Abstract
Objective
Heat shock protein 90α is involved in tumorigenesis; however, its concentration-dependent relationship and threshold with pulmonary nodules remain unclear. This study aimed to explore the association and critical threshold of plasma heat shock protein 90α levels with the prevalence of pulmonary nodules in males aged ≥40 years.
Methods
This study included 968 male participants (aged ≥40 years) undergoing low-dose computed tomography and plasma heat shock protein 90α measurement from January 2022 to June 2024. Restricted cubic splines and segmented logistic regression were used to identify were the inflection point of the concentration-response relationship. Multivariable models were adjusted for age, smoking index, high-density lipoprotein cholesterol, triglycerides, and family history of lung cancer.
Results
Pulmonary nodules were detected in 40.2% (389/968) of participants. The prevalence of pulmonary nodules increased from 28.7% to 51.0% across heat shock protein 90α quartiles (P for trend <0.001). A threshold was identified at 89.5 ng/mL (95% confidence interval: 87.2–93.1). Above this level, each 1 ng/mL increase in the heat shock protein 90α levels was associated with a 3% higher prevalence of pulmonary nodules (adjusted odds ratio = 1.03, 95% confidence interval: 1.01–1.05). The adjusted odds ratio for the highest versus lowest quartile was 1.88 (95% confidence interval: 1.26–2.80).
Conclusions
Elevated plasma heat shock protein 90α levels are associated with an increased prevalence of pulmonary nodules in asymptomatic males aged ≥40 years, with a critical threshold of 89.5 ng/mL.
Keywords
Introduction
Pulmonary nodules (PN) are the most common radiological findings in low-dose computed tomography (LDCT) screening, posing a challenge in clinical management. Despite increasing the detection rate of early-stage lung cancer by 24%, LDCT is associated with a false-positive rate that can vary widely (approximately 23%–96%), based on nodule definition and screening protocols. 1 Therefore, there is an urgent clinical need for new molecular markers to optimize risk stratification. Heat shock protein 90α (HSP90α), an evolutionarily conserved molecular chaperone, plays a crucial role in maintaining protein homeostasis within cells. In response to stressors such as oxidative stress or inflammation, it can be actively secreted into the extracellular space, where it performs distinct signaling functions. 2 Recent studies have confirmed that plasma HSP90α levels are significantly higher in patients with lung cancer than in healthy controls and are positively associated with clinical staging.3,4 Additionally, the expression level of HSP90α in benign PN has also garnered increased attention. Other studies have demonstrated that although the serum HSP90α levels in patients with benign PN are lower than those in patients with malignant PN, they are still significantly higher than those in healthy controls, 5 thereby suggesting an association between HSP90α levels and both benign and malignant PN.
However, the association between HSP90α and PN remains uncertain. Therefore, we hypothesized that a dose–response relationship exists between plasma HSP90α levels and incidentally found PN in asymptomatic males, and that a threshold concentration that can help guide clinical decisions. To test this hypothesis, we quantitatively evaluated the association between HSP90α and PN and explored a potential clinical cutoff point using restricted cubic splines (RCS) and segmented logistic regression in a cohort of 968 men who underwent LDCT and were ≥40 years old.
Methods
Population and data sources
The cross-sectional study was conducted between January 2022 and June 2024 at the Health Management Center of Shaanxi Provincial People’s Hospital. The study comprised male participants aged ≥40 years undergoing routine LDCT screening, thereby aiming to represent the general male population rather than high-risk groups. Participants were consecutively recruited during their health check-up visits. Exclusion criteria included a history of cancer, recent acute infection, chronic inflammatory or autoimmune conditions, or incomplete data. This study focused on men aged ≥40 years, as previous evidence indicates that lung cancer predominantly occurs in individuals >40 years old, with its incidence increasing remarkably after the age of 45. 6 Notably, in the Chinese population, the incidence of lung cancer is substantially higher in men than in women (age-standardized incidence rates: 50.72 vs. 26.25 per 100,000 population in 2018). 7 The final cohort consisted of 968 participants, representing a robust sample size for detecting significant clinical associations. The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cross-sectional studies. 8 The Institutional Review Board of Shaanxi Provincial People’s Hospital approved the research protocol (Approval Number: 2025R071, approved on 15 January 2025). All procedures were conducted in accordance with the Declaration of Helsinki of 1975, as revised in 2024. Written informed consent was obtained from all participants. All patient details have been deidentified.
