Search Results (68)

Background: Hepatocellular carcinoma (HCC), the most prevalent primary malignancy of the liver, is commonly treated with transarterial chemoembolization (TACE), a locoregional therapy that combines targeted intra-arterial chemotherapy with selective embolization to induce tumor ischemia and necrosis. However, current methods for monitoring the treatment response—such as the RECIST and mRECIST—often fail to detect early or subtle biological changes, such as tumor necrosis or microstructural remodeling, and therefore may underestimate the therapeutic effects, especially in cases with minimal or delayed tumor shrinkage. Thus, there is a critical need for quantitative imaging strategies that can improve early response assessment and guide more personalized treatment decision-making. The goal of this study was to assess the changes in computed tomography (CT) perfusion parameters and radiomic features in HCC before and after TACE and to evaluate the associations of these parameters/features with the tumor burden. Methods: In this retrospective, single-center study, 32 patients with histologically confirmed HCC underwent CT perfusion and radiomic analysis prior to and following TACE. Multiple quantitative perfusion parameters (arterial flow, perfusion flow, perfusion index) and radiomic features were extracted. Statistical comparisons were performed using the Wilcoxon signed-rank test and Spearman’s correlation. Radiomic feature extraction was performed in strict adherence to the Image Biomarker Standardization Initiative (IBSI) guidelines. Preprocessing steps included voxel resampling (1 × 1 × 1 mm), z-score normalization, and fixed bin-width discretization (bin width = 25). All tumor ROIs were manually segmented in consensus by two experienced radiologists to minimize inter-observer variability. Results: Arterial flow significantly decreased from a median of 56.5 to 47.7 mL/100 mL/min after TACE (p = 0.009), while nonsignificant increases in the perfusion flow (from 101.3 to 107.8 mL/100 mL/min, p = 0.44) and decreases in the perfusion index (from 38.6% to 35.7%, p = 0.25) were also observed. Perfusion flow was strongly and positively correlated with tumor size (ρ = 0.94, p < 0.001). Five radiomic texture feature values—especially those of ShortRunHighGrayLevelEmphasis (Δ = +2.11, p = 0.0001) and LargeAreaHighGrayLevelEmphasis (Δ = +75,706, p = 0.0006)—changed significantly after treatment. These radiomic feature value changes were more pronounced in tumors ≥50 mm in diameter. In addition, we performed a receiver operating characteristic (ROC) analysis of the two most discriminative radiomic features (SRHGLE and LAHGLE). We further developed a multivariable logistic regression model that achieved an AUC of 0.87, supporting the potential of these features as predictive biomarkers. Conclusions: CT perfusion and radiomics offer complementary insights into the treatment response of patients with HCC. While perfusion parameters reflect macroscopic vascular changes and are correlated with tumor burden, radiomic features can indicate microstructural changes after TACE. This combined imaging approach may improve early therapeutic assessment and support precision oncology strategies. Full article

Background/Objectives: Despite growing evidence on quantitative computed tomography (CT) analysis of coronary plaques and pericoronary adipose tissue (PCAT), their association with myocardial perfusion (MP) in patients with first acute myocardial infarction (AMI) with obstructive coronary artery disease (MICAD) and non-obstructive coronary arteries (MINOCA) remain unclear. The aim of this study was to assess the relationship between quantitative CT coronary plaque components and PCAT characteristics with MP, myocardial blood flow (MBF) and coronary flow reserve (CFR) obtained by dynamic single-photon emission computed tomography (SPECT) in patients with AMI. Methods: Patients with a first episode of AMI were included in the study. All patients underwent coronary CT angiography with quantitative assessment of plaque volume (PV) and burden (PB), as well as PCAT volume and attenuation. Dynamic SPECT was performed on cadmium–zinc–telluride gamma-camera for quantitative assessment of MP parameters, stress and rest MBF, and CFR. Results: A total of 31 patients (median age 62 [56–70] years) were analyzed, including MICAD (n = 21) and MINOCA (n = 10). MICAD patients had significantly higher total PV and PB, mainly due to non-calcified and fibrofatty components (p < 0.05), while low-attenuation (LAP) and calcified plaques (CP) did not differ between groups. PCAT volumes were higher in MICAD (p < 0.05), whereas PCAT attenuation showed no differences. Dynamic SPECT revealed lower stress MBF and CFR in MICAD (p < 0.05). Correlation analysis showed positive associations of PV and PB with MP summed stress and rest scores, except LAP or CP; PB was negatively associated with MBF. In addition, PCAT volume correlated negatively with stress and rest MBF and CFR, as well as PCAT attenuation correlated positively with stress-induced MP abnormalities. Conclusions: Patients with MICAD demonstrated a greater extent of atherosclerosis and larger PCAT volume compared with MINOCA. Moreover, PCAT volume demonstrated inverse associations with MBF and CFR, indicating a potential link between PCAT characteristics and microvascular dysfunction. Full article

