Quantitative comparative analysis of different CT modalities in systemic sclerosis-associated interstitial lung disease: correlations with lung function, serological biomarkers, and disease stratification

Introduction

Systemic sclerosis-associated interstitial lung disease (SSc-ILD) is the most severe and life-threatening complication of systemic sclerosis (SSc), affecting 40–80% of patients with SSc and serving as the leading cause of SSc-related mortality as a result of progressive pulmonary fibrosis (1,2). Pathologically, early-stage SSc-ILD is characterized by subclinical alveolitis and abnormal collagen deposition, which progresses to irreversible lung structural damage, impaired ventilation, and reduced gas exchange (3). Clinically, diagnosis relies on high-resolution computed tomography (HRCT), the gold standard for visualizing characteristic lesions, together with pulmonary function test (PFT) for disease staging and prognosis (4-6).

However, HRCT has limitations in quantifying fibrosis extent and evaluating disease activity or treatment response (7). Given the wide variability in SSc-ILD prognosis, more sensitive imaging strategies based on quantitative CT parameters are needed to enable earlier and more precise intervention (8). Novel imaging modalities like dual-energy CT (DECT) and virtual non-contrast CT (VNC CT) show promise. DECT can generates virtual monochromatic images (e.g., 70 keV) which reduce beam hardening artifacts, enhance tissue contrast, and improve the detection (9-12); VNC CT uses material decomposition techniques to simulate unenhanced images (13-15). Although their reliability in regions with coexisting fibrosis and inflammation remains to be validated.

This study systematically compared attenuation, standard deviation (SD), skewness, and kurtosis across HRCT, MonoE 70 keV CT, and VNC CT, integrating imaging, pulmonary function, and clinical data. It explored correlations between these CT parameters and lung function and clinical characteristics in SSc-ILD to provide a basis for more precise diagnosis and treatment. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2278/rc).

Methods

Study participants

This retrospective cross-sectional study included consecutive systemic sclerosis (SSc) patients admitted to the Department of Rheumatology and Immunology, West China Hospital, Sichuan University, from 2022 to 2024. A total of 43 patients were enrolled in this study. All met the 2013 ACR/EULAR SSc classification criteria (16), and SSc-ILD was confirmed by HRCT findings of interstitial fibrosis (17). Exclusion criteria: age <18 years; history of cancer/infectious disease; inability to complete CT or PFTs due to severe comorbidities. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of West China Hospital (2022 approval No. 341), and individual informed consent was waived due to the retrospective nature of the study.

Collection of clinical features

Clinical data (demographics, symptoms, laboratory results) at first hospitalization were retrieved from electronic medical records. Hematology indices [hemoglobin, white blood cell (WBC), lymphocyte, monocyte, eosinophil, basophil, platelet], autoantibody titers [anti-Scl-70, anti-antinuclear antibody (ANA), anti-Sjögren’s syndrome A antibody (SSA); categorized as negative =0, suspicious =0.5, 1:100–1:320 =1, 1:320–1:1,000 =2, 1:1,000–1:3,200 =3], and hepatic/renal function markers [alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), alkaline phosphatase (ALP), albumin, creatinine, uric acid (UA)] were extracted.

Image processing and quantification rules

Quantitative image analysis was performed using a standardized protocol. All images were analyzed on an IntelliSpace Portal workstation (Philips Healthcare) using identical reconstruction parameters. Regions of lobe measurements were independently performed by two experienced radiologists who were blinded to the clinical and pulmonary function data. For each parameter, measurements from the two readers were averaged for subsequent analysis. To minimize measurement variability, all radiological parameters were extracted using predefined settings and consistent segmentation rules across all CT modalities.

Pulmonary function testing

All patients underwent PFTs within 1 week of CT using a calibrated spirometer in a temperature (20–25 ℃) and humidity (40–60%) controlled room. Patients were seated with a nose clip to ensure mouth-only breathing.

Cluster analysis methodology

K-means clustering was performed on a multidimensional matrix of CT parameters (attenuation, SD, skewness, and kurtosis) derived from HRCT, MonoE 70 keV CT, and VNC CT images. Data were Z-score standardized to eliminate dimensional bias. Euclidean distance was used to measure sample similarity, and patients were classified into two clusters (K=2) via iterative centroid optimization.

Statistical analysis

Pearson correlation was used for normally distributed CT-PFT parameter relationships, and Spearman correlation for non-normal distributions (significance: P<0.05; r and 95% CI reported). Inter-cluster comparisons used t-tests, analysis of variance (ANOVA), or Kruskal-Wallis tests as appropriate. Missing data were assessed for all study variables. The proportion of missing values was less than 5% for all variables. Missing data accounted for less than 5% of variables and were handled by listwise deletion in SPSS 26.0.

