential biomarkers for diagnosis of RB ( Figure 2d). Furthermore, according to selected metabolic biomarkers, the potential pathways to differentiate early and advanced RB were plotted with the p value and pathway impact, including alanine, aspartate and glutamate metabolism, synthesis, and degradation of ketone bodies, and arginine and proline metabolism Mass spectrometry enables screening of metabolic biomarkers for diagnosis of retinoblastoma (RB). a) Schematic for screening metabolic biomarkers for diagnosis of RB. Biosamples were harvested from retinoblastoma patients and microarrayed with matrix for enhanced MALDI metabolic fingerprinting, and subsequent metabolic biomarkers could be identified by MS value. b) The SEM characterization of inorganic nanoparticles as the matrix. The scale bar is 100 nm. c) High-angle annular dark-field (HAADF) photograph and elemental mapping with O in purple and Fe in green validated the composition of inorganic nanoparticles. The scale bar is 200 nm. d) Normalized intensity of seven metabolites (Nor, Ace, Asp, Lev, Lac, Val, and Prc) demonstrated pronounced differences between the early and advanced retinoblastoma, which could sever as the potential biomarkers for diagnosis of RB. e) According to selected metabolic biomarkers, the potential pathway to differentiate the early and advanced RB would be plotted with the p value and pathway impact, as well as alanine, aspartate, and arginine and proline metabolism. Reproduced with permission. [1] Copyright 2022, Wiley-VCH GmbH.
( Figure 2e), which may reveal the molecular mechanism of RB. [1] Early diagnosis of myocardial infarction (MI) and screening of satisfactory biomarkers are significant but challenging. [84] Screening of MI biomarkers based on MALDI were demonstrated (Figure 3a-c). Typical MS spectra of biosamples from health control (HC), myocardial infarction (MI), without myocardial infarction (non-MI), were analyzed ( Figure 3b) and serum metabolic profiles revealed the heatmap analysis with the up-regulated and down-regulated biomarkers for discrimination of MI patients. Subsequently, five biomarkers including LA, MIAA, cis-AC, diacylglycerol (14:1/24:1) (DAG-1), and diacylglycerol (24:1/20:4) (DAG-2) was screened (Figure 3c). [85] Since early diagnosis of lung adenocarcinoma (LA) greatly increases the chances of successful treatment, 481 clinical biosamples with no significant difference were harvested to discriminate LA Figure 3. Mass spectrometry enables screening of metabolic biomarkers for diagnosis of coronary heart disease (CHD) and early-stage lung adenocarcinoma (LA). a) Typical MS spectra of biosamples from health control (HC), myocardial infarction (MI), without myocardial infarction (non-MI), respectively. b) Serum metabolic profiles demonstrated the heatmap analysis with the up-regulated and down-regulated biomarkers for discrimination of CHD patients. c) Five biomarkers including LA, MIAA, cis-AC, diacylglycerol (14:1/24:1) (DAG-1), and diacylglycerol (24:1/20:4) (DAG-2) was screened. Reproduced with permission. [85] Copyright 2021, Wiley-VCH GmbH. d) 481 clinical biosamples with no significant difference to discriminate LA from HC, benign disease and other lung cancer. e) Seven selected metabolites in serum with strong Pearson correlation were demonstrated. f) Two down-regulated metabolites and five upregulated metabolites were screened to differentiate early-stage LA from HC. g) Potential pathways were analyzed including fatty acid metabolism, sulfur metabolism, histidine metabolism, cysteine and methionine metabolism. Reproduced with permission. [27] Copyright 2020, Springer Nature Limited.
from HC, benign disease, and other lung cancers (Figure 3d). [27] Seven selected metabolites in serum with strong Pearson correlation were demonstrated (Figure 3e), and two down-regulated metabolites and five up-regulated metabolites were screened to differentiate early-stage LA from HC (Figure 3f). [27] Poten-tial pathways were analyzed including fatty acid metabolism, sulfur metabolism, histidine metabolism, cysteine, and methionine metabolism (Figure 3g), which may reveal molecular mechanism of LA progression from HC. [27] We summarized several proof-of-concept investigations that accomplished potential screening of metabolic biomarkers for diagnosis of major disease (RB, MI, LA), which will have far-reaching implications for the screening of next generation metabolic biomarkers.
