Abstract
A highly sensitive, non‐invasive, and rapid HBV (Hepatitis B virus) screening method combining membrane protein purification with silver nanoparticle‐based surface‐enhanced Raman scattering (SERS) spectroscopy was developed in this study. Reproducible serum protein SERS spectra were obtained from cellulose acetate membrane‐purified human serum from 94 HBV patients and 89 normal groups. Tentative assignments of serum protein SERS spectra showed that the HBV patients primarily led to specific biomedical changes of serum protein. Principal components analysis and linear discriminate analysis were introduced to analyse the obtained spectra, with the diagnostic sensitivity of 92.6% and specificity of 77.5% were achieved for differentiating HBV patients from normal groups.
Inspec keywords: patient diagnosis, surface enhanced Raman scattering, proteins, biomembranes, principal component analysis, purification, silver, nanoparticles, nanomedicine, diseases
Other keywords: serum analysis method, cellulose acetate membrane purification, surface‐enhanced Raman spectroscopy, noninvasive HBV screening, rapid HBV screening method, Hepatitis B virus, membrane protein purification, silver nanoparticle‐based surface‐enhanced Raman scattering spectroscopy, reproducible serum protein SERS spectra, cellulose acetate membrane‐purified human serum, linear discriminate analysis, diagnostic sensitivity, HBV patient, principal components analysis
1 Introduction
Hepatitis B virus (HBV) infection remains a significant worldwide public health problem, and more than 6,800,000 people died of HPV infection or related complications [1]. The most common screening methods for chronic HBV infection were based on the determination of HBV serology (ELISA), liver enzymes and HBV DNA. However, the HBV serology may face the problems of false‐negative diagnosis or misdiagnosis when patients were in the window period [2], chronic occult HBV infection (OBI) stage [3] or the interference of interferon and lamivudine for chronic HBV infection therapies [4, 5]. Furthermore, the quantitative detection of HBV DNA has been used as the most significant and reliable marker for HBV diagnosis and evaluation of HBV infection stage, but this method requires expensive equipment and faces the problem of time‐consuming, which indicates that a rapid, highly sensitive and cost‐effective screening method for HBV is urgently needed [6].
Raman spectroscopy (RS) is a powerful analytical tool that is capable of providing specific molecular fingerprinting information for the different structure and conformation of macromolecules with high sensitivity. Although RS features the merits of low sample volume and free from the interference of water due to its relatively small Raman cross‐section. The major drawback of Raman spectroscopy is the inherently weak signal intensity and strong fluorescence background, which is a great challenge for practical clinical applications [7]. Fortunately, surface‐enhanced Raman scattering (SERS) could overcome the aforementioned drawback and provide dramatically enhanced intensity (enhancement factors up to 1013 – 1015) of the adsorbing molecule on the surface of metallic nanoparticles, and the features of high sensitivity, inherent molecular specificity enable SERS as an excellent diagnostic tool for disease [8, 9]. Recently, we utilised label‐free SERS spectroscopy and blood plasma to discriminate cancer patients from healthy subjects and obtained exciting preliminary results [10, 11, 12]. The successful result of our previous works demonstrated its significant potential in clinical application.
Herein, the merits of serum protein detection and SERS were combined to establish a non‐invasive method for clinical HBV screening. However, as a traditional testing object, the serum is a complex bio‐matrix and the content of biochemical composition may effect by many endogenous biomolecules secreted from lesion tissue or the interference of exogenous substances from food and medicine. The diversity, as well as the content of protein and metabolites, showed high relation to disease stages, rendering the serum protein as an attractive source for disease biomarkers, which can be acted as a highly‐sensitive and reliable indicator for disease diagnosis [13]. Furthermore, the composition of HBV (e.g. core, X proteins, S proteins etc.) and the serological markers of HBV are closely related to serum protein. In addition, the content of volatile substances contained in serum may alter with the long‐term preservation of samples, thus, which may be challenging for the extraction and interpretation of useful biomedical information from the obtained serum spectra [14].
The aim of this study was to assess the feasibility of serum protein‐based SERS technology for label‐free HBV detection. To improve serum protein‐based HBV detection, we have developed a serum sample pretreatment method to reduce external interference by membrane protein purification. Firstly, cellulose acetate (CA) membrane was utilised to extract and purify total protein from serum; the spectral interference from most endogenous plasma constituents and exogenous substances was eliminated. Then using silver nanoparticles as SERS substrate, we can get the purified total protein SERS spectra without any external labels. Principal component analysis (PCA) and linear discriminate analysis (LDA) were used to analyse and classify the serum protein SERS spectra acquired from HBV patients and healthy samples. Thus, a facile method combining serum membrane protein purification with silver nanoparticle‐based SERS was developed for non‐invasive and sensitive HBV screening.
