Sang Min Lee

Quantitative evaluation of chronic obstructive pulmonary disease (COPD) using computed tomography (CT) requires soft‐kernel reconstructions to accurately measure the emphysema index (EI) and airway wall thickness. However, many institutions archive only single‐kernel CT images, limiting retrospective analysis. This study assessed whether a diffusion‐based image translation model could generate reliable soft‐kernel equivalents from routinely stored sharp‐kernel CT images. In this retrospective study, paired soft‐ and sharp‐kernel CT scans were collected from a multi‐detector CT scanner. Three approaches—a convolutional neural network (CNN), a generative adversarial network (GAN), and the diffusion‐based Image‐to‐Image Schrödinger Bridge (I ² SB)—were trained to convert sharp‐kernel to soft‐kernel images. Performance was evaluated using peak signal‐to‐noise ratio (PSNR), structural similarity index measure (SSIM), and root mean square error (RMSE); difference‐map analysis; Dice similarity coefficients (DSC) for low‐attenuation areas (LAA) and airway lumen; agreement in quantitative biomarkers (EI and standardized airway wall thickness, Pi10); and a blinded reader study involving radiologists and radiology residents. A total of 150 patients were included (mean age, 63.3 ± 13.7 years; 68.0% male). I ² SB achieved higher PSNR and SSIM and lower RMSE than CNN and GAN (all p < 0.001). I ² SB also produced higher DSC for LAA and airway lumen. For quantitative biomarkers, pix2pixHD showed a lower absolute error for EI, whereas I ² SB showed a lower absolute error for Pi10. Reader evaluations also showed a preference for I ² SB‐generated images. I ² SB demonstrated strong overall performance for sharp‐to‐soft CT kernel conversion, although its relative advantage varied across quantitative biomarkers. These findings support the technical feasibility of I ² SB‐based kernel conversion for retrospective quantitative COPD assessment.

In mild idiopathic pulmonary fibrosis, quantitative CT–derived fibrosis burden and vascular remodeling were associated with acute exacerbation risk, with venous and arterial remodeling mediating fibrosis effects, supporting the value of vascular metrics for risk stratification.

Rationale: Validated thresholds for clinically meaningful change in quantitative computed tomography (QCT) for fibrosis extent across algorithms remain unclear in idiopathic pulmonary fibrosis (IPF). Objectives: To evaluate minimal clinically important difference (MCID) for 1-year change in QCT metrics in IPF using different algorithms than previously reported and to identify prognostically meaningful thresholds for effective risk stratification. Methods: This multicenter retrospective study included IPF patients with baseline and 1-year follow-up CT. Minimal clinically important differences for change in fibrosis score (ΔFS) derived from fully automated CT quantification were estimated using an anchor-based approach with 1-year changes in FVC and diffusing capacity for carbon monoxide (DLco), and a prognostic threshold was estimated using maximally selected log-rank statistics. Prognostic significance of ΔFS was evaluated in discovery and validation cohorts. Results: The discovery cohort included 524 patients (mean age = 66.8 years; 79% men), and the external validation cohort included 224 patients (mean age = 69.6 years; 83% men). Anchor-based MCIDs were 2.72% for FVC and 4.52% for DLco, with a prognostic threshold of 4.05%. In the discovery cohort, 1-year ΔFS were independently associated with transplant-free survival (TFS), with greater risk when ΔFS exceeded MCID or prognostic thresholds (all, P < .001). In the validation cohort, ΔFS remained associated with TFS (adjusted hazard ratio [aHR] = 1.11; 95% CI, 1.04-1.18), and all prespecified ΔFS thresholds were prognostic for overall TFS, with the prognostic threshold of ΔFS ≥ 4.05% showing the most consistent associations with overall (aHR = 2.78; 95% CI, 1.36-5.68) and 3-year TFS (aHR = 2.88; 95% CI, 1.11-7.48). Conclusion: One-year change in FS and its thresholds were significantly associated with TFS, supporting annual CT follow-up with predefined QCT thresholds for monitoring and risk stratification in IPF.

