Association Between Maternal Metabolic Parameters and Quantitative Lung CT Texture Features in Pregnancy: A Retrospective Cross-Sectional Study

Association Between Maternal Metabolic Parameters and Quantitative Lung CT Texture Features in Pregnancy: A Retrospective Cross-Sectional Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Association Between Maternal Metabolic Parameters and Quantitative Lung CT Texture Features in Pregnancy: A Retrospective Cross-Sectional Study Luman Li, Kun Yang, Ruo-xi Ran, Jie Chen, Yating Li, Hui Xie, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9085820/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Objective To investigate the association between maternal metabolic parameters and quantitative CT texture features of the lung during pregnancy. Methods This multi-center retrospective cross-sectional study included 898 pregnant women. Maternal metabolic parameters including triglycerides (TG), total cholesterol (TC), fasting plasma glucose (FPG), TyG index, TyGBMI index etc. Lung CT texture features were extracted by quantitative CT texture analysis including mean attenuation (Mean), standard deviation (Std), and skewness (Skewness), etc. Spearman correlation was used for preliminary variable selection. Multivariable linear regression was employed to evaluate independent associations between metabolic parameters and CT texture features. LASSO regression and variance inflation factor (VIF) were used to assess model robustness and multicollinearity. Results Multivariable linear regression revealed that TG (β = 6.23, 95%CI: 4.46–8.00, P < 0.001) and FPG (β = 8.18, 95%CI: 4.81–11.55, P < 0.001) were independently positively associated with Std, while TC was independently negatively associated with Std (β=-3.70, 95%CI: -4.94–-2.46, P < 0.001). TyGBMI index showed independent associations with both Mean (β = 0.39, 95%CI: 0.11–0.68, P = 0.007) and Skewness (β=-0.006, 95%CI: -0.009–-0.002, P = 0.002). LASSO cross-validation selected identical variables as the primary models, and all VIF values were below 5, indicating robust models without severe multicollinearity. Conclusions this study is the first to systematically evaluate the association between maternal metabolic parameters and lung CT texture features. TG and FPG were independent risk factors for lung heterogeneity, TC was negatively correlated with heterogeneity, and TyGBMI, as a composite metabolic index, had a robust predictive value for mean and skewness.This study provides new evidence for understanding the pulmonary imaging manifestations of metabolic disorders during pregnancy. Computed tomography CT texture features Metabolic Parameters pregnancy Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Currently Computed tomography (CT) technology has become an important tool in clinical practice, and widely used in recent years, CT imaging technique is commonly used to diagnose and evaluate a wide range of medical conditions. It has particular value in the assessment of lung diseases, such as pneumonia and lung cance[ 1 ]. Radiomics collects high-throughput quantitative data from medical images, such as CT scans. These features are extracted from segmented lesions or regions of tissue. This approach offers hold significant potential for supporting clinical decision-making[ 2 , 3 ]. Changes seen on CT images can be measured by radiomic features using first-order statistics in the field of radiomics, First-order features are based on the global gray-level histogram. These include features include mean, standard deviation (Std), skewness, and kurtosis etc. Among them, Std and skewness are two important features. Skewness indicates the asymmetry of the data distribution curve relative to the mean. Kurtosis describes the "tailedness" of the distribution compared to a Gaussian distribution, which is influenced by the presence of outliers[ 4 , 5 ]. Heterogeneity in CT images reflects the spatial variation in tissue density. This characteristic is an important factor for evaluating disease progression, treatment response, and prognosis[ 6 – 8 ]. In addition, Mean reflects the overall density characteristics of the tissue. It is often used as a predictor for assessing pathological sub-types and clinical outcomes[ 9 , 10 ]. Nevertheless, current research has mainly focused on exploring the relationship between CT imaging features and various diseases. Systematic investigations into the relationship between metabolic parameters and these CT features are still limited. Current evidence suggests that metabolic disorders may alter the internal metabolic environment, such as gestational diabetes mellitus (GDM), hypertensive disorders in pregnancy (HDP), and lipid metabolism disturbances[ 11 – 14 ]. Such changes could affect the CT radiomic features in turn, especially among pregnant women. Metabolic shifts during pregnancy have profound implications for both maternal and fetal health. However, there is still a lack of systematic research on how these metabolic factors might be reflected through imaging markers, and what they reveal about maternal health. Currently, systematic research on how metabolic factors influence maternal health remains limited, and imaging techniques offer a potential assessment approach for exploring these changes. Triglycerides (TG), total cholesterol (TC), fasting plasma glucose (FPG), body mass index (BMI), TyG index, and TyG‑BMI index are reliable biochemical indicators of metabolic syndrome, reflecting lipid metabolism status. They are closely associated with cardiovascular disease, diabetes, and other conditions[ 15 – 17 ]. Recent studies have revealed a close relationship between these metabolic parameters and CT imaging features, which has gradually emerged as a new research focus, particularly in patients with metabolic disorders. For instance, patients with diabetes and metabolic syndrome are associated with CT features, and the combination of PET/CT can assess the metabolic characteristics and anatomical structures of tumors[ 18 , 19 ]. However, studies focusing on pregnant women remain relatively scarce. Especially considering the complexity of metabolic changes during pregnancy and the impact of related complications (such as GDM and HDP) on CT images. Existing literature has not yet thoroughly investigated the specific influence of metabolic indicators on the features of CT images in pregnant women. Generally, GDM and HDP, as common pregnancy complications, have a significant impact on maternal health[ 20 , 21 ]. However, whether these diseases influence CT imaging features in pregnant women through the alteration of metabolic parameters (such as TG, TC, FPG, BMI, TyG, TyGBMI etc.) remains inadequately validated. Furthermore, existing studies have largely focused on the relationship between individual metabolic indicators and CT imaging features, lacking a systemic analysis of the integrated association between metabolic parameters and imaging features, particularly within the specific context of pregnant populations. Therefore, this study seeks to explore the association between metabolic factors and CT features in pregnant women, aiming to fill the gap between metabolic factors and CT imaging features. Provide a theoretical basis for future clinical imaging assessment. Methods Data resources This retrospective study was yielded at two health clinical centers dedicated for COVID-19 pandemic from February 1, 2020, to April 30 2020. At the time 2019 COVID-19 with the severe pandemic outbreak in Wuhan, China, when pregnant women have suspected of coronavirus infection, lung CT examination is required to facilitate timely medical treatment. Clinical information and chest CT imaging data of 1122 pregnant patients were recruited from Hubei Maternal and Child Health Hospital and Zhongnan hospital of Wuhan University. Participants were enrolled during a single time point in their pregnancy, and 898 pregnant women who meet the following criteria will be included in the study: (1) Pregnant weeks ≥ 37 weeks, aged 18–45, singleton pregnancy; (2) thorax CT examination was performed for emergency usage; (3) CT scan showed no signs of pneumonia, tumor and other diseases, CT images quality met the software analyze requirements; (4) each woman contributed only one CT scan dataset to the analysis to avoid bias from duplicate recruitment. Exclusion criteria were as follows: (1) pulmonary related diseases, chest and body shape abnormalities at present or history; (2) thorax trauma or surgical history. Routine laboratory information was collected including Triglycerides (TG), Triglyceride-glucose index (TyG), Triglyceride-glucose-body mass index (TyGBMI), Total Cholesterol (TC), Fasting Plasma Glucose (FPG), Red Blood Cell count (RBC), White Blood Cell count (WBC), Platelet count (PLT), Prothrombin Time (PT), Thrombin Time (TT), Total Protein (TP), Albumin (ALB), Blood Urea Nitrogen (BUN), Creatinine (Cr), Uric Acid (UA), High-Density Lipoprotein (HDL), Low-Density Lipoprotein (LDL). Our study had also collected clinical data, including age, pre-pregnancy BMI, increased BMI (ΔBMI), systolic/diastolic blood pressure, heart rate and body temperature, gestational ages, parity history, delivery mode, pregnancy complications, fetal outcomes. Among them, Previous studies have established that both pre-pregnancy BMI and gestational weight gain (GWG) are critical and interrelated factors influencing pregnancy outcomes[ 22 , 23 ]. As such, considering both height and GWG, ΔBMI acts as a direct indicator of the total incremental change in BMI attributable to dynamic gestational weight changes. ΔBMI was calculated as the difference between BMI at term and pre-pregnancy BMI (ΔBMI = BMI at term - pre-pregnancy BMI). CT imaging Low-dose imaging protocol was followed, the possible effects of radiation exposure on the fetus were explained in detail and written consent was obtained from each patient. According to the American College of Radiology (ACR) guidelines (American College of Radiology & Society for Pediatric Radiology, 2022), diagnostic CT scans are performed in only 0.3–2.9% of pregnancies due to justified clinical indications, it estimates low Incidence of CT Exposure During Pregnancy. Chest computed tomography was performed on all patients using 64-section CT scanners (GE discovery, Philips Ingenuity or Siemens SOMATOM Definition Flash CT Scanners). Conventional CT scan was obtained with the case in the supine position at inspiratory phase. The lung algorithm settings for the pregnant participants were as follows: 80 kV and automatic tube current 50 mAs; slice thickness, 1.5 mm; slice interval, 1.5 mm; a noise index, 16; DFOV, 36.0, and matrix, 512 x 512. The thyroid, abdomen, and pelvis of pregnant were special protected by the lead sheath. And the dose-length product (DLP) was 80–100 mGy. Both centers used identical scanner models with harmonized acquisition protocols. We took Quality Control Measures to eliminate potential inter-scanner variability: Weekly phantom tests ensured consistent performance (CTDlvol variations < 5%). A dedicated radiologist reviewed all scans for protocol compliance. Data analysis The reconstructed images were transported to the Quantitative Evaluation System of CT (YT-CT-Lung, YITU Healthcare Technology Co., Ltd., China) for quantitative analysis. Histogram-based measurements refer to Hounsfield Units (HU, also referred as CT-numbers or CT-values) frequency distribution (i.e., physical density distribution). Therefore, information on air levels in specific lung regions can be derived from lung density histograms. Quantitative measurements were obtained from the low-dose chest CT images. The average parenchyma density on CT, defined as the mode of the histogram of CT pixel values in the whole lung. The entire lung volume is obtained by adding the pixels within pre-defined CT attenuation limits in volumes of each lung tissue. The total lung volume of both lung (tissue and airspace) was measured by summing the voxel dimensions in each slice. Quantitative parameters were collected including Mean, Median, Mode, Std, Skewness, Hellinger, IoU, Volume. Statistical Analysis All analyses were performed using R software. Statistical analysis was performed with SPSS 23.0 and R software (version 4.0.0 R Statistical Computing Foundation, Vienna, Austria), p < 0.05 indicated a statistically significant difference. The distribution of continuous variables was checked initially. Normally distributed variables were expressed as mean ± standard deviation, while non-normally distributed variables were reported as median with interquartile range. Categorical variables were presented as counts and percentages. The outlier values of lung density and lung volume rule out by method contour boxplots with SPSS[ 24 ] [ 25 ], boxplots use the interquartile method with fences to find outliers, this study take the interquartile range (IQR, the middle 50% of dataset), several quartile values and an adjustment factor to calculate boundaries for what constitutes minor and major outliers. Major outliers are 3 times greater than the IQR of all remaining values. Our study had eliminated the major outliers. To assess the associations between metabolic parameters and quantitative lung CT metrics, Spearman’s rank correlation coefficient (ρ) was employed. Given the potential non-normal distribution and non-linear nature of the relationships under investigation, Spearman’s correlation was selected over Pearson’s correlation. All reported correlation results include the ρ value, the corresponding two-sided p -value, and the false discovery rate (FDR)-adjusted q -value to account for multiple testing and ensure the statistical significance and reliability of the results. q values below 0.05 considered statistically significant. Covariates were chosen prior to analysis based on their clinical relevance and previous research. Maternal age, pre-pregnancy body mass index, increased body mass index, gestational diabetes mellitus, HDP and other complications were included as potential confounders, as they are associated with metabolic status and systemic or pulmonary changes during pregnancy. Multivariable linear regression models were created to assess the independent associations between metabolic parameters and CT metrics. Not all pregnancy complications were included in the primary models to prevent overadjustment and unstable estimates. The study evaluated additional pregnancy complications in sensitivity analyses. Results Demographic and epidemiologic characteristics and Clinical findings Our study retrospectively recruited 1122 patients of two clinical centers (Hubei Provincial Maternal and Child Health Hospital, Zhongnan Hospital of Wuhan University) from February 1, 2020, to April 30 2020. The entire patients were admitted to hospital for antenatal care or delivery. 