Quantum-enhanced biosensing for early detection of neurodegenerative disorders
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Journal of Clinical and Translational Research |
| Published | 2026-07-02 |
| Authors | Innocent Ojeba Musa, Miracle Uwa Livinus, Mustapha Abdulsalam, Stephen Olaide Aremu, Sunday Zeal Bala |
| DOI | 10.36922/jctr025440076 |
A diamond nitrogen-vacancy biosensor reached detection limits of 12.6 fM for Aβ42 and 18.2 fM for α-synuclein in plasma and CSF. These limits were 182.5-fold and 307.7-fold below the parallel enzyme-linked immunosorbent assay, or ELISA, values. In 60 patients with clinical Alzheimer’s disease or Parkinson’s disease and 60 controls, the three-marker assay reached AUCs of 0.94 to 0.96, used 50 µL of sample, and returned results in 25 min.
Sensor architecture and quantum readout
Section titled “Sensor architecture and quantum readout”The sensor combined antibody recognition with optically detected magnetic resonance in a synthetic diamond substrate. Controlled nitrogen-ion implantation and annealing created near-surface nitrogen-vacancy centers, referred to as NV centers. Center density was approximately 10/μm², and implantation depth was 5 to 10 nm to couple the spins to surface-bound analytes.
Gold nanoparticles were immobilized through thiol-silane linker chemistry. Monoclonal antibodies for Aβ42, p-tau181, and α-synuclein were coupled through 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and N-hydroxysuccinimide chemistry. Labeled secondary-antibody measurements gave conjugation efficiencies of 78% to 85% across batches.
Optically detected magnetic resonance, or ODMR, measured the resonance shift. The authors attributed the shift to local electromagnetic perturbations after capture on the antibody-functionalized nanostructures. These included mass loading and charge redistribution near the surface. The optical setup used a 532 nm continuous-wave laser at 50 mW, a 0.9 numerical-aperture objective, and a 637 to 800 nm bandpass filter. A patterned coplanar waveguide supplied microwave excitation.
Lock-in detection separated weak resonance changes from background. Time gating rejected plasma autofluorescence against the approximately 12 ns NV fluorescence lifetime. On-resonance and off-resonance measurements corrected baseline drift, while polyethylene glycol passivation reduced nonspecific adsorption. Lorentzian fits supplied the resonance shifts. Patient readouts used unspiked samples, so the frequencies reflected endogenous biomarker concentrations.
Clinical cohort and assay design
Section titled “Clinical cohort and assay design”The cohort comprised 120 participants. Sixty had a clinical diagnosis of Alzheimer’s disease or Parkinson’s disease, and 60 were age-matched controls. Mean ages were 68.4 ± 7.2 years for the disease group and 66.8 ± 6.9 years for controls. Fifty-two percent were female.
Cerebrospinal fluid and plasma were tested for endogenous Aβ42, p-tau181, and α-synuclein. Interference studies included albumin, immunoglobulin G, and fibrinogen at physiological concentrations. Each biomarker was also measured by commercial ELISA using kit KHB3441.
Calibration used standards and triplicate measurements, with weighted regression in log-log space. The detection limit used a three-standard-deviation blank criterion. Quantitation required a coefficient of variation of 20% or less and recovery of 80% to 120%. Precision used eight replicates within a run. Inter-assay testing used six procedures over three days with sensors from different fabrication batches.
Detection limits and signal-to-noise
Section titled “Detection limits and signal-to-noise”The measured detection limits were lower than the parallel ELISA limits in plasma and CSF.
| Biomarker | Quantum LOD | ELISA LOD | Improvement |
|---|---|---|---|
| Aβ42 | 12.6 fM | 2,300 fM | 182.5-fold |
| p-tau181 | 15.8 fM | 4,200 fM | 265.8-fold |
| α-synuclein | 18.2 fM | 5,600 fM | 307.7-fold |
Thus, the study measured 182.5-fold, 265.8-fold, and 307.7-fold improvements for Aβ42, p-tau181, and α-synuclein. ELISA limits were 2.3 to 5.6 pM, while the quantum sensor reached 12.6 to 18.2 fM.
The platform increased signal-to-noise ratio by approximately 22 dB in patient specimens. The scaling comparison gave quantum and classical values of 10 and 3.2 at 10² collected photons. At 10⁶ photons, the values were 1,000 and 31.6. The authors related this behavior to square-root photon scaling for quantum readout and fourth-root scaling for classical fluorescence.
Time gating, differential ODMR, and polyethylene glycol passivation supported performance in plasma. These measures addressed autofluorescence, baseline drift, and nonspecific protein adsorption.
Diagnostic performance and multiplexed workflow
Section titled “Diagnostic performance and multiplexed workflow”The sensor separated clinically diagnosed patients from controls more accurately than the parallel ELISA measurements.
| Biomarker | Quantum AUC | ELISA AUC |
|---|---|---|
| Aβ42 | 0.96 | 0.78 |
| p-tau181 | 0.94 | 0.72 |
| α-synuclein | 0.95 | 0.81 |
Each difference in AUC was significant at p < 0.001. Sensitivity and specificity exceeded 90% for all three biomarkers, compared with 65% to 75% for ELISA.
The multiplex assay measured three biomarkers in 50 µL of plasma or CSF and finished in 25 min. Three separate ELISAs required 120 µL and 90 min. The authors reported a 60% sample-volume reduction and more than threefold faster turnaround.
