Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Feb 1.
Published in final edited form as: Small. 2023 Sep 26;20(6):e2305110. doi: 10.1002/smll.202305110

Quantitative detection of thyroid-stimulating hormone in patient samples with a nanomechanical single-antibody spectro-immunoassay

Peng Zheng 1,†, Piyush Raj 1,†, Lintong Wu 1, Takayuki Mizutani 2, Miklos Szabo 2, William A Hanson 2, Ishan Barman 1,3,4,*
PMCID: PMC10922205  NIHMSID: NIHMS1934917  PMID: 37752776

Abstract

Functional disorders of the thyroid remain a global challenge and have profound impacts on human health. Serving as the barometer for the thyroid function, thyroid-stimulating hormone (TSH) is considered the single most useful test of the thyroid function. However, the prevailing TSH immunoassays rely on two types of antibodies in a sandwich format. The requirement of repeated incubation and washing further complicates the issue, making it unable to meet the requirement of the shifting public health landscape that demands rapid, sensitive, and low-cost TSH test. Herein, we performed a systematic study to investigate the clinical translational potential of a single antibody-based biosensing platform for TSH test. The biosensing platform leverages Raman spectral variations induced by the interaction between a TSH antigen and a Raman molecule-conjugated TSH antibody. In conjunction with machine learning, it allows TSH concentrations in various patient samples to be predicted with high accuracy and precision, which is robust against substrate-to-substrate, intra-substrate, and day-to-day variations. We envision the simplicity and generalizability of this single-antibody immunoassay coupled with the demonstrated performance in patient samples pave the way for it to be widely applied in clinical settings for low-cost detection of hormones, other molecular biomarkers, DNA, RNA, and pathogens.

Keywords: Thyroid-stimulating hormone, Patient sample, SERS, Machine learning, Immunoassay

Graphical Abstract

We presented a single antibody-based biosensing platform for detecting TSH in patient samples. The biosensing platform leverages the Raman spectral variations induced by the interaction between a TSH antigen and a Raman molecule-conjugated TSH antibody. In conjunction with machine learning, it allows TSH concentrations in various patient samples to be predicted with high accuracy and precision, which are robust against substrate-to-substrate, intra-substrate, and day-to-day variations.

graphic file with name nihms-1934917-f0005.jpg

1. Introduction

Functional disorders of the thyroid are widespread.1–2 It is estimated that, about 200 million people worldwide are affected by thyroid diseases;3 in the United States alone, there are about 20 million people affected by thyroid disorders, and over 12 percent of the population are also expected to develop a thyroid condition during their lifetime.4 Clinically, thyroid dysfunctions have profound impacts on human health.5–7 They can adversely affect childhood development, pregnancy outcome, mental health, metabolic functions, and even the cardiovascular risk factors.8–9 Serving as the barometer for thyroid function, thyroid-stimulating hormone, or TSH, is considered the single most useful test of the thyroid function.10 TSH is released from the pituitary gland, which in turn regulates the growth of the thyroid gland and the production of thyroid hormones, such as free thyroxine (T4) and free tri-iodothyronine (T3).8, 10–11 Importantly, TSH is negatively regulated by free T4 and free T3 in a log-linear manner. Subtle changes in the levels of free T4 and free T3 can lead to significant variations in the level of TSH. Hyperthyroidism and hypothyroidism, which occur when TSH levels vary beyond the normal reference range, are typical manifestations of a certain thyroid condition. As the first-line test to diagnose thyroid disease, TSH test dominates global thyroid function test market, the size of which was valued at USD 1.68 billion in 2021.12 Catalyzed by the increasing prevalence of thyroid-related diseases such as Graves’ disease, Hashimoto’s disease, and thyroid cancer, the global thyroid function test market size is expected to continue to grow and is projected to reach USD 2.56 billion by 2028.12