133 patients which diagnosed with pulmonary diseases were eliminated from our study. 91 patients were excluded from the study due to factors such as multiple pregnancies, preterm pregnancies, thorax trauma or surgical history. Among them, 898 patients were enrolled into our study (Fig. 1 ), main data were shown in Table 1 . Among the recruited patients, the mean age of all the participants was 30.39 ± 4.14 years, and the mean gestational age was 38.57 ± 1.13 weeks. For physical measurements, mean pre-pregnancy body mass index (BMI) was 23.46 ± 3.93 kg/m², and increased BMI (ΔBMI) was 4.32 ± 1.55 kg/m². In terms of metabolic parameters, the mean concentrations of triglyceride (TG), total cholesterol (TC) and fasting plasma glucose (FPG) were 5.80 ± 3.57 mmol/L, 6.01 ± 1.52 mmol/L and 4.94 ± 0.65 mmol/L, respectively. The TyG index and TyGBMI index were 2.95 ± 0.63 and 82.00 ± 21.83, respectively. There were 180 (20.04%) had gestational diabetes, 127 (14.25%) patients had gestational HDP, 117 (13.03%) had anemia, 79 (8.80%) had endocrine diseases, such as thyroid, pituitary problems, etc. Other conditions were 521 (58.02%), including uncomplicated pregnancies and other obstetric disorders as premature rupture of membrane (79, 8.80%), scarred uterus pregnant (154, 17.15%), placental-related obstetrics complications (41, 4.57%), fetal-related diseases (91, 10.13%), etc. Some participants may have experienced multiple complications. Table 1 Baseline characteristics of the study population Demographic characteristics All participants (n = 898) Age, years 30.39 ± 4.14 Gestational weeks 38.57 ± 1.13 Anthropometric measures Pre-pregnancy BMI, kg/m² 23.46 ± 3.93 ΔBMI, kg/m² 4.32 ± 1.55 Metabolic parameters Triglycerides (TG), mmol/L 5.80 ± 3.57 Total cholesterol (TC), mmol/L 6.01 ± 1.52 Fasting plasma glucose (FPG), mmol/L 4.94 ± 0.65 TyG inex 2.95 ± 0.63 TyGBMI index 82.00 ± 21.83 Pregnancy complications (%) GDM 180 (20.04) Hypertension 127 (14.25%) anemia 117 (13.03%) Endocrine disease 79 (8.80%) Others 521 (58.02%) Continuous variables are presented as mean ± SD or median (IQR). Categorical variables are presented as n (%). Pregnancy complications were not mutually exclusive; some participants experienced more than one complication. Spearman Correlation Analysis of Metabolic Parameters and Quantitative Lung CT Texture Features Spearman correlation analysis demonstrated statistically significant associations between metabolic parameters and quantitative lung CT texture features. Specifically, all assessed metabolic markers, including triglycerides (TG), total cholesterol (TC), fasting plasma glucose (FPG), the triglyceride-glucose (TyG) index, and the TyG-BMI index, exhibited significant correlations with multiple quantitative CT-derived parameters, namely Mean, Std and Skewness (all p < 0.001; false discovery rate–adjusted q < 0.001). (Table 2 and Table S1 ) Precisely, Mean, Std and Skewness of TG and lung CT showed a significant positive correlation (ρ = 0.175, 0.618, -0.395), respectively, among which Std had the strongest correlation with TG. TyG was also significantly correlated with Mean, Std and Skewness (ρ = 0.197, 0.639, -0.419), especially with Std and Skewness. TyGBMI was also significantly correlated with all CT indices (including Mean, Std and Skewness), especially Std and Skewness (ρ = 0.235, 0.586, -0.423). In addition, TC was negatively correlated with Mean and Std (ρ=-0.130, -0.447), while FPG was positively correlated with Mean and Std (ρ = 0.171, 0.325). The statistical significance of all results was FDR corrected by the Benjamini-Hochberg method with p and q values below 0.05. To visually demonstrate visualize the correlation between metabolic parameters and quantitative lung CT features, we used heat maps to present the results of Spearman correlation analysis (Fig. 2 and Figure S1 ). The color depth of the heat map reflects the strength of the correlation, where dark colors indicate strong correlation and light colors indicate weak or no correlation. The correlation between TG, TyG, and TyGBMI with CT standard deviation (Std) and Skewness (Skewness) is significant and strong (ρ values close to 0.6 or higher) as can be clearly seen from the heat map, where the relationship between these variables is presented in dark colors. At the same time, other metabolic measures such as TC and FPG showed relatively weak correlations with CT imaging, and in particular, TyG and TyGBMI showed consistent strong positive correlations with multiple CT measures, underlining their possible important role in the relationship between metabolic health and lung structure. Table 2 Spearman correlations between maternal metabolic parameters and quantitative CT metrics. Exposure Mean Std Skewness TG 0.175[ q < 0.001] 0.618[ q < 0.001] -0.395[ q < 0.001] TC -0.130[ q < 0.001] -0.447[ q < 0.001] 0.282[ q < 0.001] FPG 0.171[ q < 0.001] 0.325[ q < 0.001] -0.257[ q < 0.001] TyG 0.197[ q < 0.001] 0.639[ q < 0.001] -0.419[ q < 0.001] TyGBMI 0.235[ q < 0.001] 0.586[ q < 0.001] -0.423[ q < 0.001] Values are Spearman’s ρ. q values were adjusted using Benjamini-Hochberg false discovery rate (FDR). * q < 0.05, ** q < 0.01, *** q < 0.001. Multiple linear regression analysis of metabolic parameters and CT indexes Multiple linear regression analysis was performed using the forced entry method with Mean, Std and Skewness of lung CT texture parameters as dependent variables and five metabolic indexes (TG, TyG, TyGBMI, TC and FPG) screened by Spearman correlation analysis as independent variables, and the results are shown in Table 3 . The results of the Mean analysis showed that TyGBMI (β = 0.39, 95% CI: 0.11–0.68, P = 0.007) and FPG (β = 9.44, 95% CI: 2.72–16.17, P = 0.006) were independently and positively correlated with Mean; however, the associations between TG, TyG, TC and Mean were not statistically significant (all P > 0.05). The results of Std analysis showed that TG (β = 6.23, 95%CI: 4.46–8.00, P < 0.001) and FPG (β = 8.18, 95%CI: 4.81–11.55, P < 0.001) were independently positively correlated with Std. TC was independently negatively associated with Std (β=-3.70, 95%CI: -4.94–2.46, P 0.05). ( Table S2 ) Skewness analysis showed that TyGBMI (β=-0.006, 95%CI: -0.009–0.002, P = 0.002) and FPG (β=-0.13, 95%CI: -0.22–0.05, P = 0.002) were independently negatively correlated with Skewness. The association between TG and Skewness was marginal significant (β=-0.04, 95%CI: -0.09–0.00, P = 0.062). There was no significant correlation between TyG, TC and Skewness (all P > 0.05). Figure 3 presents the results of multiple linear regression analysis of the association of metabolic parameters with CT metrics (Mean, Std, Skewness). In the figure, square blocks (■) represent the unstandardized regression coefficient (β) of each variable, and the error line represents the 95% confidence interval (95% CI). The vertical dashed line represents the null line (β = 0), and if the error line intersects the dashed line, the association is not statistically significant. Table 3 Multivariable Regression Analysis of Metabolic Factors and lung CT Parameters Outcome Variable β (95% CI) P value Standardize Mean (HU) TyGBMI 0.39 (0.11, 0.68) 0.007 0.21 FPG (mmol/L) 9.44 (2.72, 16.17) 0.006 0.24 Std (HU) TG (mmol/L) 6.23 (4.46, 8.00) < 0.001 0.35 FPG (mmol/L) 8.18 (4.81, 11.55) < 0.001 0.28 TC (mmol/L) -3.70 (-4.94, -2.46) < 0.001 -0.29 Skewness TyGBMI -0.006 (-0.009, -0.002) 0.002 -0.22 FPG (mmol/L) -0.13 (-0.22, -0.05) 0.002 -0.23 TG †(mmol/L) -0.04 (-0.09, 0.00) 0.062 -0.15 *† Marginally significant (P = 0.062)* β: unstandardized coefficient; CI: confidence interval; Standardized β: standardized coefficient. All models were adjusted simultaneously for all variables shown in the table. Sensitivity Analysis of the LASSO Regression Model To evaluate the robustness of the model and the potential risk of overfitting, LASSO regression (10-fold cross-validation) was further used for variable selection on the same set of independent variables. The results showed that under the optimal penalty parameter (λ), the non-zero coefficient variables screened by LASSO model were exactly the same as those included in the main analysis (multiple linear regression), which were TG, FPG, TC and TyGBMI. This result indicates that the variable selection strategy of this study is robust and there is no significant overfitting of the model. To assess the robustness of the model and the risk of overfitting, LASSO regression (10-fold cross-validation) was used to select variables from 22 candidate variables. Figure 4 A shows the cross-validation error curve, showing an optimal λ of 0.16 (λ.min). Figure 4 B shows the coefficient path plot, showing that TG, FPG, and TC retain nonzero coefficients at λ.min. Figure 5 shows the selected variables under the λ.min criterion, and the direction of their coefficients is exactly the same as in Table 3 . In conclusion, the variables selected automatically by LASSO were completely consistent with our main model-TG, FPG, and TC were the most robust metabolic correlates of lung CT texture. TyGBMI was compressed under the more stringent criteria, consistent with its properties as a composite measure. In addition, variance inflation factor (VIF) was used to diagnose collinearity in each multiple linear regression model ( Table S3 ). The results showed that the VIF values of all independent variables ranged from 1.24 to 3.13, all well below the alert threshold of 10, and all below the strict criterion of 5. FPG (VIF = 1.24) and TC (VIF = 1.27) were almost completely independent of other variables. TG (VIF = 3.06) and TyGBMI (VIF = 3.13) were moderately correlated, but still within the acceptable range. The above results indicate that the regression model in this study does not suffer from serious multicollinearity problems and the parameter estimates are stable and reliable. Figure 4 A Ten-fold cross-validation error curve as a function of log(λ). The left vertical dashed line indicates λ.min (0.15, logλ=-1.86), the value minimizing mean-squared error. The right vertical dashed line indicates λ.1se (34.28, logλ = 3.53), the largest λ within one standard error of the minimum. Figure 5 Bar plot showing the coefficients of variables selected by LASSO at λ.min = 0.155. Positive coefficients (red) indicate positive associations, while negative coefficient (blue) indicates negative association. The selected variables (TG, FPG, TC, TyGBMI) and their coefficient directions are identical to those in the primary multivariable regression models (Table 3 ). Discussion This study systematically evaluated the association between maternal metabolic parameters and CT texture features of lung parenchyma based on a large sample of chest CT quantitative parameters during pregnancy. The main findings were as follows: First, triglyceride (TG) and fasting plasma glucose (FPG) were independent positive predictors of lung CT standard deviation (Std), suggesting that disorders of glucose and lipid metabolism are closely related to increased lung tissue density heterogeneity; Second, total cholesterol (TC) was independently and inversely associated with Std, suggesting its possible protective role in the maintenance of lung microstructure during pregnancy. Third, TyG‑BMI was significantly and independently associated with both CT Mean and Skewness, and the effect was more robust at the composite level than that of single metabolic parameter. These findings suggest that maternal metabolic status has an independent and multidimensional effect on imaging phenotypes of lung during pregnancy. Previous studies on the relationship between metabolism and lung structure during pregnancy mostly focused on clinical outcomes such as asthma and sleep apnea, and few used imaging methods to directly evaluate the changes of lung parenchyma[ 26 – 28 ]. In our study, TG had the highest standardized regression coefficient on Std (standardized coefficient β = 0.35) (Table 3 ), suggesting that triglycerides had the strongest correlation with lung density heterogeneity, indicating that lipid metabolism disorders are closely related to lung microstructural heterogeneity. This consistent with the results of a cohort study based on a nonpregnant population, the incidence of metabolic syndrome is increased in COPD patients, and the TG levels of COPD patients are higher than those of non-pregnant healthy people[ 29 ]. The reason may be that systemic inflammation plays an important role in both COPD and dyslipidemia, and inflammatory factors may break the balance of lipid[ 30 , 31 ]. Alterations in lipid molecules are associated with impaired lung function and inversely correlated with emphysema and interstitial lung abnormalities[ 32 ]. Lipid metabolism can initiate a variety of cellular processes through lysosomal inhibitors. Serum lysophospholipids are negatively correlated with pulmonary interstitial abnormalities, and the underlying mechanism may be related to lung function and systemic inflammatory markers[ 32 ]. In COPD patients, lung CT texture features have been demonstrated to correlate with systemic inflammatory markers. Therefore, we speculate that the association of lung CT texture abnormalities with TG observed in this study may be partially mediated by a common 'systemic inflammatory state'. Future studies incorporating inflammatory markers into mediation analyses would be helpful to further test this