Analytical validation and reproducibility
Section titled “Analytical validation and reproducibility”Calibration curves were linear, with R² > 0.99 across six orders of magnitude. The stated range was 50 to 100,000 fM for each biomarker. Limits of quantitation were 40 fM for Aβ42, 45 fM for p-tau181, and 50 fM for α-synuclein.
Intra-assay coefficients of variation ranged from 3.5% to 7.5%, and inter-assay coefficients ranged from 5.8% to 10.8%. Mean recovery in plasma and CSF was 92% to 105%. Cross-reactivity stayed below 3% in tests with high excess albumin, immunoglobulin G, and fibrinogen.
After seven days at 4 °C, the chips retained 91%, 93%, and 90% signal for Aβ42, p-tau181, and α-synuclein, respectively. Each estimate used five chips. Between-batch coefficients were 7.2%, 8.1%, and 7.8% across three fabrication cycles. Each cycle produced 24 chips, with interchip variation below 8%.
Comparison with other assays and translation limits
Section titled “Comparison with other assays and translation limits”The detection limits are comparable to the digital immunoassays discussed by the authors. Simoa was cited for subfemtomolar serum detection and, under ideal conditions, concentrations near 10⁻¹⁹ M. The NV sensor instead uses spin resonance perturbations rather than enzymatic amplification and digital counting.
Graphene FET sensors were reported at 1 to 10 fM for tau and amyloid proteins under controlled conditions, but the paper notes reproducibility problems in clinical plasma. SPR and nanophotonic resonators were described as label-free methods with low-picomolar sensitivity. Mass spectrometry has established reproducibility and accuracy, but the paper notes its specialized equipment and centralized laboratory infrastructure.
Seed amplification assays detect aggregation-competent α-synuclein seeds. The paper treats seed amplification as a test of aggregation competence and quantum sensing as a total concentration measurement.
In the reported cohort, ROC analysis separated clinically diagnosed patients from controls. It did not establish presymptomatic detection, longitudinal conversion, or separate performance for Alzheimer’s disease and Parkinson’s disease. The noninvasive claim applies to the plasma route, not lumbar-puncture-derived CSF. The population was 61% Caucasian, and the authors call for multicenter validation with broader populations and longitudinal sampling.
Production must preserve antibody stability and orientation while scaling near-surface NV fabrication. Surface chemistry must limit nonspecific adsorption without degrading spin coherence. The authors also identify optical miniaturization, point-of-care integration, and microfluidic sample preparation as unresolved tasks.
For material selection, the device needs a diamond sensing surface supporting shallow NV centers, gold nanoparticle attachment, surface passivation, and antibody coupling. The bench system combines laser excitation, photon counting, microwave delivery, and high-numerical-aperture collection. A portable system must retain the measured recovery, precision, and seven-day stability. The authors suggest using the assay as an initial screen and confirmatory mass spectrometry for follow-up, but the study did not test a clinical screening pathway.
Materials and support
Section titled “Materials and support”For Alzheimer’s and Parkinson’s biomarker biosensing, the paper needs a synthetic diamond substrate with a shallow NV layer, surface gold for antibody-conjugated nanoparticles, and a passivated interface. 6CCVD’s diamond materials catalog lists single-crystal and polycrystalline diamond grown by MPCVD and metallization with Au, Pt, Pd, Ti, W, or Cu. The engineering support service can assess custom fabrication of the substrate and metal layers, while the paper’s implantation, annealing, and antibody-conjugation steps remain part of the device build.
View Original Abstract
Background: Early detection of neurodegenerative diseases is an important and unmet clinical need because traditional assays are not sensitive enough to detect low concentrations of biomarkers at the prodromal stages. Objective: This study presents a novel biosensing system based on quantum measurement principles and biomolecular recognition to achieve ultrasensitive detection of biomarkers for Alzheimer’s disease (AD) and Parkinson’s disease (PD) in patient samples. Methods: We developed a hybrid sensor that uses antibody-functionalized nanostructures for selective biomarker capture and nitrogen-vacancy centers in diamond for quantum spin-based readout. The performance was evaluated against standard fluorescence immunoassays of cerebrospinal fluid (CSF) and plasma collected from a clinical cohort (n = 120; 60 AD/PD, 60 controls) in endogenous amyloid-β42 (Aβ42), phosphorylated tau (p-tau181), and α-synuclein. Results: The quantum-enhanced platform detection limits of 12.6 fM for Aβ42, 15.8 fM for p-tau181, and 18.2 fM for α-synuclein are 150–300-fold better than enzyme-linked immunosorbent assay (ELISA). The signal-to-noise ratio increased by ~22 dB, within the quantum-limited scaling. Diagnostic accuracy was high, with an area under the receiver operating characteristic curve ranging from 0.94 to 0.96, superior to ELISA (0.72–0.81). In addition to sensitivity, analytical validation has excellent reproducibility (intra-assay coefficient of variation [CV] ≤ 7.5%, inter-assay CV ≤ 10.8%), high recovery in patients’ plasma and CSF (92–105%), low cross-reactivity (<3%), and a stable sensor performance of seven days of storage. Multiplexed detection reduced the sample volume by 60% and the assay turnaround time by more than threefold compared to ELISA. Conclusion: Quantum-enhanced biosensing is a powerful diagnostic method for detecting biomarkers at clinically relevant concentrations, providing a scalable, noninvasive approach to early diagnosis of neurodegenerative diseases in affected populations. Relevance for patients: Clinical implementation of quantum biosensing has the potential to enable pre-symptomatic disease detection, facilitating earlier therapeutic intervention and improved disease management.
High Concentration Quantum Grade Single Crystal Diamond
View Specs →Products
View Specs →