Results

Baseline demographics and clinical characteristics

This study included 43 patients with physician-diagnosed SSc-ILD. Baseline characteristics of the study cohort are summarized in Table 1. Notably, the cohort was predominantly female (97.7%, n=42), with a median age of 53 years (range, 19–76 years). Autoantibody positivity rates were 60.5% (n=26) for ANAs, 37.2% (n=16) for anti-Scl-70 antibodies, and 14.0% (n=6) for anti-Ro/SSA antibodies. PFTs demonstrated varying degrees of functional impairment. The mean forced vital capacity (FVC) was 91.6%±16.8% of the predicted value (range, 55.6–132.9%), with 20.9% (n=9) of patients exhibiting an FVC <80% of the predicted value. The mean hemoglobin-corrected diffusing capacity for carbon monoxide (DLco) was 70.5%±19.2% of the predicted value (range, 32.3–112.6%), with 51.2% (n=22) exhibiting a DLco <70% of the predicted value (Table S1). Some patients had missing DLco or VA data due to inadequate respiratory cooperation; however, the proportion of missing data was minimal (<5%). Cases with incomplete data were excluded on a case-by-case basis using SPSS.

Intergroup comparison of CT parameters via one-way ANOVA

To compare HRCT, MonoE 70 keV CT, and VNC CT in assessing SSc-ILD pulmonary lesions, 43 patients underwent these CT examinations. Due to 6-month reexaminations in 10 patients and 1-year follow-up examinations in 6 patients, a total of 57 imaging datasets were collected, and quantitative data were generated for each CT modality. Significant differences were observed across the three CT modalities (Table S2, Figure 1).

Figure 1 Comparison of CT parameters across lung lobes. (A) Attenuation values in HRCT, MonoE 70 keV CT and VNC CT. (B) Standard deviation, (C) kurtosis, and (D) skewness in the three CT modalities. HRCT, high-resolution computed tomography; LLL, left lower lobe; LUL, left upper lobe; MonoE 70 keV CT, monoenergetic 70 keV computed tomography; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; VNC CT, virtual non-contrast computed tomography.

Analysis of 57 CT scans from 43 patients revealed SD in the right upper lobe (RUL), right middle lobe (RML), left upper lobe (LUL) and left lower lobe (LLL) exhibited marked variations across four lung lobes (Figure 1B). For the RUL and LUL, SD values derived from HRCT were higher than those from MonoE 70 keV CT, which were, in turn, higher than those from VNC CT; VNC CT showed significantly lower SD values than both HRCT (P<0.001) and MonoE 70 keV CT (P=0.008 and P=0.004, respectively). For the RML and LLL, reductions in SD were significant only when compared with HRCT (P<0.01).

In addition, we explored the monitoring capabilities of different CT imaging modalities for pulmonary lesions in patients with SSc-ILD. Patients were stratified into groups based on the scan interval, including 6 months (n=10) and 1 year (n=6). However, under the same CT modality and analysis approach, no obvious temporal changes were observed in quantitative indicators before and after follow-up examinations (Figure S1).

Analysis of the correlation between CT parameters and pulmonary function

To clarify links between quantitative CT parameters and lung function in SSc-ILD, correlation analyses were performed. Quantitative parameters derived from all three modalities, including attenuation, SD, skewness, and kurtosis, showed significant correlations with multiple lung function indices (Figure 2).

Figure 2 Spearman correlations between parameters of three CT modalities and pulmonary function indicators. (A-C) Correlation heat maps for HRCT, MonoE 70 keV CT, and VNC CT, respectively. BF, breathing frequency; DLCO SB, single-breath diffusing capacity for carbon monoxide; ERV, expiratory reserve volume; FEV1, forced expiratory volume in 1 second; FRCpleth, functional residual capacity by plethysmography; FVC, forced vital capacity; HRCT, high-resolution computed tomography; IC, inspiratory capacity; LLL, left lower lobe; LUL, left upper lobe; MMEF, maximal mid-expiratory flow; MonoE 70 keV CT, monoenergetic 70 keV computed tomography; MV, minute ventilation; MVV, maximal voluntary ventilation; PEF, peak expiratory flow; R at 20Hz, respiratory resistance at 20 Hertz; R at 5Hz, respiratory resistance at 5 Hertz; Reff, effective respiratory resistance; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; RVTLC, residual volume of total lung capacity; SD, standard deviation; SGeff, effective specific airway conductance; TLC, total lung capacity; VC MAX, vital capacity max; VNC CT, virtual non-contrast computed tomography; VT, tidal volume; X at 5Hz, respiratory reactance at 5 Hertz; Z at 5Hz, total respiratory impedance at 5 Hertz.