hypothesis[ 33 ]. Another related research confirmed that areas of structural heterogeneity on lung CT showed a significant positive correlation with glucose metabolic activity and pointed to the presence of inflammation in most areas of pulmonary heterogeneity. This finding reveals that structural alterations in the lung may be accompanied by abnormal activation of local glucose metabolism[ 34 ]. We speculate that blood glucose may indirectly affect the imaging features of lung CT by driving the inflammatory response in the lung. It was consistent with our study in pregnancy women, FPG was independently and positively correlated with the standard deviation of lung CT (Std) (β = 8.18, P < 0.001) (Table 3 ), and Std is precisely an indicator reflecting the heterogeneity of lung density. Meanwhile, the independent positive correlation between FPG and lung CT Mean (Mean) (β = 9.44, P = 0.006) suggested that blood glucose may also affect the overall density of the lung. This may reflect diffuse infiltration of inflammatory cells, alveolar wall thickening, or altered interstitial components. Notably, the negative correlation between FPG and Skewness (β=-0.13, P = 0.002) further supported the change in distribution characteristics, with a decrease in left skewness indicating a shift in lung density distribution toward higher values, consistent with an increase in Mean. (Table 3 ) The negative association of TC with Std is another important finding of our study. Although a high TC level is considered a risk factor for cardiovascular disease[ 35 ], in the special physiological state of pregnancy, cholesterol transporters involved in some pregnancy pathologies such as preeclampsia, gestational diabetes mellitus, and intrauterine growth retardation[ 36 ], hypercholesterolemia participated in the regulation of cholesterol transport across the human placenta and lipoprotein profiles in maternal and neonatal circulation[ 37 ]. Different form TG and FPG, TC showed a consistent negative correlation with other lung CT parameters: higher TC was associated with lower Mean (Mean) and standard deviation (Std) and higher Skewness (Skewness) in lung CT. The direction of association between TC and lung CT was opposite to that of other CT parameters (seen in Table 2 , Table 3 ), suggesting that cholesterol may affect lung tissue by different biological mechanisms. cholesterol is a key substrate for placental synthesis of progesterone and fetal pulmonary surfactant. Studies have shown that moderately elevated maternal cholesterol levels during pregnancy are positively correlated with neonatal lung development maturity. Therefore, the negative association between TC and lung density heterogeneity observed in this study may reflect its potential protective role in maintaining lung tissue homogeneity and structural integrity. Further longitudinal studies and in vitro experiments are needed to verify this hypothesis. In the quantitative analysis of lung CT, BMI is a variable that cannot be ignored. BMI is not only a metabolic index, but also closely related to the image acquisition and interpretation of lung CT. Image noise was significantly increased in the high BMI group (average 39.8 kg/m²)[ 38 ]. A pioneering study combined PET/CT radiomics features with glucose metabolism-related genes to construct a prognostic model of non-small cell lung cancer, and found that imaging features could effectively reflect the metabolic remodeling of lung tumors[ 39 ]. TyG index and its composite index of BMI (TyGBMI) reflect insulin resistance and multiple risk predictors of chronic diseases[ 40 – 42 ], TyG and TyGBMI generally have stronger correlations with lung CT parameters than single metabolic index, suggesting that the coupling of glucose and lipid metabolism may have a more significant effect on lung structure (Table 2 ). This suggests that insulin resistance may play a central role in the lung-metabolic axis. Notably, although TyG index was strongly correlated with all three CT texture parameters in univariate analysis, its independent effects disappeared after adjusting for TG, FPG and BMI. However, TyG‑BMI index still maintained significant predictive power for Mean and Skewness in multiple regression. (Table 2 , Table 3 ) This finding suggests that the overall burden of the metabolic syndrome may be the central factor driving phenotypic changes on lung imaging during pregnancy, rather than insulin resistance or elevated lipids alone. TyG‑BMI can better reflect the true metabolic status than single metabolic parameter. This finding is consistent with the findings of recent studies in the assessment fo insulin resistance for GDM, it supports the use of composite metabolic index in research of metabolic-imaging during pregnancy[ 43 , 44 ]. Our study has several limitations. First, as a cross-sectional study, mechanistic hypothesis has been proposed, but the direction of causality still needs to be verified by prospective cohort studies. Second, although we adjusted for multiple potential confounders, unmeasured residual confounding remains possible (e.g., dietary intake, physical activity, environmental exposure, etc.). Thirdly, our study included only pregnant women, and extrapolation to the nonpregnant population should be cautious. Forth, some metabolic measures (e.g., insulin, glycated hemoglobin) were not included in the analyses, limiting the depth of mechanistic exploration. Despite the limitations mentioned above, this study revealed for the first time the association between metabolic indicators and lung CT parameters in the pregnant population through strict FDR correction and multivariate regression analysis, which provided new clues and directions for the study of the lung-metabolic axis. In conclusion, this study first report the independent associations between maternal TG, FPG, TC, and TyGBMI indices and lung CT texture parameters during pregnancy. TG and FPG were the main independent metabolic risk factors for lung heterogeneity. TC was negatively correlated with lung heterogeneity independently, which was opposite to other metabolic indicators, suggesting that TC may affect lung structure through different biological mechanisms. TyGBMI is a robust marker to reflect the effect of metabolic burden on lung imaging phenotype, which correlated with multiple parameters of lung CT. This study provides novel imaging evidence linking metabolic abnormalities to structural changes in the lungs during pregnancy. These findings offer quantifiable indicators for the early identification and intervention of high-risk populations. Prospective cohort studies and mechanistic experiments are needed to further clarify the biological pathways. Abbreviations CT Computed tomography Std Standard deviation GDM Gestational diabetes mellitus HDP Hypertensive disorders in pregnancy TG Triglycerides TC Total cholesterol FPG Fasting plasma glucose BMI Body mass index VIF Variance inflation factor Declarations Ethic statement This study was approved by the Institutional Ethics Committees of Zhongnan Hospital of Wuhan University and the Maternal (Approval No. 2020072K). This study is a retrospective analysis of non-identifiable researcher-collected data. The requirement for informed consent from the study subjects was waived by the IRB of Ethical Approval of Ethics Committee of Zhongnan Hospital of Wuhan University due to the retrospective study design. We confirm that: The research does not involve any human experimentation; No identifiable subject data or biological materials were utilized; No personal privacy concerns or commercial interests are implicated in this work. Consent for publication Not applicable Author contributions the following authors contributed equally and should be considered co-first authors: Luman Li, Kun Yang, Ruo-xi Ran, and Jie Chen. All authors participated in the multicenter and multidisciplinary collaboration，contributed equally to this article and should be considered as first authors with specialized roles in Obstetrics, Clinical laboratory, Radiography, respectively Luman Li (Affiliations 1, 2, 3): conceptualization, study design, manuscript drafting, data interpretation. Kun Yang (Department of Pharmacy, Zhongnan Hospital of Wuhan University): sample collection, statistical analysis, data acquisition Ruo-xi Ran (Department of Clinical Laboratory Medicine, Maternal and Child Health Hospital of Hubei Province): laboratory testing, quality control, data analysis. Jie Chen (Department of Nuclear Medicine, Zhongnan Hospital of Wuhan University): imaging experiments, data interpretation Yating Li (Department of Obstetrics, Zhongnan Hospital of Wuhan University), data collection, data analysis. Hui Xie (Department of Radiography, Maternal and Child Health Hospital of Hubei Province): statistical analysis, imaging experiments, methodology support. Meiyan Liao (Department of Nuclear Medicine, Zhongnan Hospital of Wuhan University): experimental support, data validation Huijun Chen (Department of Obstetrics, Zhongnan Hospital of Wuhan University): supervision, clinical guidance, critical revision of manuscript. Yuanzhen Zhang (Affiliations 1, 2, 3): overall study supervision, funding acquisition, main corresponding author, manuscript final approval. All authors who contributed to the manuscript revision, read, and approved the submitted version. FUNDING This study was supported by Hubei Science and Technology Plan (grant number 2020FCA011), Wuhan Technology and Innovation Plan (grant number 2020020201010010), Zhongnan Hospital Major Project (grant number ZNJC202503) Competing interests The authors have no competing interests to disclose. Acknowledgements We thank all the patients who have participated in this study. Data availability The data that support the findings of this study are not publicly available due to privacy or ethical restrictions. Data are available from the corresponding author upon reasonable request. References Zhou TH, et al. CT whole lung radiomic nomogram: a potential biomarker for lung function evaluation and identification of COPD. Mil Med Res. 2024;11(1):14. Mayerhoefer ME, et al. Introduction to Radiomics. J Nucl Med. 2020;61(4):488–95. Choi H, et al. Extension of Coronavirus Disease 2019 on Chest CT and Implications for Chest Radiographic Interpretation. Radiol Cardiothorac Imaging. 2020;2(2):e200107. Aerts HJ, et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun. 2014;5:4006. Zwanenburg A, et al. The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology. 2020;295(2):328–38. Park BW, et al. Reliability of CT radiomic features reflecting tumour heterogeneity according to image quality and image processing parameters. Sci Rep. 2020;10(1):3852. Gillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563–77. Burrell RA, et al. The causes and consequences of genetic heterogeneity in cancer evolution. Nature. 2013;501(7467):338–45. Liu N, et al. Acute necrotising pancreatitis: measurements of necrosis volume and mean CT attenuation help early prediction of organ failure and need for intervention. Eur Radiol. 2021;31(10):7705–14. Koezuka S, et al. Combination of mean CT value and maximum CT value as a novel predictor of lepidic predominant lesions in small lung adenocarcinoma presenting as solid nodules. Sci Rep. 2022;12(1):5450. Mokkala K, et al. Metagenomics analysis of gut microbiota in response to diet intervention and gestational diabetes in overweight and obese women: a randomised, double-blind, placebo-controlled clinical trial. Gut. 2021;70(2):309–18. Chakraborty S, et al. Metabolites and Hypertension: Insights into Hypertension as a Metabolic Disorder: 2019 Harriet Dustan Award. Hypertension. 2020;75(6):1386–96. Klag K, et al. Dietary fat disrupts a commensal-host lipid network that promotes metabolic health. Cell Metab. 2026;38(1):157–e1739. Li L, et al. Establishment and validation of a predictive nomogram for gestational diabetes mellitus during early pregnancy term: A retrospective study. Front Endocrinol (Lausanne). 2023;14:1087994. Khan SS, et al. Body Mass Index, Adverse Pregnancy Outcomes, and Cardiovascular Disease Risk. Circ Res. 2023;133(9):725–35. Freaney PM, et al. The Role of Sex-Specific Risk Factors in the Risk Assessment of Atherosclerotic Cardiovascular Disease for Primary Prevention in Women. Curr Atheroscler Rep. 2020;22(9):46. He Q et al. Association between the cumulative atherogenic index of plasma and cardiometabolic multimorbidity: the mediating effects of the TyG index and body mass index. Lipids Health Dis, 2026. 25(1). Warner JD, et al. Correlation of HbA1c levels with CT-based body composition biomarkers in diabetes mellitus and metabolic syndrome. Sci Rep. 2024;14(1):21875. Czernin J, Benz MR, Allen-Auerbach MS. PET/CT imaging: The incremental value of assessing the glucose metabolic phenotype and the structure of cancers in a single examination. Eur J Radiol. 2010;73(3):470–80. Simmons D, et al. Regression From Early GDM to Normal Glucose Tolerance and Adverse Pregnancy Outcomes in the Treatment of Booking Gestational Diabetes Mellitus Study. Diabetes Care. 2024;47(12):2079–84. Rosenberg EA, Seely EW. Update on Preeclampsia and Hypertensive Disorders of Pregnancy. Endocrinol Metab Clin North Am. 2024;53(3):377–89. Teede HJ, et al. Association of Antenatal Diet and Physical Activity-Based Interventions With Gestational Weight Gain and Pregnancy Outcomes: A Systematic Review and Meta-analysis. JAMA Intern Med. 2022;182(2):106–14. Girchenko P, et al. Associations of polymetabolic risk of high maternal pre-pregnancy body mass index with pregnancy complications, birth outcomes, and early childhood neurodevelopment: findings from two pregnancy cohorts. BMC Pregnancy Childbirth. 2024;24(1):78. Shiffler RE. Maximum Z Scores and Outliers. Am Stat. 1988;42(1):79–80. Whitaker RT, Mirzargar M, Kirby RM. Contour boxplots: a method for characterizing uncertainty in feature sets from simulation ensembles. IEEE Trans Vis Comput Graph. 2013;19(12):2713–22. Denton E, et al. Asthma Pregnancy Allergy. 