Accurate segmentation of abdominal aortic aneurysm (AAA) from computed tomography (CT) images is critical for early diagnosis and treatment planning of vascular diseases. However, noise in CT images obscures vessel boundaries, reducing segmentation accuracy. U-Net is widely used for medical image segmentation, where noise removal is critical. This study applied various denoising filters for U-Net segmentation and classified the severity of segmented AAA images to evaluate accuracy. Poisson–Gaussian noise was added to AAA CT images, and then average, median, Wiener, and median-modified Wiener filters (MMWF) were applied. U-Net-based segmentation was performed, and the segmentation accuracy of the output images obtained per filter was quantitatively assessed. Furthermore, the Hough circle algorithm was applied to the segmented images for diameter measurement, enabling severity classification and evaluation of classification accuracy. MMWF application improved the Matthews correlation coefficient, Dice score, Jaccard coefficient, and mean surface distance by 31.09%, 34.25%, 53.99%, and 3.70%, respectively, compared with images with added noise. Moreover, classification based on the output images obtained after MMWF application demonstrated the highest accuracy, with sensitivity, precision, and accuracy reaching 100%. Thus, U-Net-based segmentation yields more accurate results when images are processed with the MMWF and analyzed using the Hough circle algorithm. Full article

Background/Objectives: Preclinical chest computed tomography (CT) imaging in small animals is often limited by low resolution due to scan time and dose constraints, which hinders accurate detection of subtle lesions. Traditional super-resolution (SR) metrics, such as peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM), may not adequately reflect clinical interpretability. We aimed to evaluate whether deep learning-based SR models could enhance image quality and lesion detectability in rat chest CT, balancing quantitative metrics with radiologist assessment. Methods: We retrospectively analyzed 222 chest CT scans acquired from polyhexamethylene guanidine phosphate (PHMG-p) exposure studies in Sprague Dawley rats. Three SR models were implemented and compared: single-image SR (SinSR), segmentation-guided SinSR with lung cropping (SinSR3), and omni-super-resolution (OmniSR). Models were trained on rat CT data and evaluated using PSNR and SSIM. Two board-certified thoracic radiologists independently performed blinded evaluations of lesion margin clarity, nodule detectability, image noise, artifacts, and overall image quality. Results: SinSR1 achieved the highest PSNR (33.64 ± 1.30 dB), while SinSR3 showed the highest SSIM (0.72 ± 0.08). Despite lower PSNR (29.21 ± 1.46 dB), OmniSR received the highest radiologist ratings for lesion margin clarity, nodule detectability, and overall image quality (mean score 4.32 ± 0.41, κ = 0.74). Reader assessments diverged from PSNR and SSIM, highlighting the limited correlation between conventional metrics and clinical interpretability. Conclusions: Deep learning-based SR improved visualization of rat chest CT images, with OmniSR providing the most clinically interpretable results despite modest numerical scores. These findings underscore the need for reader-centered evaluation when applying SR techniques to preclinical imaging. Full article

Background/Objective: Computed tomography (CT) can be used to non-invasively assess the health of the liver; however, radiologist evaluation and simple thresholding alone are insufficient for diagnosis of hepatic steatosis, necessitating biopsies. This study explored CT radiomics and machine learning to enable non-invasive, objective, and quantitative prediction of steatosis severity across the macaque liver. Methods: In this retrospective study, CT images of 42 crab-eating macaques (age [yr] = 6.1 ± 1.7; sex [male/female] = 26/16) with varying degrees of hepatic steatosis were analyzed, and the results were compared to histology-based steatosis scores of livers from the same animals. After extracting radiomic features, a thorough array of statistical analyses, feature selection techniques, and machine learning models were applied to identify a distinct radiomic signature of histologically defined hepatic steatosis. Results: We identified 12 radiomic features that correlated with steatosis scores, and hierarchical clustering based on radiomic attributes alone revealed clusters roughly aligning with steatosis severity groups. The k-nearest neighbors model architecture best predicted histopathologic steatosis scores in both classification and regression tasks (area under the receiver operating characteristic curve [AUC ROC] = 0.89 ± 0.09; root-mean-square error [RMSE] = 0.60 ± 0.10). Feature analyses identified seven key radiomic features (six first-order features and one gray-level co-occurrence matrix feature) that were most important when predicting steatosis. Conclusions: We identified a CT radiomic signature of steatosis and demonstrated that histology-based steatosis scores can be predicted non-invasively and objectively using machine learning and CT radiomics as a potential alternative to invasive core biopsies. Given the strong similarities in liver structure, liver function, and hepatic steatosis pathophysiology between macaques and humans, these findings have the potential to translate to humans. Full article