In the assessment of ventilatory function, attenuation and SD exhibited negative correlations with FVC. The correlation strength of attenuation was greater than that of SD, with attenuation showing a moderate correlation (−0.41<R<−0.40, P<0.05), whereas SD showed a weak correlation (−0.21<R<−0.19, P<0.05). In contrast, skewness and kurtosis showed moderate positive correlations with FVC (0.45<R<0.48, P<0.05). Similarly, attenuation, SD, skewness, and kurtosis showed consistent correlations with lung volume, vital capacity max (VCmax), and total lung capacity (TLC). Compared with FVC, these correlations, with the exception of SD, were stronger and more significant (0.41<|R|<0.56, P<0.001).

Notably, certain CT parameters showed pronounced site-specific correlations with MMEF75/25 and DLco. Specifically, SD negatively correlated with MMEF75/25 exclusively in the LLL and RLL (−0.31<R<−0.28, P<0.05). In contrast, attenuation showed weak negative correlations with DLco, predominantly in the left lung (LUL and LLL; −0.39<R<−0.26, P<0.05). In these lobes, skewness and kurtosis exhibited moderate positive correlations with DLco, residual volume (RV), and functional residual capacity measured by FRCpleth (0.20<R<0.45, P<0.05).

To explore longitudinal associations between CT parameters and lung function in SSc-ILD, we analyzed relative changes in these parameters. Follow-up data (Figure S2) demonstrated that significant correlations were mainly observed in lung volume indices. Attenuation showed strong negative correlations with changes in FRCpleth (−0.98<R<−0.92, P<0.05) and, in the 6-month follow-up group, with changes in FEV1/FVC and MMEF (−0.93<R<−0.68, P<0.05).

Characterization of the CT cluster analysis results

To assess whether quantitative CT features could stratify patients with SSc-ILD, K-means clustering was performed using z-score-normalized attenuation, SD, skewness, and kurtosis derived from HRCT, MonoE 70 keV CT, and VNC CT images. Patients were classified into two clusters (K=2) with distinct quantitative CT profiles (Figure 3). The contribution of individual CT parameters to cluster separation was confirmed by one-way ANOVA, with detailed results summarized in Table S3.

Figure 3 Cluster analysis of three CT modalities based on respective parameters. (A-C) PCA plots and final cluster center distances for HRCT, MonoE 70 keV CT, and VNC CT, respectively. HRCT, high-resolution computed tomography; LLL, left lower lobe; LUL, left upper lobe; MonoE 70 keV CT, monoenergetic 70 keV computed tomography; PCA, principal component analysis; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; SD, standard deviation; VNC CT, virtual non-contrast computed tomography.

To further evaluate the clinical relevance of CT-based clustering, lung function and clinical indicators were compared between Cluster 1 and Cluster 2. HRCT-based clustering identified the largest number of pulmonary function indices with significant intergroup differences (n=9), followed by VNC CT (n=7), whereas MonoE 70 keV CT identified fewer differential indices (n=4) (Figure 4). Notably, despite identifying fewer indices, MonoE 70 keV CT uniquely identified differences in RV and functional residual capacity measured by FRCpleth. These four indices [TLC, alveolar volume (VA), residual volume (RV), and functional residual capacity by plethysmography (FRCpleth)] were all related to lung volume, highlighting the advantage of MonoE 70 keV CT in assessing volume alterations and pulmonary status in SSc-ILD.

Figure 4 Pulmonary function indicators across clusters in three CT modalities. (A-C) Heat maps and t-test bar charts for lung function in two clusters of HRCT, MonoE 70 keV CT, and VNC CT, respectively. *, P<0.05; **, P<0.01; ***, P<0.001. BF, breathing frequency; DLCO SB, single-breath diffusing capacity for carbon monoxide; ERV, expiratory reserve volume; FEV1, forced expiratory volume in 1 second; FRCpleth, functional residual capacity by plethysmography; FVC, forced vital capacity; HRCT, high-resolution computed tomography; IC, inspiratory capacity; LLL, left lower lobe; LUL, left upper lobe; MEF, maximal mid-expiratory flow; MMEF, maximal mid-expiratory flow; MonoE 70 keV CT, monoenergetic 70 keV computed tomography; MV, minute ventilation; MVV, maximal voluntary ventilation; PEF, peak expiratory flow; R at 20Hz, respiratory resistance at 20 Hertz; R at 5Hz, respiratory resistance at 5 Hertz; Reff, effective respiratory resistance; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; RV % TLC, residual volume percentage of total lung capacity; SD, standard deviation; SGeff, effective specific airway conductance; TLC, total lung capacity; VC MAX, vital capacity max; VNC CT, virtual non-contrast computed tomography; VT, tidal volume; X at 5Hz, respiratory reactance at 5 Hertz; Z at 5Hz, total respiratory impedance at 5 Hertz.