Despite the clinical importance and enormous market prospect of TSH test, the prevailing TSH immunoassay design, which relies on two types of antibodies in a sandwich format and requires repeated incubation and washing, lags behind the shifted landscape of public health that demands rapid, sensitive, and low-cost test of TSH.13 Efforts have been devoted to developing lateral flow immunoassays for rapid TSH test based on colorimetric14 and surface-enhanced Raman scattering (SERS)15–16 transducers. In spite of the ease of operation on a lateral flow platform, colorimetric assays lack the desired sensitivity to detect analytes at a lower concentration beyond the cutoff threshold, nor can they quantify the concentration of detected analytes. SERS-based lateral flow assays can in principle simultaneously enable rapid and ultrasensitive detection owing to the remarkably enhanced SERS signals by the intense plasmonic fields. Nevertheless, even the state-of-the-art intensity only-based assays,15–18 chemiluminescent assays19–20 included, suffer from non-negligible intensity fluctuations at a high magnitude and/or false signals owing to non-specifical binding and signal interferences. Therefore, it is imperative to develop innovative methods that enable robust, rapid, sensitive, and low-cost TSH test.

Inspired by the recent discovery of SERS frequency shift21 of antibody-conjugated Raman molecules after antibody-antigen binding, we have recently developed a dual-modal single-antibody spectro-immunoassay for small molecule detection by harnessing such unique SERS frequency shift along with the SERS intensity change.22 By combining the SERS frequency shift with chemometric analysis, we have also invented a single antibody-based serum biomarker detection platform, which, under controlled conditions, displayed the desired performance for TSH detection in goat serum, where TSH was artificially spiked into the goat serum matrix.23 Success in these studies highlight the clinical translational potential of the SERS frequency shift method24–26 that relies on the nanomechanical perturbations to Raman molecules because of the interaction between antibody and antigen, which ends up transducing a robust signal in the form of frequency shift instead of the peak intensity variation.

In this study, we implemented a systematic study to investigate the clinical translational potential of the SERS frequency shift method that relies on only a single antibody for detecting TSH in patient samples. We standardized the plasmonic substrate preparation protocol in combination with 3D printing to prepare compartmentalized cells on gold nanopyramid arrays, which would allow high-throughput, cross-interference-free, and scalable bioanalytical applications. We established the protocol for studying patient samples with various TSH concentrations across different plasmonic substrates over multiple days. By combining the SERS peak shift-based analysis with machine learning (ML) analysis, we investigated the capability and detection precision of this biosensing platform for TSH test in patient samples. The robustness of this method was tested by selecting data from one of the three runs as the calibration spectra to do the prediction based on the data from an independent run on a different day. Informed by the importance of monitoring the thyroid functions, this work presented a clinically applicable and simplified immunoassay for TSH test. More importantly, with TSH as the benchmark biomarker, and with further optimization through studies featuring additional patient samples, we expect the single antibody-based immunoassay platform could be made ready for clinical translation.

2. Results and discussion

2.1. Experimental protocol

We leveraged an experimental protocol22–23 we recently established to prepare the plasmonic gold nanopyramid array substrate for studying TSH patient samples. The gold nanopyramid array is a proven plasmonic substrate with superior capability for enhancing SERS and fluorescence signals.22–23, 27–31 Detailed experimental methods and scanning electron microscopy images for the gold nanopyramid array plasmonic substrate were supplied in the supporting information (Section S2 and Fig. S1). Briefly, as schematically shown in Fig. 1a, the plasmonic substrate was first functionalized with the Raman molecule 4-mercaptobenzoic acid (4-MBA) and TSH monoclonal antibody. We subsequently employed 3D printing to compartmentalize the plasmonic substrate into smaller 3 × 3 cells (Fig. S2). Each cell could thus be devoted to studying a given TSH patient sample without interference from a different sample in other cells. It is worth noting that, the prevailing methods for SERS biosensing require repeated analyte incubation and washing on the same plasmonic substrate, which imposes a considerable challenge for subsequent accurate measurements on the substrate owing to the changing surface conditions, also called the SERS memory effect.32 Therefore, the 3D printing method for substrate compartmentalization provides a high-throughput, cross-interference-free, and scalable strategy for studying TSH patient samples. Subsequently, TSH patient samples with varying levels of TSH antigens (5 μL for each) was pipetted into each cell. After incubation at 37 °C for 20 min, Access Wash Buffer was used to remove excessive or non-specifically adsorbed TSH antigens and other substances. Afterwards, Raman mapping was performed to scan each cell.

Figure 1. Experimental protocols.

Figure 1.