Introduction: Coronary artery disease remains one of the most prevalent causes of hospital cardiac arrest (OHCA). Although the benefit of early coronary angiography is well stablished in patients with ST-segment elevation, the benefit and the timing of performing it in other patients remain a matter of debate. This is due to the difficulty of identifying those in which an infarction with non-ST-segment elevation is the cause of the OHCA. Coronary artery calcium (CAC) emerges as a reliable predictor of coronary disease and adverse cardiovascular events, detectable even in non-gated chest computed tomography (CT) scans commonly used in OHCA etiological studies, showcasing potential for streamlined risk assessment and management. Aim: The aim of this study was to evaluate if CAC in non-gated CT scans performed in OHCA survivors could act as a good predictor of coronary artery disease on coronary angiography. Methods: This is a single-center, retrospective study of OHCA survivors without ST-segment elevation at presentation. We selected patients for whom a non-gated chest CT was performed and underwent coronary angiography due to the clinical, electrocardiogram (ECG), or echocardiographic suspicion of acute coronary syndrome. An investigator, blinded to the coronary angiography report, evaluated CAC both quantitively (with Agatston score) and qualitatively (visual assessment: absent, mild, moderate, or severe). Results: A total of 44 consecutive patients were included: 70% male, mean age of 60 ± 13 years old. The mean Agatston score was 396 ± 573 AU (Agatston units). Regarding the qualitative assessment, CAC was classified as mild, moderate, and severe in 11%, 25%, and 20% of patients, respectively. The coronary angiography revealed significant coronary lesions in 15 patients (34%), of which 87% were revascularized (80% underwent PCI and 7% CABG). The quantitative CAC assessment accurately predicted the presence of significant lesions on coronary angiography (AUC = 0.90, 95% CI 0.81–0.99, p < 0.001). The presence of moderate or severe CAC by visual assessment also predicted significant lesions on coronary angiography (OR 2.66, 95% CI 1.87–109.71, p = 0.01). There was also a good and significant correlation between the vessel with severe calcification in the CT scan and the culprit vessel evaluated by coronary angiography. CAC was reported in only 16% of the reviewed CTs, most of them with severe calcification. Conclusion: The assessment of CAC in non-gated chest CT scans proved to be feasible and displayed a robust correlation with the presence, severity, and location of coronary artery disease. Its routine use upfront was shown to be an important complement to CT scan reports, ensuring more precise and personalized OHCA management. Full article

Background/Objectives: Head and neck squamous cell carcinoma (HNSCC) diagnosis and treatment rely heavily on computed tomography (CT) imaging to evaluate tumor characteristics and lymph node (LN) involvement, crucial for staging, prognosis, and therapy planning. Conventional LN evaluation methods based on morphological criteria such as size, shape, and anatomical location often lack sensitivity for early metastasis detection. This study leverages radiomics to extract quantitative features from CT images, addressing the limitations of subjective assessment and aiming to enhance LN classification accuracy while potentially reducing reliance on invasive histopathology. Methods: We analyzed 234 LNs from 27 HNSCC patients, deriving 120 features per node, resulting in over 25,000 data points classified into reactive, pathologic, and pathologic with extracapsular spread classes. Considering the challenges of high dimensionality and limited dataset size, more than 44,000 experiments systematically optimized machine learning models, feature selection methods, and hyperparameters, including ensemble approaches to strengthen classification robustness. A Pareto front strategy was employed to balance diagnostic accuracy with significant feature reduction. Results: The optimal model, validated via 5-fold cross-validation, achieved a balanced accuracy of 0.90, an area under the curve (AUC) of 0.93, and an F1-score of 0.88 using only five radiomics features. Conclusions: This interpretable approach aligns well with clinical radiology practice, demonstrating strong potential for clinical application in noninvasive LN classification in HNSCC. Full article

