Surrogate virus neutralization test as a scalable and reliable tool for immune surveillance in booster-vaccinated populations Authors: Lin Cai, Zhaoyang Fang, Renyun Zha, Zhifeng Zhou, Ruiyin Zhang, Xueyun Li, Cheng Guo, Conghui Liao Published: 2026 ABSTRACT With waning vaccine-induced immunity and the continued emergence of immune-evasive SARS-CoV-2 variants, booster vaccination has become essential for sustaining population-level protection. However, scalable and reliable serological tools for monitoring post-booster humoral immunity across different vaccine platforms remain insufficiently evaluated. We conducted a cross-sectional study and a booster cohort study to systematically compare the performance and concordance of surrogate virus neutralization test (sVNT), pseudovirus neutralization test (pVNT), and binding IgG antibody assays. In the cross-sectional analysis, serum samples from 259 individuals collected 1–2 months after a third dose of inactivated vaccine were used to compare sVNT and pVNT. In the booster cohort, paired sera from 288 participants receiving a fourth dose with different vaccine platforms were analyzed to evaluate the relationship between sVNT and IgG responses. Correlation and agreement were assessed using Spearman correlation and Bland–Altman analyses. In the cross-sectional study, sVNT showed strong concordance with pVNT, demonstrated by a very high correlation (r = 0.97) and good agreement in Bland–Altman analysis. In the booster cohort study, IgG antibody levels correlated strongly with sVNT overall (r = 0.91). The correlation strength remained highly consistent across different vaccine platforms, with no statistically significant differences observed. Collectively, these findings demonstrate that sVNT closely reflects functional neutralizing activity, supporting its utility for large-scale immune monitoring. sVNT offers an optimal balance of analytical performance and scalability for high-throughput monitoring without requiring cell-based assays. Consequently, we propose a tiered immune monitoring strategy: utilizing sVNT for broad population-level surveillance and reserving pVNT for targeted, precise functional assessment.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-49973-3. Introduction Vaccination is one of the most effective and cost-efficient public health interventions for controlling infectious diseases1. Vaccines elicit pathogen-specific humoral and cellular immune responses, thereby markedly decreasing the risk of severe disease and mortality in individuals while simultaneously curtailing widespread pathogen transmission via the attainment of herd immunity2. Since the emergence of SARS-CoV-2, multiple COVID-19 vaccines spanning diverse technological platforms have been developed and deployed at unprecedented speed, and real-world evidence indicates that vaccination has substantially reduced COVID-19-associated hospitalization, thereby helping to alleviate pressure on healthcare systems worldwide3,4. Despite the effectiveness of primary immunization, substantial evidence indicates that vaccine-induced immune responses wane progressively over time5,6. Meanwhile, the persistent emergence of SARS-CoV-2 variants, especially those exhibiting substantial immune-evasion capabilities, has additionally undermined the protective efficacy of preexisting immunity7,8. In this context, booster vaccination has emerged as a critical strategy for sustaining and strengthening population-level immune protection. Accumulating evidence indicates that booster doses markedly elevate antibody responses, with particularly pronounced increases in neutralizing antibody titers against prevalent variants9,10. Currently deployed vaccines incorporate diverse technological platforms, such as inactivated, recombinant protein, and adenovirus-vectored approaches, which results in considerable heterogeneity in antigen design, delivery systems, and immune activation mechanisms. Therefore, the establishment of scientifically robust, reliable, and scalable methodologies for evaluating vaccine efficacy is essential for enhancing immunization strategies, guiding vaccination policies, and protecting vulnerable populations. Presently, serological methods for evaluating post-vaccination humoral immune responses can be broadly categorized into binding antibody assays and neutralizing antibody assays11. Binding antibody assays are typically based on immunoassay platforms such as enzyme-linked immunosorbent assays (ELISA) or chemiluminescent immunoassays, and primarily focus on the quantitative or semi-quantitative detection of serum antibodies targeting the spike protein and its receptor-binding domain (RBD), or the nucleocapsid protein12. While these assays offer advantages in terms of high throughput and ease of automation, they fail to directly reflect the functional capacity of antibodies to block viral entry into host cells. For the assessment of neutralizing antibodies, conventional virus neutralization test (cVNT) is widely regarded as the gold standard for quantifying serum neutralizing antibody titers. However, their dependence on biosafety level-3 facilities along with labor-intensive procedures