Magnetically localized and wash-free fluorescent immuno-assay: From a research platform (MLFIA) to a multiplexed POC system (MagIA)

technology has the potential to address these needs, as the MLFIA 18-chamber microfluidic cartridge and the MLFIA Analyzer were previously characterized and evaluated with plasma and serum from patients infected with HIV, Hepatitis B (Hep B) or C (Hep C). Here, we present the efforts to transfer this research platform (MLFIA) to a fully integrated multi-analysis solution (MagIA). First, we present the design changes of the consumable enabling to perform multiple assays in parallel, a fast filling of the cartridge with patient samples, and a homogeneous re-agent/sample incubation. Second, we describe the development a piezoelectric actuator integrated into the Analyzer: this mixing module allows for an automated, fully integrated and portable workflow, with homogeneous in-situ mixing capabilities. The obtained MagIA platform was further characterized and validated for immunoassays (LOD, cartridge stability over time), using various biological models including OVA and IgG. We discuss the performances of the MLFIA and MagIA platforms for the detection of HIV / Hep B / Hep C using results from 102 patient plasma samples. Lastly, we assessed the compatibility of the MagIA platform with veinous and capillary blood samples as a final step towards its POC validation.


Introduction
Recent pandemics have underscored the necessity of point-of-care testing (POCT) and its role as valuable complement to conventional lab testing.It facilitates the reach out to key populations outside of conventional healthcare pathways and helps relieving the burden on biological labs and hospitals during peak contamination periods and when monitoring epidemics [1][2][3].Driven by the COVID-19 pandemic, accessible testing centres like pharmacies have contributed to increased testing rates [4,5].This trend has influenced healthcare practices, promoting rapid and self-testing as a strategy to combat infectious diseases [6].In the fields of microbiology and molecular biology, a syndromic approach-testing defined infectious panels together-has garnered considerable attention [7][8][9][10].This approach shall be expended beyond symptomatic diseases, such as respiratory conditions, to encompass asymptomatic diseases, such as HIV, Hep B and Hep C, that share contamination pathways and impact same key populations.
With 40 million people infected with HIV, 58 million infected by Hepatitis C and 296 million chronically infected by Hepatitis B, blood borne viruses remain one of the global public health priorities, not only in emerging countries but also in western countries [11][12][13].The WHO and governments committed to eradicate HIV and viral hepatitis B & C (considered among the most lethal STIs) by 2030 [14,15].Consequently, prioritizing screening of these diseases is a critical step to ensure effective linkage to care ("Finding the Missing Millions").As an alarming statistic, less than 20 % of people infected with viral Hepatitis B and C were aware of their infection in Europe, and about 40 % in the US [16].These viral infections continue to increase, particularly among key population with high risk profile due to their sexual practices, origins, or drug abuse [17][18][19][20].
As these diseases share the same infection pathways (blood exposure, sexual and perinatal transmissions), numerous guidelines and studies conducted in various countries [21,[23][24][25][26][27] and by several public health scientific boards (HAS [28], ECDC [21]) have concluded on the benefits of combined testing for these 3 diseases.Routine tests for these diseases are particularly recommended for pregnant women, blood donors and high-risk populations [22].Yet, this combined approach is currently limited to laboratory settings.To conduct all three serologies at once, testing within high-risk populations is organized by STI clinics and community centres.Typically, they collect blood samples and send them to a central lab for analysis.In the face of increasing demand and frozen budgets, these screening consultations are currently hindered by extended waiting time (3-4H), overcrowding and up to 30 % of results non-given to the patients due to this centralized laboratory approach [29].
The cited facilities also perform outreach testing campaigns using Rapid Diagnostic Test (RDT), using a dedicated assay for each virus [30,31].As of today, only mono-disease rapid tests have received CE marking (or FDA approval) for screening of HIV, hepatitis B and C. HIV, HCV, HBV RDT are designed to detect in capillary blood respectively anti-HIV, anti-HCV antibodies and HBsAg, with sensitivities and specificities close to the WHO standards [32][33][34].We can mention companies such as Abbott (USA), Chembio (USA), Biosynex (France), Biolytical (Canada) and Orasure (USA).HIV 3 rd generation tests (Ab detection only) allow for the detection of potential seropositivity starting from 3 months following the infection [35].Abbott's ALERE HIV Combo is the only HIV 4 th generation rapid test to detect p24 antigen, enabling an earlier detection of primary HIV infection.This test is CLIA waived for fingerstick allowing its use in non-laboratory settings [36].The cited rapid tests present various incubation times, ranging from 1 to 40 min, with various sampling volumes ranging from 5 to 75 µL.Such variations in sample sizes, incubation times and reading times between different rapid tests make it challenging to manage them in parallel.As a result, although the multiplex approach is advantageous [37,38], rapid testing commonly employed in outdoor settings is usually limited to a single virus (mainly HIV) [29,39], causing missed opportunities for Hepatitis diagnostics.Few HIV, HBV, HCV triplex tests are available on the market.However, most of them are only compatible with serum and plasma taken with venepuncture and not capillary blood, which drastically limits their use for point-of-care purposes [40].Among existing tests, the one of Boson detects anti-HIV, anti-HCV and HBsAg in whole blood within the same cassette but requires injecting the sample 3 times for each parameter detected.The Exacto® Triplex of Biosynex is the only test allowing for the simultaneous diagnostic of HIV-1, 2, Hepatitis C antibodies, and HBsAg in the same sample of capillary blood.The evaluation of the Exacto® with serum sample in Central African Republic demonstrated high sensitivity for HIV and HBsAg (99 %) and relatively high sensitivity (96 %) for HCV [41].Still, none of these tests are certified yet (CE nor FDA).
All cited tests are lateral flow assays, that share as main limitations the absence of traceability and the risk of false results due to inaccurate interpretation by the healthcare professional [42].While lateral flow assay readouts are straightforward, they can be prone to misinterpretation, leading to false negative results, especially when dealing with low signals: in a stressful environment, the thin lines of a low titration may be overseen [43].Additionally, their visual interpretation prevents the offering of quantitative results.Thus, they are not suitable for quantitative titration.In terms of traceability, the manual reporting can induce errors and makes the process time consuming.Overall, these limitations underscore the necessity for technologies using Analyzers, that provide objective and traceable results.Analyzer devices can furthermore address the challenges related to traceability and offer enhanced connectivity in reporting a read out.Therefore, there remains a need for an immunological POC platform for STI screening, that can deliver laboratory-level performance and traceability, with multiplexing capabilities, using capillary blood.Such platform should consider battery enabled autonomy, portability, and ideally cost-effectiveness in production.
We developed the MLFIA (Magnetically Localized and wash-free Fluorescent Immuno-Assay) technology specifically to meet these requirements [44][45][46].MLFIA involves differential measurement, comparing the fluorescence of immune complexes specifically bound to Magnetic Nano Particles (MNPs) which are locally captured by micro-magnets, with the residual background fluorescence of unbound fluorescent probes in the sample.We obtain an immunoassay that locally concentrates targeted biomolecules (antibodies or antigens) on the micro-magnet lines, with the number of captured biomolecules quantitatively measured without any washing step.The 18-chamber microfluidic cartridge for MLFIA, along with the MLFIA Analyzer, were previously introduced, characterized, and successfully evaluated using plasma and serum samples from patients infected with HIV, Hep B or Hep C [45].Several limitations were however emphasized with this first platform.The preanalytical phase requires manual and human controlled steps prone to inaccuracies: mixing of reagents and sample in a tube, 15 min incubation, subsequent cartridge filling and time sensitive magnetic capture.As a result, the analysis of large number of samples distributed in multiples cartridges results in an increased variability thus slightly reduced performance.Furthermore, the platform demonstration was realized using serum and plasma, which requires a blood preparation step such as centrifugation, hence a laboratory setting with skilled technicians.A cartridge compatible with whole blood, allowing for multiple parallel assays, would be preferable for POC applications.
Here, we present the efforts and challenges involved in the transfer of this technology from a research platform (MLFIA) to a fully integrated multi-analysis solution (MagIA).First, we present the different design changes made on the consumable, enabling multiple parallel assays.Second, we detail the different efforts to add, characterize and validate a piezoelectric actuator within the Analyzer.This actuator allows for an automated and fully integrated workflow, with in-situ mixing in-situ.Third, we present the comprehensive characterization and validation of the MagIA platform for immunoassays, utilizing various biological models like OVA and IgG.We report a side-by-side comparison between the MLFIA and the MagIA platforms for the detection of HIV / Hep B / Hep C, using patient plasma samples.Lastly, we evaluate the compatibility of the MagIA platform with both veinous and capillary blood samples, the final step towards its POC validation.