2 Materials and methods
2.1 Preparation of serum samples
A total of 183 blood serum samples were collected from 89 normal healthy volunteers and 94 HBV patients. Each blood sample was ∼6 ml in volume, extracted intravenously from subjects after 12 h of overnight fasting to avoid the interference of food. Each of the serum samples was divided into two aliquots: 3 ml for clinical testing and another 3 ml for SERS measurement. All these samples were provided by Quanzhou Blood Centre, and serum samples were drawn from donors who signed an informed consent form in accordance with the ethical guidelines published by the ethical committee at that institution. Clinical diagnoses of the HBsAg subjects were given in Table 3.
2.2 Silver colloid preparation
Silver (Ag) nanoparticle solutions were synthesised using the process developed by Leopold and Lendl [15]. The reagents were all of the analytical grades and without further purification. Ultrapure water (>18.0 MΩ·cm) was purified using a Millipore Milli‐Q gradient system throughout the experiment. Briefly, 4.5 ml of sodium hydroxide (0.1 M) and 5 ml of hydroxylamine hydrochloride (0.06 M) were uniformly mixed. Then quickly addition of 90 ml silver nitrate aqueous solution (0.0011 M) accompanied by stirring gently, then the solution turned into grey‐brown. At last, the absorption spectra of the colloid solution had a maximum absorption peak at 416 nm. The size of silver nanoparticles normally distributes with an average diameter of 35 ± 5 nm. The silver colloidal solution was concentrated by centrifugation at 10,000 rpm for 10 min, discarding a portion of the supernatant and bringing the final concentration for later use. The corresponding SEM of the prepared Ag colloid is given in Fig. 8 of the Appendix.
2.3 Serum protein purification
As shown in Fig. 1, CA membrane (2 × 8 cm) was utilised to extract and purify serum proteins, and firstly immersed in a barbital buffer solution for 30 min [16, 17]. Then the CA membrane was transferred onto a filter paper to remove the excess buffer solution. After sample collection, 5 μl of each prepared serum sample was suctioned out and placed on the CA membrane for 10 min until the serum was completely dry. To further eliminating the interference of other components contained in serum, the membrane was dipped in 100 ml of the rinse solution for 6 min (composed of 5 ml glacial acetic acid, 45 ml 95% ethanol, and 50 ml distilled water). Then the CA membranes with serum protein absorbed were cut into pieces and placed in an Eppendorf tube. And after the addition of 150 μl of acetic acid into the tube, the membrane fragment was dissolved and turn into a transparent gel. Each tube was mixed with 300 μl of prepared silver colloid and incubated at 37°C, with kept evenly stirred for 30 min. Then, a drop of supernatant solution (serum protein‐Ag NP mixture) was transferred onto an aluminium plate until the sample was naturally dry.
Fig. 1.
Schematic diagram of membrane protein purification and SERS spectroscopy detection
2.4 Instrument setting and SERS spectroscopy
The SERS spectra of serum protein (protein–Ag NPs mixture) were measured by Renishaw Raman micro‐spectrometer (InVia Raman microscope, RENISHAW, United Kingdom) using 785 nm diode laser excitation in the range of 400–1800 cm−1. All SERS spectra were obtained by a 50 × objective using a Peltier cooled charge‐coupled device (CCD) detector and a single‐grating spectrograph with an 1800 lines/mm grating allowing a spectral resolution of 2 cm−1. Furthermore, the 520 cm−1 band of a silicon wafer was used for frequency calibration before SERS measurement. The SERS spectra were acquired from three random locations of samples for average with illuminated by 5 mW of incident laser power and 10‐s exposure time. The software package WIRE 2.0 was employed for spectral acquisition and analysis.
2.5 Data processing and multivariate statistical analysis
Since the complex biochemical composition of biofluids results in SERS spectra always accompanied by fluorescence interference, Vancouver Raman Algorithm (based on the fifth‐order polynomial fitting method) was utilised to fit the autofluorescence and noise signals in the original SERS spectrum [18]. For further extraction and interpretation of useful biochemical information from the pretreated SERS spectra, protein SERS spectra were then normalised to the integrated area under the curve in the 400–1800 cm−1 wavenumber range of each spectrum to reduce the influence of small changes in experimental condition, such as laser power fluctuations, sample volume or concentration fluctuations and so on, enabling a better comparison of the peak intensities among different serum protein group's analysis.
To evaluate the feasibility of serum protein SERS spectra to differentiate normal groups from HBV patients, principal component analysis (PCA) combined with linear discriminant analysis (LDA) were performed on the measured SERS spectra for automatically analyse and differentiate by SPSS statistical software (SPSS Inc., Chicago, IL, USA). PCA was performed on pretreated spectra to extract useful diagnostic information from the multidimensional dataset, and determine the key variables by an independent‐samples t ‐test (p ‐value <0.05). Then, the significant PCs were input into the LDA model, together with the leave‐one spectrum‐out, cross‐validation method, for HBV discrimination. Finally, the receiver operating characteristic (ROC) curve was calculated to verify the performance of the diagnostic algorithm.