Open in a new tab

(a) Method for substrate preparation: the gold nanopyramid array substrate was first functionalized with Raman molecule 4-mercaptobenzoic acid (4-MBA) and TSH antibody; 3D printing was used to compartmentalize the substrate into smaller cells; TSH patient samples (5 μL for each cell) were then pipetted into each cell; after incubation at 37 °C for 20 min, Access Wash Buffer was used to remove excessive or non-specifically adsorbed TSH antigens and other substances; Raman mapping was subsequently conducted, where each cell was mapped by 5×5 Raman spectra. (b) Protocol for studying TSH patient samples: nine different levels of TSH patient samples were pipetted onto the 3D-printed cells on three different 4-MBA and TSH monoclonal antibody-functionalized gold nanopyramid array substrates. For each run, which was completed on the same day, there were three repeats. A total of three runs were performed on three consecutive days.

To assess the performance of this method for studying TSH patient samples, we implemented a systematic protocol, as schematically depicted in Fig. 1b, which includes three runs on nine different plasmonic substrates over three consecutive days. Each run consists of three repeats. Therefore, variables, such as substrate-to-substrate variation, intra-substrate variation, and day-to-day variation, have all been taken into account in this protocol. For each studied sample in the 3D-compartmentalized cell, a total of 5×5 Raman spectra were collected over an area of 20 μm × 20 μm with a laser excitation wavelength of 785 nm.

2.2. SERS frequency shift-based analysis for TSH patient samples

Following the protocol in Fig. 1b for SERS spectra collection for various TSH patient samples, we first plotted the mean SERS spectra focusing on the peak at about 1580 cm−1. This characteristic Raman peak originates from the vibrational mode of the benzene ring in 4-MBA.21 It has been found to be vulnerable to nanomechanical perturbation, which can induce a subtle frequency shift under deformation.21, 33–38 Indeed, the mean SERS spectra as shown in Fig. 2a displayed a gradual peak shift with an increase of the TSH level in the patient samples, marked from C1 to C9. To provide a global picture on the evolution of the SERS frequency, we presented, in Fig. 2b, the concentration-dependent heat map reconstructed from the SERS frequency shift averaged for each run as well as averaged for all three runs. The SERS frequency for the 1580 cm−1 peak was found to be strongly correlated with the TSH concentrations in the patient samples. Regression analysis based on the SERS frequency shift from all the three runs further confirmed such a correlation with a R2 of 0.95, as shown in Fig. 2c.

Figure 2. SERS frequency shift-based analysis for TSH patient samples.

Figure 2

Open in a new tab

(a) Measured SERS spectra averaged from the three runs for TSH patient samples with varying levels of TSH antigens from C1 to C9. The spectra were intentionally vertically offset for easy visualization. (b) Mapping of the SERS peak frequency shift at about 1580 cm−1 for varying levels of TSH antigens as specified. For the top three rows SERS frequency shift mapping, they were respectively averaged from each run. The bottom row of SERS frequency shift mapping was obtained by averaging all runs. The relative SERS frequency shift was defined as Δν=|νc-νm|, where νc is the measured SERS peak frequency for the TSH patient sample with a concentration of C, νm is the measured SERS peak frequency that have a maximum wavenumber. (c) Regression analysis based on the frequency shift of the SERS peak at about 1580 cm−1.

In this study, for the same run, samples marked as C1 to C3 were tested on the same plasmonic substrate; C4 to C6 were tested on another substrate; C7 to C9 were also tested on another independent substrate. Each run was completed on the same day; all the three runs were completed over three consecutive days. It is important to note that, despite different substrates used and different days on which the experiments were conducted, the trend of the TSH concentration-dependent SERS frequency shift was largely consistent. Such observations demonstrated that the test results were not significantly affected by those variables, such as the substrate-to-substrate variation, intra-substrate variation, and day-to-day variation.