2026;81(1):84–108. Tasali E, et al. Obstructive Sleep Apnea and Cardiometabolic Disease: Obesity, Hypertension, and Diabetes. Circ Res. 2025;137(5):764–87. Tang WZ, et al. Obstructive sleep apnea-associated hypertensive disorders in pregnancy: a literature review and clinical management strategies. Reprod Biol Endocrinol. 2025;23(1):114. Xuan L, et al. Association between chronic obstructive pulmonary disease and serum lipid levels: a meta-analysis. Lipids Health Dis. 2018;17(1):263. Novgorodtseva TP, et al. Modification of the fatty acid composition of the erythrocyte membrane in patients with chronic respiratory diseases. Lipids Health Dis. 2013;12:117. Esteve E, Ricart W, Fernández-Real JM. Dyslipidemia and inflammation: an evolutionary conserved mechanism. Clin Nutr. 2005;24(1):16–31. Zhou Q, et al. Lysophospholipid metabolism, clinical characteristics, and artificial intelligence-based quantitative assessments of chest CT in patients with stable COPD and healthy smokers. Sci Rep. 2025;15(1):26376. Zeng Q, et al. CT-derived abdominal adiposity: Distributions and better predictive ability than BMI in a nationwide study of 59,429 adults in China. Metabolism. 2021;115:154456. Cressoni M, et al. Lung inhomogeneities, inflation and [18F]2-fluoro-2-deoxy-D-glucose uptake rate in acute respiratory distress syndrome. Eur Respir J. 2016;47(1):233–42. Zhang G, et al. LDL-C and TC Mediate the Risk of PNPLA3 Inhibition in Cardiovascular Diseases. J Clin Endocrinol Metab. 2025;110(2):e231–8. Kallol S, Albrecht C. Materno-fetal cholesterol transport during pregnancy. Biochem Soc Trans. 2020;48(3):775–86. Cantin C, Fuenzalida B, Leiva A. Maternal hypercholesterolemia during pregnancy: Potential modulation of cholesterol transport through the human placenta and lipoprotein profile in maternal and neonatal circulation. Placenta. 2020;94:26–33. Sourlos N, et al. Does BMI influence AI and human reader lung nodule detection in low-dose chest CT? Eur J Radiol. 2025;193:112453. Wang C, et al. Development and validation of a radiogenomics prognostic model integrating PET/CT radiomics and glucose metabolism-related gene signatures for non-small cell lung cancer. Eur J Nucl Med Mol Imaging. 2025;52(13):4924–38. Jiang L, et al. Assessment of six insulin resistance surrogate indexes for predicting stroke incidence in Chinese middle-aged and elderly populations with abnormal glucose metabolism: a nationwide prospective cohort study. Cardiovasc Diabetol. 2025;24(1):56. Liu H, et al. Triglyceride-glucose index correlates with the occurrence and prognosis of acute myocardial infarction complicated by cardiogenic shock: data from two large cohorts. Cardiovasc Diabetol. 2024;23(1):337. He D, et al. Prospective associations of triglyceride-glucose related indices with cardiovascular disease and mortality in individuals with metabolic syndrome: evidence from the UK biobank. Cardiovasc Diabetol. 2026;25(1):53. Zhang L, et al. Assessment of first-trimester insulin resistance indices for gestational diabetes mellitus: a prospective cohort study. J Endocrinol Invest. 2025;48(9):2167–76. Liang X, et al. Association between triglyceride glucose-body mass index and gestational diabetes mellitus: a prospective cohort study. BMC Pregnancy Childbirth. 2025;25(1):170. Additional Declarations No competing interests reported. Supplementary Files Table1S.doc tableS2.doc TableS3.doc Figure1S.jpg Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 08 Apr, 2026 Editor assigned by journal 06 Apr, 2026 Editor invited by journal 18 Mar, 2026 Submission checks completed at journal 17 Mar, 2026 First submitted to journal 17 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9085820","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":623380625,"identity":"d5ced910-d961-4404-95f1-b34e786dce16","order_by":0,"name":"Luman Li","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Luman","middleName":"","lastName":"Li","suffix":""},{"id":623380626,"identity":"7e180006-1f47-4fb5-8b12-8aa517db632f","order_by":1,"name":"Kun Yang","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Kun","middleName":"","lastName":"Yang","suffix":""},{"id":623380627,"identity":"ad4ae284-829f-48ec-9eb0-e697e9c2c5e4","order_by":2,"name":"Ruo-xi Ran","email":"","orcid":"","institution":"Hubei Provincial Women and Children's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ruo-xi","middleName":"","lastName":"Ran","suffix":""},{"id":623380628,"identity":"4beafeab-23e8-41d8-82e9-ca8096641d5f","order_by":3,"name":"Jie Chen","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Jie","middleName":"","lastName":"Chen","suffix":""},{"id":623380629,"identity":"82be64d1-59ca-4bea-8333-1b23f136136c","order_by":4,"name":"Yating Li","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Yating","middleName":"","lastName":"Li","suffix":""},{"id":623380630,"identity":"1428765b-609e-4be8-86bc-54032fd35911","order_by":5,"name":"Hui Xie","email":"","orcid":"","institution":"Hubei Provincial Women and Children's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Xie","suffix":""},{"id":623380631,"identity":"ed688f32-ba0a-429b-b3f0-6c07e435a69e","order_by":6,"name":"Meiyan Liao","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Meiyan","middleName":"","lastName":"Liao","suffix":""},{"id":623380632,"identity":"b0b7d4d6-26a4-4059-bb51-54d7a49618ff","order_by":7,"name":"Huijun Chen","email":"","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":false,"prefix":"","firstName":"Huijun","middleName":"","lastName":"Chen","suffix":""},{"id":623380633,"identity":"d19cedfa-9d6d-4dfd-989b-5a9649ea785f","order_by":8,"name":"Yuanzhen Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAUlEQVRIiWNgGAWjYDACCTiLB4grGBJQBQlrOUOyFsY2IrTIz25+9vBr22F5c/6zBx8XzjucZ3CA+eBtHga7PFxaGOccMzeWbTtsuLPhXLLxzG2Hiw0OsCVb8zAkF+PSwiyRYCYt2XaYccPBHjNp3m23Ezcc4DGT5mE4kNiAQwubRPo3kBb7DYd5zH/zzgFp4f+GVwuPRI6Z5Me2w4kbjvGYMfM2gG1hw6tFQiKnTJrhXHryhjM8xtI8x/4nzjzMZmw5xyAZpxb5GenbJH+UWdtuOH/G8DNPTVpi3/HmhzfeVNjh1AIOAl42FC6IMMCjHggYf/zBr2AUjIJRMApGOAAAnZxYEboafkAAAAAASUVORK5CYII=","orcid":"","institution":"Zhongnan Hospital of Wuhan University","correspondingAuthor":true,"prefix":"","firstName":"Yuanzhen","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2026-03-10 15:38:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9085820/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9085820/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107245419,"identity":"0097341c-6665-41be-9612-eb58d57e30fc","added_by":"auto","created_at":"2026-04-19 08:05:11","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":459588,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of study participant selection.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/acc28e64623444b2cd5ea5a0.jpg"},{"id":107482638,"identity":"e2417044-0dba-4646-9c2a-477ed8ed4446","added_by":"auto","created_at":"2026-04-22 02:24:16","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":174571,"visible":true,"origin":"","legend":"\u003cp\u003eheat maps of Spearman correlation analysis of Metabolic Parameters and Quantitative Lung CT Texture Features\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/92fc8488f79aaca6b754d2b1.jpg"},{"id":107482740,"identity":"0d5e5fb9-764f-4197-82db-48bda09e9e5a","added_by":"auto","created_at":"2026-04-22 02:24:43","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":226818,"visible":true,"origin":"","legend":"\u003cp\u003emultiple linear regression analysis of the association of metabolic parameters with CT features\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/e701eb443a4525e65b10d264.jpg"},{"id":107245425,"identity":"1e412bda-ce21-4c4e-9e4a-9a1ec9e48f2c","added_by":"auto","created_at":"2026-04-19 08:05:11","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":311735,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA \u003c/strong\u003eTen-fold cross-validation error curve as a function of log(λ). The left vertical dashed line indicates λ.min (0.15, logλ=-1.86), the value minimizing mean-squared error. The right vertical dashed line indicates λ.1se (34.28, logλ=3.53), the largest λ within one standard error of the minimum.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e Coefficient paths for TG, FPG, and TC across different log(λ) values. The vertical dashed line indicates λ.min, at which these variables retain non-zero coefficients.\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/c451b7cb7a1d0a43e18bb664.jpg"},{"id":107482854,"identity":"a14af7a8-7296-4d2c-8328-3bb010f9cad7","added_by":"auto","created_at":"2026-04-22 02:25:09","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":120413,"visible":true,"origin":"","legend":"\u003cp\u003eBar plot showing the coefficients of variables selected by LASSO at λ.min=0.155. Positive coefficients (red) indicate positive associations, while negative coefficient (blue) indicates negative association. The selected variables (TG, FPG, TC, TyGBMI) and their coefficient directions are identical to those in the primary multivariable regression models (Table 3).\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/213d5fdec2cf601605408a21.jpg"},{"id":107705324,"identity":"c2e2db1e-45bb-4e97-b3ac-344ea8e03107","added_by":"auto","created_at":"2026-04-24 09:11:29","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1648598,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/c58e0a80-6e1f-4f7c-8bc9-8f91c57ba064.pdf"},{"id":107484383,"identity":"fa59d315-1b05-451c-9704-1dccfc854d18","added_by":"auto","created_at":"2026-04-22 02:31:47","extension":"doc","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":412672,"visible":true,"origin":"","legend":"","description":"","filename":"Table1S.doc","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/61c9e13352b7b6316b7992a7.doc"},{"id":107482860,"identity":"4e39a925-42e1-48b7-969b-d33ea84a0a93","added_by":"auto","created_at":"2026-04-22 02:25:11","extension":"doc","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":83456,"visible":true,"origin":"","legend":"","description":"","filename":"tableS2.doc","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/0a587ae2a1ee95ad6dfbfcd0.doc"},{"id":107245423,"identity":"8e117fcd-3431-46a5-b52f-198ba108bd72","added_by":"auto","created_at":"2026-04-19 08:05:11","extension":"doc","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":42496,"visible":true,"origin":"","legend":"","description":"","filename":"TableS3.doc","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/c09f0b164268e59fc92dbf27.doc"},{"id":107484370,"identity":"549f87de-5e67-42cd-ae57-9731fe3bbeab","added_by":"auto","created_at":"2026-04-22 02:31:44","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":244104,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1S.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9085820/v1/e2decbaf72d414f355291d18.jpg"}],"financialInterests":"No competing interests reported.","formattedTitle":"Association Between Maternal Metabolic Parameters and Quantitative Lung CT Texture Features in Pregnancy: A Retrospective Cross-Sectional Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eCurrently Computed tomography (CT) technology has become an important tool in clinical practice, and widely used in recent years, CT imaging technique is commonly used to diagnose and evaluate a wide range of medical conditions. It has particular value in the assessment of lung diseases, such as pneumonia and lung cance[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Radiomics collects high-throughput quantitative data from medical images, such as CT scans. These features are extracted from segmented lesions or regions of tissue. This approach offers hold significant potential for supporting clinical decision-making[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Changes seen on CT images can be measured by radiomic features using first-order statistics in the field of radiomics, First-order features are based on the global gray-level histogram. These include features include mean, standard deviation (Std), skewness, and kurtosis etc. Among them, Std and skewness are two important features. Skewness indicates the asymmetry of the data distribution curve relative to the mean. Kurtosis describes the \"tailedness\" of the distribution compared to a Gaussian distribution, which is influenced by the presence of outliers[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Heterogeneity in CT images reflects the spatial variation in tissue density. This characteristic is an important factor for evaluating disease progression, treatment response, and prognosis[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. In addition, Mean reflects the overall density characteristics of the tissue. It is often used as a predictor for assessing pathological sub-types and clinical outcomes[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eNevertheless, current research has mainly focused on exploring the relationship between CT imaging features and various diseases. Systematic investigations into the relationship between metabolic parameters and these CT features are still limited. Current evidence suggests that metabolic disorders may alter the internal metabolic environment, such as gestational diabetes mellitus (GDM), hypertensive disorders in pregnancy (HDP), and lipid metabolism disturbances[\u003cspan additionalcitationids=\"CR12 CR13\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Such changes could affect the CT radiomic features in turn, especially among pregnant women. Metabolic shifts during pregnancy have profound implications for both maternal and fetal health.\u003c/p\u003e \u003cp\u003eHowever, there is still a lack of systematic research on how these metabolic factors might be reflected through imaging markers, and what they reveal about maternal health. Currently, systematic research on how metabolic factors influence maternal health remains limited, and imaging techniques offer a potential assessment approach for exploring these changes.