Electrochemical wearable biosensor allows highly sensitive, cost-effective, and noninvasive metabolite quantification, promising practical wearable device platform for healthcare monitoring. A uric acid and tyrosine biosensor was mounted on neck and can be bent to fit the skin (Figure 4a,b). [3] A three electrodes-system (counter electrode (CE), working electrode (WE), reference electrode (RE)) LEG-based flexible sensor was presented for simultaneous urine acid (UA) and tyrosine detection (Figure 4c). The metabolites were quantified from the oxidation peak via differential pulse voltammetry strategy (DPV, Figure 4d). The wearables achieved direct biosensing of UA and Tyr with low detection limits of 0.74 and 3.6 μm, with sensitivities of 3.50 and 0.61 μA μm −1 cm −2 , respectively (Figure 4e,f).
The micro/nanomaterials play a significant role for improving the performances of wearable electrochemical sensors in following aspects: a) Improvement of electronic transfer capability; b) Expansion of the bioactive site for electrochemical reactions; c) Interface-adaptation (skin-affinity, biosafety, mechanical compliance, and signal fidelity). [50,[103][104][105][106] We summarized MXene, hydrogels, and PEDOT:PSS micro/nanomaterials for enhancement of performances of electrochemical wearable biosensors. As a typical two-dimensional material, a new class of MXene was reported, which was composed of transition metal carbides and either nitrides or carbonitrides. [107,108] Generally, the MXene could be synthesized by selectively etching Al atomic layers from original MAX phase in the presence of hydrofluoric acid (HF). [108][109][110][111][112][113]  Originally, MXene was extensively investigated for the energy conversion and storage because of its high gravimetric capability. Meanwhile, MXene has been widely applied for the design of biosensor due its unique characteristics of favorable electrical conductivity, rich functional group (e.g., OH) and the high surface area (due to the multilayer architecture). [114,115] Herein, the 2-dimensional architecture as validated by scanning electron microscopy (SEM, Figure 5a,b) demonstrated unique intercalation architecture for enhanced electrochemical activity. [116] Meanwhile, tNPs/NPC/MXene-based wearable biosensor displayed a favorable and improved sensitivity of 82.68 μA mm −1 cm −2 (correlation coefficient of 0.9956).
Hydrogel is a class of polymer-material with threedimensional networked architecture, which has attracted intensive attentions in the healthcare field. [117] The advantages of hydrogel for design of wearable biosensors can be illustrated as follows: i) As excellent compatibility with majority of biomolecules, hydrogel-based wearable biosensor demonstrated favorable interface-safety, especially skin biocompatibility; [50,118] ii) Hydrogels are composed of flexible materials such as Reproduced with permission. [116] Copyright 2022, Wiley-VCH GmbH. The layer-by-layer illustration demonstrated the fabrication procedures from top to bottom. The glucose would be oxidized by GOx. c) The SEM characterization of P1T0.75Pa0.25 hydrogel for wearable metabolite biosensor toward d) noninvasive detection of glucose. e) The hierarchical hybrid hydrogel included first level (microstructure) and second level (nanostructure). f) The electrochemical current as a function of glucose concentration for three hydrogels. Reproduced with permission. [49] Copyright 2021, Elsevier. g,h) The SEM characterization of PEDOT:PSS/DF/PB hydrogel and its magnified images on the silicon wafer demonstrated its high surface area. i) On-body glucose monitoring was carried out on two volunteers for real-time noninvasive biosensing of glucose. Reproduced with permission. [50] Copyright 2022, Elsevier.
polymer, elastomer, biocompatible molecules, carbohydrates, and additives. Therefore, hydrogels are suitable for design of flexible/elastic devices with acceptable mechanical flexibility and stability; [119] iii) Hydrogels with three-dimensional polymer network architecture have merits of large surface area, easy-to-functionalize sites, which is beneficial for biochemical functionalizations; iv) These are a class of emerging conductive hydrogel materials including conducting polymers, carbon nanostructures, and conductive additives, whose electrical conductivity can be tunable, resulting in amplified signal transduction and enhanced sensitivity and lower limit of detection (LOD). [120] A P1T0.75Pa0.25 hydrogel demonstrated hierarchical hybrid hydrogel included first level (microstructure) and second level (nanostructure), which guarantee the close adhesion between the biosensor and the epidermal interface (Figure 5ce). [49] Subsequently, the hydrogel-based biosensor afforded a detection range of glucose concentrations (0.2 μm-10 mm, Figure 5f). Considering typical glucose concentration in sweat with concentration of 200 μm-0.6 mm, the proposed hydrogel-based metabolite biosensor covered the practical quantification of glucose detection range in sweat. [121,122] The PEDOT:PSS is a conductive polymer with excellent electrical conductivity. [123][124][125][126] The SEM characterization of PEDOT:PSS/DF/PB hydrogel and its magnified photograph on the silicon wafer demonstrated its high surface area (Figure 5g,h), [50] which is beneficial for enhancement of electrochemical activity. Furthermore, on-body glucose monitoring was performed on volunteers by attaching the hydrogel-based bioelectronics to their arm for the real-time noninvasive biosensing of glucose (Figure 5i). The PEDOT:PSSbased wearable biosensor afforded a satisfactory glucose detection range of 1-3242 μm with a low detection limit of 0.85 μm.