2.3. Leave-one-concentration-out partial least square (PLS) analysis for TSH patient samples

We went on to leverage a ML model we recently developed23 by exploiting the subtle but robust SERS frequency shift and other encoded TSH concentration-dependent spectral variations for quantitative prediction of TSH antigens in patient samples. In the ML model, we first implemented outlier rejection by robust principal component analysis (ROBPCA) and then performed the wavenumber reduction using the significant multivariate correlation algorithm. It is important to note that, in classical PCA, the first component is along the direction in which the projected observation has the largest variance. Thereafter, the second component is orthogonal to the first component and again maximizes the variance of the data point. Continuing in this way produces all the principal components, which correspond to the eigenvectors of the empirical covariance matrix. Unfortunately, both the classical variance (which is being maximized) and the classical covariance matrix (which is being decomposed) are very sensitive to anomalous observations. Consequently, the first components are often attracted toward outlying points, and may not capture the variation of the regular observations. Therefore, data reduction based on classical PCA (CPCA) becomes unreliable if outliers are present in the data. ROBPCA combines ideas of projection pursuit and robust covariance estimation to give more accurate estimates at non-contaminated datasets and more robust estimates at contaminated data, as shown in Fig. S3. Therefore, in this study, we implemented ROBPCA to detect and reject spectral outliers that could be produced by background signals, or those irrelevant Raman spectral features not induced by capturing TSH antigens.

Subsequently, the leave-one-concentration-out PLS analysis was conducted to predict the TSH concentration. In the leave-one-concentration-out PLS analysis, which was schematically depicted in Fig. 3a, all the spectra collected for a particular level of TSH patient samples were used as the test spectra while the rest were used as the calibration spectra to build the ML model, which was used to predict the concentration for that particular level. The process was repeated till all levels of TSH patient samples were tested.

Figure 3. Leave-one-concentration-out PLS analysis for TSH patient samples.

Figure 3.

Open in a new tab

(a) Schematic for the leave-one-concentration-out PLS analysis, where all the spectra collected for a particular level of TSH patient samples were used as the test spectra while the rest were used as the calibration spectra to build the model to predict the concentration for that particular level. The process was repeated till all levels of TSH patient samples were tested. (b) Measured SERS spectra (same as Fig. 2a but with an expanded wavenumber range) averaged from the three runs for TSH patient samples with varying levels of TSH antigens from C1 to C9. The spectra were intentionally vertically offset for easy visualization. (c) Correlation between the predicted and true TSH concentrations. (d) Comparison between the predicted and true TSH concentrations. (e) The coefficient of variation (CV) for the predicted TSH concentrations.

In this case, instead of focusing on a characteristic SERS peak, we employed the collected SERS spectra (Fig. 3b) with an expanded wavenumber range from 550 cm−1 to 1800 cm−1 for the ML analysis. The ML analysis revealed that the predicted concentration was strongly correlated with the true TSH concentrations with a slope of 0.88 and a R2 of 0.89, as shown in Fig. 3c. We compared the predicted and true TSH concentration side by side for all the studied nine patient samples. As shown in Fig. 3d, the results confirmed the capability of our ML model to accurately predict the TSH concentrations in various patient samples. Analysis of the coefficient of variation (CV) further demonstrated the precision for the predicted TSH concentrations, as shown in Fig. 3e.

In addition to the leave-one-concentration-out PLS analysis, we also conducted the leave-one-spectrum-out PLS analysis as a comparison. In the leave-one-spectrum-out PLS analysis, a particular spectrum was used as the test spectrum, while all the remaining spectra were used as the calibration spectra to create the ML model. The ML model was then used to predict the TSH concentration corresponding to the test spectrum. The process was repeated till all the collected spectra were tested. The leave-one-spectrum-out PLS analysis as shown in Fig. S4 largely confirmed the results based on the leave-one-concentration-out PLS analysis.

It is important to note that, for both types of PLS analysis, we were able to perform TSH detection from 0.082 μIU/mL to 1.021 μIU/mL with high accuracy and precision. It is worth noting that normal serum TSH concentration generally falls between 0.35 to 4.5 μIU/mL.39–40 While we focused on measurements of patient samples containing TSH at the lower end of the clinically relevant range, given the linear dynamic response of SERS frequency shift to the TSH antigen-induced nanomechanical perturbation, it can be reasonably expected that both the SERS frequency shift- and the PLS-based ML analysis could be applied to study patient samples containing TSH with a concentration higher than the currently studied range. The results obtained based on studying serum-based patient samples (Fig. 2, 3, S2) suggest that the single-antibody spectro-immunoassay developed in this study holds potential for clinical translation with further optimization through studies featuring additional patient samples.