3 Results
To investigate the silver colloids enhancement effect on the Raman scattering of serum protein, the SERS spectra of proteins added to the CA membrane with silver colloids as SERS substrate were firstly measured. At the same time, the SERS spectra of CA membrane with Ag NPs added and silver colloid as control data, which were measured under the same instrument settings and experimental conditions. As shown in Fig. 2, the observed vibrational bands at 1050 and 1000 cm−1 could, respectively, assign as the background signal of Ag NPs and CA membrane in SERS spectra of Ag NPs–CA membrane without serum addition. As compared with the blank control group, the corresponding background peak of Ag NPs is weaker than that of blank groups, indicating that the addition of Ag NPs and CA membrane does not introduce significant influence signal in the interested spectral range.
Fig. 2.
Comparison of SERS spectrum of blank groups of Ag NPs–CA membrane without serum addition, Ag NPs and Ag NPs–CA membrane with serum addition
For verifying the advantages of membrane protein purification, Fig. 3 shows the comparison of serum and serum protein SERS spectra from the fresh sample (black line) and the sample after three months of storage at −4°C (red line). Fig. 3 a shows serum SERS spectra and Fig. 3 b shows the serum protein SERS spectra. In the serum SERS spectra of Fig. 3 a, the spectra intensity significantly changed after three months storage, especially in the wavenumber ranges of 800–1000 cm−1 and for the peaks at 725 and 1332 cm−1, as well as the peak at 1070 cm−1 shifts to the 1094 cm−1. And there were almost no SERS peaks were observed at 1447 cm−1 in the fresh serum sample, while higher peak intensity appeared after three months storage. This may introduce unnecessary spectral vibration and make it difficult for the interpretation of vibrational spectroscopic information of HBV, thus, leading to misdiagnosis for HBV serum samples with different storage time. However, at the bottom of serum protein SERS spectra after membrane protein purification, only the three SERS peaks at 885, 1445, and 1680 cm−1 showed a minor change of spectral intensities after three months storage. There was almost no other change in the other SERS peaks intensity of serum protein derived from the serum sample stored for three months. Moreover, we also used correlation coefficients to evaluate the changes in the two SERS spectra with different storage times. The correlation coefficient of serum SERS spectra obtained in different storage times is only 0.818, while the corresponding correlation coefficient of serum protein SERS spectra increases to 0.927. Thus, the blood contained exogenous substances, and some variable endogenous serum components can be well removed by membrane protein purification, which may provide a unique and reliable serum protein analysis method.
Fig. 3.
Comparison of the SERS spectral changes from fresh samples and sample preserved three month
(a) Surface enhanced Raman spectra of the fresh serum sample and the serum sample stored for 3 months, (b) Surface enhanced Raman spectra of the fresh serum protein and the serum protein derived from serum samples stored for 3 months
Fig. 4 illustrates the average SERS spectra of serum protein obtained from normal (black line, n = 89) and HBV patients (red line, n = 94) groups after the area of serum protein SERS spectra was normalised in the 400–1800 cm−1 wavenumber range. It can be seen that while significant SERS spectral differences exist between normal and HBV serum protein samples, prominent SERS bands of serum protein were observed as follows: 513, 621, 642, 759, 828, 853, 1003, 1030, 1123, 1160, 1178, 1207, 1261, 1445 and 1680 cm−1 in the difference spectra, with the strongest peaks at 1003, 1030, 1261 and 1680 cm−1.
Fig. 4.
Average SERS spectra of serum protein obtained from the normal group (black line, n = 89) and HBV patients (red line, n = 94) groups in the range of 400–1800 cm−1. The shaded areas represent the standard deviations of the means spectra (mean ± SD). The difference spectra were obtained by subtracting the normal group from HBV patients group
Moreover, at the bottom of difference spectra, the normalised intensities of SERS peaks at 1003, 1160, 1178, 1207, 1261 and 1680 cm−1 showing higher intensities in HBV patients groups, while lower at 513 and 1030 cm−1, respectively. The most obvious differences between normal and HBV patients can be found at 513, 1003, 1030, 1207 and 1680 cm−1.
Meanwhile, by comparing the assignments of average SERS spectra between serum (Table 4) and serum protein, it can be observed that the serum SERS spectra showed a large SD value relative to serum protein SERS spectra. This phenomenon is in agreement with our previous works [19]. This may be due to the interference of composition changes with different serum storage time. The result shows that this protein purification method can reduce the interference of long‐term storage on SERS spectra and final statistical analysis.
To better understand the relationship between the biomedical mechanism and the changes in quantity or conformation of some biomolecules presented by normalised SERS spectra of serum protein, the SERS bands are assigned to protein Raman bands. Table 1 presents the tentative assignments of biochemical corresponding to prominent SERS bands in normal and HBV patients groups of specific serum protein [16, 20]. For example, the SERS peak at 1123 cm−1 corresponding to protein content (protein assignment) is negative in the difference spectra, which indicating that the content of proteins is decreased with HBV infection or the degree of diseases.