Theory
The previously developed MLFIA platform worked as illustrated in Supp.Fig. 1: The sample was mixed off-chip, in a microtube with the reagents, consisting of functionalized MNPs with antigens or antibodies to capture the target molecule and fluorescent detection antibodies.The mixture was incubated off-chip to let the antibody-antigen binding occur and immune complexes to form on the surface of the MNPs.After incubation, the mixture was injected into a cartridge comprising 18 individual chambers, each of which integrated micro-magnets (MLFIA Cartridge).The cartridge was subsequently placed above a magnetic array (known as MagActivator), which featured an alternation of cmsized magnets.This setup expedited the capture of MNPs by the micro-magnets.The high magnetic field gradients present at the junction between the micro-magnets lead to the local capture of MNPs, forming a distinct stripe pattern.The cartridge was finally inserted in the MLFIA Analyzer, allowing for the sequential imaging of each of the 18 chambers.The localized capture of magnetic immunocomplexes along the micro-magnetic stripes enabled a discriminative detection of free and bound detection antibodies within minutes.The quantification in MLFIA consisted in differentially measuring the fluorescence specifically associated with the micro-magnet (on the stripes) and the unbound fluorescent probes (between the stripes), hence the name of Magnetically Localized Fluorescent Immuno-Assay or MLFIA (Fig. 1C).The peak heights of the fluorescence signals were proportional to the concentration of antibodies / antigens present in the analyte.The preceding incubation step lasted 15 min.The overall process, including sample deposition, did not necessitate any washing step and still provided quantitative and sensitive results.
Our wish was to optimize the MLFIA platform towards a commercial POC assay, easily transportable on site and to be performed without extensive lab skills and minimal user error.We first addressed the usability aspect through three main strategies: (1) Equipment minimization: Our aim was to limit the number of required equipment.With MLFIA, the mixing and incubation were conducted off-chip, while the MNPs would be captured on the micro-magnets using a separate platform -the MagActivator.We successfully implemented these three steps (mixing, incubation, capture) within a portable analyzer (weight: 3,4 kg; dimension: 24 × 15 × 19 cm) (Fig. 1A).(2) Usage simplification: We focused on reducing the process for the user, aiming to minimize pipetting, thus making it suitable for a point-of-care usage.The MagIA platform now conveniently integrates reagents (functionalized MNPs and detection Ab) directly into the cartridge.(3) Volume reduction and parallelization: By minimizing pipetting and washing steps, we achieved a reduction in reagent volume, enabling parallel analyses for a given sample.We explored the feasibility of conducting multiple parallel immunoassays using a consumable with 5 reaction chambers connected to a single sample injection port.
The detailed MagIA workflow is presented in Fig. 1B.The reagents are dried into the cartridge, with a specific combination of functionalized MNPs and detection antibodies present in each of the chamber (depending on the target molecule).The micro-magnets are directly embedded at the bottom of the cartridge.❶ A diluted patient sample (e. g. a dilution of 1:10), containing the analytes targeted, is injected into the cartridge, where it can flow towards the reaction chambers.The flow is not sufficient for mixing the analytes and the reagents with each other.❷ The user loads the cartridge into the MagIA Analyzer.Inside, a piezoelectric module sequentially agitates each reaction chambers for one minute, ensuring thorough mixing of the reagents with the sample.The process re-suspends previously dried reagents, allowing them to interact with the relevant targets in the sample.Subsequently, the piezo actuation ceases, ending the mixing process.❸ The suspension incubates for 10 min to allow the formation of the immune complexes.❹ Following the passive incubation, an external magnet is used to capture the MNPs onto the junctions of the micro-magnets in each reaction chamber successively (1 min/chamber).❺ The cartridge is ready for its fluorescence wash-free detection.The Analyzer can process two cartridges concurrently in a 30 min timeframe.

Consumable: towards multiple parallelized assays in a single cartridge
The cartridge is the consumable where the actual multiparametric immunoassay takes place.To enhance usability, the cartridge was designed with 2 Sections; (1) a handling side, allowing the user to grasp the cartridge, and (2) an assay side (Fig. 2A).The assay side comprises a single sample injection port that directs to 5 reaction chambers, each of which terminates in a venting port.Each chamber allows to perform a specific immunoassay (Fig. 2B, C).In this study, 3 chambers are allocated for target-specific assays (HIV, HCV, HBV), while the remaining two are designed for control assays (positive and negative).In each reaction chamber, specific reagents (functionalized MNPs and fluorescent detection Abs) required for the assay are dried.Fig. 2A illustrates the different layers of the cartridge, further described in the Experimental section.The minimum dimension of the fluidic system is set at a channel height of 250 µm and a minimum width of 400 µm.This design helps preventing clogging that could result from the high number of blood cells flowing within the channels.

Integration of reagents into the cartridge
To incorporate reagents into the cartridge, 2 µL of functionalized MNPs [10 μg/mL of coated protein] and 2 µL of detection antibodies [2 μg/mL] were pipetted onto each chamber of the preassembled cartridge substrate.Afterwards, the reagents were dried before the final assembly of the cartridge.This is illustrated in Fig. 2B, where the dried reagent drops in the reaction chambers are showed with coloured circles.