\u003c/p\u003e \u003cp\u003eTriglycerides (TG), total cholesterol (TC), fasting plasma glucose (FPG), body mass index (BMI), TyG index, and TyG‑BMI index are reliable biochemical indicators of metabolic syndrome, reflecting lipid metabolism status. They are closely associated with cardiovascular disease, diabetes, and other conditions[\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Recent studies have revealed a close relationship between these metabolic parameters and CT imaging features, which has gradually emerged as a new research focus, particularly in patients with metabolic disorders. For instance, patients with diabetes and metabolic syndrome are associated with CT features, and the combination of PET/CT can assess the metabolic characteristics and anatomical structures of tumors[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. However, studies focusing on pregnant women remain relatively scarce. Especially considering the complexity of metabolic changes during pregnancy and the impact of related complications (such as GDM and HDP) on CT images. Existing literature has not yet thoroughly investigated the specific influence of metabolic indicators on the features of CT images in pregnant women.\u003c/p\u003e \u003cp\u003eGenerally, GDM and HDP, as common pregnancy complications, have a significant impact on maternal health[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. However, whether these diseases influence CT imaging features in pregnant women through the alteration of metabolic parameters (such as TG, TC, FPG, BMI, TyG, TyGBMI etc.) remains inadequately validated. Furthermore, existing studies have largely focused on the relationship between individual metabolic indicators and CT imaging features, lacking a systemic analysis of the integrated association between metabolic parameters and imaging features, particularly within the specific context of pregnant populations.\u003c/p\u003e \u003cp\u003eTherefore, this study seeks to explore the association between metabolic factors and CT features in pregnant women, aiming to fill the gap between metabolic factors and CT imaging features. Provide a theoretical basis for future clinical imaging assessment.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eData resources\u003c/h2\u003e \u003cp\u003eThis retrospective study was yielded at two health clinical centers dedicated for COVID-19 pandemic from February 1, 2020, to April 30 2020. At the time 2019 COVID-19 with the severe pandemic outbreak in Wuhan, China, when pregnant women have suspected of coronavirus infection, lung CT examination is required to facilitate timely medical treatment. Clinical information and chest CT imaging data of 1122 pregnant patients were recruited from Hubei Maternal and Child Health Hospital and Zhongnan hospital of Wuhan University. Participants were enrolled during a single time point in their pregnancy, and 898 pregnant women who meet the following criteria will be included in the study: (1) Pregnant weeks\u0026thinsp;\u0026ge;\u0026thinsp;37 weeks, aged 18\u0026ndash;45, singleton pregnancy; (2) thorax CT examination was performed for emergency usage; (3) CT scan showed no signs of pneumonia, tumor and other diseases, CT images quality met the software analyze requirements; (4) each woman contributed only one CT scan dataset to the analysis to avoid bias from duplicate recruitment. Exclusion criteria were as follows: (1) pulmonary related diseases, chest and body shape abnormalities at present or history; (2) thorax trauma or surgical history.\u003c/p\u003e \u003cp\u003eRoutine laboratory information was collected including Triglycerides (TG), Triglyceride-glucose index (TyG), Triglyceride-glucose-body mass index (TyGBMI), Total Cholesterol (TC), Fasting Plasma Glucose (FPG), Red Blood Cell count (RBC), White Blood Cell count (WBC), Platelet count (PLT), Prothrombin Time (PT), Thrombin Time (TT), Total Protein (TP), Albumin (ALB), Blood Urea Nitrogen (BUN), Creatinine (Cr), Uric Acid (UA), High-Density Lipoprotein (HDL), Low-Density Lipoprotein (LDL). Our study had also collected clinical data, including age, pre-pregnancy BMI, increased BMI (ΔBMI), systolic/diastolic blood pressure, heart rate and body temperature, gestational ages, parity history, delivery mode, pregnancy complications, fetal outcomes. Among them, Previous studies have established that both pre-pregnancy BMI and gestational weight gain (GWG) are critical and interrelated factors influencing pregnancy outcomes[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. As such, considering both height and GWG, ΔBMI acts as a direct indicator of the total incremental change in BMI attributable to dynamic gestational weight changes. ΔBMI was calculated as the difference between BMI at term and pre-pregnancy BMI (ΔBMI\u0026thinsp;=\u0026thinsp;BMI at term - pre-pregnancy BMI).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCT imaging\u003c/h3\u003e\n\u003cp\u003eLow-dose imaging protocol was followed, the possible effects of radiation exposure on the fetus were explained in detail and written consent was obtained from each patient. According to the American College of Radiology (ACR) guidelines (American College of Radiology \u0026amp; Society for Pediatric Radiology, 2022), diagnostic CT scans are performed in only 0.3\u0026ndash;2.9% of pregnancies due to justified clinical indications, it estimates low Incidence of CT Exposure During Pregnancy. Chest computed tomography was performed on all patients using 64-section CT scanners (GE discovery, Philips Ingenuity or Siemens SOMATOM Definition Flash CT Scanners). Conventional CT scan was obtained with the case in the supine position at inspiratory phase. The lung algorithm settings for the pregnant participants were as follows: 80 kV and automatic tube current 50 mAs; slice thickness, 1.5 mm; slice interval, 1.5 mm; a noise index, 16; DFOV, 36.0, and matrix, 512 x 512. The thyroid, abdomen, and pelvis of pregnant were special protected by the lead sheath. And the dose-length product (DLP) was 80\u0026ndash;100 mGy. Both centers used identical scanner models with harmonized acquisition protocols. We took Quality Control Measures to eliminate potential inter-scanner variability: Weekly phantom tests ensured consistent performance (CTDlvol variations\u0026thinsp;\u0026lt;\u0026thinsp;5%). A dedicated radiologist reviewed all scans for protocol compliance.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cp\u003eThe reconstructed images were transported to the Quantitative Evaluation System of CT (YT-CT-Lung, YITU Healthcare Technology Co., Ltd., China) for quantitative analysis. Histogram-based measurements refer to Hounsfield Units (HU, also referred as CT-numbers or CT-values) frequency distribution (i.e., physical density distribution). Therefore, information on air levels in specific lung regions can be derived from lung density histograms. Quantitative measurements were obtained from the low-dose chest CT images. The average parenchyma density on CT, defined as the mode of the histogram of CT pixel values in the whole lung. The entire lung volume is obtained by adding the pixels within pre-defined CT attenuation limits in volumes of each lung tissue. The total lung volume of both lung (tissue and airspace) was measured by summing the voxel dimensions in each slice. Quantitative parameters were collected including Mean, Median, Mode, Std, Skewness, Hellinger, IoU, Volume.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eAll analyses were performed using R software. Statistical analysis was performed with SPSS 23.0 and R software (version 4.0.0 R Statistical Computing Foundation, Vienna, Austria), \u003cem\u003ep\u0026thinsp;\u0026lt;\u003c/em\u003e\u0026thinsp;0.05 indicated a statistically significant difference. The distribution of continuous variables was checked initially. Normally distributed variables were expressed as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation, while non-normally distributed variables were reported as median with interquartile range. Categorical variables were presented as counts and percentages. The outlier values of lung density and lung volume rule out by method \u003cem\u003econtour boxplots\u003c/em\u003e with SPSS[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e] [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], boxplots use the interquartile method with fences to find outliers, this study take the interquartile range (IQR, the middle 50% of dataset), several quartile values and an adjustment factor to calculate boundaries for what constitutes minor and major outliers. Major outliers are 3 times greater than the IQR of all remaining values. Our study had eliminated the major outliers. To assess the associations between metabolic parameters and quantitative lung CT metrics, Spearman\u0026rsquo;s rank correlation coefficient (ρ) was employed. Given the potential non-normal distribution and non-linear nature of the relationships under investigation, Spearman\u0026rsquo;s correlation was selected over Pearson\u0026rsquo;s correlation. All reported correlation results include the ρ value, the corresponding two-sided \u003cem\u003ep\u003c/em\u003e-value, and the false discovery rate (FDR)-adjusted \u003cem\u003eq\u003c/em\u003e-value to account for multiple testing and ensure the statistical significance and reliability of the results. q values below 0.05 considered statistically significant.\u003c/p\u003e \u003cp\u003eCovariates were chosen prior to analysis based on their clinical relevance and previous research. Maternal age, pre-pregnancy body mass index, increased body mass index, gestational diabetes mellitus, HDP and other complications were included as potential confounders, as they are associated with metabolic status and systemic or pulmonary changes during pregnancy. Multivariable linear regression models were created to assess the independent associations between metabolic parameters and CT metrics. Not all pregnancy complications were included in the primary models to prevent overadjustment and unstable estimates. The study evaluated additional pregnancy complications in sensitivity analyses.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eDemographic and epidemiologic characteristics and Clinical findings\u003c/h2\u003e \u003cp\u003eOur study retrospectively recruited 1122 patients of two clinical centers (Hubei Provincial Maternal and Child Health Hospital, Zhongnan Hospital of Wuhan University) from February 1, 2020, to April 30 2020. The entire patients were admitted to hospital for antenatal care or delivery. 133 patients which diagnosed with pulmonary diseases were eliminated from our study. 91 patients were excluded from the study due to factors such as multiple pregnancies, preterm pregnancies, thorax trauma or surgical history. Among them, 898 patients were enrolled into our study (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), main data were shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Among the recruited patients, the mean age of all the participants was 30.39\u0026thinsp;\u0026plusmn;\u0026thinsp;4.14 years, and the mean gestational age was 38.57\u0026thinsp;\u0026plusmn;\u0026thinsp;1.13 weeks. For physical measurements, mean pre-pregnancy body mass index (BMI) was 23.46\u0026thinsp;\u0026plusmn;\u0026thinsp;3.93 kg/m\u0026sup2;, and increased BMI (ΔBMI) was 4.32\u0026thinsp;\u0026plusmn;\u0026thinsp;1.55 kg/m\u0026sup2;. In terms of metabolic parameters, the mean concentrations of triglyceride (TG), total cholesterol (TC) and fasting plasma glucose (FPG) were 5.80\u0026thinsp;\u0026plusmn;\u0026thinsp;3.57 mmol/L, 6.01\u0026thinsp;\u0026plusmn;\u0026thinsp;1.52 mmol/L and 4.94\u0026thinsp;\u0026plusmn;\u0026thinsp;0.65 mmol/L, respectively. The TyG index and TyGBMI index were 2.95\u0026thinsp;\u0026plusmn;\u0026thinsp;0.63 and 82.00\u0026thinsp;\u0026plusmn;\u0026thinsp;21.83, respectively.\u003c/p\u003e \u003cp\u003eThere were 180 (20.04%) had gestational diabetes, 127 (14.25%) patients had gestational HDP, 117 (13.03%) had anemia, 79 (8.80%) had endocrine diseases, such as thyroid, pituitary problems, etc. Other conditions were 521 (58.02%), including uncomplicated pregnancies and other obstetric disorders as premature rupture of membrane (79, 8.80%), scarred uterus pregnant (154, 17.15%), placental-related obstetrics complications (41, 4.57%), fetal-related diseases (91, 10.13%), etc. Some participants may have experienced multiple complications.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline characteristics of the study population\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cb\u003eDemographic characteristics\u003c/b\u003e\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAll participants (n\u0026thinsp;=\u0026thinsp;898)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge, years\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30.39\u0026thinsp;\u0026plusmn;\u0026thinsp;4.14\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGestational weeks\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38.57\u0026thinsp;\u0026plusmn;\u0026thinsp;1.13\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAnthropometric measures\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePre-pregnancy BMI, kg/m\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e23.46\u0026thinsp;\u0026plusmn;\u0026thinsp;3.93\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eΔBMI, kg/m\u0026sup2;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.32\u0026thinsp;\u0026plusmn;\u0026thinsp;1.55\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMetabolic parameters\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTriglycerides (TG), mmol/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5.80\u0026thinsp;\u0026plusmn;\u0026thinsp;3.57\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal cholesterol (TC), mmol/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e6.01\u0026thinsp;\u0026plusmn;\u0026thinsp;1.52\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFasting plasma glucose (FPG), mmol/L\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.94\u0026thinsp;\u0026plusmn;\u0026thinsp;0.65\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTyG inex\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2.95\u0026thinsp;\u0026plusmn;\u0026thinsp;0.63\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTyGBMI index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e82.00\u0026thinsp;\u0026plusmn;\u0026thinsp;21.83\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePregnancy complications (%)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGDM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e180 (20.04)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127 (14.25%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eanemia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e117 (13.03%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEndocrine disease\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79 (8.80%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOthers\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e521 (58.02%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eContinuous variables are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD or median (IQR). Categorical variables are presented as n (%). Pregnancy complications were not mutually exclusive; some participants experienced more than one complication.