CT protocol and nodule definition
Participants underwent noncontrast LDCT scans according to a standardized protocol to ensure consistency and reproducibility. The scans were performed using 120 kVp tube voltage, automatic tube current modulation (30–100 mAs), and 1-mm slice thickness to optimize image quality and minimize radiation. CT images were interpreted by two experienced radiologists who were blinded to the participants’ biomarker data to avoid bias. PNs were defined, according to the Fleischner Society guidelines, as focal, rounded, or irregular opacities, either solid or subsolid, with a diameter of up to 30 mm. 1 Typically benign calcified nodules were not included in this definition.
Laboratory assays
Fasting venous blood samples were collected from the participants in the morning on the day of the CT scan to minimize diurnal variations and ensure consistency in sample collection. Plasma HSP90α levels were measured using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (Cloud-Clone Corp.; Wuhan, China) according to the manufacturer’s instructions. The assay’s had a detection range of 0.156–10 ng/mL, with a lower limit of quantification of 0.156 ng/mL. All samples were measured in duplicate, and the mean value was used for analysis. Samples were processed immediately after collection to avoid multiple freeze-thaw cycles. The coefficient of variation (CV) was maintained below 8% to ensure reliable measurements. The manufacturer reported no significant cross-reactivity with other HSP90 isoforms. Conventional biochemical tests and tumor markers, including carcinoembryonic antigen (CEA), cytokeratin 19 fragment 21-1 (CYFRA21-1), and neuron-specific enolase (NSE), were processed simultaneously to provide a comprehensive health assessment.
Statistical analysis
Continuous variables were presented as mean ± SD or median (interquartile range (IQR)) to provide a clear description of the data distribution. Categorical variables were expressed as n (%). HSP90α was analyzed both as a continuous variable (per 1 ng/mL increment) and as quartiles (categorical). Logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CIs) for the association between plasma HSP90α levels and the presence of PN. A multivariable model was built to adjust for potential confounders, which were selected a priori based on their known or suspected association with PN or HSP90α levels from the literature. These included age, smoking index (pack-years), high-density lipoprotein cholesterol (HDL-C), triglycerides (TGs), and family history of lung cancer. Given the relatively high prevalence of PN (>40%), the ORs may overestimate the relative risk, which should be considered when interpreting the findings. To explore the potential nonlinear relationship, a RCS regression with 3 knots placed at the 10th, 50th, and 90th percentiles of the HSP90α distribution was fitted. When nonlinearity was observed, a two-piecewise logistic regression model was applied to identify the threshold (inflection point) where the association changed. Subgroup analyses were performed to explore potential effect modification. These analyses are considered exploratory, and no adjustment for multiple testing was applied; therefore, results should be interpreted with caution. All statistical analyses were performed using R 4.3.1 software, and two-sided P values <0.05 was considered statistically significant.
Results
Baseline characteristics of the study participants
This study enrolled 968 male participants aged ≥40 years. The average age of the participants was 52.3 years, with ages ranging from 40 to 70 years. A comprehensive summary of the baseline characteristics of the study population is presented in Table 1. The prevalence of PNs among the participants was 40.2%, corresponding to 389 participants out of the 968 enrolled. Additionally, the distribution of participants across quartiles of HSP90α, labeled Q1 through Q4, is detailed in Table 1.
Baseline characteristics of participants.