Distancing of the micro-magnet structure from the MNPs
In contrast to the MLFIA approach, the micro-magnets are covered with a single-sided adhesive tape layer, referred to as the "spacing layer", which forms a gap between the sample and the micro-magnet surface.Thus, in the absence of an external magnetic field, the magnetic field generated by the micro-magnets is insufficient to suspend MNPs in the sample.This enables a homogeneous incubation within the cartridge, undisturbed by the stray field of the micro-magnets.

Optimization of the wettability of the fluidic channels in the cartridge
Without surface treatment, the blood sample may eventually enter the cartridge; however, the filling process would be slow and nonuniform.Therefore, a hydrophilization step is essential to facilitate the filling of the cartridge and ensure a smooth wetting of the fluidic path.We chose a 1 % Pluronic® solution, sprayed on the surface of the shell, that would come into contact with the sample.In this study, we conducted tests involving 1 and 5 spray coatings, each spray coating followed by a drying step between each pulverization.We evaluated their effect with static and dynamic analysis, as well as optical visualization.
In the static analysis (Fig. 2D), a 5 µL drop of water was deposited on the cartridge shell for each condition.The drop angle was measured at 63.2 • ±1,7 • without hydrophilization, and 31.5 • ±11,5 • with 1 or 5 applications of Pluronic® 1 %.These results strongly suggest that the coating significantly improved the surface wettability.This finding was further validated with 10X diluted capillary blood loaded into the sample loading port of cartridges treated under similar conditions (Fig. 2D, Supp.Video).Without hydrophilization, the blood sample did not enter the cartridge (left photo).With 5 applications of 1 % Pluronic® (right photo), the blood sample entered the cartridge smoothly, allowing continuous visualization of blood flowing withing upon filling.Optical observations of the cartridge chambers provide more explanation (Supp.Fig. 2).When drying, the Pluronic pulverization creates some circular stains (dried droplets), not evenly spaced.With the application of 1 % Pluronic®, the liquid moves forward slowly, accelerates suddenly when meeting a dried spot, then slows down until it meets another dried spot.Having 5 pulverizations of Pluronic® enables a greater density of dried spots over the fluidic surface, thus helping to fill in the cartridge much faster: The complete filling of the cartridge was achieved in 10 s.
Supp.Video 1. Injection of a blood sample in 3 cartridges.The first one, without Pluronic treatment, illustrates the difficulty to inject the blood through the fluidic channels: the blood drop stays on top of the channel inlet port.In the second cartridge, containing 1 pulverization of 1 % Pluronic, the wettability of the channel is enhanced, facilitating the blood capillary flow within the channels, but this still takes some time (40 s).With 5 pulverizations of 1 % Pluronic, however, the blood flows in quickly and homogenously (10 s).

Instrument: integration of the mixing process
Our aim was to incorporate the mixing process into the Analyzer.This is required for resuspension of dried reagents in the sample.A dedicated mixing module was engineered and positioned beneath the cartridge to effectively induce vibrations within the fluidic reaction chambers.The mixing module comprises a Langevin Piezoelectrical Transducer (LPT) and an actuator (Fig. 3A).The actuator serves two functions: (i) pushing the transducer up against the bottom face of the cartridge substrate (magnetic substrate), enabling it to generate vibration within the cartridge, and (ii) driving down the transducer to allow for the lateral translation of the cartridge (in order to position the different chambers above the transducer).The transducer induces vibrations within the cartridge to resuspend and mix the reagents.To enhance vibration efficiency, we dimensioned the transducer to match the resonance frequencies of the MagIA cartridge.
First, we assessed the resonance frequencies of the fluidic chambers by simulating their response to a vibrational frequency sweep.To achieve this, the chamber geometry was simplified according to its symmetric axes (a quarter of the chamber was simulated) (Fig. 3B).We detected four resonance frequencies: 25 kHz, 52 kHz, 79 kHz and 110 kHz (Fig. 3C).These frequencies were consistent with a separate 3D modelling of the entire cartridge.The lowest frequency (25 kHz) was disregarded due to its higher potential to induce cavitation, e.g. the formation of bubbles within the reaction mix.The highest frequency (110 kHz) was set aside as no suitable off-the-shelf electronic driver could be found for frequencies above 100 kHz.Two separate Langevin Bolt clamped transducers were mounted for the remaining frequencies, resonating at 52 kHz (with a height of 49.5 mm) and 74 kHz (with a height of 35.5 mm), and their performances compared side-by-side using 3 distinct parameters: (1) The optimal push force applied by the actuator to the cartridge: The mixing necessitates regular connections and disconnections of the transducer from the cartridge (e.g. when moving the cartridge to place another chamber above the transducer).For this connection, a certain amount of force has to be applied to press the transducer against the cartridge magnetic substrate.Hence, we determined the maximum achievable transducer push force without compromising the structural integrity of the cartridge: The estimated optimal push force was in the range of 11-12 N for the 74 kHz actuator and 14 N for the 52 kHz actuator (Fig. 3D).At higher forces (indicated with an empty orange rhombus), measurements became unstable, and cartridges were damaged.This caused both fluidic leakage and electronic control instability (Fig. 3F).( 2) The (maximum) vertical deformation of the cartridge under applied vibration (Fig. 3E).(i) We first applied forces of 14 N and 11 N between the empty cartridge and the transducer tip, while measuring the shell displacement for each of the 5 chambers (empty blue circles).(ii) This was repeated with cartridges filled with water (full blue circles).Finally (iii), we measured the displacement of the surface spacing layer on top of the magnetic substrate/spacing layer, e.g. the bottom of the cartridge well (grey triangles).This position is of particular interest since it is where the dried reagents, requiring resuspension and mixing, are situated.The atop within the magnetic substrate plus spacing layer section should, therefore, exhibit a high degree of reproducibility.The displacement of the shell depends on the applied frequencies but not on whether the cartridge is filled or empty.Therefore, it can be assessed that the loading of the sample does not alter the vibrational behaviour of the cartridge at the measured frequencies, and that the displacement of an empty cartridge is representative of a full cartridge (Fig. 3E, grey triangles).The results show a higher chamber-to-chamber variability at 52 kHz, with a coefficient of variation (CV) of 24 %, compared to 12 % at 74 kHz.(3) The cartridge usability.Experiments showed that the higher vibrational deformation on the 52 kHz (Fig. 3E) transducer may result in irreversible deformation in the cartridge.Furthermore, some cartridges experienced random liquid expulsion through the cartridge outlets when actuated at 52 kHz.Both phenomena were less observed at 74kHz, which was therefore chosen for the system.