and limited throughput, substantially restricts their use in large-scale studies13. Pseudovirus neutralization test (pVNT), while conducted under biosafety level-2 (BSL-2) conditions and offering improved biosafety, remains constrained by the complexity of cell-based culture systems14. In contrast, surrogate virus neutralization test (sVNT), which is based on competitive inhibition of key molecular interactions, offers advantages in biosafety, cost-effectiveness, and operational simplicity. Nevertheless, its ability to accurately reflect authentic neutralizing antibody titers and to capture neutralization mechanisms beyond RBD-targeted responses requires further validation15, particularly in individuals with complex vaccination histories. Consequently, there is an urgent need to evaluate the performance of sVNT across diverse cohorts to establish its utility as a reliable surrogate for functional immunity. In summary, sVNT, pVNT, and binding antibody assays each exhibit distinct strengths and limitations in evaluating vaccine-induced humoral immune responses. However, within booster-vaccinated populations receiving COVID-19 vaccines based on different technological platforms, systematic head-to-head comparisons of multiple serological assays within the same cohort remain limited. In particular, robust empirical evidence is lacking regarding the potential of sVNT as a surrogate for functional neutralizing antibody assessment and its concordance with pVNT and binding antibody measurements. In practice, selecting appropriate serological assays while accounting for analytical accuracy, biosafety requirements, operational feasibility, and scalability, remains of critical importance. On this basis, we systematically compared the performance and concordance of sVNT and pVNT in individuals who received a single booster dose of an inactivated vaccine following primary immunization. Building on this analysis, we further included populations receiving a second booster dose using different vaccine platforms to evaluate the performance of sVNT relative to IgG measurements in capturing booster-induced humoral immune responses. This study aims to provide a testing strategy that balances analytical accuracy and operational feasibility for evaluating SARS-CoV-2 vaccine-induced immunity under different booster vaccination strategies, thereby offering experimental evidence to support routine immune surveillance and public health decision-making. Results Characteristics of the study population after vaccination The overall study design and the principles of the serological assays are illustrated in Fig. 1. A total of 259 participants were enrolled in Group 1 and 288 in Group 2. Table 1 shows the demographic data of individuals in two groups. The median age was 38 years (IQR, 30–51) in Group 1 and 39 years (IQR, 30–53) in Group 2, with similar age distributions across predefined age categories. In Group 2, participants received different booster vaccine platforms, including V01, ZF2001, Ad5-nCoV, and inactivated vaccines. The median interval between vaccination and sample collection was 27 days (IQR, 25–44) in Group 1 and 33 days (IQR, 31–38) in Group 2.Fig. 1(A) Illustration of the overall research framework of the study. The study comprises a cross-sectional study and a booster cohort study. In the cross-sectional component, individuals who had received three doses of inactivated COVID-19 vaccines (n = 259) were evaluated using pVNT and sVNT. In the booster cohort component, participants previously vaccinated with three doses of inactivated vaccines (n = 288) received different booster vaccines (inactivated vaccine, ZF2001, V01, or Ad5-nCoV), and their post-booster immune responses were assessed using sVNTs and IgG antibody assays. (B) Graphical illustration for the principle of pVNT, sVNT and IgG antibody assays. (a) Mechanism of pVNT where anti-SARS-CoV-2 nAbs block the binding of SARS-CoV-2 spike (S) protein to human ACE2 receptor on the host cell surface. (b) Principle of sVNT where anti-SARS-CoV-2 nAbs block the binding of HRP-conjugated RBD protein to the precoated hACE2 protein on the ELISA plate. (c) The IgG antibody assay is based on an indirect chemiluminescence immunoassay, in which RBD-coated magnetic microparticles capture anti-SARS-CoV-2 IgG antibodies, followed by detection using labeled anti-human IgG antibodies.Table 1Demographic data of individuals in two groups.CharacteristicGroup 1 (n = 259)Group 2 (n = 288)Age group 18–3067 (25.87%)73 (25.35%) 31–4074 (28.57%)86 (29.86%) 41–5052 (20.08%)44 (15.28%) 51–6033 (12.74%)42 (14.58%) ≥ 6133 (12.74%)43 (14.93%)Median age, years (IQR)38 (30–51)39 (30–53)Gender Female116 (44.79%)137 (47.57%) Male143 (55.21%)151 (52.43%)Vaccine platforms (Group 2 only) Ad5-nCoV-126 (43.75%) V01-87 (30.21%) ZF2001-54 (18.75%) Inactivated-21 (7.29%)Interval between vaccination and sampling, days (IQR)27 (25–44)33 (31–38) (A) Illustration of the overall research framework of the study. The study comprises a cross-sectional study and a booster cohort study. In the cross-sectional component, individuals who had received three doses of inactivated COVID-19 vaccines (n = 259) were evaluated