For the analysis of clinical indicators, we systematically evaluated autoantibodies, routine blood parameters, and indicators of liver and kidney function. Anti-Scl-70 titers were slightly higher in Cluster 1 than in Cluster 2 (Table 2); although the difference was not statistically significant, VNC CT showed a lower P value (P=0.08) than HRCT (P=0.68) and MonoE 70 keV CT (P=0.77), suggesting greater sensitivity to association trends. HRCT identified lower creatinine levels in Cluster 1, while MonoE 70 keV CT detected lower ALT and AST levels in Cluster 1 (Figure 5A-5C). Both VNC CT and MonoE 70 keV CT identified higher white blood cell counts in Cluster 1 (Figure 5D-5F). These results indicate distinct advantages of each CT modality in assessing fibrosis severity and associated clinical features in SSc-ILD, thereby supporting more precise diagnosis. Representative CT images illustrating the distinct imaging patterns between Cluster 1 and Cluster 2 across HRCT, MonoE 70 keV CT, and VNC CT are shown in Figure 6.

Figure 5 Clinical laboratory indicators in two clusters of three CT modalities. (A-C) Liver and kidney function bar charts for HRCT, MonoE 70 keV CT, and VNC CT, respectively; (D-F) corresponding blood routine bar charts. *, P<0.05; **, P<0.01. Alb, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BASO, basophil; Cre, creatinine; eGFR, estimated glomerular filtration rate; EOS, eosinophil; GGT, gamma-glutamyl transferase; HB, hemoglobin; HRCT, high-resolution computed tomography; LYM, lymphocyte; MONO, monocyte; MonoE 70 keV CT, monoenergetic 70 keV computed tomography; PLT, platelet; UA, uric acid; VNC CT, virtual non-contrast computed tomography; WBC, white blood cell.

Figure 6 Representative CT lung images of patients from the two clusters. The illustrated patients were consistently assigned to the same cluster across all three CT modalities, demonstrating the reproducibility of clustering results. CT, computed tomography.

Discussion

SSc-ILD presents a critical clinical challenge due to its aggressive progression and heterogeneous outcomes, necessitating robust monitoring tools (1,18,19). While PFT guides disease assessment, its clinical utility is limited by patient intolerance, suboptimal effort, and reduced reliability in advanced disease stages (20,21). Quantitative CT parameters have gained attention for their correlation with PFT and fibrosis extent, but evidence in SSc-ILD remains largely limited to conventional HRCT (22-26). This gap highlights the need to explore CT as a potential surrogate for pulmonary function testing and to compare the utility of different CT modalities. Such an approach may help validate CT-based alternatives for patients unable to perform PFTs and provide additional insights into disease pathophysiology. This study aimed to verify whether quantitative CT parameters reflect lung function in SSc-ILD and to characterize the relative strengths of different CT modalities. Specifically, MonoE 70 keV CT demonstrated stronger associations with lung volume-related parameters, including RV and FRCpleth, while VNC CT-based clustering demonstrated a trend toward higher anti-Scl-70 antibody levels in the more severe cluster compared with the milder cluster (P=0.08), suggesting a potential association between VNC-derived quantitative features and serological disease activity (Figures 2,4; Table 2).

In this study, we systematically investigated, for the first time, the disparities among HRCT, MonoE 70 keV CT, and VNC CT in quantifying the extent of pulmonary fibrosis in patients with SSc-ILD. The three CT modalities showed similar trends in attenuation, skewness, and kurtosis, but distinct differences in SD, which is a crucial metric for quantifying the dispersion of voxel gray values and reflecting the inhomogeneity of local tissue density (27). The findings of this study indicated that SD values obtained from the three CT imaging techniques were significantly increased and were mainly distributed within the range of 136–201 Hounsfield units (HU) (Figure 1B). These findings suggest that lung tissue structure exhibits a high degree of heterogeneity, which is consistent with the mixed pathological features observed within fibrotic regions of SSc-ILD, where ground-glass opacities, reticular patterns, and honeycombing cysts coexist (28).