MagIA characterization with anti-OVA model in PBS
In the initial biological proof of concept, we evaluated MagIA for the detection of an anti-ovalbumin monoclonal antibody produced in mice (anti-OVA mAb) (Fig. 4A).Hence, we deposited and dried the same reagents in all 5 cartridge chambers: (1) MNPs coated with ovalbumin, and (2) anti-mouse IgG APC, serving as the fluorescent detection antibodies.Following cartridge assembly, we loaded a suspension of anti-OVA mAb in a PBS buffer as sample (Fig. 4B).
We assessed the analytical sensitivity of the detection by varying the concentration of anti-OVA mAb in buffer from 1 to 10,000 ng/mL (Fig. 4C).The linear range extended from 40 ng/mL to 3,300 ng/mL, demonstrating a high correlation coefficient of R 2 =0.99.To determine the Limit of Detection (LOD), we processed three blank buffer samples (N=3) and used the mean value + 3σ (3 times the standard deviation) as a cut-off to identify a sample as positive.In this experimental configuration, the LOD ranged between 14 and 40 ng/mL, displaying a similar curve trend as observed with the MLFIA platform.
Next, we conducted a direct comparison of the signal obtained with the MagIA technology alongside the MLFIA technology.Both signals exhibited linear proportionality with a high correlation coefficient of 0.99.The CV ranged from 8 to 25 % for MagIA, and 7 to 28 % for MLFIA, affirming the successful integration of the MLFIA protocol into an automatized immunoassay.
Finally, we evaluated the stability of the OVA cartridges stored at room temperature.We selected three concentrations of anti-OVA within the previously assessed linear range (1000 ng/ml; 100 ng/ml and 0 ng/ ml), and compared the results obtained with cartridges tested immediately after manufacturing with those tested 6 months later (Fig. 4E).After 6 months, cartridges exhibited signals like the ones obtained at Month 0. The slightly higher signals at Month 6 are likely due to variability between the 2 Analyzers used.Indeed, image analysis suggests that the Analyzer used at Month 6 produces brighter images, potentially due to a variation of the power or position of the LED.

MagIA characterization with human IgG model in serum
The previous experiment involved a simple biological model in PBS buffer.In this experiment, we adapted the protocol to detect IgG antibodies in serum samples, aiming to evaluate reliability and reproducibility of MagIA detection in a typical complex sample (Fig. 5A).Hence, we deposited and dried again appropriated reagents in all 5 cartridge chambers: (1) MNPs coated with anti-human IgG, and (2) anti-human IgG+M APC, serving as the fluorescent detection antibodies.Following cartridge assembly, we loaded 10X diluted human serum as sample (Fig. 5B).To minimize variations related to the sample, user and instrument, all the experiments detailed here were conducted by the same person, using the same instrument, and the same sample.We processed 18 cartridges (N=18) in pairs on the same Analyzer (Fig. 5C).
No significative difference in signal distribution was observed between cartridges inserted at the same time in the same analyzer (position 1 and 2) in the 1st (orange dots) or 2nd position (blue dots) (Fig. 5C).When considering the same chamber over time, we observe a CV of 8 % for chambers 1 and 4 when cumulating the runs on cartridges 1 and 2, 7 % for chambers 2 and 3, but a CV of 16 % for chamber 5.The high deviation obtained in one experiment in chamber 5, explaining the CV of 16 %, is likely related to some sample leakages that occurred in one cartridge in this chamber.
This reproducibility surpasses that of MLFIA, once again confirming the successful integration of our manual platform into an automated system with multiple assays in a single consumable.Potential factors contributing to variability may include variations in the volume of reagents deposited between cartridges, as well as the need for adjustment to ensure homogeneity in the mixing and capturing of targets/probes.

MagIA validation for combined HIV HCV HBV detection in patient serum
After demonstrating the capability of MagIA to detect human IgGs in serum, we proceeded to assess its suitability for detecting clinically relevant targets, including both antibodies and antigens, in patient samples.We dried reagents in the cartridge chambers as followed (Fig. 6A-C): In chamber 2, for HCV detection, MNPs coated with a recombinant HCV fusion protein (NS3/core/NS4/NS5), and anti-human Immunoglobulin G+M coupled to APC (anti-human IgG+M APC).In chamber 3, for HIV detection, MNPs coated with an HIV recombinant gp41 protein, and anti-human IgG+M APC.In chamber 4, for the detection of Hepatitis B surface antigen (HBsAg), MNPs coated with a monoclonal anti-HBsAg antibody, and a second anti-HBsAg coupled to APC. Chamber 1 was used as a positive control, intended for the detection of total IgG antibodies in the sample, as previously described.Chamber 5 was used as a negative control using MNPs coated with OVA and anti-human IgG+M APC.We processed 102 plasma samples from patients infected with either HCV, HBV or HIV (in singlicate), all of which had previously been analyzed using MLFIA [45] and by the blood bank (Fig. 6B).The selection of these 102 plasmas aimed to encompass various genotypes, subtypes, and concentration levels of the targeted molecules.
Each cartridge tested provided controls signals.The signals obtained for negative controls ranged from 2430 to 5998, whereas the signals for the positive control ranged from 50867 to 1377728.Supp.Fig. 3 presents the results from the control chambers of the 102 patient samples analyzed.
For each clinical parameter, we established an arbitrary threshold, which was calculated as the mean of the signals of the negative samples + 3σ.Positive samples with signals lower than this threshold were classified as false negatives, while negative samples above the threshold   were classified as false positives.
For HCV (Fig. 6D), all 56 (out of 56) negative samples were correctly identified as negative by MagIA.43 out of 46 positive samples were detected as positive by MagIA, resulting in a diagnostic sensitivity of 0.93.Further investigation revealed that 2 among the 3 false negative samples had previously been confirmed as DNA negative by the blood bank using a confirmation test (Procleix Ultrio assay from Novartis).This indicates that these patients had been cured of hepatitis C. Therefore, these samples are not relevant for a screening application intended to detect currently infected patients, to provide them with appropriate care and to limit further contaminations.Thus, when compared with traditional blood laboratory tests, MagIA demonstrated an overall concordance of 97 %, a diagnostic sensitivity of 0.93, and a specificity of 1.These results can be compared with previously published results of 87.2 % concordance, 0.8 sensitivity, and 1.0 specificity obtained with the MLFIA approach.
For HIV (Fig. 6E), all 62 (out of 62) negative samples were correctly identified as negative by MagIA.Among the patient samples, all 40 (out of 40) positive samples were detected as positive by MagIA, resulting in a diagnostic sensitivity of 1.When compared with the provided blood laboratory tests, MagIA demonstrated an overall concordance of 100 %, a diagnostic sensitivity of 1, and a specificity of 1.These results can be compared with previously published results of 82.5 % concordance, 0.75 sensitivity and 0.9 specificity obtained with the MLFIA approach.
For HBV (Fig. 6F), 85 (out of 86) negative samples were correctly identified as negative by MagIA, and 1 was considered false positive.Among the patient samples, all 16 (out of the 16) positive samples, with HBV viral loads ranging from 35 to 90,000 UI/mL were successfully detected as positive by MagIA, resulting in a diagnostic sensitivity of 1. Notably, MagIA could detect HBsAg from Hepatitis B genotypes A, B, C and D (Supp.Fig. 4).The detection seemed overall quantitative (R 2 =0,89).When compared with the blood laboratory tests, MagIA demonstrated an overall concordance of 99 %, a diagnostic sensitivity of 1, and a specificity of 0.99.These results can be compared with previously published results of 94.3 % concordance, 0.9 sensitivity and 1.0 specificity obtained with the MLFIA approach.This improvement is visually represented in Fig. 6G, which provides a benchmark comparison between MLFIA and MagIA for HBsAg detection in serum samples against blood laboratory tests.We can further observe that MagIA is more effective at distinguishing between samples with low-titer HBsAg positivity and negative samples when compared to MLFIA.
To demonstrate the ability of MagIA to detect the 3 parameters within a single sample, we combined 3 positive plasmas to create a sample positive for HIV, HBV and HCV.This sample is detected positive for the 3 parameters as illustrated by the purple dots on the Fig. 6 D, E,  and F. This sample was not taken into account for the calculation of the sensitivity and specificity, since it is not coming from a real patient.
In Fig. 6H, we conducted further benchmarking by using a WHOaccredited HBsAg reference sample diluted from 0.15 to 47.3 UI/mL.Relying on the proportionality of the MLFIA signal with the international standard dilution (R 2 =0.99), one could envision quantitative detection capabilities of MagIA for HBsAg.We compared the analytical sensitivity of MagIA for HBsAg to three widely recognized commercial Rapid Diagnostic Tests -RDTs (VIKIA from Biomérieux, Toyo, and Determine TM from Alere).MagIA demonstrated a limit of detection around 0.15 IU/mL, as indicated by the blue dotted line, which is competitive with the analytical sensitivity of 0.1 IU/mL claimed by Toyo and Alere.
The concordance with MagIA detection, when compared to MLFIA, increased by 9.8 %, 17.5 %, and 4.7 % for the three assessed parameters HCV/HIV/HBV, respectively.These results confirm the overall enhancement in our detection performance enabled by the updated consumable and instrument.