HSP90α: plasma heat shock protein 90α; WBC: white blood cell count; NEU: neutrophil count; CRP: C-reactive protein; CEA: carcinoembryonic antigen; CYFRA21-1: cytokeratin 19 fragment 21-1; ALT: alanine aminotransferase; AST: aspartate aminotransferase; ALB, albumin; BUN: blood urea nitrogen; CRE: creatinine; GLU: glucose; HDL-C: high-density lipoprotein cholesterol; TC: total cholesterol; LDL-C: low-density lipoprotein cholesterol; TG: triglycerides; FEV1: forced expiratory volume in 1 second; FEV1/FVC: forced expiratory volume in 1 second/forced vital capacity; PN: pulmonary nodules.
Univariate analysis
Univariate analysis identified significant associations between PNs and multiple variables (Table 2). Elevated plasma HSP90α (OR = 1.00, 95% CI: 1.002–1.010, P = 0.002), older age (OR = 1.03, 95% CI: 1.01–1.05, P = 0.002), and a higher smoking index (OR = 1.02, 95% CI: 1.01–1.03, P < 0.001) were positively associated with the presence of nodules. Increased white blood cell (WBC) count (OR = 1.17, 95% CI: 1.05–1.30, P = 0.005) and neutrophil (NEU) count (OR = 1.22, 95% CI: 1.06–1.40, P = 0.006) also demonstrated significant association. Metabolic markers, including lower HDL-C (OR = 0.55, 95% CI: 0.32–0.95, P = 0.032) and higher TG (OR = 1.12, 95% CI: 1.02–1.23, P = 0.020), were associated with increased risk. Additionally, family history of lung cancer was associated with higher risk (OR = 1.57, 95% CI: 1.01–2.46, P = 0.04).
Results of univariate analysis.
HSP90α: plasma heat shock protein 90α; WBC: white blood cell count; NEU: neutrophil count; CRP: C-reactive protein; CEA: carcinoembryonic antigen; CYFRA21-1: cytokeratin 19 fragment 21-1; ALT: alanine aminotransferase; AST: aspartate aminotransferase; ALB: albumin; BUN: blood urea nitrogen; CRE: creatinine; GLU: glucose; HDL-C: high-density lipoprotein cholesterol; TC: total cholesterol; LDL-C: low-density lipoprotein cholesterol; TG: triglycerides; FEV1: forced expiratory volume in 1 second; FEV1/FVC: forced expiratory volume in 1 second/forced vital capacity; PN: pulmonary nodules.
Relationship between baseline HSP90α levels and risk of PNs in male participants aged ≥40 years
Multivariable logistic regression analysis revealed a significant positive association between the HSP90α levels and the prevalence of PN in males aged more than 40 years, as detailed in Table 3. In the fully adjusted Model II, which accounted for potential confounders including age, smoking index, HDL-C, TG, and family history of lung cancer, each 1 ng/mL increment in the plasma HSP90α concentration was associated with a 5% higher odds of having PN. The adjusted OR for this association was 1.05, with a 95% CI ranging from 1.01 to 1.10, and statistical significance was achieved at a P-value of 0.012. When stratified by quartiles of HSP90α levels, individuals in the highest quartile (Q4) demonstrated a substantially elevated risk compared with those in the lowest quartile (Q1). Particularly, participants in Q4 exhibited an 88% greater odds of having PN, with an adjusted OR of 1.88 (95% CI: 1.26–2.80, P = 0.002). A significant concentration-dependent relationship between increasing HSP90α levels and PN prevalence was further confirmed through a trend test across quartiles, yielding a P-value for trend of 0.001. These results collectively suggest that elevated plasma HSP90α concentrations are independently associated with the presence of PNs in this specific demographic group.
Association between plasma HSP90α levels and prevalence of PN in males aged ≥40 years.
Crude model adjusted for none; Model I adjusted for age and smoking index; Model II adjusted for age, smoking index, HDL-C, TG, and family history of lung cancer.
HSP90α: plasma heat shock protein 90α; HDL-C: high-density lipoprotein cholesterol; TG: triglycerides; PN: pulmonary nodules; OR: odds ratio; CI: confidence interval.