using pVNT and sVNT. In the booster cohort component, participants previously vaccinated with three doses of inactivated vaccines (n = 288) received different booster vaccines (inactivated vaccine, ZF2001, V01, or Ad5-nCoV), and their post-booster immune responses were assessed using sVNTs and IgG antibody assays. (B) Graphical illustration for the principle of pVNT, sVNT and IgG antibody assays. (a) Mechanism of pVNT where anti-SARS-CoV-2 nAbs block the binding of SARS-CoV-2 spike (S) protein to human ACE2 receptor on the host cell surface. (b) Principle of sVNT where anti-SARS-CoV-2 nAbs block the binding of HRP-conjugated RBD protein to the precoated hACE2 protein on the ELISA plate. (c) The IgG antibody assay is based on an indirect chemiluminescence immunoassay, in which RBD-coated magnetic microparticles capture anti-SARS-CoV-2 IgG antibodies, followed by detection using labeled anti-human IgG antibodies. Table 1Demographic data of individuals in two groups.CharacteristicGroup 1 (n = 259)Group 2 (n = 288)Age group 18–3067 (25.87%)73 (25.35%) 31–4074 (28.57%)86 (29.86%) 41–5052 (20.08%)44 (15.28%) 51–6033 (12.74%)42 (14.58%) ≥ 6133 (12.74%)43 (14.93%)Median age, years (IQR)38 (30–51)39 (30–53)Gender Female116 (44.79%)137 (47.57%) Male143 (55.21%)151 (52.43%)Vaccine platforms (Group 2 only) Ad5-nCoV-126 (43.75%) V01-87 (30.21%) ZF2001-54 (18.75%) Inactivated-21 (7.29%)Interval between vaccination and sampling, days (IQR)27 (25–44)33 (31–38) Demographic data of individuals in two groups. Overall antibody responses after vaccination In group 1, the median neutralizing antibody titer measured by sVNT was 51 (IQR: 23–106), whereas the median titer determined by pVNT was 78 (IQR: 36–157). A Wilcoxon signed-rank test indicated a statistically significant difference between the titers obtained by the two methods (P < 0.0001; Fig. 2A). Based on the established cutoff criteria for each assay, the seropositivity rates were 91.12% (236/259) for sVNT and 92.66% (240/259) for pVNT. In group 2, the median IgG antibody level was 5.935 IU/mL (IQR: 1.510–27.58) before booster vaccination and increased to 283.4 IU/mL (IQR: 117.8–553.1) at the post-booster follow-up visit. This represents an approximate 48-fold increase, which was statistically significant (Wilcoxon signed-rank test, P < 0.0001). Similarly, sVNT50 was 10 (IQR: 5–29) before booster vaccination and rose to 402 (IQR: 144.8–975.3) after booster vaccination. This approximate 40-fold elevation in antibody levels was statistically significant, as evidenced by a Wilcoxon signed-rank test (p < 0.0001; Fig. 2B).Fig. 2(A) Post-vaccination antibody levels determined by sVNT and pVNT in the cross-sectional study. Dashed lines indicate the median. (B) Changes in antibody levels pre- and post-vaccination measured by sVNT and IgG antibody assays in the booster cohort study. Dashed lines indicate the median, while connected lines represent the shifts in median titers between the pre- and post-vaccination time points, as measured by sVNT and IgG assays, respectively. NT50, 50% neutralization titer (representing sVNT50 or pVNT50). Comparisons of paired values were performed using the Wilcoxon signed-rank test. ****, P < 0.0001. (A) Post-vaccination antibody levels determined by sVNT and pVNT in the cross-sectional study. Dashed lines indicate the median. (B) Changes in antibody levels pre- and post-vaccination measured by sVNT and IgG antibody assays in the booster cohort study. Dashed lines indicate the median, while connected lines represent the shifts in median titers between the pre- and post-vaccination time points, as measured by sVNT and IgG assays, respectively. NT50, 50% neutralization titer (representing sVNT50 or pVNT50). Comparisons of paired values were performed using the Wilcoxon signed-rank test. ****, P < 0.0001. Booster cohort study shows that booster vaccination significantly increased both sVNT50 and IgG antibody levels across all vaccine platforms (Fig. 3). In recipients of the V01, ZF2001, Ad5-nCoV and Inactivated vaccines, antibody levels at the post-booster follow-up were significantly higher than that at baseline according to both assays (P < 0.05). Furthermore, based on the established cutoff criteria for each assay, the overall seropositivity rates were 97.22% (280/288) for IgG and 97.92% (282/288) for sVNT.Fig. 3Antibody changes from pre- to post-booster vaccination across booster vaccine platforms, assessed by sVNT and IgG antibody assays: (A) V01; (B) ZF2001; (C) Ad5-nCoV; and (D) Inactivated vaccine. Comparisons of paired values were performed using the Wilcoxon signed-rank test. Connected lines indicate the median. *, P < 0.05; **, P < 0.01; ****, P < 0.0001. Antibody changes from pre- to post-booster vaccination across booster vaccine platforms, assessed by sVNT and IgG antibody assays: (A) V01; (B) ZF2001; (C) Ad5-nCoV; and (D) Inactivated vaccine. Comparisons of paired values were performed using the Wilcoxon signed-rank test. Connected lines indicate the median. *, P < 0.05; **, P < 0.01; ****, P < 0.0001. Correlation, agreement, and diagnostic performance of sVNT versus pVNT, with stratified evaluation by sex and age To ensure comparability