MagIA platform validation across various biological sample matrices
To validate the compatibility of the MagIA platform with a point-ofcare workflow and demonstrate its flexibility with respect to sample matrices, we concurrently processed both serum and venous blood samples diluted 10 times from the same patient (Fig. 7A).Regardless of the patient's condition, such as a patient positive for HBsAg (B) or a patient positive for HCV and HIV (C), the results obtained are similar between serum and venous blood.The positive controls are always positive and similar for both matrices.The negative controls are always negative and concordant as well.We compared the results obtained for HBV, HIV, and HCV with the thresholds defined in previous experiments.As anticipated, samples from patients infected with Hepatitis B yield signals surpassing the defined threshold, and the same holds true for samples from patients infected with HIV and hepatitis C and their respective thresholds.
We also compared venous blood and capillary blood.Therefore, we collected 50 µL of capillary blood from a healthy donor's fingerstick, diluted it 10-fold with PBS in a dedicated dropper bottle, and introduced four drops of this sample (100 µL in total) into a MagIA consumable for immediate analysis as illustrated in Figure .The signal was compared with the results obtained from his venous blood.Fig. 7E shows that both matrices give similar results: the positive control is positive, while the chambers 2 to 5 remain negative.
These findings serve as an initial proof of concept for the compatibility of the MagIA platform with venous and capillary blood.

Blood samples
Plasma samples were collected from blood donors who were recruited in accordance with the relevant clinical protocols and provided informed consent.The samples were obtained through collaboration with the French Blood Bank (EFS Tours).Venous blood samples were acquired from patients who were recruited as part of the study.Informed consent was obtained from these patients in accordance with the appropriate clinical protocols at Hôpital Européen in Marseille.For each donor or patient, peripheral blood was collected and decanted in dry tubes to obtain sera and in EDTA coated tubes to obtain plasma.Blood samples were analyzed for HIV, HBV and HCV in central blood labs using chemiluminescent immunoassays (Abbott Prism or Architect) as a benchmark for MagIA.Capillary blood was collected from fingerstick, using a 18 gauge lancet (Sarstedt, 85.1017) and a minivette of 50 µL (Sarstedt, 17.2111.050),diluted 10X with PBS into a small dropper bottle (Measom Freer, 6103 PEX).

MNP functionalization and characterization
MNP were functionalized and characterized as described previously by the authors [45].In summary, after an activation step, magnetic nanoparticles (MNPs) were incubated with 100 μL of the antigen or antibody for 1 h at room temperature, then blocked in a solution of PBS, supplemented with BSA (100 mg/mL), Tween 20 (0.05 %) and proclin (0.01 %).The functionalization of MNPs with ovalbumin, anti-human immunoglobulins G, HIV-1 gp41 recombinant protein, HCV fusion recombinant protein and anti-HBsAg mAb strictly followed the same protocol, using the same amounts of proteins.Functionalized MNPs are stable for up to 1 year when stored at 4 • C [45].

MLFIA cartridge, composition and assembly
The composition and assembly of the MLFIA cartridge were previously described in a publication by the authors [45] and reminded in Supp.Fig. 1.

MagIA cartridge composition
As described in Fig. 2A, the MagIA multiplex cartridges consist of several components, including a magnetic substrate (including a micromagnet surface) as described in [45], a polymer spacing layer, a channel layer with chambers and channels, and a molded polycarbonate shell.The channel layer contains 5 chambers, each measuring 7 mm x 2.4 mm in size, which are connected by 400 µm wide channels to a sample injection port.The polycarbonate shell is patterned with chambers, channels, a sample injection port, and 5 venting ports.This design allows for the efficient processing of samples in the MagIA system.

MagIA cartridge assembly
The polymer, magnetic, spacing and channel layers were assembled using adhesive to form the cartridge substrate.Being made of inert materials, this cartridge substrate can be stored indefinitely under normal storage conditions.Then, reagents were deposited on top of the cartridge substrate, following the method described hereinafter.Once the reagents were dried, the shell was assembled with the cartridge substrate using the channel layer double-sided tape.Cartridges were finally packaged into an aluminium foil containing desiccant.

Reagent integration in MagIA cartridge
In each of the 5 chambers, a combination of functionalized magnetic particles and detection antibodies were deposited (Fig. 2B).Among the 5 chambers, one was dedicated to the measure of patient IgG, as a positive control to ensure the proper functioning of the test.The 4 remaining chambers were functionalized specifically for the assay targeted.To do so, 2 µL of grafted MNPs and 2 µL of detection antibodies were pipetted on the preassembled cartridge substrate, into each chamber of the cartridge (2 drops per chamber, dropped with a home-made dispenser).Then, reagents were dried up at 40 • C during 2 h before final cartridge assembly.