Nonlinear threshold effect of HSP90α on PN prevalence
Graphical representation in Figure 1 illustrates the correlation between plasma HSP90α levels and PN prevalence, demonstrating a critical threshold of approximately 89.5 ng/mL. Beyond this inflection point, the relationship between HSP90α and PN prevalence becomes statistically significant, with an adjusted OR of 1.03 (95% CI: 1.01–1.05, P = 0.008). Therefore, each 1 ng/mL increase in HSP90α levels is associated with a 3% higher odds of having PN, as detailed in Table 4.
Nonlinear relationship between plasma HSP90α levels and PN prevalence. HSP90α: heat shock protein 90α; PN: pulmonary nodule.
Results of two-piecewise linear regression model.
OR: odds ratio; CI: confidence interval; LRT: likelihood ratio test.
Subgroup analysis results and interaction effect testing
Analyses demonstrated that the relationship between elevated HSP90α and PN prevalence was significantly influenced by various demographic and clinical factors (Table 5). The strength of this association increases notably with advancing age, reaching its peak in individuals aged ≥60 years, with an OR of 2.78 (95% CI: 1.67–4.62). Additionally, a clear concentration-dependent pattern emerged with smoking exposure; heavy smokers, defined as those with more than 20 pack-years of smoking history, exhibited substantially higher odds of PNs than nonsmokers, with an OR of 2.53 (95%CI: 1.61–3.98). Inflammatory and metabolic abnormalities were also found to significantly increase the effect of elevated HSP90α. Participants with specific biomarker abnormalities, including elevated NEU counts (≥4.1 × 109/L), low HDL-C levels (<1.10 mmol/L), or high TG concentrations (>2.0 mmol/L), exhibited consistently stronger associations, with ORs ranging from 2.45 to 2.58 (all P for interaction < 0.05). Furthermore, the presence of elevated traditional tumor markers, including CEA levels exceeding 5.0 ng/mL or CYFRA21-1 levels >3.3 ng/mL, as well as impaired lung function characterized by a forced expiratory volume in one second/forced vital capacity (FEV1/FVC) ratio <0.70, were also associated with a higher odds of PN in the presence of elevated HSP90α levels.
Stratified analysis and interaction effects of plasma HSP90α level on the prevalence of PN among asymptomatic men aged ≥40 years.
HSP90α: plasma heat shock protein 90α; WBC: white blood cell count; NEU: neutrophil count; CRP: C-reactive protein; CEA, carcinoembryonic antigen; CYFRA21-1, cytokeratin 19 fragment; ALT, alanine aminotransferase; AST: aspartate aminotransferase; ALB: albumin; BUN: blood urea nitrogen; CRE, creatinine; GLU, glucose; HDL-C, high-density lipoprotein cholesterol; TC: total cholesterol; LDL-C: low-density lipoprotein cholesterol; TG: triglycerides; FEV1: forced expiratory volume in 1 second; FEV1/FVC: forced expiratory volume in 1 second/forced vital capacity; PN: pulmonary nodules; pred: predicted.
Discussion
This cross-sectional study demonstrated a significant concentration-dependent association between elevated plasma HSP90α levels and the prevalence of incidental PN in asymptomatic males aged ≥40 years who underwent LDCT screening. Importantly, we identified a potential threshold concentration of 89.5 ng/mL. Above this threshold, each 1 ng/mL increase in HSP90α levels was associated with a 3% increased odds of having PNs, independent of known risk factors including age, cumulative smoking exposure, hyperlipidemia, and family history of lung cancer. The observed concentration-dependent relationship strongly supports the well-established role of HSP90α in pulmonary pathobiology. Once released into the extracellular space, extracellular HSP90α (eHSP90α) acts as a damage-associated molecular pattern (DAMP) molecule. It exerts its biological effects by binding to specific cell surface receptors such as low-density lipoprotein receptor-related protein 1(LRP-1)/cluster of differentiation (CD) 91 (CD91) and integrins, which subsequently activates nuclear factor-kappa B(NF-κB) signaling pathways.9,10 The activation of NF-κB signaling initiates a cascade of downstream events, including the induction of pro-inflammatory responses, promotion of epithelial–mesenchymal transition (EMT), and stimulation of angiogenesis. Although these processes are crucial in driving both the formation of inflammatory nodules and the early stages of carcinogenesis,11,12 it is important to note that our observational study cannot establish a causal link between these pathways and the presence of PN. Notably, our findings show that HSP90α maintains significant independent predictive value even after adjusting for systemic markers of inflammation, including C-reactive protein (CRP) and WBC count.