Background/Objectives: To assess quantitatively the correlation between the lung ultrasound severity scores (LUSSs) and chest CT severity scores (CTSSs) derived from low-dose computed tomography (LDCT) for evaluating pulmonary inflammation in COVID-19 patients. Methods: In this prospective observational study, from an initial cohort of 1000 patients, 555 adults (≥18 years) with confirmed COVID-19 were enrolled based on inclusion criteria. All underwent LDCT imaging, scored by the CTSS (0–25 points), quantifying involvement across five lung lobes. Lung ultrasound examinations using standardized semi-quantitative scales for the B-line (LUSS B) and consolidation (LUSS C) were performed in a subgroup of 170 patients; 110 had follow-up imaging after one week. Correlation analyses included Spearman’s and Pearson’s coefficients. Results: Significant positive correlations were found between the CTSS and both the LUSS B (r = 0.32; p < 0.001) and LUSS C (r = 0.24; p = 0.006), with the LUSS B showing a slightly stronger relationship. Each incremental increase in the LUSS B corresponded to an average increase of 0.18 CTSS points, whereas a one-point increase in the LUSS C corresponded to a 0.27-point CTSS increase. The mean influence of the LUSS on CTSS was 8.0%. Neither ultrasound score significantly predicted ICU admission or mortality (p > 0.05). Conclusion: Standardized lung ultrasound severity scores show a significant correlation with low-dose CT in assessing pulmonary involvement in COVID-19, particularly for the B-line artifacts. Lung ultrasound represents a valuable bedside tool, complementing—but not substituting—CT in predicting clinical severity. Integrating both imaging modalities may enable the acquisition of complementary bedside information and facilitate dynamic monitoring of disease progression. Full article

Background: This study aims to identify predictive factors for pulmonary fibrosis development in COVID-19 patients by analyzing thorax CT (computed tomography) findings, serum prolidase activity, MMP-1, MMP-7, TGF-β values, laboratory findings, and demographic characteristics. Materials and methods: The investigation involved 68 patients, both male and female, aged 18 years and older, who were volunteers and had been diagnosed with confirmed COVID-19. The pulmonologist and the radiologist evaluated the thorax CT by consensus. Patients were evaluated in two categories, group 1 and group 2, based on the status of fibrotic changes, and 3-month fibrosis scores were calculated. Findings in both lungs were calculated and noted for the lobes, considering lobar spread. Correlations between quantitative parameters were assessed with Spearman’s rho correlation coefficient. Comparisons between independent samples were evaluated using either the independent sample t-test or the Mann–Whitney U test. We evaluated the relationship between categorical variables using the Pearson chi-square test and Fisher’s exact test. Results: Serum prolidase activity, MMP-1, MMP-7, and TGF-β biomarkers were not statistically significant among groups. LDH was found to be significantly high in the group with fibrotic changes. Additionally, the group with fibrotic changes also had higher levels of fibrinogen. The percentage of neutrophils, the severity of the disease, muscle–joint pain and fatigue symptoms, and the length of hospitalization stay were correlated with the total scores of fibrosis at the third month. In the group with fibrotic changes, the duration of muscle–joint pain and fatigue symptoms and the length of hospitalization were longer than in the other group. Conclusions: The group with fibrotic changes showed an increase in biomarkers. However, this increase did not reach a statistically significant level, suggesting that the third month may be an early period for these changes. The group with fibrotic changes showed high levels of LDH, one of the most important laboratory parameters of pulmonary fibrosis risk factors, along with fibrinogen, suggesting that these parameters are valuable in predicting pulmonary fibrosis. Patients with fibrotic changes can experience specific symptoms, commonly seen in COVID-19. Full article