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eSpearman Correlation Analysis of Metabolic Parameters and Quantitative Lung CT Texture Features\u003c/h3\u003e\n\u003cp\u003eSpearman correlation analysis demonstrated statistically significant associations between metabolic parameters and quantitative lung CT texture features. Specifically, all assessed metabolic markers, including triglycerides (TG), total cholesterol (TC), fasting plasma glucose (FPG), the triglyceride-glucose (TyG) index, and the TyG-BMI index, exhibited significant correlations with multiple quantitative CT-derived parameters, namely Mean, Std and Skewness (all \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; false discovery rate\u0026ndash;adjusted \u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cb\u003eTable \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e)\u003c/p\u003e \u003cp\u003ePrecisely, Mean, Std and Skewness of TG and lung CT showed a significant positive correlation (ρ\u0026thinsp;=\u0026thinsp;0.175, 0.618, -0.395), respectively, among which Std had the strongest correlation with TG. TyG was also significantly correlated with Mean, Std and Skewness (ρ\u0026thinsp;=\u0026thinsp;0.197, 0.639, -0.419), especially with Std and Skewness. TyGBMI was also significantly correlated with all CT indices (including Mean, Std and Skewness), especially Std and Skewness (ρ\u0026thinsp;=\u0026thinsp;0.235, 0.586, -0.423). In addition, TC was negatively correlated with Mean and Std (ρ=-0.130, -0.447), while FPG was positively correlated with Mean and Std (ρ\u0026thinsp;=\u0026thinsp;0.171, 0.325). The statistical significance of all results was FDR corrected by the Benjamini-Hochberg method with \u003cem\u003ep\u003c/em\u003e and \u003cem\u003eq\u003c/em\u003e values below 0.05.\u003c/p\u003e \u003cp\u003eTo visually demonstrate visualize the correlation between metabolic parameters and quantitative lung CT features, we used heat maps to present the results of Spearman correlation analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cb\u003eFigure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). The color depth of the heat map reflects the strength of the correlation, where dark colors indicate strong correlation and light colors indicate weak or no correlation. The correlation between TG, TyG, and TyGBMI with CT standard deviation (Std) and Skewness (Skewness) is significant and strong (ρ values close to 0.6 or higher) as can be clearly seen from the heat map, where the relationship between these variables is presented in dark colors. At the same time, other metabolic measures such as TC and FPG showed relatively weak correlations with CT imaging, and in particular, TyG and TyGBMI showed consistent strong positive correlations with multiple CT measures, underlining their possible important role in the relationship between metabolic health and lung structure.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eSpearman correlations between maternal metabolic parameters and quantitative CT metrics.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExposure\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eStd\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSkewness\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.175[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.618[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.395[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.130[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.447[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.282[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFPG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.171[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.325[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.257[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTyG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.197[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.639[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.419[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTyGBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.235[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.586[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.423[\u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eValues are Spearman\u0026rsquo;s ρ. \u003cem\u003eq\u003c/em\u003e values were adjusted using Benjamini-Hochberg false discovery rate (FDR). * \u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, ** \u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01, *** \u003cem\u003eq\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eMultiple linear regression analysis of metabolic parameters and CT indexes\u003c/h3\u003e\n\u003cp\u003eMultiple linear regression analysis was performed using the forced entry method with Mean, Std and Skewness of lung CT texture parameters as dependent variables and five metabolic indexes (TG, TyG, TyGBMI, TC and FPG) screened by Spearman correlation analysis as independent variables, and the results are shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eThe results of the Mean analysis showed that TyGBMI (β\u0026thinsp;=\u0026thinsp;0.39, 95% CI: 0.11\u0026ndash;0.68, P\u0026thinsp;=\u0026thinsp;0.007) and FPG (β\u0026thinsp;=\u0026thinsp;9.44, 95% CI: 2.72\u0026ndash;16.17, P\u0026thinsp;=\u0026thinsp;0.006) were independently and positively correlated with Mean; however, the associations between TG, TyG, TC and Mean were not statistically significant (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eThe results of Std analysis showed that TG (β\u0026thinsp;=\u0026thinsp;6.23, 95%CI: 4.46\u0026ndash;8.00, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and FPG (β\u0026thinsp;=\u0026thinsp;8.18, 95%CI: 4.81\u0026ndash;11.55, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) were independently positively correlated with Std. TC was independently negatively associated with Std (β=-3.70, 95%CI: -4.94\u0026ndash;2.46, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). There was no significant association between TyG, TyGBMI and Std (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). (\u003cb\u003eTable \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e\u003c/b\u003e)\u003c/p\u003e \u003cp\u003eSkewness analysis showed that TyGBMI (β=-0.006, 95%CI: -0.009\u0026ndash;0.002, P\u0026thinsp;=\u0026thinsp;0.002) and FPG (β=-0.13, 95%CI: -0.22\u0026ndash;0.05, P\u0026thinsp;=\u0026thinsp;0.002) were independently negatively correlated with Skewness. The association between TG and Skewness was marginal significant (β=-0.04, 95%CI: -0.09\u0026ndash;0.00, P\u0026thinsp;=\u0026thinsp;0.062). There was no significant correlation between TyG, TC and Skewness (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the results of multiple linear regression analysis of the association of metabolic parameters with CT metrics (Mean, Std, Skewness). In the figure, square blocks (■) represent the unstandardized regression coefficient (β) of each variable, and the error line represents the 95% confidence interval (95% CI). The vertical dashed line represents the null line (β\u0026thinsp;=\u0026thinsp;0), and if the error line intersects the dashed line, the association is not statistically significant.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMultivariable Regression Analysis of Metabolic Factors and lung CT Parameters\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eStandardize\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean\u0026nbsp;(HU)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTyGBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.39 (0.11, 0.68)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.21\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFPG (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9.44 (2.72, 16.17)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.24\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStd\u0026nbsp;(HU)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTG (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6.23 (4.46, 8.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.35\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFPG (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.18 (4.81, 11.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.28\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTC (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-3.70 (-4.94, -2.46)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.29\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSkewness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTyGBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.006 (-0.009, -0.002)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.22\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFPG (mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.13 (-0.22, -0.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.23\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTG \u0026dagger;(mmol/L)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.04 (-0.09, 0.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.062\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e-0.15\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e*\u0026dagger; Marginally significant (P\u0026thinsp;=\u0026thinsp;0.062)*\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eβ: unstandardized coefficient; CI: confidence interval; Standardized β: standardized coefficient.\u003c/p\u003e \u003cp\u003eAll models were adjusted simultaneously for all variables shown in the table.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity Analysis of the LASSO Regression Model\u003c/h2\u003e \u003cp\u003eTo evaluate the robustness of the model and the potential risk of overfitting, LASSO regression (10-fold cross-validation) was further used for variable selection on the same set of independent variables. The results showed that under the optimal penalty parameter (λ), the non-zero coefficient variables screened by LASSO model were exactly the same as those included in the main analysis (multiple linear regression), which were TG, FPG, TC and TyGBMI. This result indicates that the variable selection strategy of this study is robust and there is no significant overfitting of the model. To assess the robustness of the model and the risk of overfitting, LASSO regression (10-fold cross-validation) was used to select variables from 22 candidate variables.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA shows the cross-validation error curve, showing an optimal λ of 0.16 (λ.min). Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB shows the coefficient path plot, showing that TG, FPG, and TC retain nonzero coefficients at λ.min. Figure\u0026nbsp;5 shows the selected variables under the λ.min criterion, and the direction of their coefficients is exactly the same as in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eIn conclusion, the variables selected automatically by LASSO were completely consistent with our main model-TG, FPG, and TC were the most robust metabolic correlates of lung CT texture. TyGBMI was compressed under the more stringent criteria, consistent with its properties as a composite measure.\u003c/p\u003e \u003cp\u003eIn addition, variance inflation factor (VIF) was used to diagnose collinearity in each multiple linear regression model (\u003cb\u003eTable \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e\u003c/b\u003e). The results showed that the VIF values of all independent variables ranged from 1.24 to 3.13, all well below the alert threshold of 10, and all below the strict criterion of 5. FPG (VIF\u0026thinsp;=\u0026thinsp;1.24) and TC (VIF\u0026thinsp;=\u0026thinsp;1.27) were almost completely independent of other variables. TG (VIF\u0026thinsp;=\u0026thinsp;3.06) and TyGBMI (VIF\u0026thinsp;=\u0026thinsp;3.13) were moderately correlated, but still within the acceptable range. The above results indicate that the regression model in this study does not suffer from serious multicollinearity problems and the parameter estimates are stable and reliable.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA Ten-fold cross-validation error curve as a function of log(λ). The left vertical dashed line indicates λ.min (0.15, logλ=-1.86), the value minimizing mean-squared error. The right vertical dashed line indicates λ.1se (34.28, logλ\u0026thinsp;=\u0026thinsp;3.53), the largest λ within one standard error of the minimum.