Table 1.
Peak positions and tentative assignments of major vibrational bands observed in serum protein samples
| Peak position, cm−1 | Vibrational mode | Major assignments |
|---|---|---|
| 513 | ν (S–S) | cysteine |
| 621 | C–C twisting mode | phenylalanine |
| 642 | ν (C–S) | tyrosine |
| 759 | ring breathing mode | tryptophan |
| 828/853 | ring breathing mode | tyrosine |
| 1003 | C–C twisting mode | phenylalanine |
| 1030 | ν (C–C) skeletal | keratin |
| 1123 | ν (C–N) | proteins |
| 1160/1178 | δ (C–H) | tyrosine |
| 1207 | — | hydroxyproline |
| 1261 | α ‐helix | amide III |
| 1446 | δ (C–H) | — |
| 1680 | ν (C = O) | amide I |
The prominent SERS band at 1680 cm−1 can be attributed to the amide I band of proteins in the α ‐helix conformation, which associated with the amide stretching modes characterise of the globular section of the protein, suggesting that an increase in the percentage of amide I contents relative to the total serum SERS‐active components in HBV patients groups. In addition, the SERS band of amide III at 1261 cm−1 in serum protein of HBV patients groups shows higher signal intensity than in normal samples, because the amide III band has been shown to be sensitive to protein secondary structure [21]. The prominent SERS peak located at around 1207 cm−1 is ascribed to the appearance of hydroxyproline, as an important product of collagen metabolism, whose content is associated with the development of liver fibrosis (as a complication of HBV infection) [22, 23]. The SERS peak of phenylalanine at 1007 cm−1 and the ring breathing of tyrosine at 1160 and 1178 cm−1 showed higher normalised intensity in the HBV patients groups. Moreover, we found that the vibrational bands of phenylalanine and tyrosine were obviously enhanced in the serum SERS spectra and serum protein SERS spectra of hepatocellular carcinoma cancer (HCC) [20, 24], which is consistent with the result of our research.
In contrast, the SERS peak at 1030 cm−1 assigned to keratin exhibited lower signal intensity in HBV patients groups than in the normal group, suggesting that the content of keratin obtained in the serum protein decreases with the degree of HBV infection [25]. Because keratin is a kind of protective protein that plays an important role in liver functioning [26, 27]. In contrast, the SERS peak of cysteine at 513 cm−1 was found to be lower in serum protein of HBV patient groups. This may be due to the fact that cysteine synthesis of glutathione (maintaining immune system functioning) is associated with a decreased risk of HBV infection [28]. The high expression levels of proteins, amide I, amide III, hydroxyproline, phenylalanine and tyrosine and the low expression levels of keratin and cysteine in HBV patients groups compared with the normal group could clearly reflect the molecular changes of amino acid metabolism with hepatitis B infection.
To evaluate the ability of serum protein SERS spectra for differentiating normal from HBV patients groups, PCA‐LDA was performed on the measured serum protein SERS spectra for further analysis and differentiating. An independent‐sample t ‐test on all the PC scores was utilised to identify diagnostically significant PCs (PC < 0.05) for each spectrum of protein SERS. There were four PCs (PC3, PC6, PC8 and PC11, p < 0.05) that were the most significant PCs for discriminating HBV patients groups from the normal group. Fig. 5 shows the four significant PCs loadings calculated using PCA together with the percentage of spectral variation captured by each corresponding PCs (PC3 – 8.4%; PC6 – 4.4%; PC8 – 2.9%; PC11 – 2.0%), whereas the successive PCs describe the spectral features that contribute progressively smaller variances. Some PC features, such as peaks, troughs, and spectral shapes are similar to those of serum protein SERS spectra in Fig. 3 b. PC3, PC6, PC8 and PC11 primarily represent the molecule‐specific Raman signals (e.g. at 513, 621, 642, 759, 828, 853, 1003, 1030, 1123, 1160, 1178, 1207, 1261, 1445 and 1680 cm−1). Some new diagnostic peak positions, although the assignment of these peaks is not yet known. We note that one of the most significant PC scores (e.g. PC3) only describes a small amount (8.4%) of the total variance. This indicates that some PC scores with small variances can still contain useful HBV diagnostic information.
Fig. 5.