Further comparison revealed significant differences in SD among the three CT modalities, with the order HRCT > MonoE 70 keV CT > VNC CT (Figure 1B). This discrepancy likely stems from differences in noise-processing mechanisms inherent to each imaging technique. Edge enhancement in HRCT amplifies image noise, whereas MonoE 70 keV CT and VNC CT achieve a balance between noise and contrast (29-31). Notably, SD values in the lower lobes were significantly elevated, with values approximately 30–50 HU higher than those observed in the middle and upper lung regions. This distribution pattern closely aligns with the known clinical manifestation of pulmonary fibrosis in SSc-ILD, which predominantly affects the lower lung fields, particularly the basal segments (32). Its lower SD reflects reduced noise-driven variability, which is critical for consistent monitoring of SSc-ILD and supports its use in tracking disease progression and treatment response.

Alan C. Best et al. demonstrated that, as pulmonary fibrosis progresses and lung function deteriorates, skewness and kurtosis values derived from HRCT exhibit a significant downward trend. Follow-up data further indicated that the extent of fibrosis, reductions in TLC, and changes in skewness and kurtosis may serve as important predictive indicators of mortality risk in affected patients (23,33). The findings of the present study are consistent with these previously reported conclusions (Figure 2). Correlation analyses revealed that skewness and kurtosis parameters derived from HRCT, MonoE 70 keV CT, and VNC CT exhibited moderate positive correlations with pulmonary function indices. Notably, this study further demonstrated that skewness and kurtosis were also significantly correlated with lung volume–related parameters, including RV, VA, and FRCpleth. These findings add a quantitative dimension to imaging-based evaluation of pulmonary fibrosis.

Attenuation and SD, key parameters for quantifying tissue density heterogeneity, increase in fibrotic regions as a result of collagen deposition and interstitial thickening (34). Our data showed that these parameters were negatively correlated with lung function, particularly TLC and VA, suggesting that lung structural damage contributes to volumetric functional impairment.

To aid the evaluation of clinical disease progression, we grouped patients with follow-up examinations at 6-month and 1-year intervals and analyzed differences in CT parameters as well as their correlations with changes in lung function (Figure S2). No significant temporal trends were observed, which may be attributable to several factors. The absence of significant short-term changes likely reflects clinical stability under treatment and the presence of established fibrosis, which typically evolves over years rather than months. Clinical observations suggest that meaningful functional or radiological deterioration in well-managed SSc-ILD cohorts often occurs over years rather than months (35,36). Consequently, CT parameters reflecting established fibrotic changes are unlikely to show marked variation over short follow-up intervals (37). SSc-ILD often begins with a phase of subclinical alveolitis, during which rapid radiological changes may occur (38-40); however, many participants in this analysis may have already transitioned to a fibrotic phase characterized by stable, chronic alterations. At this stage, CT parameters tend to be less dynamic, making acute fluctuations more difficult to capture. This highlights that short-term CT assessments may be insufficient to fully characterize SSc-ILD disease dynamics, underscoring the need for longer follow-up intervals or more sensitive biomarkers.

The correlation results indicate that attenuation, SD, skewness, and kurtosis across the three CT modalities exhibit high sensitivity for detecting changes in lung volume, thereby supporting the potential utility of these parameters in longitudinal disease evaluation. Notably, a strong negative correlation was observed between attenuation and certain pulmonary ventilation function indices, including FEV1/FVC and maximal mid-expiratory flow (MMEF), in the 6-month follow-up group. FEV1/FVC and MMEF are sensitive indicators of airway function that often show alterations at the early stage of SSc-ILD (41,42). These findings provide an imaging-based basis for dynamic monitoring of disease progression in patients with SSc-ILD.

This study used K-means clustering to stratify patients with SSc-ILD using quantitative CT parameters, integrating multiple metrics to capture phenotypic patterns beyond single-parameter correlations (43). This approach successfully categorized the cohort into two distinct subgroups. Cluster 1 was characterized by high attenuation values and SD, accompanied by low skewness and kurtosis, whereas Cluster 2 displayed the opposite pattern (Figure 3).

Notably, Cluster 1 exhibited significantly lower FVC and diffusing capacity for DLco than Cluster 2, indicating more severe fibrosis and functional impairment (Figure 4), which was consistent with the corresponding CT parameter patterns. With respect to modality-based stratification, HRCT and VNC CT showed high consistency in reflecting variations in lung function, with high sensitivity to intergroup differences. Although MonoE 70 keV CT identified only four significant parameters, it uniquely captured differences in RV and FRCpleth, highlighting its advantage in quantifying lung volume. In contrast, HRCT and VNC CT provided broader coverage, encompassing volume-related measurements and discriminating ventilatory function.