across different detection methods, all antibody measurements were log10-transformed. Spearman’s rank correlation analysis revealed a significant positive correlation between the sVNT50 and pVNT50 values (r = 0.9729, P < 0.0001, Fig. 4A), with the scatter plot exhibiting minimal dispersion. Bland–Altman analysis indicated good overall agreement and comparability between the log10-transformed sVNT50 and pVNT50 titers (Fig. 4B). The mean difference was 0.1728, with 95% limits of agreement (LOA) ranging from −0.0522 to 0.3978.Fig. 4(A) Correlation between neutralizing antibody titers measured by sVNT and pVNT. (B) Agreement between neutralizing antibody titers measured by sVNT and pVNT. Data were obtained from the cross-sectional study 1–2 months after the third dose of inactivated SARS-CoV-2 vaccine. All antibodies are presented using base 10 log scale. (A) Correlation between neutralizing antibody titers measured by sVNT and pVNT. (B) Agreement between neutralizing antibody titers measured by sVNT and pVNT. Data were obtained from the cross-sectional study 1–2 months after the third dose of inactivated SARS-CoV-2 vaccine. All antibodies are presented using base 10 log scale. Furthermore, using pVNT as the reference standard, 236 of the 240 pVNT-positive samples tested positive by sVNT, yielding a sensitivity of 98.33%. Among the 19 pVNT-negative samples, sVNT consistently yielded negative results, demonstrating a specificity of 100%. The positive predictive value (PPV) and negative predictive value (NPV) for sVNT were 100% and 82.61%, respectively. To evaluate the robustness of the sVNT method across different demographic groups, we conducted stratified analyses based on gender and age, and the results are summarized in Supplementary Table S1. Differences in sensitivity between groups were assessed using Fisher’s exact test. No significant differences were observed between females and males (P = 0.6271) or between participants aged < 40 years and ≥ 40 years (P = 0.3383). Among pVNT-positive individuals, the sensitivity of sVNT was 99.09% in females and 97.69% in males, with no significant difference between the two groups (Fisher’s exact test, P = 0.6271). Similarly, the sensitivity was 99.22% in participants aged < 40 years and 97.30% in those aged ≥ 40 years, with no significant difference between age groups (Fisher’s exact test, P = 0.3383). Evaluation of correlation and concordance between IgG antibody assays and sVNT. Spearman’s rank correlation analysis revealed a significant positive correlation between IgG antibody levels and sVNT50 (r = 0.9059, P < 0.0001; Fig. 5A). Given that all sVNT-negative results were uniformly assigned a value of 5, a sensitivity analysis was conducted to evaluate whether this imputation approach influenced the overall findings. Specifically, we excluded negative samples and re-performed the correlation analysis. The strong positive correlation remained significant (r = 0.92, P < 0.0001), supporting the robustness of the association between IgG antibody levels and sVNT50. Analysis of the overall dataset showed a systematic difference between IgG antibody levels and sVNT50, with a mean difference (bias) of 0.2530 (log10) and 95% LOA ranging from −0.6536 to 1.160 (Fig. 5B). These findings indicate that, although the two assays exhibited a strong positive correlation, quantitative differences persisted. Consequently, the assays are not directly interchangeable at the individual level.Fig. 5(A) Correlation between sVNT50 and IgG antibodies. (B) Agreement between sVNT50 and IgG antibodies. Data were derived from the full booster cohort study including all samples from recipients of the four different vaccine platforms. All antibodies are presented using base 10 log scale. (A) Correlation between sVNT50 and IgG antibodies. (B) Agreement between sVNT50 and IgG antibodies. Data were derived from the full booster cohort study including all samples from recipients of the four different vaccine platforms. All antibodies are presented using base 10 log scale. Across different vaccine platforms, the correlation trends between sVNT50 and IgG antibody remained largely consistent in cohorts receiving booster immunization(r ranging from 0.79 to 0.92; Fig. 6). To further compare whether correlation strength differed between platforms, and to ensure independence of observations, correlations between the two assay methods were analyzed separately for pre-vaccination and post-vaccination samples, followed by Fisher’s r-to-z transformation for each time point. For pre-vaccination samples, correlation coefficients ranged from 0.58 to 0.79 across the four groups. Fisher’s r-to-z transformation showed no statistically significant differences in correlation strength between groups after Benjamini–Hochberg correction (all adjusted P > 0.05). Similarly, for post-vaccination samples, coefficients ranged from 0.81 to 0.90 and no significant differences between groups were observed after multiple comparison correction.Fig. 6Upper panels show correlation between IgG antibodies levels and sVNT50 across different booster vaccine platforms in cohorts. Lower panels present agreement between IgG antibodies levels and