Organic Electrochemical Transistors
The OECTs has drawn tremendous attentions in last decades, which was developed by Wrighton and colleagues in the mid-1980s. [127,128] Briefly, an organic semiconducting film, an Figure 6. The OECTs enables wearable biosensors for metabolite quantifications. a) The biosensing mechanism of OECTs. Reproduced with permission. [136] Copyright 2018, American Chemical Society. b) EDLs of OECTs. c) A prototype of OECTs-based wearable biosensor for detection of metabolite could be illustrated, where the Pt gate electrode was modified with UOx-GO/PANI/Nafion-graphene. d) The OECTs-based biosensor has been adopted for noninvasive detection of urine acid in saliva testing with favorable mechanical compliance and skin-affinity no matter the changes of skin morphology. Reproduced with permission. [137] Copyright 2015, Wiley-VCH GmbH. e) OECTs-based cortisol biosensor would seamlessly connect with skin in different locations (e.g., arm, forehead). f) Wearable OECTs cortisol biosensor afforded comparable results with LC-MS/MS measurements in real sweat applications. Reproduced with permission. [138] Copyright 2022, American Chemical Society. electrolyte and metal interconnections compose of a typical OECTs (Figure 6a,b). The organic semiconducting film is seamlessly contacted with metal interconnection electrodes including source and drain and immersed in electrolyte (Figure 6a). [129][130][131] The gate electrode would modulate the parameters of channel via holes or electrons flow from the source to the drain. The OECTs relies on ions that are injected from the electrolyte into the organic film, thus adjusting its doping-dedoping state and enhanc-ing its mobility. [132,133] The small signal would be amplified into considerable readout via OECTs amplifier. The working mechanism of the OECTs can be summarized by: [134,135]   Reproduced with permission. [147] Copyright 2022, American Chemical Society. g) A molecularly selective nanoporous membrane-based wearable organic electrochemical device for noninvasive cortisol biosensing. h,i) Selective polymer-based OECTs patch is comfortable to wear and easy to peel off and afforded a detection range of cortisol from 0.01 μm-5 mm. Reproduced with permission. [148] Copyright 2018, American Association for the Advancement of Science.
where V p is the pinch-off voltage; q and are the electronic charge and hole mobility; V G eff is the effective voltage of gate; V offset is the offset voltage; V G-E and V E-C are the voltages on the gate and electrolyte and electrolyte and channel, respectively.
Meanwhile, the amplification factor of OECTs would be measured by transconductance (G m ) and described as the following equation: The C* and the V th are the volumetric capacitance and voltage of threshold.
The PEDOT:PSS is a typical p-type semiconducting materials for OECTs biosensor (Figure 6b), [41,[139][140][141][142] and a prototype of OECTs-based wearable biosensor for detection of metabolite could be illustrated in Figure 6c, where the Pt gate electrode was modified UOx-GO/PANI/Nafion-graphene. Once the gate electrode monitored the oxidized signal of uric acid, the electrical double layers (EDLs, Figure 6b) would be transferred and amplify the signal into considerable readout. [137] The OECTsbased biosensor has been adopted for noninvasive detection of urine acid in saliva testing with favorable mechanical compliance and skin-affinity no matter the changes of skin morphology (Figure 6d). [138] Meanwhile, OECTs-based cortisol biosensor would be seamlessly connected to skin in different locations (e.g., arm, forehead, Figure 6e). The wearable OECTs cortisol biosensor afforded comparable results with LC-MS/MS measurements in real sweat applications (Figure 6f). [138] Selections and optimizations of micro/nanomaterials decide the capability of OECTs biosensor. The material engineering for OECTs-based biosensor can be considered from the following aspects: 1) Substitutes for metal interconnections; 2) Chemical functionalization or doping of semiconducting polymer; 3) Selective membrane for specific binding. [134,[143][144][145] To provide more alternatives for metal interconnection (e.g., Au), an organic electrochemical transistors (OECTs) based on PVA-co-PE nanofibers (NFs) and polypyrrole (PPy) nanofiber networks for noninvasive monitoring of dopamine was presented and demonstrated superior performance to the gold and platinum (Pt) wires (Figure 7a). [146] Moreover, the commercial PA6 fiber (85D) would be fabricated and assembled into woven fabrics in the type of wearable biosensor (Figure 7b), resulting in wide detection range of dopamine of 1-1000 nm and higher sensitivity than the performance of gold and Pt electrode (Figure 7c). For the optimizations of semiconducting film, the chemical engineering of semiconducting is a tailored strategy to adjust performance of wearable OECTs biosensor. Biochemical