Clinical laboratory analysis further clarified modality-specific diagnostic features. Anti-Scl-70, a key serological marker associated with fibrosis progression in SSc-ILD (44), were higher in VNC CT Cluster 1 than in Cluster 2, showing a trend toward association with more severe fibrosis (P=0.08) (Table 2). Although liver and kidney function indices (ALT, AST, creatinine) and routine blood (WBC) showed intergroup differences, all values remained normal (Figure 5). HRCT stratification revealed lower creatinine levels in Cluster 1 compared with Cluster 2, accompanied by inverse trends in creatine kinase (CK) (Figure S3). Elevated CK levels have been associated with inflammatory SSc phenotypes, end-organ damage, and immunosuppressive treatment exposure (45,46). Creatinine is the end product of creatine and phosphocreatine metabolism, which can be interconverted by CK (47). The reciprocal changes in CK and creatinine levels observed in this study may indicate potential dysregulation of the creatine metabolic pathway during the pathogenesis of SSc-ILD, thereby providing potential insights into underlying disease mechanisms.

From a clinical perspective, the integration of quantitative CT into existing SSc-ILD management frameworks may provide practical benefits across multiple stages of disease evaluation. HRCT remains the cornerstone for initial diagnosis and structural assessment by providing high-resolution visualization of fibrotic patterns and disease extent. Building upon conventional HRCT assessment, MonoE 70 keV CT may serve as a valuable complementary tool for functional evaluation, particularly for assessing lung volume-related parameters such as RV and FRCpleth, which are often difficult to accurately assess in patients with poor PFT performance. In addition, quantitative features derived from VNC CT may offer supplementary information regarding disease activity, as suggested by their association with serological markers such as anti-Scl-70 antibodies. Taken together, these findings support a multimodal quantitative CT approach, in which different CT techniques contribute complementary information to improve disease stratification, monitoring, and individualized clinical decision-making in patients with SSc-ILD.

This study has several limitations, including a relatively small sample size (n=43), which may have limited statistical power and increased the risk of false-positive or false-negative findings. Accordingly, the results should be interpreted with caution and warrant validation in larger, prospective, multicenter cohorts. Second, the selection of CT parameters may have omitted some relevant fibrotic features, and the use of predefined clustering algorithms may introduce grouping bias. Third, unmeasured confounders, such as comorbidities and prior treatments, were not included in multivariate analyses and may have influenced the observed relationships between CT parameters and pulmonary function. Future multicenter studies involving larger cohorts, refined methodologies, and improved control of confounding factors are needed to further validate these findings.

Conclusions

In patients with SSc-ILD, for whom pulmonary function testing may be limited by feasibility and patient tolerance, quantitative CT provides valuable complementary information for disease assessment. Among the three modalities evaluated, HRCT remains indispensable for comprehensive structural characterization and overall assessment of disease severity. MonoE 70 keV CT demonstrates advantages in lung volumetric evaluation, particularly in capturing alterations in RV and FRCpleth, highlighting its sensitivity to volume-related functional impairment. VNC CT shows potential in reflecting serological disease activity, as suggested by its closer association with anti-Scl-70 antibody levels in clustering analyses. Taken together, these findings indicate complementary strengths among CT modalities, supporting a multimodal quantitative CT approach for individualized SSc-ILD evaluation. Nevertheless, these observations should be interpreted with caution and warrant validation in larger, prospective cohorts.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2278/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2278/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2278/prf