To prospectively evaluate positional variability in quantitative CT (qCT)-derived measurements of interstitial lung abnormality (ILA) using the same-day supine and prone CT. In this prospective study (February 2024-February 2025), participants with ILA underwent sequential non-contrast supine and prone CT scans using identical acquisition parameters. A commercially available deep learning-based software quantified fibrotic (reticulation and honeycombing) and nonfibrotic (ground-glass opacity) ILA components in accordance with Fleischner Society definitions. qCT differences between supine and prone measurements were assessed using paired t tests, Bland-Altman analysis with 95% limits of agreement (LOA), and concordance correlation coefficients (CCC). Of 47 consented participants, 38 (mean age, 70.9 ± 6.4 years; 27 men) were included in the final analysis. Mean total ILA extent was greater on supine than on prone scans (1.76% vs 1.39%, p = 0.02), largely due to a greater extent of fibrotic ILAs on supine images (1.43% vs 1.12%, p = 0.007). The 95% LOA between supine and prone scans were -1.49% to 2.23% for total ILA extent, -0.97% to 1.58% for the fibrotic component, and -0.89% to 1.03% for the nonfibrotic component. Agreement between positions was moderate for the fibrotic component (CCC = 0.770), and poor for the nonfibrotic component (CCC = 0.431), with the lowest reproducibility observed in dependent lung zones. The measurement variability of qCT result of ILA between supine and prone scans was approximately 1.9%, with moderate agreement for the fibrotic component but poor agreement for the nonfibrotic component between the same-day supine and prone CT scans. Question Whether qCT-derived measurements of ILA are interchangeable between supine and prone CT scans remains unknown. Findings qCT measurements showed approximately 1.9% positional variability, with moderate agreement for fibrotic components but poor agreement for nonfibrotic components, especially in dependent lung zones. Clinical relevance Supine and prone scans should not be used interchangeably for qCT analysis of ILA. Consistent patient positioning is essential to ensure accurate longitudinal assessment.

Objectives: To investigate how recurrence patterns in oesophageal squamous cell carcinoma (ESCC) are influenced by primary tumour location, treatment modality, and pathologic stage, and distinguish true recurrence from frequent secondary primary malignancies. Methods: A total of 953 ESCC patients (mean age, 62.6 ± 7.7 years, 883 [92.7%] male) who underwent curative-intent oesophagectomy between 2010 and 2020 were retrospectively analysed. Patients were categorized by tumour location (upper/mid vs lower), pathologic stage (0-I vs II-IV), and treatment modality (upfront surgery vs neoadjuvant chemoradiation). Recurrence patterns were classified as loco-regional, distant, or mixed. Secondary primary malignancies were also documented. Associations between recurrence patterns and post-recurrence survival (PRS) were assessed using multivariable Cox regression. Results: Recurrence occurred in 23.4% (223/953) of patients, mostly within 2 years post-surgery (83.4%). Among early pathologic stage (0-I) patients treated with upfront surgery, upper/mid oesophageal cancers had a higher rate of loco-regional recurrence compared with lower oesophageal cancers (16.3% vs 6.7%; P = .003), particularly in the mediastinal (10.9% vs 4.8%) and supraclavicular lymph nodes (6.3% vs 1.4%). No location-related differences were observed in advanced pathologic stage (II-IV) or neoadjuvant chemoradiation groups. Distant recurrence predominated in the neoadjuvant chemotherapy group (19.6% vs 9.7%). The lung was the most frequent site of distant metastasis (7.6%-8.6%), presenting as solitary nodules (24/45). Secondary lung cancers occurred later than pulmonary recurrences (median 58.3 vs 6.9 months; P < .001). Mixed recurrence was associated with worse PRS (adjusted hazard ratio, 1.77; 95% confidence interval, 1.18-2.66; P = .006). Conclusions: ESCC recurrence patterns vary by tumour location, treatment modality, and pathologic stage. The lung is the most common site of distant recurrence and presents earlier than secondary lung cancer.

Regional lung ventilation imaging using xenon-enhanced dual-energy CT (Xe-DECT) offers valuable insight into obstructive pulmonary diseases but remains limited in clinical use due to technical and logistical constraints. In this study, a multi-task conditional generative adversarial network (GAN) was developed to generate deep learning–generated ventilation images (DL-Vent) from virtual non-contrast (VNC) images. A total of 269 scans from 177 patients with COPD or asthma-COPD overlap syndrome (ACOS) were used to train, validate, and test the model. The architecture was designed to simultaneously predict ventilation images and emphysema masks using paired inspiratory and expiratory VNC input images. DL-Vent demonstrated strong similarity to measured Xe-DECT ventilation images (Xe-Vent), with dice similarity coefficients of 0.56 for ventilation defects and 0.88 for ventilation regions. The ventilation defect percentages (VDP) of DL-Vent and Xe-Vent showed a high correlation (r_s = 0.82), and both were similarly correlated with pulmonary function test results, including FEV1 (p = 0.71). Radiologists rated DL-Vent images as “fair to good” (mean score 3.9/5) and reliably differentiated defect patterns between COPD and ACOS (Cramer’s V = 0.41, p = 0.03). The proposed model provides a promising alternative for functional lung imaging without requiring xenon administration.