functionalization of poly (EDOT-COOH-co-EDOT-EG3) nanotubes were performed by simple template-free electropolymerization in the channel area of an OECTs device featuring a PEDOT:PSS underlayer for the detection of sweat cortisol. Morphological characterization  [136] Copyright 2018, American Chemical Society. b) As a typical prototype of wearable FETs biosensor for quantification of cortisol. The gate channel with the semiconducting material of graphene was modified with the cortisol antibody. c) The proposed biosensor exhibited a sufficiently low LOD of 10 pg mL −1 , and the current decreased linearly with the concentration of the cortisol. The LOD was calculated at the signal-to-noise ratio of 3.0. d) Integration of FETs biosensor into smart contact lens (inset: close-up outer image of the smart contact lens. The scale bar is 1 cm). e) Photograph of an adult woman wearing the smart contact lens in her left eye (inset: close-up image of the smart contact lens on the eye). f) Cortisol concentration was measured using the contact lens sensor as a function of the cortisol concentration that was dropped into the eye. Reproduced with permission. [167] Copyright 2020, American Association for the Advancement of Science.
of engineered poly(EDOT-COOH-co-EDOT-EG3) nanotubes and schematic illustration presented a noninvasive, sensitive detection of cortisol using biochemical functionalization strategy (Figure 7d,e). [147] The real-time monitoring of cortisol using the poly(EDOT-COOH-co-EDOT-EG3) nanotube-embedded OECT immunosensor afforded sensitive quantification of cortisol ranging from 1 fg mL −1 to 1 μg mL −1 (Figure 7f). For the enhancement of selectivity, a molecularly selective nanoporous membrane-based wearable organic electrochemical device was reported for noninvasive cortisol sensing. [148] Synthetic and biomimetic polymeric membrane with SEM characterization (Figure 7g), which acts as a molecular memory layer to facilitate the stable and selective molecular recognition of the human stress hormone cortisol. With person-oriented design, selective polymer-based OECTs patch is comfortable to wear and easy to peel off (Figure 7h) and afforded a detection range of cortisol from 0.01 μm-5 mm (Figure 7i). Abnormal levels of cortisol have a significant impact on the regulation of physiological status such as blood pressure, glucose levels. Abovementioned strategies have optimized the capability of OECTs-based wearable biosensors in terms of sensitivity, stability, and mechanical properties.

Field Effect Transistors
Affordable, quick-response, sensitive, and user-friendly diagnostic tools are required particularly in less-developed nations and resource-limited districts. [149][150][151] However, existing advanced technologies usually are incapable to meet the various requirements of medical care due to dependency on advanced clinical configurations, complicated procedures, and high costs. [152][153][154] As a result, researchers spare no effort to develop alternative strategies for disease diagnosis that are cost-effective, wearable, and without sophisticated and expensive equipment. [155][156][157] The field effect transistors biosensor (FETs) are found to be cost-effective, flexible, highly sensitive, and easily integrated for real-time diagnostics. FETs have emerged as powerful platform for ultrasensitive, direct electrical readout, and label-free www.advancedsciencenews.com www.advsensorres.com Figure 9. The micro/nanomaterials for FETs-based wearable biosensors toward noninvasive quantification of metabolic biomarkers. a) Architecture of a In 2 O 3 -based FETs. b,c) The biosensor on flexible polyimide demonstrated mechanical bending, and corresponding transfer curves from a representative FETs sensor exhibited negligible changes at pH 6.8 or 7.0 under different bending radius. d) Cortisol-aptamer-FETs detected cortisol concentrations over six orders of magnitude (1 pm to 1 um). Reproduced with permission. [168] Copyright 2022, American Association for the Advancement of Science. e) Schematic of GFETs on an ultrathin film with mechanical bending was illustrated with the graphene as the semiconductor. f) GFET flexible device was conformably attached to the human wrist. g,h) Stretchable biosensor can be stretched with the activity of the human body and demonstrated a detection range of 0.03 to 500 nm for TNF-. Reproduced with permission. [169] Copyright 2020, Multidisciplinary Digital Publishing Institute. i,j) Carbon nanotubes would work as the metal interconnection with mechanical stability and deformable biosensor would be gently and seamlessly mounted on the hand. The scale bars in j are 3 and 1 cm, respectively. k) The proposed array biosensor would be bent with flexibility and achieve wide glucose detection range of 0.01-1000 μm in artificial tears, sweat and saliva. Reproduced with permission. [170] Copyright 2018, American Chemical Society.