First four significant PC scores calculated from PCA on serum protein SERS spectra (PC3 – 8.4%; PC6 –4.4%; PC8 – 2.9%; PC11 – 2.0%), revealing the significant spectral features for serum protein classification
To illustrate the utility of PC scores for classification of serum protein SERS spectra, direct comparisons between normal and HBV patients groups are presented in Figs. 6 a –f. The scatter plot of the third PC versus the sixth PC of the serum protein SERS spectra for the normal and HBV patients groups is shown in Fig. 6 a. In this analysis, 89 normal serum protein SERS spectra were compared with 94 HBV protein SERS spectra. An analogous comparison for the discrimination of normal and HVB groups is shown in Figs. 6 b –f. It shows the scatter plots of the two selected PC scores with different symbols for the normal group (black circular) and HBV patients groups (red triangle). The data points for the HBV patients groups and the normal groups are clustered into two separate groups based on different combinations of the significant PCs, and the corresponding separation lines in Figs. 6 a –f classify normal from HBV serum protein with the sensitivity of 77.7, 83.0, 83.0, 74.5, 69.1, and 78.7%; specificity of 61.8, 61.8, 71.9, 68.5, 74.2, and 74.2%, respectively. These results show that the selection of different combinations of significant PC scores will produce different levels of accuracy for HBV detection.
Fig. 6.
Scatter plots of the diagnostically significantly PC scores for HBV patients groups and normal groups derived from SERS spectra
(a) PC3 versus PC6, separation lines (PC6 = 2.028 PC3 + 0.038), (b) PC3 versus PC8, separation lines (PC8 = 1.167 PC3 + 0.030), (c) PC3 versus PC11, (d) PC6 versus PC8, (e) PC6 versus PC11, (f) PC8 versus PC11. The corresponding separation lines (PC6 = 2.028 PC3 + 0.038; PC8 = 1.167 PC3 + 0.030; PC11 = −0.977 PC3−0.030; PC8 = −0.575 PC6 + 0.017; PC11 = 0.482 PC6−0.018; and PC11 = 0.838 PC8−0.026) classify HBV serum protein from normal with the sensitivity of 77.7, 83.0, 83.0, 74.5, 69.1, and 78.7%; specificity of 61.8, 61.8, 71.9, 68.5, 74.2, and 74.2%, respectively
To prevent over‐training, then only the most four significant PC scores (PC3, PC6, PC8, and PC11) were loaded into the LDA model for discriminant classification in cross‐validation methods. Fig. 7 shows the posterior probability values belonging to HBV patients and normal serum protein‐based on PCA‐LDA algorithms together with cross‐validation, and the corresponding separation line is 0.5. The diagnostic sensitivity, specificity, and accuracy reached 92.6% (87/94), 77.5% (69/84), and 85.2% (156/183), respectively. To evaluate the performance of the PCA‐LDA diagnostic algorithms derived from the most four significant PC scores of serum protein SERS data set, the ROC curve (in Fig. 7 b) was further utilised to evaluate the performance of the PCA‐LDA‐based diagnostic algorithm. The integration areas under the ROC curve is 0.920, which further demonstrates that the usage of serum protein SERS spectra combined with PCA‐LDA diagnostic algorithms can be more robust and powerful in distinguishing HBV patients groups from normal groups. These results demonstrated an exciting diagnostic efficiency achieved by the membrane protein purification method integrated with serum SERS spectra.
Fig. 7.
The corresponding posterior probability values and ROC curves of HBV patients and normal subject
(a) Scatter plot of the posterior probability values belonging to HBV patients and normal serum protein‐based on PCA‐LDA modelling together with cross‐validation. The separate dotted line gives a diagnostic sensitivity and specificity of 92.6% (87/94), 77.5% (69/89) in serum protein for separating HBV patients groups from normal groups, (b) The ROC curve of the discrimination results for the PCA‐LDA‐based SERS spectral classification. AUC: the integration areas under the ROC curves
Besides, we utilised the first four diagnostically significant PCs (p ⩽ 0.05) to build the LDA model. 60 samples from each sample group were taken as a training set and the rest part as a test set. The diagnostic sensitivity and specificity for classifying HBV patients from the serum protein SERS were 94.1 and 82.8%, respectively (Table 2). This exploratory study demonstrated that serum protein SERS technique, in conjunction with PCA‐LDA provided a novel strategy for the clinical diagnosis of HBV.
Table 2.
Sensitivity and specificity of classification results executed by LDA algorithms
| Observed | Predicted | N test /N train | Sensitivity, % | Specificity, % | ||
|---|---|---|---|---|---|---|
| Normal | HBV | |||||
| SERS spectrum | HBV | 2 | 32 | 34/60 | 94.1 | 82.8 |
| normal | 24 | 5 | 29/60 | |||
4 Conclusion
In summary, a simple method combining serum membrane protein purification with silver nanoparticle‐based SERS was developed for non‐invasive and highly‐sensitive HBV screening. We found that the composition and infection of HBV and the serological markers of HBV infection are closely related to serum protein. Moreover, by utilising the advantages of the membrane protein purification method, the interference of exogenous substances and some variable endogenous serum component can be eliminated. Diagnostic sensitivity of 92.6% and specificity of 77.5% can be achieved by the PCA‐LDA algorithm for differentiating HBV patients groups from normal groups in serum protein. This exploratory study demonstrated that serum protein SERS technique, in conjunction with PCA‐LDA provided a novel strategy for the screening of HBV. Our next goal is to take more detailed relevant studies to verify the reliability of this potential HBV detection method, and even further studies on the early diagnosis of liver cancer and liver fibrosis.