Aortic stenosis (AS), the most common valvular heart disease, is traditionally graded based on several echocardiographic quantitative parameters, such as aortic valve area (AVA), mean pressure gradient (MPG), and peak jet velocity (Vmax). This systematic review evaluates the clinical significance and prognostic implications of discordant high-gradient AS (DHG-AS), a distinct hemodynamic phenotype characterized by elevated MPG despite a preserved AVA (>1.0 cm²). Although often overlooked, DHG-AS presents unique diagnostic and therapeutic challenges, as high gradients remain a strong predictor of adverse outcomes despite moderately reduced AVA. Sixty-three studies were included following rigorous selection and quality assessment of the key studies. Prognostic outcomes across five key studies were discrepant: some showed better survival in DHG-AS compared to concordant high-gradient AS (CHG-AS), while others reported similar or worse outcomes. For instance, a retrospective observational study including 3209 patients with AS found higher mortality in CHG-AS (unadjusted HR: 1.4; 95% CI: 1.1 to 1.7), whereas another retrospective multicenter study including 2724 patients with AS observed worse outcomes in DHG-AS (adjusted HR: 1.59; 95% CI: 1.04 to 2.56). These discrepancies may stem from delays in intervention or heterogeneity in study populations. Despite the diagnostic ambiguity, the presence of high gradients warrants careful evaluation, aggressive risk stratification, and timely management. Current guidelines recommend a multimodal approach combining echocardiography, computed tomography (CT) calcium scoring, transesophageal echocardiography (TEE) planimetry, and, when needed, catheterization. Anatomic AVA assessment by TEE, CT, and cardiac magnetic resonance imaging (CMR) can improve diagnostic accuracy by directly visualizing valve morphology and planimetry-based AVA, helping to clarify the true severity in discordant cases. However, these modalities are limited by factors such as image quality (especially with TEE), radiation exposure and contrast use (in CT), and availability or contraindications (in CMR). Management remains largely based on CHG-AS protocols, with intervention primarily guided by transvalvular gradient and symptom burden. The variability among the different guidelines in defining severity and therapeutic thresholds highlights the need for tailored approaches in DHG-AS. DHG-AS is clinically relevant and associated with substantial prognostic uncertainty. Timely recognition and individualized treatment could improve outcomes in this complex subgroup. Full article

With the depletion of high-grade iron ore resources, the efficient utilization of low-grade iron ore has become a critical demand in the steel industry. Due to its uniform particle size and chemical composition, pelletized iron ore significantly enhances both the utilization rate of iron ore and the efficiency of metallurgical processes. This paper presents a deep learning model based on ResUNet, which integrates three-dimensional CT images obtained through industrial computed tomography (ICT) to precisely segment hematite, liquid phase, and porosity. By incorporating residual connections and batch normalization, the model enhances both robustness and segmentation accuracy, achieving F1 scores of 98.37%, 95.10%, and 83.87% for the hematite, pores, and liquid phase, respectively, on the test set. Through 3D reconstruction and quantitative analysis, the volume fractions and fractal dimensions of each component were computed, revealing the impact of the spatial distribution of different components on the physical properties of the pellets. Systematic evaluation of model robustness demonstrated varying sensitivity to different CT artifacts, with the strongest resistance to beam hardening and highest sensitivity to Gaussian noise. Multi-scale resolution analysis revealed that segmentation quality and fractal dimension estimates exhibit phase-dependent responses to resolution changes, with the liquid phase being the most sensitive. Despite these dependencies, the relative complexity relationships among phases remained consistent across resolutions, supporting the reliability of our qualitative conclusions. The study demonstrates that the deep learning-based image segmentation method effectively captures microstructural details, reduces human error, and enhances automation, providing a scientific foundation for optimizing pellet quality and improving metallurgical performance. It holds considerable potential for industrial applications. Full article

Background: Upper tract urothelial carcinoma (UTUC) often presents with aggressive behaviour, demanding accurate preoperative assessment to guide management. Radiomics-based approaches have shown promise in extracting quantitative features from imaging, yet few studies have explored whether perirenal fat (PRF) radiomics can augment tumour-only models. Methods: A retrospective cohort of 103 UTUC patients undergoing radical nephroureterectomy was analysed. Tumour regions of interest (ROI) and concentric PRF expansions (10–30 mm) were segmented from computed tomography (CT) scans. Radiomic features were extracted using PyRadiomics, filtered by correlation and intraclass correlation coefficients, and integrated with clinical variables (e.g., age, BMI, multifocality). Multiple machine learning models, including MLPClassifier and CatBoost, were evaluated via repeated cross-validation. Performance was assessed using the area under the ROC curve (AUC), sensitivity, specificity, F1-score, and DeLong tests. Results: The best tumour grade model (AUC = 0.961) merged tumour-derived features with a 10 mm PRF margin, exceeding PRF-only (AUC = 0.900) and tumour-only (AUC = 0.934) approaches. However, the improvement over tumour-only was not always statistically significant. For stage prediction, combining tumour and 15 mm PRF features yielded the top AUC of 0.852, surpassing the tumour-alone model (AUC = 0.802) and outperforming PRF-only (AUC ≤ 0.778). PRF features provided an additional predictive value for both grade and stage models. Conclusions: Integrating PRF radiomics with tumour-based analyses enhances predictive accuracy for UTUC grade and stage, suggesting that the tumour microenvironment contains complementary imaging cues. These findings, pending external validation, support the potential for radiomics-driven risk stratification and personalised treatment planning in UTUC. Full article