sVNT50 for each booster platform in cohorts. (A) V01; (B) ZF2001; (C) Ad5-nCoV; (D) Inactivated vaccine. All antibodies are presented using base 10 log scale. Upper panels show correlation between IgG antibodies levels and sVNT50 across different booster vaccine platforms in cohorts. Lower panels present agreement between IgG antibodies levels and sVNT50 for each booster platform in cohorts. (A) V01; (B) ZF2001; (C) Ad5-nCoV; (D) Inactivated vaccine. All antibodies are presented using base 10 log scale. Discussion The critical importance of serological assays in public health responses has been firmly established since the emergence of COVID-19. Given the demand for large-scale population-level immune monitoring, there is an urgent need for a detection method that offers high throughput and cost-effectiveness while accurately reflecting neutralizing antibody activity. Based on two distinct study populations, this study evaluated the performance of sVNT relative to pVNT, and IgG antibody assays relative to sVNT, for assessing humoral immune responses following COVID-19 vaccination. As early as 2020, a study introduced the sVNT method and validated the strong correlation between sVNT and pVNT results using serum samples from naturally infected COVID-19 patients15. Furthermore, subsequent studies have corroborated the application of this method in vaccinated individuals16. Importantly, evidence from vulnerable populations further supports the utility of sVNT. In patients with solid tumors receiving mRNA- or adenoviral vector-based COVID-19 vaccines, a multiplex ACE2 inhibition-based sVNT reliably captured variant-specific neutralizing responses and showed a robust correlation with IgG titers, supporting its value for scalable immune monitoring in high-risk populations17. The results of this study revealed a strong positive correlation between sVNT₅₀ and pVNT₅₀ (r = 0.9729), suggesting a strong association between the two assays in measuring neutralizing activity. Although sVNT and pVNT also showed high qualitative agreement, the absolute neutralizing antibody titers differed significantly. Consistent with previous literature, this suggests that absolute values from sVNT and pVNT are not directly interchangeable, though the qualitative determination remains highly credible18. When pVNT was used as the reference standard, sVNT exhibited a sensitivity of 98.33% and a specificity of 100%, demonstrating excellent diagnostic performance. Stratified analyses were performed to evaluate the impact of age and sex on sVNT reliability within Group 1 of this cross-sectional study. The absence of statistically significant differences between subgroups suggests that sVNT performance remained stable across these demographic strata. This observation contrasts with previous reports using serum from SARS-CoV-2 infected individuals, which indicated that IgG-based serological assays may be influenced by demographic variables19,20. One possible explanation is that serological assay performance may differ between convalescent and vaccine cohorts, as antibody responses following natural infection are generally more heterogeneous than those induced under standardized vaccination conditions21,22. These findings are consistent with previous studies, further supporting the reliability of sVNT as a surrogate assay for detecting neutralizing antibodies23,24. The Bland–Altman plots demonstrated low variability between the results of the two methods. This robust correlation can be attributed to the mechanistic similarity between the assays, as both focus on RBD–ACE2 interaction, though they capture neutralization in different ways. sVNT indirectly assesses neutralization capacity by measuring antibody-mediated inhibition of RBD binding to the ACE2 receptor on host cells. In contrast, pVNT directly evaluates the ability of antibodies to prevent pseudovirus entry into cells14,15. This shared mechanistic foundation accounts for the high agreement observed between the two methods. However, several methodological studies have demonstrated that sVNT exhibits high sensitivity for samples with elevated neutralizing antibody titers, yet it is more susceptible to false-negative results at lower titers25,26. Additionally, previous studies have demonstrated the consistency between IgG and sVNT in serological applications, which aligns with the findings of the present study27–29. In this study, IgG antibody levels were significantly correlated with sVNT50 (r = 0.9052). Because a substantial proportion of baseline samples had titers below the lower limit of detection, we further performed a sensitivity analysis by excluding pVNT-negative samples.. The correlation, indicating that the observed association was not driven solely by the imputed low baseline values. Furthermore, the strength of correlation between IgG and sVNT was lower compared to that observed between sVNT50 and pVNT50. This discrepancy may arise because IgG assays quantify the total amount of binding antibodies targeting the S-protein RBD but cannot distinguish between antibodies with neutralizing activity and non-neutralizing antibodies30. Consequently, IgG assays may, to a certain