Funding: This work was supported by Sichuan Science and Technology Program (No. 2023YFQ0095).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2278/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of West China Hospital (2022 approval No. 341), and individual informed consent was waived due to the retrospective nature of the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Perelas A, Silver RM, Arrossi AV, et al. Systemic sclerosis-associated interstitial lung disease. Lancet Respir Med 2020;8:304-20. [Crossref] [PubMed]
2. Bussone G, Mouthon L. Interstitial lung disease in systemic sclerosis. Autoimmun Rev 2011;10:248-55. [Crossref] [PubMed]
3. Mouawad JE, Feghali-Bostwick C. The Molecular Mechanisms of Systemic Sclerosis-Associated Lung Fibrosis. Int J Mol Sci 2023;24:2963. [Crossref] [PubMed]
4. Tan A, Denton CP, Mikhailidis DP, et al. Recent advances in the diagnosis and treatment of interstitial lung disease in systemic sclerosis (scleroderma): a review. Clin Exp Rheumatol 2011;29:S66-74.
5. Hoffmann-Vold AM, Distler O, Crestani B, et al. Recent advances in the management of systemic sclerosis-associated interstitial lung disease. Curr Opin Pulm Med 2022;28:441-7. [Crossref] [PubMed]
6. Hoffmann-Vold AM, Petelytska L, Fretheim H, et al. Predicting the risk of subsequent progression in patients with systemic sclerosis-associated interstitial lung disease with progression: a multicentre observational cohort study. Lancet Rheumatol 2025;7:e463-71. [Crossref] [PubMed]
7. Poerio A, Carlicchi E, Zompatori M. Diagnosis of interstitial lung disease (ILD) secondary to systemic sclerosis (SSc) and rheumatoid arthritis (RA) and identification of ‘progressive pulmonary fibrosis’ using chest CT: a narrative review. Clin Exp Med 2023;23:4721-8. [Crossref] [PubMed]
8. Morrisroe K, Hansen D, Stevens W, et al. Progressive pulmonary fibrosis and its impact on survival in systemic sclerosis-related interstitial lung disease. Rheumatology (Oxford) 2024;63:1874-81. [Crossref] [PubMed]
9. Takane Y, Sato K, Kageyama R, et al. Accuracy of virtual non-contrast images with different algorithms in dual-energy computed tomography. Radiol Phys Technol 2022;15:234-44. [Crossref] [PubMed]
10. Dettmer S, Scharm S, Shin HO. Radiological features of interstitial lung diseases. Pathologe 2021;42:86-94. [Crossref] [PubMed]
11. Liu LP, Shapira N, Halliburton SS, et al. Spectral performance evaluation of a second-generation spectral detector CT. J Appl Clin Med Phys 2024;25:e14300. [Crossref] [PubMed]
12. Rassouli N, Etesami M, Dhanantwari A, et al. Detector-based spectral CT with a novel dual-layer technology: principles and applications. Insights Imaging 2017;8:589-98. [Crossref] [PubMed]
13. Yu L, Leng S, McCollough CH. Dual-energy CT-based monochromatic imaging. AJR Am J Roentgenol 2012;199:S9-S15. [Crossref] [PubMed]
14. Papadakis AE, Damilakis J. Technical Note: Quality assessment of virtual monochromatic spectral images on a dual energy CT scanner. Phys Med 2021;82:114-21. [Crossref] [PubMed]
15. Parakh A, Lennartz S, An C, et al. Dual-Energy CT Images: Pearls and Pitfalls. Radiographics 2021;41:98-119. [Crossref] [PubMed]
16. van den Hoogen F, Khanna D, Fransen J, et al. 2013 classification criteria for systemic sclerosis: an American college of rheumatology/European league against rheumatism collaborative initiative. Ann Rheum Dis 2013;72:1747-55. [Crossref] [PubMed]
17. Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med 2011;183:788-824. [Crossref] [PubMed]
18. Ngo K. Systemic Sclerosis and Pulmonary Disease. Adv Exp Med Biol 2021;1303:173-82. [Crossref] [PubMed]
19. Hoffmann T, Aringer M, Koschel D, et al. Pulmonary Involvement in Autoimmune-Mediated Disease. Dtsch Arztebl Int 2025;arztebl.m2025.0162.
20. Cottin V, Brown KK. Interstitial lung disease associated with systemic sclerosis (SSc-ILD). Respir Res 2019;20:13. [Crossref] [PubMed]
21. Martín-López M, Carreira PE. The Impact of Progressive Pulmonary Fibrosis in Systemic Sclerosis-Associated Interstitial Lung Disease. J Clin Med 2023;12:6680. [Crossref] [PubMed]
22. Loeh B, Brylski LT, von der Beck D, et al. Lung CT Densitometry in Idiopathic Pulmonary Fibrosis for the Prediction of Natural Course, Severity, and Mortality. Chest 2019;155:972-81. [Crossref] [PubMed]
23. Best AC, Meng J, Lynch AM, et al. Idiopathic pulmonary fibrosis: physiologic tests, quantitative CT indexes, and CT visual scores as predictors of mortality. Radiology 2008;246:935-40. [Crossref] [PubMed]
24. Ash SY, Harmouche R, Vallejo DL, et al. Densitometric and local histogram based analysis of computed tomography images in patients with idiopathic pulmonary fibrosis. Respir Res 2017;18:45. [Crossref] [PubMed]