2.4. Robustness test for the leave-one-concentration-out PLS analysis

To test the robustness for the leave-one-concentration-out PLS analysis-based ML model, we selected data from one run as the calibration spectra to build the ML model, and then used it to predict the TSH concentrations corresponding to the test spectra collected from another run, as schematically shown in Fig. 4a. The predicted concentrations were found to be similarly strongly correlated with the true TSH concentrations with a slope of 0.97 and a R2 of 0.86, as shown in Fig. 4b. The prediction precision was also confirmed by the side-by-side comparison between the predicted and true TSH concentrations (Fig. 4c) as well as the CV analysis (Fig. 4d). We also noted that the training model can be built from as little as three unique concentrations of calibration spectra and yet we are able to achieve R2 > 0.9 on sample tested on different substrates and different day. This highlights our superior SERS substrate performance and the measurement robustness. It is important to note that, in this robustness test, the ML model was built based on spectra collected across different substrates over different days; the test spectra were also collected across different substrates and on a different day. Despite all these variables involved, the prediction accuracy and precision were found to be largely unaffected (except for the lowest concentration) by those variables, such as the substrate-to-substrate variation, intra-substrate variation, and day-to-day variation. In this study, the lowest concentration studied is 0.082 μIU/mL, which is far below the clinically relevant TSH concentration range and previously remained undetectable41–42. Therefore, it is unsurprising to note that such a low concentration was affected the most by those variables. Nevertheless, the above robustness test broadly underscores the strong capability and superior performance of the reported nanomechanical single-antibody method in conjunction with ML for quantitative detection of TSH in patient samples.

Figure 4. Robustness test for the leave-one-concentration-out PLS analysis.

Figure 4.

Open in a new tab

(a) Schematic for the robustness test method: data from a random run was trained to establish the ML model, which was then used for prediction based on data from another independent run. (b) Correlation between the predicted and true TSH concentrations. (c) Comparison between the predicted and true TSH concentrations. (d) The coefficient of variation (CV) for the predicted TSH concentrations. In this study, data from a random run was used as the calibration spectra to build the ML model. The ML model was implemented to predict the TSH concentrations corresponding to the test spectra from another independent run.

3. Conclusion

In this study, we systematically assessed the performance of a nanomechanical single-antibody spectro-immunoassay for detecting TSH in patient samples. By a combination of the SERS frequency shift analysis and leave-one-concentration-out PLS analysis, we were able to accurately predict the TSH concentrations in the patient samples. Robustness study revealed that variables, such as substrate-to-substrate variation, intra-substrate variation, and day-to-day variation, barely affected the performance of this biosensing platform. Success in studying patient samples suggests that this novel immunoassay can be potentially translated for clinical applications. Given the simplicity of the immunoassay design based on a single antibody, this platform could be widely used for studying various protein biomarkers, nucleic acids, pathogens directly in clinical samples.

Supplementary Material

Supinfo

ACKNOWLEDGMENT

This research was supported by National Institute of General Medical Sciences (DP2GM128198), National Institute of General Medical Sciences (1R35GM149272), National Institute of Biomedical Imaging and Bioengineering (2-P41-EB015871), and by Beckman Coulter Inc.

Footnotes

ASSOCIATED CONTENT

Supporting Information: Sections S1-S5; Fig. S1-S4.

The authors declare no competing financial interest.