biological/chemical quantification. [158][159][160][161] The main components for the FETs biosensor include: i) A semiconducting layer; ii) The three conducting interconnection (e.g., metal) for external contacts, and iii) The dielectric layer. [162][163][164] The mechanism of FETs biosensor voltage is modulating conductivity of the channel between source and drain (Figure 8a). [165][166][167] As a typical prototype of wearable FETs biosensor for quantification of cortisol (Figure 8b), [167] the gate channel with the semiconducting material of graphene was modified with the cortisol antibody. The proposed biosensor exhibited a sufficiently low LOD of 10 pg mL −1 , and the current decreased linearly with the cortisol concentration. The LOD was calculated via the signal-to-noise ratio of 3.0 (Figure 8c). Moreover, the biosensor would be integrated into smart contact lens (inset: close-up outer image of the smart contact lens, scale bars, 1 cm, Figure 8d). Notably, photographs were demonstrated that an adult woman was wearing a smart contact lens in her left eye (Figure 8e). Cortisol concentration was quantified using the contact lens sensor as a function of the cortisol concentration that was dropped into the eye (Figure 8f). [167] The micro/nanomaterials have been intensively reported for the assisting the functionality of FETs-based wearable biosensor including following aspects: 1) N-type transparent, ultrathin semiconductor functional materials (e.g., indium oxide, In 2 O 3 ) for excellent mechanical flexibility; 2) Two-dimensional materials (e.g., graphene) for favorable semiconducting properties; 3) Alternatives of metal interconnection such as the carbon nanotubes with excellent electrical conduction and mechanical stability. For the In 2 O 3 -based FETs, a wearable and deformable affinity biosensor was presented for monitoring of metabolic cortisol in biofluids (Figure 9a). [168] The biosensor on flexible polyimide demonstrated mechanical bending (Figure 9b), and corresponding transfer curves from a representative FETs sensor exhibited negligible changes at pH 6.8 or 7.0 under different bending radius (Figure 9c). Cortisol-aptamer-FETs detected cortisol concentrations over six orders of magnitude (1 pM to 1 uM, Figure 9d). [169] For the wearable graphene-based field effect transistors (GFETs) biosensors, favorable semiconducting properties, ultra-flexibility with mechanical stability and high surface area are unique advantages that make graphene a candidate for practical noninvasive biosensing. Schematic of GFETs on an ultrathin film with mechanical bending was illustrated with the graphene as the semiconductor (Figure 9e). [169] The GFET flexible device was conformably attached to the human wrist (Figure 9f). The stretchable biosensor can be stretched with the human activity (Figure 9g) and demonstrated a detection range of 0.03-500 nm for metabolic TNF- (Figure 9h). [169] The carbon nanotubes would work as the conductive interconnection with mechanical adaptation and stability ( Figure 9i) and deformable biosensor would be gently and seamlessly mounted on the hand (Figure 9j). [170] The proposed array biosensor would be bent www.advancedsciencenews.com www.advsensorres.com with flexibility and achieve wide glucose detection range of 0.01-1000 μm in artificial tears, sweat and saliva (Figure 9k). [170] Abovementioned strategies including EC-, OECTs-, FETs-based wearable biosensors were summarized for noninvasive quantification of metabolic biomarkers.

Conclusions and Prospective
With the trend of aging population and the increasing number of sub-healthy people in China, there is a huge opportunity for the wearable devices in the field of healthcare filed, which would not only solve the problem of insufficient and imbalanced medical resources, but also enable people could monitor their healthcare without the limitations of time, location, and environment. Herein, we described the screening and quantification of metabolic biomarkers based on micro/nanomaterials toward wearable device applications. The future prospective of micro/nanomaterials-based screening/quantification analysis can be illustrated by material engineering, metabolic analysis, device development. For the material engineering, size, morphology, and functionalization are key indicators to consider when selecting and optimizing micro/nanomaterials. For the metabolic analysis, quantified model, and algorithm to decipher the massive MS data dramatically influences the screening of metabolite biomarkers. For the development of wearable devices, the following aspects may be beneficial to be considered: i) Acceptable accuracy, sensitivity, reliability, reproducibility, robustness, and signal fidelity; ii) Biocompatibility and skin-safety; iii) Bio-interfacial compliance. The integration of screening and noninvasive quantification techniques may bright prospects for creating next generation wearable devices toward large-scale home healthcare.