5 Acknowledgments
This work was supported by the National Natural Science Foundation of China (nos. 61575043, and U1605253), the scientific research innovation team construction program of Fujian Normal University (no. IRTL1702), the program for Changjiang Scholars and Innovative Research Team in University (no. IRT1115) and new century excellent talent in Fujian province university (J1‐1160), Natural Science Foundation of Fujian Province of China (grant nos. 2016J01292 and 2015J01436). Yunchao Xu, Yunyi Wang, Huijin Lin, Xiaokun Liu, Zuci Zheng contributed equally to this work.
If the measured value of hepatitis B surface antigen (HBsAg) is >0.18 ng/ml, the level of HBsAg of the patient is regarded as a positive result, then it means that the patient has hepatitis B virus, and vice versa (Table 3).
Table 3.
Clinical diagnosis of the HBsAg subjects
| Number | HBsAg, ng/ml | Number | HBsAg, ng/ml | Number | HBsAg, ng/ml | Number | HBsAg, ng/ml | Number | HBsAg, ng/ml |
|---|---|---|---|---|---|---|---|---|---|
| Positive | |||||||||
| 1 | 2132 | 11 | 45,479 | 21 | 366 | 31 | 3.56 | 41 | 1471 |
| 2 | 428 | 12 | 27,415 | 22 | 3288 | 32 | 4.38 | 42 | 0.3 |
| 3 | 0.29 | 13 | 3524 | 23 | 14 | 33 | 73,932 | 43 | 22.65 |
| 4 | 2288 | 14 | 6751 | 24 | 6792 | 34 | 3972 | 44 | 66.16 |
| 5 | 14,508 | 15 | 1175 | 25 | 30 | 35 | 4482 | 45 | 1.28 |
| 6 | 44 | 16 | 697 | 26 | 3387 | 36 | 34,821 | 46 | 887 |
| 7 | 11,897 | 17 | 1396 | 27 | 13,921 | 37 | 1032 | 47 | 1637 |
| 8 | 493 | 18 | 2403 | 28 | 2940 | 38 | 578 | 48 | 276 |
| 9 | 43 | 19 | 12 | 29 | 936 | 39 | 12.85 | 49 | 1366 |
| 10 | 13,003 | 20 | 17 | 30 | 581 | 40 | 691 | 50 | 2019 |
| 51 | 2.1 | 61 | 406 | 71 | 204 | 81 | 8429 | 91 | 2178 |
| 52 | 1984 | 62 | 393 | 72 | 2.6 | 82 | 4609 | 92 | 3850 |
| 53 | 1326 | 63 | 12,231 | 73 | 806 | 83 | 2659 | 93 | 2865 |
| 54 | 155 | 64 | 944 | 74 | 4950 | 84 | 556 | 94 | 2578 |
| 55 | 2424 | 65 | 192 | 75 | 5773 | 85 | 2781 | — | — |
| 56 | 1513 | 66 | 7.59 | 76 | 2347 | 86 | 5.43 | — | — |
| 57 | 171 | 67 | 87.61 | 77 | 5148 | 87 | 33,197 | — | — |
| 58 | 768 | 68 | 13.04 | 78 | 1227 | 88 | 248 | — | — |
| 59 | 46 | 69 | 3.05 | 79 | 1633 | 89 | 7873 | — | — |
| 60 | 390 | 70 | 51.02 | 80 | 97.05 | 90 | 67,230 | — | — |
| Negative | |||||||||
| 1 | 0 | 11 | 0.01 | 21 | 0.01 | 31 | 0 | 41 | 0.02 |
| 2 | 0.02 | 12 | 0.01 | 22 | 0.01 | 32 | 0.01 | 42 | 0.01 |
| 3 | 0.01 | 13 | 0.02 | 23 | 0 | 33 | 0.01 | 43 | 0 |
| 4 | 0.01 | 14 | 0 | 24 | 0 | 34 | 0 | 44 | 0 |
| 5 | 0 | 15 | 0.01 | 25 | 0 | 35 | 0 | 45 | 0.01 |
| 6 | 0.02 | 16 | 0 | 26 | 0.01 | 36 | 0.01 | 46 | 0.03 |
| 7 | 0 | 17 | 0.01 | 27 | 0.01 | 37 | 0.02 | 47 | 0.01 |
| 8 | 0 | 18 | 0 | 28 | 0 | 38 | 0.01 | 48 | 0.02 |
| 9 | 0 | 19 | 0.01 | 29 | 0.01 | 39 | 0 | 49 | 0.01 |
| 10 | 0.01 | 20 | 0 | 30 | 0.01 | 40 | 0.01 | 50 | 0.01 |
| 51 | 0 | 61 | 0.02 | 71 | 0.03 | 81 | 0 | — | — |
| 52 | 0.01 | 62 | 0.02 | 72 | 0.01 | 82 | 0 | — | — |
| 53 | 0 | 63 | 0.02 | 73 | 0.02 | 83 | 0 | — | — |
| 54 | 0.02 | 64 | 0 | 74 | 0 | 84 | 0.02 | — | — |
| 55 | 0.01 | 65 | 0 | 75 | 0.01 | 85 | 0 | — | — |
| 56 | 0.01 | 66 | 0.02 | 76 | 0.02 | 86 | — | — | — |
| 57 | 0.02 | 67 | 0.02 | 77 | 0.02 | 87 | 0.01 | — | — |
| 58 | 0.01 | 68 | 0.02 | 78 | 0 | 88 | 0.01 | — | — |
| 59 | 0.02 | 69 | 0.02 | 79 | 0 | 89 | 0.02 | — | — |
| 60 | 0.02 | 70 | 0 | 80 | 0.02 | — | — | — | — |
The window period should refer to after the virus infects the human body, the human immune system has not produced relative antibodies, or the antibodies are unstable. Therefore, the results of examining antiviral antibodies are negative, which may cause missed diagnosis (Table 4 and Fig. 8).