25. Tschalèr L, Jordan S, Aaløkken TM, et al. Validation of a semi-quantitative method to assess interstitial lung disease severity and progression in systemic sclerosis by standard and low-dose HRCT scans. RMD Open 2025;11:e004938. [Crossref] [PubMed]
26. Bruni C, Tofani L, Fretheim H, et al. A screening tool to detect interstitial lung disease in systemic sclerosis: the ILD-RISC score. Rheumatology (Oxford) 2025;64:6285-93. [Crossref] [PubMed]
27. Zeng R, Lin CY, Li Q, et al. Performance of a deep learning-based CT image denoising method: Generalizability over dose, reconstruction kernel, and slice thickness. Med Phys 2022;49:836-53. [Crossref] [PubMed]
28. Paone G, Di Michele L, Mattia P, et al. Pulmonary involvement in sclerodermia. Minerva Med 1994;85:293-300.
29. Lennartz S, Pisuchpen N, Parakh A, et al. Virtual Unenhanced Images: Qualitative and Quantitative Comparison Between Different Dual-Energy CT Scanners in a Patient and Phantom Study. Invest Radiol 2022;57:52-61. [Crossref] [PubMed]
30. Koetzier LR, Mastrodicasa D, Szczykutowicz TP, et al. Deep Learning Image Reconstruction for CT: Technical Principles and Clinical Prospects. Radiology 2023;306:e221257. [Crossref] [PubMed]
31. Zeng W, Zeng LM, Xu X, et al. Noise Reduction Effect of Deep-learning-based Image Reconstruction Algorithms in Thin-section Chest CT. Sichuan Da Xue Xue Bao Yi Xue Ban 2021;52:286-92. [Crossref] [PubMed]
32. Yao WZ. Pulmonary changes in progressive systemic sclerosis. Zhonghua Jie He He Hu Xi Za Zhi 1990;13:157-9, 91.
33. Palmucci S, Tiralongo F, Galioto F, et al. Histogram-based analysis in progressive pulmonary fibrosis: relationships between pulmonary functional tests and HRCT indexes. Br J Radiol 2023;96:20221160. [Crossref] [PubMed]
34. Best AC, Lynch AM, Bozic CM, et al. Quantitative CT indexes in idiopathic pulmonary fibrosis: relationship with physiologic impairment. Radiology 2003;228:407-14. [Crossref] [PubMed]
35. Ruaro B, Confalonieri M, Matucci-Cerinic M, et al. The Treatment of Lung Involvement in Systemic Sclerosis. Pharmaceuticals (Basel) 2021;14:154. [Crossref] [PubMed]
36. Behera AK, Sharma P, Tg R, et al. Real-World Clinical Profile and Safety of Nintedanib in Systemic Sclerosis-Associated Interstitial Lung Disease: A Subgroup Analysis of Interstitial Lung Disease Data From an Interstitial Lung Disease (ILD) Specialty Clinic in India. Cureus 2024;16:e65579. [Crossref] [PubMed]
37. Goldman N, Ong VH, Denton CP. Pathogenesis of interstitial lung disease in systemic sclerosis. Rheumatol Immunol Res 2024;5:141-51. [Crossref] [PubMed]
38. Remy-Jardin M, Remy J, Wallaert B, et al. Pulmonary involvement in progressive systemic sclerosis: sequential evaluation with CT, pulmonary function tests, and bronchoalveolar lavage. Radiology 1993;188:499-506. [Crossref] [PubMed]
39. Salaffi F, Carotti M, Baldelli S, et al. Subclinical interstitial lung involvement in rheumatic diseases. Correlation of high resolution computerized tomography and functional and cytologic findings. Radiol Med 1999;97:33-41.
40. De Santis M, Bosello S, La Torre G, et al. Functional, radiological and biological markers of alveolitis and infections of the lower respiratory tract in patients with systemic sclerosis. Respir Res 2005;6:96. [Crossref] [PubMed]
41. Panagopoulos PK, Goules AV, Georgakopoulou VE, et al. Small airways dysfunction in patients with systemic sclerosis and interstitial lung disease. Front Med (Lausanne) 2022;9:1016898. [Crossref] [PubMed]
42. Bjerke RD, Tashkin DP, Clements PJ, et al. Small airways in progressive systemic sclerosis (PSS). Am J Med 1979;66:201-9. [Crossref] [PubMed]
43. McLachlan GJ, Bean RW, Ng SK. Clustering. Methods Mol Biol 2017;1526:345-62. [Crossref] [PubMed]
44. Wang F, Li C, Li X, et al. Autoantibodies in systemic sclerosis overlap syndrome and their correlation with organ damage and survival. Ann Med 2024;56:2407526. [Crossref] [PubMed]
45. Fairley JL, Hansen D, Day J, et al. Proximal weakness and creatine kinase elevation in systemic sclerosis: Clinical correlates, prognosis and functional implications. Semin Arthritis Rheum 2024;65:152363. [Crossref] [PubMed]
46. Wangkaew S, Intum J, Prasertwittayakij N, et al. Elevated baseline serum creatine kinase in Thai early systemic sclerosis patients is associated with high incidence of cardiopulmonary complications and poor survival: an inception cohort study. Clin Rheumatol 2022;41:3055-63. [Crossref] [PubMed]
47. Hespel P, Eijnde BO, Derave W, et al. Creatine supplementation: exploring the role of the creatine kinase/phosphocreatine system in human muscle. Can J Appl Physiol 2001;26:S79-102. [Crossref] [PubMed]