Notably, Hu et al. 13 have demonstrated that HSP90α forms condensates and interacts with the client proteins containing arginine–glycine (RG)-rich repeat sequences, revealing novel phase separation mechanism that may be involved in its role in lung pathophysiology. This study provides important insights into how HSP90 organizes signaling complexes in the extracellular environment. Mechanistically, preclinical studies have suggested that HSP90α promotes nodule development through several key pathways. First, it activates the PI3K-Akt-mTOR signaling cascade, which accelerates cellular proliferation and enhances survival within newly forming lesions, as demonstrated in previous studies.14,15 Second, HSP90α induces matrix metalloproteinases, enzymes that facilitate tissue remodeling, thereby supporting the persistence and growth of nodules. 16 Third, it modulates the immune response by recruiting myeloid-derived suppressor cells, which create a permissive microenvironment conducive to nodule development. 17 Recent advancements in targeted therapy have further highlighted the role of extracellular HSP90α. Reynolds et al. 18 developed cell-impermeable, HSP90α-selective inhibitors that provide compelling evidence for the specific involvement of extracellular HSP90α in pathological processes. These inhibitors exhibit high selectivity for HSP90α compared with other HSP90 isoforms and importantly avoid inhibiting intracellular HSP90, a mechanism previously associated with significant detrimental side effects. This finding is consistent with that of recent studies indicating elevated levels of HSP90α in both benign pulmonary conditions and during malignant transformation,19,20 suggesting its involvement across the entire spectrum of nodule pathogenesis. Notably, Zhong et al. 21 revealed that extracellular HSP90α (eHSP90α) drives fibroblast senescence and pulmonary fibrosis through the transforming growth factor (TGF-β)/SMAD signaling pathway. This discovery establishes a direct mechanistic association between HSP90α and pulmonary remodeling processes. Furthermore, Yang et al. 22 found that HSP90α works in synergy with tumor necrosis factor-alpha (TNF-α) for maintaining chronic inflammation within lung tissues. This additional evidence supports the significant role of HSP90α in the development and progression of PN.
The threshold of 89.5 ng/mL for HSP90α represents a potentially clinically relevant benchmark. Importantly, this concentration is lower than the diagnostic cut-offs typically used for established malignancies, which generally range from 100 to 150 ng/mL,19,23. This suggests that the 89.5 ng/mL threshold may reflect an earlier, subclinical pathogenic process associated with nodule initiation or promotion rather than merely indicating an already established malignancy. However, this hypothesis requires validation in prospective studies. Conventional biomarkers such as CEA and CYFRA21-1 exhibit poor sensitivity, less than 50%, for detecting subcentimeter nodules.24–26 In contrast, HSP90α maintains a strong association even after adjusting for these traditional markers, highlighting its potential as a complementary indicator. The association between HSP90α and PN is particularly pronounced in older adults aged ≥65 years, with an OR of 1.77. This is consistent with the concept of cumulative exposure to environmental factors and the effects of immunosenescence. 27 Among smokers, each unit increase in HSP90α levels was associated with a 45% higher risk of PN, with an OR of 1.45, suggesting that tobacco-induced oxidative stress may amplify HSP90α activity. 28 Additionally, HSP90α shows metabolic interactions: there is an inverse association with HDL-C (OR = 0.55) and a positive association with TG (OR = 1.12), indicating interplay with pathways involved in metabolic syndrome. 29 Recent studies support these findings. Ambrocio et al. 30 demonstrated that HSP90α expression is upregulated in the airway epithelial cells of smokers and is associated with the progression of emphysema. Other studies have further demonstrated that HSP90α exacerbates lung injury associated with metabolic dysfunction by activating the NLRP3 inflammasome.31,32