Table 4.
Peak positions and tentative assignments of major vibrational bands observed in serum samples
| Peak position, cm−1 | Major assignments |
|---|---|
| 495 | L‐arginine |
| 639 | uric acid |
| 725 | hypoxanthine |
| 813 | uric acid |
| 889 | uric acid |
| 960 | ν (C–C) valine |
| 1007 | phenylalanine |
| 1073 | tyrosine |
| 1136 | uric acid |
| 1206 | tyrosine |
| 1332 | nucleic acid |
| 1655 | amide I, α ‐helix |
Fig. 8.
Corresponding SEM of the prepared Ag colloid
6 References
- 1. Beasley R.P.: ‘Hepatitis B virus. The major etiology of hepatocellular carcinoma’, Cancer, 1988, 61, (10), pp. 1942 –1956 [DOI] [PubMed] [Google Scholar]
- 2. Tsoi W.C. Lelie N. Lin C.K.: ‘Enhanced detection of hepatitis B virus in Hong Kong blood donors after introduction of a more sensitive transcription‐mediated amplification assay’, Transfusion, 2013, 53, (10), pp. 2477 –2488 [DOI] [PubMed] [Google Scholar]
- 3. Taira R. Satake M. Momose S. et al.: ‘Residual risk of transfusion‐transmitted hepatitis B virus (HBV) infection caused by blood components derived from donors with occult HBV infection in Japan’, Transfusion, 2013, 53, (7), pp. 1393 –1404 [DOI] [PubMed] [Google Scholar]
- 4. Jarvis B. Faulds D.: ‘Lamivudine. A review of its therapeutic potential in chronic hepatitis B’, Drugs, 1999, 58, (1), pp. 101 –141 [DOI] [PubMed] [Google Scholar]
- 5. Locarnini S. Birch C.: ‘Antiviral chemotherapy for chronic hepatitis B infection: lessons learned from treating HIV‐infected patients’, J. Hepatol., 1999, 30, (3), pp. 536 –550 [DOI] [PubMed] [Google Scholar]
- 6. Kimura T. Rokuhara A. Matsumoto A. et al.: ‘New enzyme immunoassay for detection of hepatitis B virus core antigen (HBcAg) and relation between levels of HBcAg and HBV DNA’, J. Clin. Microbiol., 2003, 41, (5), pp. 1901 –1906 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Raman Ç. M.: ‘Spectroscopy for cancer diagnosis: how far have we come?’, Bioanalysis, 2015, 7, (21), pp. 2813 –2824 [DOI] [PubMed] [Google Scholar]
- 8. Li S.X. Zhang Y.J. Xu J.F. et al.: ‘Noninvasive prostate cancer screening based on serum surface‐enhanced Raman spectroscopy and support vector machine’, Appl. Phys. Lett., 2014, 105, (9), p. 091104 [Google Scholar]
- 9. Zhang Y. Ye X.J. Xu G.X. et al.: ‘Identification and distinction of non‐small‐cell lung cancer cells by intracellular SERS nanoprobes’, RSC Adv., 2016, 6, (7), pp. 5401 –5407 [Google Scholar]
- 10. Lin D. Huang H. Qiu S.F. et al.: ‘Diagnostic potential of polarized surface enhanced Raman spectroscopy technology for colorectal cancer detection’, Opt. Express, 2016, 24, (3), pp. 2222 –2234 [DOI] [PubMed] [Google Scholar]
- 11. Lin J. Huang Z.F. Feng S.Y. et al.: ‘Label‐free optical detection of type II diabetes based on surface‐enhanced Raman spectroscopy and multivariate analysis’, J. Raman Spectrosc., 2015, 45, (10), pp. 884 –889 [Google Scholar]
- 12. Feng S. Lin D. Lin J.Q. et al.: ‘Blood plasma surface‐enhanced Raman spectroscopy for non‐invasive optical detection of cervical cancer’, Analyst, 2013, 138, (14), pp. 3967 –3974 [DOI] [PubMed] [Google Scholar]
- 13. Zhang A. Sun H. Wang X.: ‘Serum metabolomics as a novel diagnostic approach for disease: a systematic review’, Anal. Bioanal. Chem., 2012, 404, (4), pp. 1239 –1245 [DOI] [PubMed] [Google Scholar]