Background/Objectives: Lung neuroendocrine neoplasms (NENs) are a heterogeneous group of tumors requiring accurate differentiation from non-small cell lung cancer (NSCLC) for effective treatment. Conventional computed tomography (CT) lacks pathognomonic features to distinguish these subtypes. Radiomics, which extracts quantitative imaging features, offers a potential solution. Methods: This retrospective multicenter study included 301 patients with histologically confirmed lung cancer who underwent native CT scans. The dataset comprised 150 NSCLC cases (75 adenocarcinomas, 75 squamous cell carcinomas) and 151 NENs (75 SCLC, 60 carcinoids, 16 large cell neuroendocrine carcinomas). Tumors were manually segmented, and 107 radiomics features were extracted. Dimensionality reduction and feature selection were performed using Pearson correlation analysis and LASSO regression. Decision tree and random forest classifiers were trained and evaluated using a 70:30 training–testing split. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), accuracy, precision, recall, and F1-score. Results: The model differentiating NENs from NSCLC achieved an AUC of 0.988 on the test set, with an accuracy of 97.8%. The model distinguishing SCLC from other NENs attained an AUC of 0.860 and an accuracy of 82.6%. First-order and textural radiomics features were key discriminators. Conclusions: Radiomics-based machine learning models demonstrated high diagnostic accuracy in differentiating lung NENs from NSCLC and in subclassifying NENs. These findings highlight the potential of radiomics as a non-invasive, quantitative tool for lung cancer diagnosis, warranting further validation in larger multicenter studies. Full article

(1) Background: Dual-energy X-ray absorptiometry (DXA)-based parameters such areal bone mineral density (aBMD) and Trabecular Bone Score (TBS) are routinely used to evaluate participants at risk for fragility fractures (FFs). We compared the accuracy of lumbar spine aBMD and TBS to that of volumetric BMD (vBMD) by quantitative computed tomography (QCT). (2) Methods: We conducted a retrospective analysis of participants who received both a DXA scan and a chest/abdomen CT scan. BMD and TBS values were obtained from lumbar DXA and vBMD values from QCT (three vertebrae from L1 to L4). T-score values were used for DXA diagnosis; the American College of Radiology ranges were used to diagnose bone status with QCT. (3) Results: We included 105 participants (87 women, mean age 69 ± 11 years). Among them, n = 49 (46.6%) presented at least one major FF. QCT diagnosis was as follows: osteoporosis = 59 (56.2%); osteopenia = 36 (34.3%); and normal status = 10 (9.5%). DXA diagnosis was osteoporosis = 25 (23.8%); osteopenia (33.3%) = 35; and normal status = 45 (42.9%). A total of 38 participants (36.2%) showed a TBS degraded microarchitecture. Correlation was moderate between aBMD and vBMD (r = 0.446), as well as between TBS and vBMD (r = 0.524). A good correlation was found between BMD and TBS (r = 0.621). ROC curves to discriminate between participants with/without FFs showed the following areas under the curve: 0.575 for aBMD, 0.650 for TBS, and 0.748 for QCT BMD. (4) Conclusions: QCT detected a higher prevalence of osteoporosis compared to DXA. TBS performed better than aBMD from DXA in discriminating between subjects with and without FFs. Full article

Background: The advent of artificial intelligence has significantly impacted radiology, with radiomics emerging as a transformative approach that extracts quantitative data from medical images to improve diagnostic and therapeutic accuracy. This study aimed to enhance the radiomic workflow by applying deep learning, through transfer learning, for the automatic segmentation of lung regions in computed tomography scans as a preprocessing step. Methods: Leveraging a pipeline articulated in (i) patient-based data splitting, (ii) intensity normalization, (iii) voxel resampling, (iv) bed removal, (v) contrast enhancement and (vi) model training, a DeepLabV3+ convolutional neural network (CNN) was fine tuned to perform whole-lung-region segmentation. Results: The trained model achieved high accuracy, Dice coefficient (0.97) and BF (93.06%) scores, and it effectively preserved lung region areas and removed confounding anatomical regions such as the heart and the spine. Conclusions: This study introduces a deep learning framework for the automatic segmentation of lung regions in CT images, leveraging an articulated pipeline and demonstrating excellent performance of the model, effectively isolating lung regions while excluding confounding anatomical structures. Ultimately, this work paves the way for more efficient, automated preprocessing tools in lung cancer detection, with potential to significantly improve clinical decision making and patient outcomes. Full article