Several key strengths support our conclusions. We used a well-characterized cohort of approximately 1000 asymptomatic males, ensuring a robust sample base for analysis. The study employed rigorous LDCT protocols, with radiological assessments conducted by blinded evaluators following strict adherence to Fleischner Society guidelines, enhancing the reliability of imaging data. HSP90α measurements were standardized across the study, demonstrating low intra-assay variability, which minimizes technical error in biomarker quantification. Additionally, we comprehensively adjusted for major confounding variables, including detailed smoking history (such as pack-years and current smoking status) and a full metabolic profile (encompassing factors such as blood glucose levels, lipid profiles, and body mass index), reducing the impact of extraneous influences on our findings. To further strengthen our results, we applied advanced statistical methods, including RCS and piece-wise regression, which allowed us to robustly characterize the nonlinear relationship between HSP90α levels and PN and explore potential thresholds in this association.
However, certain limitations must be acknowledged. First, the cross-sectional design of the study limits the ability to establish a causal relationship between elevated HSP90α levels and the presence of PN. Therefore, causal language has been avoided, and the findings are discussed as associations only. Elevated HSP90α could either be a contributing factor of PN development or a biological consequence of existing nodules; therefore, temporal relationships cannot be definitively determined. Prospective longitudinal studies that monitor HSP90α levels over time, alongside the evolution of PNs, would be critical to clarify the direction of this association. Second, our cohort consisted solely of males enrolled from a single center in China, which may limit the generalizability of our findings. Validation of these results in more diverse populations, including women, individuals from different ethnic backgrounds, and multicenter settings, is necessary to confirm the broader applicability of our observations. Third, although we adjusted for key confounding factors, residual confounding remains a possibility due to unmeasured variables. These could include detailed environmental exposures (such as occupational hazards or air pollution levels) and genetic susceptibility factors beyond family history, which were not fully captured in our analysis. Fourth, the threshold of 89.5 ng/mL identified in this exploratory analysis has not been internally validated (e.g. through bootstrap resampling) and requires confirmation in independent cohorts. Finally, this study was conducted in participants from a health management center, which may introduce a healthy volunteer bias, limiting the generalizability of the results to the general population.
Conclusion
This study demonstrated a significant, concentration-dependent association between elevated plasma HSP90α levels and an increased prevalence of PNs in asymptomatic males aged ≥40 years, with a critical risk threshold identified at 89.5 ng/mL. These findings suggest that plasma HSP90α levels may serve as a promising complementary biomarker for improving risk stratification in lung cancer screening. Further prospective studies are warranted to validate this threshold, clarify the temporal and causal role of HSP90α in nodule development, and evaluate the clinical usefulness of integrating HSP90α measurement into strategies for managing screen-detected PN.
Footnotes
Acknowledgments
We employed artificial intelligence tools (DeepSeek) solely for language polishing during manuscript preparation. The authors retained full oversight of content generation, data interpretation, and critical editing throughout this process.
Author contributions
JD: Conceptualization and Writing–original draft. JN: Data curation and Writing–review & editing. SZ: Formal Analysis and Writing–review & editing. SW: Formal Analysis and Writing–review & editing.
Data availability
The data underlying this article will be shared on reasonable request to the corresponding author.
Declaration of conflicting interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.
Funding
The author(s) declare that financial support was received for the research and/or publication of this article. This study was financially supported by the Key Research and development program of Shaanxi Province (grant numbers: 2024SF-YBXM-108) and the Science and Technology Development Incubation Fund Program of Shaanxi Provincial People’s Hospital (grant numbers: 2023YJY-58).