- 14. Atkins C.G. Buckley K. Chen D. et al.: ‘Raman spectroscopy as a novel tool for monitoring biochemical changes and inter‐donor variability in stored red blood cell units’, Analyst, 2016, 141, (11), pp. 3319 –3327 [DOI] [PubMed] [Google Scholar]
- 15. Leopold N. Lendl B.: ‘A new method for fast preparation of highly surface‐enhanced Raman scattering (SERS) active silver colloids at room temperature by reduction of silver nitrate with hydroxylamine hydrochloride’, J. Phys. Chem. B, 2003, 107, (24), pp. 5723 –5727 [Google Scholar]
- 16. Feng S. Lin D. Lin J. et al.: ‘Saliva analysis combining membrane protein purification with surface‐enhanced Raman spectroscopy for nasopharyngeal cancer detection’, Appl. Phys. Lett., 2014, 104, (7), pp. 238 –243 [Google Scholar]
- 17. Lin J. Chen R. Feng S. et al.: ‘A novel blood plasma analysis technique combining membrane electrophoresis with silver nanoparticle‐based SERS spectroscopy for potential applications in noninvasive cancer detection’, Nanomed. Nanotechnol., 2011, 7, (5), pp. 655 –663 [DOI] [PubMed] [Google Scholar]
- 18. Huang Z. McWilliams A. Lui H. et al.: ‘Near‐infrared Raman spectroscopy for optical diagnosis of lung cancer’, Int. J. Cancer, 2003, 107, (6), pp. 1047 –1052 [DOI] [PubMed] [Google Scholar]
- 19. Movasaghi Z. Rehman S. Rehman D.I.U.: ‘Raman spectroscopy of biological tissues’, Appl. Spectrosc. Rev., 2007, 42, (5), pp. 493 –541 [Google Scholar]
- 20. Lin J. Wang J. Xu C. et al.: ‘Differentiation of digestive system cancers by using serum protein‐based surface‐enhanced Raman spectroscopy’, J. Raman Spectrosc., 2017, 48, pp. 16 –21 [Google Scholar]
- 21. Cai S. Singh B.R.: ‘A distinct utility of the amide III infrared band for secondary structure estimation of aqueous protein solutions using partial least squares methods’, Biochemistry, 2004, 43, (9), pp. 2541 –2549 [DOI] [PubMed] [Google Scholar]
- 22. Frank C.J. McCreecy R.L. Redd D.C.B.: ‘Raman spectroscopy of normal and diseased human breast tissues’, Anal. Chem., 1995, 67, pp. 777 –783 [DOI] [PubMed] [Google Scholar]
- 23. Lacoste B. Raymond V. Lapierre P. et al.: ‘Protection against acute hepatocellular injury afforded by liver fibrosis Is independent of T lymphocytes’, PLOS One, 2016, 11, (10), p. e0165360 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Xiao R. Zhang X. Rong Z. et al.: ‘Non‐invasive detection of hepatocellular carcinoma serum metabolic profile through surface‐enhanced Raman spectroscopy’, Nanomed. Nanotechnol., 2016, 12, (8), pp. 2475 –2484 [DOI] [PubMed] [Google Scholar]
- 25. Huang Z. Lui H. McLean D.I. et al.: ‘Raman spectroscopy in combination with background near‐infrared autofluorescence enhances the in vivo assessment of malignant tissues’, Photochem. Photobiol., 2005, 81, pp. 1219 –1226 [DOI] [PubMed] [Google Scholar]
- 26. Felipe‐Silva A. Wakamatsu A. Cirqueira C. et al.: ‘Immunohistochemistry panel segregates molecular types of hepatocellular carcinoma in Brazilian autopsy cases’, World J. Gastroenterol., 2016, 27, pp. 6246 –6256 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Haeryoung K. Choi G. Na C. et al.: ‘Human hepatocellular carcinomas with ‘Stemness’‐related marker expression: keratin 19 expression and a poor prognosis’, Hepatology, 2011, 54, (5), pp. 1707 –1717 [DOI] [PubMed] [Google Scholar]
- 28. Huang Z. Chen C. Zeng Z. et al.: ‘Mechanism and significance of increased glutathione level in human hepatocellular carcinoma and liver regeneration’, FASEB J., 2001, 15, (1), pp. 19 –21 [DOI] [PubMed] [Google Scholar]