REFERENCE

  • 1.Taylor PN; Albrecht D; Scholz A; Gutierrez-Buey G; Lazarus JH; Dayan CM; Okosieme OE, Global epidemiology of hyperthyroidism and hypothyroidism. Nature Reviews Endocrinology 2018, 14 (5), 301–316. [DOI] [PubMed] [Google Scholar]
  • 2.Jusufovic S; Hodzic E, Functional thyroid disorders are more common in patients on chronic hemodIalysis compared with the general population. Materia socio-medica 2011, 23 (4), 206. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.The Lancet D; Endocrinology, The untapped potential of the thyroid axis. The Lancet Diabetes & Endocrinology 2013, 1 (3), 163. [DOI] [PubMed] [Google Scholar]
  • 4.Medicine I. o., Medicare Coverage of Routine Screening for Thyroid Dysfunction. The National Academies Press: Washington, DC, 2003; p 135. [PubMed] [Google Scholar]
  • 5.Bernal J, Thyroid hormones and brain development. Vitamins & Hormones 2005, 71, 95–122. [DOI] [PubMed] [Google Scholar]
  • 6.Rovet JF, The role of thyroid hormones for brain development and cognitive function. Paediatric Thyroidology 2014, 26, 26–43. [DOI] [PubMed] [Google Scholar]
  • 7.Iwen KA; Schröder E; Brabant G, Thyroid Hormones and the Metabolic Syndrome. European Thyroid Journal 2013, 2 (2), 83–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Zhou W; Brumpton B; Kabil O; Gudmundsson J; Thorleifsson G; Weinstock J; Zawistowski M; Nielsen JB; Chaker L; Medici M; Teumer A; Naitza S; Sanna S; Schultheiss UT; Cappola A; Karjalainen J; Kurki M; Oneka M; Taylor P; Fritsche LG; Graham SE; Wolford BN; Overton W; Rasheed H; Haug EB; Gabrielsen ME; Skogholt AH; Surakka I; Davey Smith G; Pandit A; Roychowdhury T; Hornsby WE; Jonasson JG; Senter L; Liyanarachchi S; Ringel MD; Xu L; Kiemeney LA; He H; Netea-Maier RT; Mayordomo JI; Plantinga TS; Hrafnkelsson J; Hjartarson H; Sturgis EM; Palotie A; Daly M; Citterio CE; Arvan P; Brummett CM; Boehnke M; de la Chapelle A; Stefansson K; Hveem K; Willer CJ; Åsvold BO, GWAS of thyroid stimulating hormone highlights pleiotropic effects and inverse association with thyroid cancer. Nature Communications 2020, 11 (1), 3981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Khelifa L; Hu Y; Jiang N; Yetisen AK, Lateral flow assays for hormone detection. Lab on a Chip 2022, 22 (13), 2451–2475. [DOI] [PubMed] [Google Scholar]
  • 10.Sheehan MT, Biochemical testing of the thyroid: TSH is the best and, oftentimes, only test needed–a review for primary care. Clinical medicine & research 2016, 14 (2), 83–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Szkudlinski MW; Fremont V; Ronin C; Weintraub BD, Thyroid-Stimulating Hormone and Thyroid-Stimulating Hormone Receptor Structure-Function Relationships. Physiological Reviews 2002, 82 (2), 473–502. [DOI] [PubMed] [Google Scholar]
  • 12.Global Thyroid Function Test Market Size, Share, Growth Analysis, By Type (TSH tests, T3 tests), By End-use (Hospitals, Research laboratories) - Industry Forecast 2022–2028; SQMIG35E2010; 2022.
  • 13.Hoofnagle AN; Wener MH, The fundamental flaws of immunoassays and potential solutions using tandem mass spectrometry. Journal of Immunological Methods 2009, 347 (1), 3–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Bikkarolla SK; McNamee SE; McGregor S; Vance P; McGhee H; Marlow EL; McLaughlin J, A lateral flow immunoassay with self-sufficient microfluidic system for enhanced detection of thyroid-stimulating hormone. AIP Advances 2020, 10 (12), 125316. [Google Scholar]
  • 15.Kim K; Han DK; Choi N; Kim SH; Joung Y; Kim K; Ho NT; Joo S-W; Choo J, Surface-Enhanced Raman Scattering-Based Dual-Flow Lateral Flow Assay Sensor for the Ultrasensitive Detection of the Thyroid-Stimulating Hormone. Analytical Chemistry 2021, 93 (17), 6673–6681. [DOI] [PubMed] [Google Scholar]