extent, overestimate the actual level of functional neutralizing capacity. Therefore, current scientific consensus favors employing quantitative IgG assays primarily for monitoring population-level immunity and comparing the immunogenicity of various regimens, rather than as precise quantitative indicators of individual protective thresholds31,32. This study compared the correlation between IgG levels and sVNT50 in the context of booster immunizations using different vaccine platforms. While research exploring similar concepts has been published, studies specifically targeting the second booster dose (Fourth-Dose) remain limited33. Furthermore, this study encompasses booster vaccinations spanning four different platforms. Furthermore, the comparative analysis of correlation coefficients via Fisher’s r-to-z transformation showed no significant disparity across different vaccine platforms Although the observed correlation coefficients showed some numerical variation across platforms, no statistically significant between-group differences were identified at either time point after Benjamini–Hochberg correction. These findings suggest that across the four vaccine platforms, the overall relationship between the two assays was broadly consistent after booster vaccination. Bland–Altman analyses indicated a measurable systematic bias between IgG and sVNT measurements, together with relatively wide limits of agreement. This discrepancy primarily arises because the two assays capture fundamentally different biological properties, as the IgG assay quantifies the total binding antibodies, whereas sVNT measures the functional ability of serum antibodies to block the RBD-ACE2 interaction. These results are in line with prior studies indicating that while binding antibody assays cannot fully replace functional neutralization tests at the individual level, they remain valuable tools for immunogenicity monitoring and comparative vaccine evaluation34,35. Each of the three methods possesses distinct advantages and specific applications. IgG assays are characterized by operational simplicity, high throughput, and cost-effectiveness, making them well-suited for large-scale population serological screening and the monitoring of immune response trends. However, their primary limitation lies in the inability to directly assess functional neutralizing antibody levels. sVNT achieves a favorable balance between throughput, biosafety, and functional assessment. By eliminating the need for live virus handling, sVNT can be conducted in BSL-2 laboratories with rapid turnaround times. These advantages make it highly suitable for large-scale assessment of vaccine immunogenicity and initial clinical screening. Given that the ancestral SARS-CoV-2 strain has been replaced by a succession of variants, the clinical utility of sVNT increasingly depends on its adaptability to contemporary strains. Previous studies have demonstrated that sVNTs based on variant-specific RBD antigens correlate significantly with both authentic and pseudotyped virus neutralization assays and can serve as rapid functional assays to evaluate post‑vaccination neutralizing antibodies against multiple variants36–38. Furthermore, the application of multiplex sVNT platforms, such as bead-based Luminex assays, enables the incorporation of a broad diversity of variant-specific antigens. These multiplexed systems allow for the simultaneous evaluation of neutralizing responses against both dominant and emerging variants, providing a critical tool for estimating future protection in vaccinated cohorts39–42. Although pVNT is operationally complex, time-consuming, and limited in throughput, it most faithfully mimics the viral infection process. Consequently, it remains the gold standard for evaluating neutralizing antibody activity, particularly for the precise assessment of cross-protective capabilities against emerging variants and for in-depth mechanistic studies. Our findings demonstrate that sVNT serves as a reliable and efficient surrogate method for detecting neutralizing antibodies in assessments following inactivated vaccination. In public health practice, a triage testing strategy could be implemented, wherein sVNT is employed for large-scale preliminary screening, followed by confirmatory pVNT testing for samples with borderline results or for key populations. This approach would facilitate resource optimization and precise assessment. This study has several limitations. Most notably, both the pseudoviruses used in the pVNT and the antigens employed in the sVNT were derived exclusively from the ancestral (wild-type) SARS-CoV-2 strain, and did not include variants of concern, such as Omicron. Given the continued antigenic evolution of SARS-CoV-2 and the current predominance of Omicron JN.1-descendant lineages, ancestral-strain-based serological readouts may have limited ability to reflect neutralizing activity against contemporary circulating variants43. Nevertheless, The primary objective of this study is methodological in nature, aiming to establish a comparative framework for evaluating the performance of different serological assays following vaccination. Because our study cohort