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure 5\u003c/b\u003e Bar plot showing the coefficients of variables selected by LASSO at λ.min\u0026thinsp;=\u0026thinsp;0.155. Positive coefficients (red) indicate positive associations, while negative coefficient (blue) indicates negative association. The selected variables (TG, FPG, TC, TyGBMI) and their coefficient directions are identical to those in the primary multivariable regression models (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study systematically evaluated the association between maternal metabolic parameters and CT texture features of lung parenchyma based on a large sample of chest CT quantitative parameters during pregnancy. The main findings were as follows: First, triglyceride (TG) and fasting plasma glucose (FPG) were independent positive predictors of lung CT standard deviation (Std), suggesting that disorders of glucose and lipid metabolism are closely related to increased lung tissue density heterogeneity; Second, total cholesterol (TC) was independently and inversely associated with Std, suggesting its possible protective role in the maintenance of lung microstructure during pregnancy. Third, TyG‑BMI was significantly and independently associated with both CT Mean and Skewness, and the effect was more robust at the composite level than that of single metabolic parameter. These findings suggest that maternal metabolic status has an independent and multidimensional effect on imaging phenotypes of lung during pregnancy.\u003c/p\u003e \u003cp\u003ePrevious studies on the relationship between metabolism and lung structure during pregnancy mostly focused on clinical outcomes such as asthma and sleep apnea, and few used imaging methods to directly evaluate the changes of lung parenchyma[\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn our study, TG had the highest standardized regression coefficient on Std (standardized coefficient β\u0026thinsp;=\u0026thinsp;0.35) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), suggesting that triglycerides had the strongest correlation with lung density heterogeneity, indicating that lipid metabolism disorders are closely related to lung microstructural heterogeneity. This consistent with the results of a cohort study based on a nonpregnant population, the incidence of metabolic syndrome is increased in COPD patients, and the TG levels of COPD patients are higher than those of non-pregnant healthy people[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. The reason may be that systemic inflammation plays an important role in both COPD and dyslipidemia, and inflammatory factors may break the balance of lipid[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Alterations in lipid molecules are associated with impaired lung function and inversely correlated with emphysema and interstitial lung abnormalities[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Lipid metabolism can initiate a variety of cellular processes through lysosomal inhibitors. Serum lysophospholipids are negatively correlated with pulmonary interstitial abnormalities, and the underlying mechanism may be related to lung function and systemic inflammatory markers[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. In COPD patients, lung CT texture features have been demonstrated to correlate with systemic inflammatory markers. Therefore, we speculate that the association of lung CT texture abnormalities with TG observed in this study may be partially mediated by a common 'systemic inflammatory state'. Future studies incorporating inflammatory markers into mediation analyses would be helpful to further test this hypothesis[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAnother related research confirmed that areas of structural heterogeneity on lung CT showed a significant positive correlation with glucose metabolic activity and pointed to the presence of inflammation in most areas of pulmonary heterogeneity. This finding reveals that structural alterations in the lung may be accompanied by abnormal activation of local glucose metabolism[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. We speculate that blood glucose may indirectly affect the imaging features of lung CT by driving the inflammatory response in the lung. It was consistent with our study in pregnancy women, FPG was independently and positively correlated with the standard deviation of lung CT (Std) (β\u0026thinsp;=\u0026thinsp;8.18, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), and Std is precisely an indicator reflecting the heterogeneity of lung density. Meanwhile, the independent positive correlation between FPG and lung CT Mean (Mean) (β\u0026thinsp;=\u0026thinsp;9.44, P\u0026thinsp;=\u0026thinsp;0.006) suggested that blood glucose may also affect the overall density of the lung. This may reflect diffuse infiltration of inflammatory cells, alveolar wall thickening, or altered interstitial components. Notably, the negative correlation between FPG and Skewness (β=-0.13, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002) further supported the change in distribution characteristics, with a decrease in left skewness indicating a shift in lung density distribution toward higher values, consistent with an increase in Mean. (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eThe negative association of TC with Std is another important finding of our study. Although a high TC level is considered a risk factor for cardiovascular disease[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e], in the special physiological state of pregnancy, cholesterol transporters involved in some pregnancy pathologies such as preeclampsia, gestational diabetes mellitus, and intrauterine growth retardation[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e], hypercholesterolemia participated in the regulation of cholesterol transport across the human placenta and lipoprotein profiles in maternal and neonatal circulation[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eDifferent form TG and FPG, TC showed a consistent negative correlation with other lung CT parameters: higher TC was associated with lower Mean (Mean) and standard deviation (Std) and higher Skewness (Skewness) in lung CT. The direction of association between TC and lung CT was opposite to that of other CT parameters (seen in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), suggesting that cholesterol may affect lung tissue by different biological mechanisms.\u003c/p\u003e \u003cp\u003echolesterol is a key substrate for placental synthesis of progesterone and fetal pulmonary surfactant. Studies have shown that moderately elevated maternal cholesterol levels during pregnancy are positively correlated with neonatal lung development maturity. Therefore, the negative association between TC and lung density heterogeneity observed in this study may reflect its potential protective role in maintaining lung tissue homogeneity and structural integrity. Further longitudinal studies and in vitro experiments are needed to verify this hypothesis.\u003c/p\u003e \u003cp\u003eIn the quantitative analysis of lung CT, BMI is a variable that cannot be ignored. BMI is not only a metabolic index, but also closely related to the image acquisition and interpretation of lung CT. Image noise was significantly increased in the high BMI group (average 39.8 kg/m\u0026sup2;)[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. A pioneering study combined PET/CT radiomics features with glucose metabolism-related genes to construct a prognostic model of non-small cell lung cancer, and found that imaging features could effectively reflect the metabolic remodeling of lung tumors[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTyG index and its composite index of BMI (TyGBMI) reflect insulin resistance and multiple risk predictors of chronic diseases[\u003cspan additionalcitationids=\"CR41\" citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e], TyG and TyGBMI generally have stronger correlations with lung CT parameters than single metabolic index, suggesting that the coupling of glucose and lipid metabolism may have a more significant effect on lung structure (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). This suggests that insulin resistance may play a central role in the lung-metabolic axis.\u003c/p\u003e \u003cp\u003eNotably, although TyG index was strongly correlated with all three CT texture parameters in univariate analysis, its independent effects disappeared after adjusting for TG, FPG and BMI. However, TyG‑BMI index still maintained significant predictive power for Mean and Skewness in multiple regression. (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) This finding suggests that the overall burden of the metabolic syndrome may be the central factor driving phenotypic changes on lung imaging during pregnancy, rather than insulin resistance or elevated lipids alone. TyG‑BMI can better reflect the true metabolic status than single metabolic parameter. This finding is consistent with the findings of recent studies in the assessment fo insulin resistance for GDM, it supports the use of composite metabolic index in research of metabolic-imaging during pregnancy[\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eOur study has several limitations. First, as a cross-sectional study, mechanistic hypothesis has been proposed, but the direction of causality still needs to be verified by prospective cohort studies. Second, although we adjusted for multiple potential confounders, unmeasured residual confounding remains possible (e.g., dietary intake, physical activity, environmental exposure, etc.). Thirdly, our study included only pregnant women, and extrapolation to the nonpregnant population should be cautious. Forth, some metabolic measures (e.g., insulin, glycated hemoglobin) were not included in the analyses, limiting the depth of mechanistic exploration. Despite the limitations mentioned above, this study revealed for the first time the association between metabolic indicators and lung CT parameters in the pregnant population through strict FDR correction and multivariate regression analysis, which provided new clues and directions for the study of the lung-metabolic axis.\u003c/p\u003e \u003cp\u003eIn conclusion, this study first report the independent associations between maternal TG, FPG, TC, and TyGBMI indices and lung CT texture parameters during pregnancy. TG and FPG were the main independent metabolic risk factors for lung heterogeneity. TC was negatively correlated with lung heterogeneity independently, which was opposite to other metabolic indicators, suggesting that TC may affect lung structure through different biological mechanisms. TyGBMI is a robust marker to reflect the effect of metabolic burden on lung imaging phenotype, which correlated with multiple parameters of lung CT. This study provides novel imaging evidence linking metabolic abnormalities to structural changes in the lungs during pregnancy. These findings offer quantifiable indicators for the early identification and intervention of high-risk populations. Prospective cohort studies and mechanistic experiments are needed to further clarify the biological pathways.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eCT Computed tomography\u003c/p\u003e\u003cp\u003eStd Standard deviation\u003c/p\u003e\u003cp\u003eGDM Gestational diabetes mellitus\u003c/p\u003e\u003cp\u003eHDP Hypertensive disorders in pregnancy\u003c/p\u003e\u003cp\u003eTG Triglycerides\u003c/p\u003e\u003cp\u003eTC Total cholesterol\u003c/p\u003e\u003cp\u003eFPG Fasting plasma glucose\u003c/p\u003e\u003cp\u003eBMI Body mass index\u003c/p\u003e\u003cp\u003eVIF Variance inflation factor\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthic statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Institutional Ethics Committees of Zhongnan Hospital of Wuhan University and the Maternal (Approval No. 2020072K). This study is a retrospective analysis of non-identifiable researcher-collected data. The requirement for informed consent from the study subjects was waived by the IRB of Ethical Approval of Ethics Committee of Zhongnan Hospital of Wuhan University \u0026nbsp;due to the retrospective study design. We confirm that: The research does not involve any human experimentation; No identifiable subject data or biological materials were utilized; No personal privacy concerns or commercial interests are implicated in this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ethe following authors contributed equally and should be considered co-first authors: Luman Li, Kun Yang, Ruo-xi Ran, and Jie Chen. All authors participated in the multicenter and multidisciplinary collaboration，contributed equally to this article and should be considered as first authors with specialized roles in Obstetrics, Clinical laboratory, Radiography, respectively\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLuman Li\u003c/strong\u003e (Affiliations 1, 2, 3): conceptualization, study design, manuscript drafting, data interpretation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKun Yang\u003c/strong\u003e (Department of Pharmacy, Zhongnan Hospital of Wuhan University): sample collection, statistical analysis, data acquisition\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRuo-xi Ran\u003c/strong\u003e (Department of Clinical Laboratory Medicine, Maternal and Child Health Hospital of Hubei Province): laboratory testing, quality control, data analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJie Chen\u003c/strong\u003e (Department of Nuclear Medicine, Zhongnan Hospital of Wuhan University): imaging experiments, data interpretation\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eYating Li\u003c/strong\u003e (Department of Obstetrics, Zhongnan Hospital of Wuhan University), data collection, data analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHui Xie\u003c/strong\u003e (Department of Radiography, Maternal and Child Health Hospital of Hubei Province): statistical analysis, imaging experiments, methodology support.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMeiyan Liao\u003c/strong\u003e (Department of Nuclear Medicine, Zhongnan Hospital of Wuhan University): experimental support, data validation\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHuijun Chen\u003c/strong\u003e (Department of Obstetrics, Zhongnan Hospital of Wuhan University): supervision, clinical guidance, critical revision of manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eYuanzhen Zhang\u003c/strong\u003e (Affiliations 1, 2, 3): overall study supervision, funding acquisition, main corresponding author, manuscript final approval.