MLFIA Analyzer
The MLFIA Analyzer consists of a miniaturized epifluorescence microscope, adapted with a 3D printed rail and a motor to allow the automatic displacement of the cartridge under the optical module [45].

MagIA Analyzer
It consists of a miniaturized epifluorescence microscope adapted for far-red fluorescence imaging and designed to process and analyze either 1 or 2 MagIA cartridges at the same time.A mechanical module has been designed for the displacement of 2 cartridges.A mixing module, consisting in a Langevin Bolt Clamped piezoelectric transducer, allows the mixing of reagents and sample.The optical module of the Analyzer is similar to the one designed for the MLFIA and enables the chambers imaging.A magnetic module, which consists in an assembly of permanent Neodynium Iron Bore (NdFeB) magnets, allows to capture the

Simulations
To accurately design the piezoelectric actuator for the mixing, the acoustic modes and pressure fields within the fluidic channels were computed with the appropriate modules of COMSOL Multiphysics 5.5 (mechanical, acoustic, electric) and the standard equations of these modules.To save computational time, we only considered the chamber, where the mixing has to occur.The geometry of the chamber was further simplified, assuming 2 symmetric axes, according to Fig. 3B, and the meshing conditions set by Comsol.At the chamber borders, the boundary conditions were defined as followed: (1) Symmetrical conditions on the surfaces of simplification.Mechanical hypothesis: displacement normal at the surface = 0. Acoustic hypothesis: normal pressure field at the surface = 0. (2) Boundary conditions on the other surfaces.Mechanical hypothesis: free surfaces.Acoustic hypothesis: rigid wall (normal pressure field = 0).The speed of the vertical vibration of the magnetic substrate (bottom wall of the channel) induced by the transducer was set at 0.2 m/s.This is the typical speed of the transducer as measured by the laser velocimeter.Based on this model, we analyzed the average pressure in the channel as a function of the frequency, ranging from 20 kHz to 120 kHz, to identify the resonance frequencies of the channel.On each pressure mode, the longitudinal pressure field was modelled.

Bolt Clamped Langevin (BLT) Transducer
We used a Bolt Clamped Langevin Transducer (BLT); a piezoelectric ultrasound transducer, consisting of a series of piezoelectric drivers, usually ceramic disks, compressed between the two masses (front mass and back mass) by a bolt, through the centre of the assembly.BLTs have very low capacitance and high amplitude of displacement compared to other types of piezoelectric actuators, which make them ideal for highfrequency, high amplitude actuation.The back and front masses were designed with respect to the acoustic wavelength, to optimize the efficiency of energy transmission through the front mass.Back and front mass dimensions, material, and clamping torque applied to the piezoelectric drivers, determine the efficiency of coupling.The BLT was designed to maximize the displacement of the transducer when pressure is applied against the cartridge, with respect to the MagIA cartridge chamber resonance frequency of 79 kHz.Due to the manufacturing constraints, the BLT had the actual resonance frequency of ~74 kHz, as measured by the electronic control equipment (Supp.Fig. 4.B).The transducer is controlled by a generator assuring a stable operation frequency, through transducer excitation current feedback control and monitoring a multitude of electrical parameters during operation.This transducer has been mounted in contact with a controlled pressure between its tip and the substrate of the cartridge bottom (e.g. the micromagnets).

Measurement of the transducer displacement
Supp.Fig. 5.A shows a cartridge holder that can move the cartridge horizontally.The transducer is placed under the cartridge holder with a controlled pressure, using a 12 mm diameter cylinder with a 10 mm displacement course driven by an electric pressure regulator.A heightadjustable stop plate is placed on the top of the cartridge holder, drilled for laser measurement.A laser interferometer (Vibrometer laser Polytec, with treatment unit OFV-2500) is placed vertically to measure the deformation of the cartridge substrate.

Assessment of the resonance frequencies
The exact resonance frequencies of both transducers were determined experimentally.Each transducer was mounted in a non-constraining support piece (the white cylinder in Fig. 3), with no load applied to the transducer.During the test, transducers were controlled by the generator, featuring the resonance frequency search mechanism.To determine the resonance frequency, the generator sweeps through a relatively large range of frequencies that holds the theoretical resonance, implying a sweep from 50 to 55 kHz for larger transducer, and 70 to 80 kHz for the smaller one.The resonance frequency is identified by a minimal transducer impedance and an abrupt current phase shift.The results of the test are represented in Supp.Fig. 5C and indicate resonance frequencies of 52 kHz and 74 kHz respectively for two transducers considered.

Qualitative assessment of the mixing
A test bench was developed to assess the relative mixing performance of transducers (Supp.Fig. 5.D).The purpose of this test bench was to enable a rapid and qualitative testing of the mixing for a given transducer.The test bench was thus developed with high flexibility rather than precision in mind, allowing to accommodate both transducer sizes, set any desirable transducer push force, execute magnetic capture followed by the image acquisition, and observe the liquid dynamics through the microscope.The test bench was operated half-manually, which, as expected, introduces significant uncertainties when trying to quantify the obtained results.The transducer was brought into contact with the cartridge by a lever, the transducer push force F being defined by the counterweight P applied to the opposite lever end.The test bench frame can execute the magnetic capture and image acquisition, as done by the actual MagIA Analyzer.There is a multitude of factors influencing the image quality, such as microscope focus, magnetic capture dynamic, transducer angle of attack, incubation time, etc.These parameters could not be controlled in this test bench and introduced a certain amount of variability to the images.Hence, only a qualitative set of criteria could be used to assess the mixing quality: homogeneity of the background fluorescence, clearly visible signal lines, absence of air bubbles or dense fluorescent particle clouds obstructing the view.