received wild-type vaccines, utilizing homologous ancestral antigens is biologically essential to accurately assess the vaccine-induced humoral response. Secondly, all samples in the cross-sectional study were obtained exclusively from recipients of inactivated vaccines. Additionally, given the vaccination landscape in the study region during the enrollment period, mRNA vaccines were not included in this analysis. As mRNA vaccines are the primary booster platform in many parts of the world and may induce distinct humoral response profiles44, the generalizability of our findings and of the proposed testing strategy to mRNA-boosted populations remains uncertain. Further validation in cohorts receiving mRNA-based boosters is warranted. Furthermore, the uneven distribution of sample sizes across different platforms in the booster cohort study may affect statistical power. Lastly, immune protection is multifaceted, this study assessed humoral immunity exclusively and did not evaluate T-cell-mediated cellular immune responses. Conclusions sVNT provides an optimal balance between analytical performance, biosafety, and scalability by enabling reliable functional antibody assessment without the need for live virus or cell based assays, thereby supporting high throughput population level immune monitoring. Furthermore, a robust and consistent correlation between sVNT and IgG across diverse vaccine platforms was demonstrated, indicating that the relationship between the assays is platform-independent. These findings support a tiered immune monitoring strategy that prioritizes sVNT for large scale surveillance while reserving pVNT for targeted, precise functional immune assessment. Methods Study design, ethical approval, and clinical samples A cross-sectional study was conducted from December 2021 to February 2023, enrolling 259 participants in Guangzhou (Group 1). Serum samples were collected 1–2 months after the administration of a third dose of inactivated SARS-CoV-2 vaccine (Vero cells) to evaluate the performance of sVNT and pVNT. Concurrently, a cohort of individuals receiving a second SARS-CoV-2 booster dose (fourth dose overall) was established in Futian District, Shenzhen, between January and June 2022. Serum samples were collected immediately prior to administration of the second booster dose and approximately 1 month after vaccination. A total of 377 participants were recruited, of whom 288 were ultimately included in the final analysis (Group 2) based on the availability of paired serum samples from both time points. This cohort was used to evaluate the performance of sVNT and IgG binding assays. Participants were stratified by booster vaccine platform(ZF2001 [recombinant protein subunit vaccine, Anhui Zhifei Longcom Biopharmaceutical], n = 54; V01 [recombinant protein subunit vaccine, Sinovac Biotech], n = 87; Ad5-nCoV [adenovirus-vectored vaccine, CanSino Biologics], n = 126; inactivated [inactivated whole-virus vaccine, China National Biotec Group or Sinovac Biotech], n = 21). The study was conducted in accordance with the Declaration of Helsinki, following the ethics review board’s guidelines and regulations and was approved by the Ethics Committee of the School of Public Health, Sun Yat-sen University, and the Ethics Committee of the Futian District Center for Disease Control and Prevention. Written informed consent was obtained from all participants prior to their enrollment. All blood samples were centrifuged to obtain serum, which was then aliquoted and stored at −80 °C. Serological testing All samples were tested in a blinded manner, and all experimental procedures were performed strictly in accordance with the respective manufacturers’ instructions and standard operating procedures. Pseudovirus neutralization test (pVNT) For pVNT, pseudovirus particles were constructed based on a Vesicular Stomatitis Virus (VSV) backbone, carrying genes encoding a luciferase reporter and the SARS-CoV-2 spike protein. The assay was performed using a HEK-293 T cell line stably overexpressing human ACE2 (hACE2) and TMPRSS2 (HEK-293 T-ACE2-TMPRSS2). The experiments were carried out on 96‐well plates. After diluting the serum 1:10, we sequentially performed gradient dilutions of 1:3. The experimental protocol was as follows: Each batch of pseudovirus underwent pre-titration before use, with standardized positive and negative sera employed for quality control. The inoculum was normalized to the predefined working range of approximately 200–500 infectious units per well to ensure inter-assay comparability. An equal volume of pseudovirus was mixed with the serum at each dilution and neutralization was allowed to proceed at 37 °C in a 5% CO₂ incubator for 1 h. Subsequently, the virus-serum mixture was added to pre-seeded HEK-293 T-ACE2-TMPRSS2 cells and cultured for an additional 24 h. Following cell lysis, luciferase activity was measured using the Luciferase Assay System (Promega, Cat. No. E1500) and a Centro XS3 chemiluminescence plate reader (Berthold). Wells containing cells cultured without serum served as positive controls (representing 100% infection), while wells containing only culture medium served as background