  • 16.Choi S; Hwang J; Lee S; Lim DW; Joo H; Choo J, Quantitative analysis of thyroid-stimulating hormone (TSH) using SERS-based lateral flow immunoassay. Sensors and Actuators B: Chemical 2017, 240, 358–364. [Google Scholar]
  • 17.Leirs K; Dal Dosso F; Perez-Ruiz E; Decrop D; Cops R; Huff J; Hayden M; Collier N; Yu KXZ; Brown S; Lammertyn J, Bridging the Gap between Digital Assays and Point-of-Care Testing: Automated, Low Cost, and Ultrasensitive Detection of Thyroid Stimulating Hormone. Analytical Chemistry 2022, 94 (25), 8919–8927. [DOI] [PubMed] [Google Scholar]
  • 18.Bikkarolla SK; McNamee SE; Vance P; McLaughlin J. High-Sensitive Detection and Quantitative Analysis of Thyroid-Stimulating Hormone Using Gold-Nanoshell-Based Lateral Flow Immunoassay Device Biosensors [Online], 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Chen C; Tsai K, Clinical application of chemiluminescent immunoassay for thyroid stimulating hormone, free-T4 and intact-parathyroid hormone. Journal of the Formosan Medical Association= Taiwan yi zhi 1996, 95 (3), 197–202. [PubMed] [Google Scholar]
  • 20.Sarkar R, TSH Comparison Between Chemiluminescence (Architect) and Electrochemiluminescence (Cobas) Immunoassays: An Indian Population Perspective. Indian Journal of Clinical Biochemistry 2014, 29 (2), 189–195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Kho KW; Dinish US; Kumar A; Olivo M, Frequency Shifts in SERS for Biosensing. ACS Nano 2012, 6 (6), 4892–4902. [DOI] [PubMed] [Google Scholar]
  • 22.Zheng P; Wu L; Raj P; Mizutani T; Szabo M; Hanson WA; Barman I, A Dual-Modal Single-Antibody Plasmonic Spectro-Immunoassay for Detection of Small Molecules. Small 2022, 18 (18), 2200090. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zheng P; Raj P; Wu L; Szabo M; Hanson WA; Mizutani T; Barman I, Leveraging Nanomechanical Perturbations in Raman Spectro-Immunoassays to Design a Versatile Serum Biomarker Detection Platform. Small 2022, 18 (42), 2204541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Zhuang H; Zhu W; Yao Z; Li M; Zhao Y, SERS-based sensing technique for trace melamine detection – A new method exploring. Talanta 2016, 153, 186–190. [DOI] [PubMed] [Google Scholar]
  • 25.Xie D; Zhu WF; Cheng H; Yao ZY; Li M; Zhao YL, An antibody-free assay for simultaneous capture and detection of glycoproteins by surface enhanced Raman spectroscopy. Physical Chemistry Chemical Physics 2018, 20 (13), 8881–8886. [DOI] [PubMed] [Google Scholar]
  • 26.Wang Y; Yu Z; Ji W; Tanaka Y; Sui H; Zhao B; Ozaki Y, Enantioselective Discrimination of Alcohols by Hydrogen Bonding: A SERS Study. Angewandte Chemie International Edition 2014, 53 (50), 13866–13870. [DOI] [PubMed] [Google Scholar]
  • 27.Zheng P; Kasani S; Wu N, Converting plasmonic light scattering to confined light absorption and creating plexcitons by coupling a gold nano-pyramid array onto a silica–gold film. Nanoscale Horizons 2019, 4 (2), 516–525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Zheng P; Kasani S; Shi X; Boryczka AE; Yang F; Tang H; Li M; Zheng W; Elswick DE; Wu N, Detection of nitrite with a surface-enhanced Raman scattering sensor based on silver nanopyramid array. Analytica Chimica Acta 2018, 1040, 158–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Zheng P; Wu L; Raj P; Mizutani T; Szabo M; Hanson WA; Barman I, A Dual-Modal Single-Antibody Plasmonic Spectro-Immunoassay for Detection of Small Molecules. Small 2022, n/a (n/a), 2200090. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Tabatabaei M; Sangar A; Kazemi-Zanjani N; Torchio P; Merlen A; Lagugné-Labarthet F, Optical Properties of Silver and Gold Tetrahedral Nanopyramid Arrays Prepared by Nanosphere Lithography. The Journal of Physical Chemistry C 2013, 117 (28), 14778–14786. [Google Scholar]