Discussion
In this work, we presented our efforts to transfer a research platform (MLFIA) to a comprehensive rapid multi-analysis point-of-care solution (MagIA) compatible with capillary blood.The MagIA platform was first characterized using a biological model focused on the detection of anti-OVA monoclonal antibodies.The findings revealed a limit of detection ranging between 14 and 40 ng/mL, a linear range spanning from 40 to 3300 ng/mL, similar to the results obtained by MLFIA.Subsequently, the system was adapted for the detection of IgG to evaluate its reproducibility in biological samples, specifically human serum, considering parameters such as intra-cartridge coefficient of variation (CV) and inter-chamber variability.We finally adapted the MagIA platform for the detection of anti-HCV, anti-HIV and HBsAg from a same sample and analyzed 102 plasmas from positive patients.43/46, 40/40 and 16/16 samples positive for HCV, HIV and HBV respectively were detected as positive by MagIA, corresponding to overall concordances of 97 % (Se = 0,93, Spe = 1), 100 % (Se = 1, Spe = 1) and 99 % (Se = 1, Spe = 0.99).We showed with HBsAg international standard that the detection limit is close to that of CE marked and FDA approved rapid tests with the potential of being quantitative with a signal proportional to the standard dilution as demonstrated in Fig. 6H.Lastly, we assessed the compatibility of MagIA platform with veinous and capillary blood samples as a final step towards its POC validation.
Using these results, we can compare the new MagIA platform with the legacy MLFIA platform.Compared to the MLFIA cartridge, the updated MagIA consumable is better suited for POC diagnostic purposes.It incorporates dried reagents (functionalized MNPs and detection Ab), simplifying the user experience by reducing the number of pipetting events.Furthermore, this consumable is compatible with capillary blood sampling tools used in RDT, such as lancets, disposable pipettes, and dropper bottles.The need for an additional centrifugation step to separate plasma from whole blood has been removed.This eliminates the dependence on laboratory facilities, extra equipment, and skilled technicians.The cartridge exhibited promising stability at room temperature over a 6-month period using ovalbumin (OVA) model reagents.The revised cartridge design now allows for the parallel analysis of five assays -three parameters and two internal controls within five chambers, all connected by a single filling port, for enhanced robustness and reliability.The design and assembly of the cartridge remain straightforward, requiring the assembly of five compounds plus the reagents over eight assembly or transformational steps.Concordances with laboratory reference method for HCV, HIV and HBV detection increase of 9.8 %, 17.5 % and 4.7 % respectively compared to MLFIA, confirming that the integration of the technology into an automatized immunoassay was successful and knocked down the technological barriers towards a triplex POC for STI screening.These results show a pronounced improvement with MagIA compared to MLFIA, not observed in the OVA experiments.This disparity can be attributed to the operators' ability to uphold uniformity in MLFIA's OVA experiments, involving only 3 cartridges and 9 samples, thereby ensuring consistency in incubation and magnetic capture times.However, challenges arise in HIV/HBV/HCV experiments, where MLFIA necessitates the manual handling of 10 cartridges (for 80 samples), resulting in difficulties in achieving uniform incubations and magnetic captures.
With such a POC triplex capable of providing the full infectious status of patients during a single consultation, testing practices and patient care pathways can be significantly simplified, aligning with the objective of identifying the "missing millions" of infected individuals who are unaware of their infections.This advancement offers advantages for all the actors: To distinguish our approach from other existing methods (Table 1), we took into consideration from early on the ASSURED criteria defined by the WHO for point-of-care testing: Affordable, Sensitive, Specific, User-friendly, Rapid and Robust, Equipment free, and Deliverable to end-users [47].The MagIA cartridges are scalable for industrial production (Deliverable) and cost-effective (Affordable).The platform is compatible with capillary blood, and the user workflow is as straightforward as that of rapid immunochromatographic strip tests, however with the advantage of giving objective results (User friendly).MagIA can be conducted in any location, even without access to laboratory equipment (Equipment free) or power sources.Results are given in 30 min (Rapid) with performances aligning with the anticipated diagnostic requirements (Sensitive and Specific).
As stated before, most RDT (for STIs) are mono-disease.Managing multiple tests can be challenging for the medical personnel involved.In comparison, MagIA offers a syndromic approach, providing a complete panel of analyses typically conducted by STI clinics in just 30 min, with comparable performances.Such approach enables the detection of co-infections in the same testing, making testing much more efficient.When compared to the Biosynex triplex system, the MagIA Analyzer provides an objective readout and enhanced traceability, all while maintaining the simplicity of execution similar to that of a lateral flow rapid test.The cartridge, containing 5 chambers, offers the possibility to increase the number of parameters to be detected using only 50 µL of capillary blood.In place of the negative control, the detection of a protein of interest, such as p24 for early detection of HIV infection or syphilis, which is another common co-infection of HIV, could be considered.The remaining positive control will still enable us to verify the accurate execution of the test.Additionally, our workflow was thoughtfully designed to accommodate the filling of two cartridges with just 50 µL of capillary blood, while the Analyzer can simultaneously process two cartridges.This setup has the potential to increase the number of tested parameters per fingerstick to 8.
The ability of MagIA to detect both antibodies and antigens on the same panel and from the same sample, with performance levels approaching the required standards, showcases its high degree of flexibility.Such flexibility often raises concerns about specificity in multiplexed strip tests, where both detections occur simultaneously.To address this concern, the MagIA consumable was designed to perform multiple assays in parallel from the same sample, as opposed to a multiplexed approach that conducts multiple analyses from the same sample within the same reaction chamber.Our format offers several advantages by preventing cross-reactions, thereby improving overall sensitivity and specificity.It also simplifies future developments.A parallelized assay is modular; once developed and validated, a specific assay can be easily replaced with another to create custom panels tailored to different testing practices.Lastly, as demonstrated with the OVA model and HBsAg detection, the assay seems to have the potential for quantitative measurements, which can be valuable for several parameters where patient care depends on the concentration of the target molecule (such as PCT, CRP, etc.).
As part of our WHO compliance efforts, we have also factored in the environmental aspect in our engineering development.The global response to the coronavirus pandemic has highlighted the increased availability of diagnostic tests for public health, which has implications for both human but also environmental health.This impact encompasses concerns related to liquid waste and plastic waste [48,49].As an example, according to a recent WHO report, more than 140 million test kits have been shipped through the United Nations procurement portal alone during the COVID-19 pandemic, with the potential to generate 731,000 liters of chemical waste [49].Additionally, the production and disposal of single-use devices, regardless of the assay format, also present a challenge.These devices are often made from unsustainable polymeric materials derived from fossil sources, and their disposal necessitates specific safety procedures.It is crucial to address these environmental and human factors, particularly in low-income countries where waste management infrastructure is often inadequate, despite the anticipated significant increase in the use of point-of-care tests.In the development of the MagIA platform, we began by addressing issues related to biological and chemical waste.The elimination of washing steps in our workflow, minimal pipetting requirements, and compatibility with capillary blood collectively contribute to a significant reduction in liquid waste generation.Furthermore, our approach substantially reduces plastic consumption, resulting in a threefold reduction in plastic volume (1 cartridge, 1 pipette, 1 lancet, and 1 diluent bottle, all serving three assays).Furthermore, considering our initial target market for STI screening campaigns in Africa, we are exploring strategies for local consumable manufacturing, which may involve small, semi-automated manufacturing instruments that can be housed in a mobile container.The next steps involve the following: (1) Enhancing the signal quality to increase the differentiation between positive and negative signals; (2) Further expediting the overall test duration by minimizing mixing, incubation, and capture steps; (3) Reducing test variability by refining the reproducibility of reagent deposition and enhancing the Analyzer's robustness.Then, we will analyze numerous negative samples for HIV, HBV, and HCV to establish distinct threshold values.Once identified, these thresholds will be integrated into the Analyzer, enabling users to receive qualitative (positive/negative) test results instead of raw arbitrary unit data.This will then allow us to evaluate the technology in a clinical trial on a large cohort of patients, utilizing capillary fingerstick blood samples, in order to finally clinically validate the MagIA technology to initiate the CE and FDA certification.Simultaneously, to facilitate widespread implementation, the consumable must undergo testing for long-term stability and be exposed to real-world, non-laboratory conditions, including varying temperatures.In the future, we plan to expand the range of tests available on the MagIA platform, adding further assays to create a comprehensive serological panel.This expanded panel could include tests for antibodies against Hepatitis B surface antigen (anti-HBs), Syphilis, or p24 antigen for earlier HIV infection detection.Additionally, due to the versatility, user-friendliness, and the capability to use capillary blood, the technology has the potential to be applied in various other contexts.These potential applications may include managing infectious diseases in the field among at-risk populations, emergency diagnostics (e.g., in emergency rooms), or even routine monitoring by patients without the direct assistance of healthcare professionals.The ability to process small blood volumes also opens the door to applying such assays to rare samples, including but not limited to newborn testing.Beyond healthcare applications, this approach could be applicable to routine tests in the food industry or biodefense, such as the detection of toxins like Ricin or Botulinum.