controls. The 50% pseudovirus neutralization titer (pVNT50), defined as the serum dilution inhibiting 50% of luciferase activity, was calculated using a non-linear regression model. A pVNT50 ≥ 10 was considered positive. A neutralizing antibody titer < 10 was defined as titer 5. ELISA-based surrogate virus neutralization tests (sVNT) sVNT was performed using the SARS-CoV-2 Surrogate Virus Neutralization Test Kit (GenScript). This assay utilizes a competitive ELISA format to mimic the virus-host interaction. Specifically, neutralizing antibodies in the sample bind to the Horseradish Peroxidase (HRP)-conjugated receptor-binding domain (HRP-RBD), thereby inhibiting its binding to the hACE2 protein immobilized on the plate. Consequently, the intensity of the colorimetric signal is inversely proportional to the concentration of neutralizing antibodies. The ELISA plates were pre-coated with recombinant hACE2 protein. Serum samples were subjected to threefold serial dilutions (ranging from 1:10 to 1:7,200) in a separate 96-well plate and mixed with a specific concentration of HRP-RBD antigen. The mixture was incubated at 37 °C in a 5% CO₂ incubator for 1 h. Subsequently, the mixture was then transferred to the hACE2-coated ELISA plate, allowing unneutralized HRP-RBD to bind to the pre-coated hACE2 receptors. After three washes with PBST, TMB substrate was added, and the plate was incubated at 37 °C in the dark for 15 min to develop the colorimetric reaction. The reaction was terminated with 2 M H₂SO₄, and the absorbance (optical density, OD) was measured at 450 nm. The 50% surrogate virus neutralization titer (sVNT50) was calculated using the Reed-Muench method. A sVNT50 ≥ 10 was considered positive and a neutralizing antibody titer < 10 was defined as a titer of 5. Serum anti-SARS-CoV-2 IgG antibody detection Serum anti-SARS-CoV-2 IgG antibodies targeting the RBD of the spike protein were quantitatively measured using a chemiluminescence immunoassay (CLIA) based on an indirect immunoassay principle. In this assay, magnetic microparticles coated with recombinant SARS-CoV-2 spike RBD antigen were used to capture RBD-specific IgG antibodies present in serum samples. The bound human IgG antibodies were subsequently detected with acridinium ester-labeled anti-human IgG monoclonal antibodies, forming a magnetic microparticle–antigen–antibody–labeled antibody complex. After magnetic separation and washing, pre-trigger and trigger solutions were added to initiate the chemiluminescent reaction. The emitted signal, expressed as relative light units, was measured by the analyzer and was directly proportional to the concentration of anti-SARS-CoV-2 IgG antibodies in the sample. This study employed the Novel Coronavirus (2019-nCoV) Spike Protein IgG Antibody Detection Kit (Chemiluminescence Method) manufactured by RayBiotech (Guangzhou, China). The assay was conducted on a matching SMART 6500 Automatic Magnetic Particle Chemiluminescence Analyzer (Chongqing Cosmaire, China). The experimental protocol was as follows: serum samples were automatically diluted 1:40. Subsequently, 15 μL of the diluted sample, 40 μL of the magnetic microbead solution, and 80 μL of the luminescent marker were added to the reaction vessel and incubated at 37 °C for 20 min. After three cycles of magnetic washing, 100 μL each of pre-trigger and trigger solutions were added sequentially to initiate the reaction, and the chemiluminescent signal was detected. A four-parameter logistic (4PL) curve was used to fit the calibrator concentrations (range, 4.5–5,200 IU/L) and calculate the IgG antibody concentration in the samples, expressed in International Units per milliliter (IU/mL). A concentration of ≥ 10.0 IU/mL was defined as positive. Quality control samples were included in each batch to ensure experimental validity. Statistical analysis Data were analyzed using GraphPad Prism 9.2. and SPSS 26.0. Continuous data are given as a median and interquartile range, and categorical data as counts and percentages. Wilcoxon signed-rank (based on the non-normal distribution of the data of several groups) was applied to analyze changes in paired data. P < 0.05 was considered to be statistically significant. Spearman’s rank correlation analysis was employed to assess the correlation between IgG antibody titers, sVNT50, and pVNT50. Correlation interpretation was as follows: a coefficient of 0–0.39 indicates a weak correlation, 0.40–0.59 indicates a moderate correlation, 0.6–0.79 indicates a strong correlation, and 0.8–1 indicates a very strong correlation. Bland–Altman analysis was conducted to assess the agreement between methods and the 95% limits of agreement (LOA) were calculated. Differences in correlation strength between platforms were evaluated using Fisher’s r-to-z transformation. P values were adjusted for multiple comparisons using the Benjamini–Hochberg method. Using pVNT as the reference standard for neutralizing activity, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the sVNT were calculated. Fisher’s exact tests were used to compare results in the gender and age groups. Supplementary Information Supplementary Information. Supplementary Information.