\u003c/p\u003e\n\u003cp\u003eAll authors who contributed to the manuscript revision, read, and approved the submitted version.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFUNDING\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by Hubei Science and Technology Plan (grant number 2020FCA011), Wuhan Technology and Innovation Plan (grant number 2020020201010010), Zhongnan Hospital Major Project (grant number ZNJC202503)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no competing interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank all the patients who have participated in this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are not publicly available due to privacy or ethical restrictions. Data are available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eZhou TH, et al. CT whole lung radiomic nomogram: a potential biomarker for lung function evaluation and identification of COPD. Mil Med Res. 2024;11(1):14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMayerhoefer ME, et al. Introduction to Radiomics. J Nucl Med. 2020;61(4):488\u0026ndash;95.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoi H, et al. Extension of Coronavirus Disease 2019 on Chest CT and Implications for Chest Radiographic Interpretation. Radiol Cardiothorac Imaging. 2020;2(2):e200107.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAerts HJ, et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun. 2014;5:4006.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZwanenburg A, et al. The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping. Radiology. 2020;295(2):328\u0026ndash;38.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark BW, et al. Reliability of CT radiomic features reflecting tumour heterogeneity according to image quality and image processing parameters. Sci Rep. 2020;10(1):3852.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurrell RA, et al. The causes and consequences of genetic heterogeneity in cancer evolution. Nature. 2013;501(7467):338\u0026ndash;45.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu N, et al. Acute necrotising pancreatitis: measurements of necrosis volume and mean CT attenuation help early prediction of organ failure and need for intervention. Eur Radiol. 2021;31(10):7705\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoezuka S, et al. Combination of mean CT value and maximum CT value as a novel predictor of lepidic predominant lesions in small lung adenocarcinoma presenting as solid nodules. Sci Rep. 2022;12(1):5450.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMokkala K, et al. Metagenomics analysis of gut microbiota in response to diet intervention and gestational diabetes in overweight and obese women: a randomised, double-blind, placebo-controlled clinical trial. Gut. 2021;70(2):309\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChakraborty S, et al. Metabolites and Hypertension: Insights into Hypertension as a Metabolic Disorder: 2019 Harriet Dustan Award. Hypertension. 2020;75(6):1386\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlag K, et al. Dietary fat disrupts a commensal-host lipid network that promotes metabolic health. Cell Metab. 2026;38(1):157\u0026ndash;e1739.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi L, et al. Establishment and validation of a predictive nomogram for gestational diabetes mellitus during early pregnancy term: A retrospective study. Front Endocrinol (Lausanne). 2023;14:1087994.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhan SS, et al. Body Mass Index, Adverse Pregnancy Outcomes, and Cardiovascular Disease Risk. Circ Res. 2023;133(9):725\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFreaney PM, et al. The Role of Sex-Specific Risk Factors in the Risk Assessment of Atherosclerotic Cardiovascular Disease for Primary Prevention in Women. Curr Atheroscler Rep. 2020;22(9):46.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe Q et al. Association between the cumulative atherogenic index of plasma and cardiometabolic multimorbidity: the mediating effects of the TyG index and body mass index. Lipids Health Dis, 2026. 25(1).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWarner JD, et al. Correlation of HbA1c levels with CT-based body composition biomarkers in diabetes mellitus and metabolic syndrome. Sci Rep. 2024;14(1):21875.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCzernin J, Benz MR, Allen-Auerbach MS. PET/CT imaging: The incremental value of assessing the glucose metabolic phenotype and the structure of cancers in a single examination. Eur J Radiol. 2010;73(3):470\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimmons D, et al. Regression From Early GDM to Normal Glucose Tolerance and Adverse Pregnancy Outcomes in the Treatment of Booking Gestational Diabetes Mellitus Study. Diabetes Care. 2024;47(12):2079\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosenberg EA, Seely EW. Update on Preeclampsia and Hypertensive Disorders of Pregnancy. Endocrinol Metab Clin North Am. 2024;53(3):377\u0026ndash;89.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeede HJ, et al. Association of Antenatal Diet and Physical Activity-Based Interventions With Gestational Weight Gain and Pregnancy Outcomes: A Systematic Review and Meta-analysis. JAMA Intern Med. 2022;182(2):106\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGirchenko P, et al. Associations of polymetabolic risk of high maternal pre-pregnancy body mass index with pregnancy complications, birth outcomes, and early childhood neurodevelopment: findings from two pregnancy cohorts. BMC Pregnancy Childbirth. 2024;24(1):78.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShiffler RE. Maximum Z Scores and Outliers. Am Stat. 1988;42(1):79\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhitaker RT, Mirzargar M, Kirby RM. Contour boxplots: a method for characterizing uncertainty in feature sets from simulation ensembles. IEEE Trans Vis Comput Graph. 2013;19(12):2713\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDenton E, et al. Asthma Pregnancy Allergy. 2026;81(1):84\u0026ndash;108.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTasali E, et al. Obstructive Sleep Apnea and Cardiometabolic Disease: Obesity, Hypertension, and Diabetes. Circ Res. 2025;137(5):764\u0026ndash;87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTang WZ, et al. Obstructive sleep apnea-associated hypertensive disorders in pregnancy: a literature review and clinical management strategies. Reprod Biol Endocrinol. 2025;23(1):114.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXuan L, et al. Association between chronic obstructive pulmonary disease and serum lipid levels: a meta-analysis. Lipids Health Dis. 2018;17(1):263.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNovgorodtseva TP, et al. Modification of the fatty acid composition of the erythrocyte membrane in patients with chronic respiratory diseases. Lipids Health Dis. 2013;12:117.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEsteve E, Ricart W, Fern\u0026aacute;ndez-Real JM. Dyslipidemia and inflammation: an evolutionary conserved mechanism. Clin Nutr. 2005;24(1):16\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Q, et al. Lysophospholipid metabolism, clinical characteristics, and artificial intelligence-based quantitative assessments of chest CT in patients with stable COPD and healthy smokers. Sci Rep. 2025;15(1):26376.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeng Q, et al. CT-derived abdominal adiposity: Distributions and better predictive ability than BMI in a nationwide study of 59,429 adults in China. Metabolism. 2021;115:154456.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCressoni M, et al. Lung inhomogeneities, inflation and [18F]2-fluoro-2-deoxy-D-glucose uptake rate in acute respiratory distress syndrome. Eur Respir J. 2016;47(1):233\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang G, et al. LDL-C and TC Mediate the Risk of PNPLA3 Inhibition in Cardiovascular Diseases. J Clin Endocrinol Metab. 2025;110(2):e231\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKallol S, Albrecht C. Materno-fetal cholesterol transport during pregnancy. Biochem Soc Trans. 2020;48(3):775\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCantin C, Fuenzalida B, Leiva A. Maternal hypercholesterolemia during pregnancy: Potential modulation of cholesterol transport through the human placenta and lipoprotein profile in maternal and neonatal circulation. Placenta. 2020;94:26\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSourlos N, et al. Does BMI influence AI and human reader lung nodule detection in low-dose chest CT? Eur J Radiol. 2025;193:112453.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang C, et al. Development and validation of a radiogenomics prognostic model integrating PET/CT radiomics and glucose metabolism-related gene signatures for non-small cell lung cancer. Eur J Nucl Med Mol Imaging. 2025;52(13):4924\u0026ndash;38.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang L, et al. Assessment of six insulin resistance surrogate indexes for predicting stroke incidence in Chinese middle-aged and elderly populations with abnormal glucose metabolism: a nationwide prospective cohort study. Cardiovasc Diabetol. 2025;24(1):56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu H, et al. Triglyceride-glucose index correlates with the occurrence and prognosis of acute myocardial infarction complicated by cardiogenic shock: data from two large cohorts. Cardiovasc Diabetol. 2024;23(1):337.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe D, et al. Prospective associations of triglyceride-glucose related indices with cardiovascular disease and mortality in individuals with metabolic syndrome: evidence from the UK biobank. Cardiovasc Diabetol. 2026;25(1):53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang L, et al. Assessment of first-trimester insulin resistance indices for gestational diabetes mellitus: a prospective cohort study. J Endocrinol Invest. 2025;48(9):2167\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiang X, et al. Association between triglyceride glucose-body mass index and gestational diabetes mellitus: a prospective cohort study. BMC Pregnancy Childbirth. 2025;25(1):170.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Computed tomography, CT texture features, Metabolic Parameters, pregnancy","lastPublishedDoi":"10.21203/rs.3.rs-9085820/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9085820/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjective\u003c/h2\u003e \u003cp\u003eTo investigate the association between maternal metabolic parameters and quantitative CT texture features of the lung during pregnancy.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eThis multi-center retrospective cross-sectional study included 898 pregnant women. Maternal metabolic parameters including triglycerides (TG), total cholesterol (TC), fasting plasma glucose (FPG), TyG index, TyGBMI index etc. Lung CT texture features were extracted by quantitative CT texture analysis including mean attenuation (Mean), standard deviation (Std), and skewness (Skewness), etc. Spearman correlation was used for preliminary variable selection. Multivariable linear regression was employed to evaluate independent associations between metabolic parameters and CT texture features. LASSO regression and variance inflation factor (VIF) were used to assess model robustness and multicollinearity.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eMultivariable linear regression revealed that TG (β\u0026thinsp;=\u0026thinsp;6.23, 95%CI: 4.46\u0026ndash;8.00, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and FPG (β\u0026thinsp;=\u0026thinsp;8.18, 95%CI: 4.81\u0026ndash;11.55, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) were independently positively associated with Std, while TC was independently negatively associated with Std (β=-3.70, 95%CI: -4.94\u0026ndash;-2.46, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). TyGBMI index showed independent associations with both Mean (β\u0026thinsp;=\u0026thinsp;0.39, 95%CI: 0.11\u0026ndash;0.68, P\u0026thinsp;=\u0026thinsp;0.007) and Skewness (β=-0.006, 95%CI: -0.009\u0026ndash;-0.002, P\u0026thinsp;=\u0026thinsp;0.002). LASSO cross-validation selected identical variables as the primary models, and all VIF values were below 5, indicating robust models without severe multicollinearity.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003ethis study is the first to systematically evaluate the association between maternal metabolic parameters and lung CT texture features. TG and FPG were independent risk factors for lung heterogeneity, TC was negatively correlated with heterogeneity, and TyGBMI, as a composite metabolic index, had a robust predictive value for mean and skewness.This study provides new evidence for understanding the pulmonary imaging manifestations of metabolic disorders during pregnancy.\u003c/p\u003e","manuscriptTitle":"Association Between Maternal Metabolic Parameters and Quantitative Lung CT Texture Features in Pregnancy: A Retrospective Cross-Sectional Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-19 08:05:06","doi":"10.21203/rs.3.rs-9085820/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-04-08T17:14:26+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-06T11:38:58+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-18T11:53:46+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-17T21:45:21+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Pregnancy and Childbirth","date":"2026-03-17T15:06:00+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"785aaf6b-3971-42ed-8233-f02127dfd871","owner":[],"postedDate":"April 19th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-19T08:05:06+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-19 08:05:06","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9085820","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9085820","identity":"rs-9085820","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}