  • 31.Zheng P; Kasani S; Tan W; Boryczka J; Gao X; Yang F; Wu N, Plasmon-enhanced near-infrared fluorescence detection of traumatic brain injury biomarker glial fibrillary acidic protein in blood plasma. Analytica Chimica Acta 2022, 1203, 339721. [DOI] [PubMed] [Google Scholar]
  • 32.Plou J; Charconnet M; García I; Calvo J; Liz-Marzán LM, Preventing Memory Effects in Surface-Enhanced Raman Scattering Substrates by Polymer Coating and Laser-Activated Deprotection. ACS Nano 2021, 15 (5), 8984–8995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Guerrini L; Pazos E; Penas C; Vázquez ME; Mascareñas JL; Alvarez-Puebla RA, Highly Sensitive SERS Quantification of the Oncogenic Protein c-Jun in Cellular Extracts. Journal of the American Chemical Society 2013, 135 (28), 10314–10317. [DOI] [PubMed] [Google Scholar]
  • 34.Ma H; Liu S; Zheng N; Liu Y; Han XX; He C; Lu H; Zhao B, Frequency Shifts in Surface-Enhanced Raman Spectroscopy-Based Immunoassays: Mechanistic Insights and Application in Protein Carbonylation Detection. Analytical Chemistry 2019, 91 (15), 9376–9381. [DOI] [PubMed] [Google Scholar]
  • 35.Zhu W; Wang Y; Xie D; Cheng L; Wang P; Zeng Q; Li M; Zhao Y, In Situ Monitoring the Aggregation Dynamics of Amyloid-β Protein Aβ42 in Physiological Media via a Raman-Based Frequency Shift Method. ACS Applied Bio Materials 2018, 1 (3), 814–824. [DOI] [PubMed] [Google Scholar]
  • 36.Zhang J; Dong Y; Zhu W; Xie D; Zhao Y; Yang D; Li M, Ultrasensitive Detection of Circulating Tumor DNA of Lung Cancer via an Enzymatically Amplified SERS-Based Frequency Shift Assay. ACS Applied Materials & Interfaces 2019, 11 (20), 18145–18152. [DOI] [PubMed] [Google Scholar]
  • 37.Cheng L; Zhang Z; Zuo D; Zhu W; Zhang J; Zeng Q; Yang D; Li M; Zhao Y, Ultrasensitive Detection of Serum MicroRNA Using Branched DNA-Based SERS Platform Combining Simultaneous Detection of α-Fetoprotein for Early Diagnosis of Liver Cancer. ACS Applied Materials & Interfaces 2018, 10 (41), 34869–34877. [DOI] [PubMed] [Google Scholar]
  • 38.Zhu W-F; Cheng L-X; Li M; Zuo D; Zhang N; Zhuang H-J; Xie D; Zeng Q-D; Hutchison JA; Zhao Y-L, Frequency Shift Raman-Based Sensing of Serum MicroRNAs for Early Diagnosis and Discrimination of Primary Liver Cancers. Analytical Chemistry 2018, 90 (17), 10144–10151. [DOI] [PubMed] [Google Scholar]
  • 39.Mendes D; Alves C; Silverio N; Batel Marques F, Prevalence of Undiagnosed Hypothyroidism in Europe: A Systematic Review and Meta-Analysis. European Thyroid Journal 2019, 8 (3), 130–143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Goel A; Shivaprasad C; Kolly A; Pulikkal A; Boppana R; Dwarakanath C, Frequent occurrence of faulty practices, misconceptions and lack of knowledge among hypothyroid patients. Journal of clinical and diagnostic research: JCDR 2017, 11 (7), OC15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Association AT; Hyperthyroidism A. A. o. C. E. T. o.; Thyrotoxicosis O. C. o.; Bahn RS; Burch HB; Cooper DS; Garber JR; Greenlee MC; Klein I; Laurberg P; McDougall IR; Montori VM, Hyperthyroidism and other causes of thyrotoxicosis: management guidelines of the American Thyroid Association and American Association of Clinical Endocrinologists. Thyroid 2011, 21 (6), 593–646. [DOI] [PubMed] [Google Scholar]
  • 42.Garber JR; Cobin RH; Gharib H; Hennessey JV; Klein I; Mechanick JI; Pessah-Pollack R; Singer PA; Endocrinologists W. f. t. A. A. o. C.; American Thyroid Association Taskforce on Hypothyroidism in Adults KA, Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Thyroid 2012, 22 (12), 1200–1235. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo
NIHMS1934917-supplement-Supinfo.docx (692.8KB, docx)

RESOURCES