Fig. 1 .
Fig. 1.MagIA technology and protocol.(A) MagIA workflow.(B) Schematic description of MagIA technology.Functionalized MNPs and detection antibodies are dried in the 5-chambers cartridge.The sample is injected, and the cartridge inserted into the MagIA Analyzer, where analytes and reagents (MNPs and detection Ab) are mixed and incubated for 10 min, to enable the formation of the immune complexes.These complexes are then captured by the micro-magnets after magnetic activation, and their fluorescence signal detected.(C) Integrating those lines yields a signal proportional to the immune complexes formed.

Fig. 2 .
Fig. 2. MagIA cartridge.(A) Schematic of the cartridge composition, and its different layers.(B, C) Zoom on the assay side, with 1 sample injection port for sample injection, 5 reaction chambers for 5 assays led in parallel, and 5 venting ports to enable sample capillary migration within the fluidic channels.(B) Cross section of the cartridge observed with a polarized bright field microscope.(C) Schematic view of the cartridge assay side and dimensions.(D) Hydrophilization of the fluidic channels for better wettability; (top) brightfield imaging of the channels without or with 5 pulverizations of 1 % Pluronic®, (middle) contact angle measurements, (bottom) visualization of a blood injection in the chambers.

Fig. 3 .
Fig. 3. (A) Description of the mixing module, with the cartridge in contact with the Langevin Bolt Clamp piezoelectric transducer (blue and purple).(B) Design geometry selected for the acoustic simulation.(C) Simulation of the channel vibration modes for the geometric model described above.(D) Representation of the cartridge displacement as a function of the contact force applied, respectively for transducers at resonance frequencies of 52 kHz (left) and 74 kHz (right).(E) Cartridge displacement as a function of the fluidic chambers.The displacement of the micro-magnets surface could not be measured in chamber 2. (F) Example of a cartridge damaged (here; melted) by too much contact force applied.

Fig. 4 .
Fig. 4. First validation of MagIA technology on a simple biological target (anti-OVA mAb), in a PBS buffer.(A) Schematic of the reaction.(B) Schematic of the fluidic cartridge functionalization.(C) MLFIA signal, linear range of detection and LOD, indicated with a blue dotted line.(D) MagIA performance versus MLFIA approach, at similar experimental conditions, for the detection of anti-OVA mAb in PBS (N=5 for MagIA, N=2 for MLFIA).(E) Stability of MagIA cartridges over time, with MagIA signal for different concentrations of anti-OVA Ab, at Month 0 and Month 6 following cartridge fabrication.

Fig. 5 .
Fig. 5. Variability characterization of MagIA technology with a simple biological target (human IgG), in a complex biological fluid (serum).(A) Schematic of the reaction.(B) Schematic of the fluidic cartridge functionalization.(C) MagIA signal as a function of cartridge chambers (N=18), for intra-cartridge variability.(D) MagIA signal as a function of cartridge number (N=18), for inter-cartridge variability.

Fig. 6 .
Fig. 6.Validation of our technology on clinical targets; HCV, HIV, HBV.(A) Schematic of the fluidic cartridge composition.(B) Illustration of the experimental workflow.(C) Functionalization pattern per chamber.(D, E, F) MagIA signal obtained for 102 patient serum samples, with comparison to MLFIA results.For all benchmarking experiments, MagIA cut-off is defined as Mean (Neg controls) + 3σ and indicated with a grey dotted line.The orange dots indicate the samples corresponding to patients identified as positive by the blood central lab.The blue dots are samples identified as being negative for the pathology considered.The purple dots indicate a sample made of 3 positive plasma to create a sample positive for HIV, HBV and HCV.(G) Benchmarking of MLFIA (in orange) and MagIA (in blue) versus blood lab test for HBsAg detection in serum samples.Cut-offs are indicated with dotted lines.(H) Benchmarking of MagIA versus commercial RDTs using a standard sample from NIBSC WHO; VIKIA test, from Biomerieux, and Toyo and Alere tests.Each cut-off is indicated with an orange dotted line.A blue dot represents the blank value for MagIA.For both benchmarking experiments, MagIA cut-off is defined as Mean (Neg controls) + 3σ, and indicated with a blue dotted line.

Fig. 7 .
Fig. 7. First assessment of our technology on veinous blood samples.(A) Schematic of the fluidic cartridge composition.(B, C) MagIA signals obtained for serum versus veinous blood, collected from 2 different patients; a patient positive for HBsAg (B), and a patient positive for HCV and HIV (C).(D, E) Application of MagIA detection workflow to capillary blood from a healthy donor.The capillary blood is collected on-site and processed right away (D), to provide a MagIA signal in 30 min including the sampling time (E).The dash dotted lines indicate the positivity cut-off defined for each reagent.

( 1 )
Key Populations: It allows healthcare providers to reach key populations (often unaware of their risks of infection) during outdoor campaigns, that are often HIV-focused.MagIA would reduce missed opportunities for detecting co-infections and increase the chances of early diagnosis and treatment.(2) Healthcare Providers: Community centers are aiming to enhance their time and budget efficiency while expanding their outreach to more patients during outdoor campaigns.STI clinics are looking to minimize patient waiting times, ensuring that everyone entering their facilities receives prompt results and appropriate treatment when testing positive.(3) Public Health: Public health organizations can explore innovative solutions for combined STI testing, aligning with the World Health Organization's (WHO) goal of eradicating HIV, HBV, and HCV by 2030, especially in high-risk communities with limited budgets.

Table 1
Comparison of RDTs, laboratory tests, MLFIA and MAGIA according to the ASSURED criteria defined by the WHO for point-of-care testing.