Automated evaluation of pulmonary lesion changes on chest radiograph during follow-up using semantic segmentation - Diagnostic and Interventional Radiology

Objectives: The rate and extent of hemithorax volume reduction on the operated side, corresponding compensatory volume increase on the contralateral side, and mediastinal angle shift following pneumonectomy have not been quantitatively analysed. This study aimed to quantify these time-dependent changes in thoracic anatomy after pneumonectomy. Methods: Patients who underwent pneumonectomy between April 2003 and June 2019 were retrospectively included. The volume of each hemithorax and the mediastinal shift angle were quantified using computed tomography scans taken pre- and postoperatively (6 months, 1, 2, and 3 years after surgery). Results: Finally, 119 patients (median age, 58.5 years; 95 men) were evaluated. The pneumonectomy site volume decreased to 25% and 37% of the original volume in the left and right pneumonectomy, respectively, over 3 years, with 21%-24% hyperinflation in the non-operated hemithorax. Mediastinal shifting towards the pneumonectomy side was observed as an increased angle (from 7.1° to 21.4° for the left pneumonectomy and from 4.4° to 20.3° for the right pneumonectomy). Volume reduction was significant until the first 6 months for right pneumonectomy and 1 year for left pneumonectomy (all P < .05), while hyperinflation and mediastinal shifting stabilized after 6 months. Left pneumonectomy showed a more abrupt volume decrease than right pneumonectomy during the observed period (all P < .05). Conclusions: Significant anatomical changes occur within the first 6 months after a pneumonectomy. Thereafter, the magnitude of change decreases, and the thoracic structures reach a steady state by 2 to 3 years postoperatively. Clinical Registration Number: None.

In patients undergoing sublobar resection for pathologic stage I lung cancer, preoperative percutaneous transthoracic biopsy was associated with an increased risk of local-regional recurrence, especially in the remnant lung.

Objectives To evaluate the effect of CT progression of interstitial lung abnormality (ILA) and ILA subtypes on pulmonary function and mortality in lung cancer screening (LCS) participants with preserved pulmonary function. Materials and methods Consecutive participants who met the 2021 United States Preventive Services Task Force guidelines for LCS during a medical check-up between 2012 and 2014 were retrospectively analyzed. Forced vital capacity (FVC) ≥ 80% at baseline was considered indicative of preserved pulmonary function. CT progression and ILA subtype were evaluated for their association with an FVC decline to < 80% and mortality using multivariable time-dependent Cox analysis. Results Among the 6332 LCS participants, 133 with baseline FVC ≥ 80% and follow-up CT and FVC data were included. CT progression was observed in 81.8% (54/66) of those with ILA and 67.2% (45/67) of those with equivocal ILA, with median follow-ups of 61.0 and 76.0 months, respectively. FVC decline to < 80% occurred in 21.1% (28/133) with a median time of 61.4 months. It was associated only with baseline FVC (hazard ratio (HR), 0.78; p < 0.001), while CT progression (p = 0.720) and fibrotic ILA (p = 0.066) were not. For mortality, both CT progression (HR, 8.74; p < 0.001) and a relative FVC decline ≥ 10% (HR, 10.30; p < 0.001) were independent risk factors, whereas fibrotic ILA was not (p = 0.254). Conclusion CT progression was a risk factor for mortality, although it was not associated with a decline in FVC to below 80% in participants with preserved lung function. Monitoring CT progression in LCS would be helpful for risk stratification of participants with ILA. Key Points Question Does interstitial lung abnormality (ILA) CT progression affect pulmonary function decline and mortality, even when pulmonary function is preserved? Findings When pulmonary function was preserved, CT progression was not associated with forced vital capacity decline but was an independent risk factor for mortality. Clinical relevance Monitoring CT progression during lung cancer screening adherence could aid in risk stratification for participants with ILA. Graphical Abstract