Radiologia Medica最新文献

Purpose: Chest high-resolution computed tomography (HRCT) is crucial for diagnosing and monitoring pulmonary diseases involving parenchymal, vascular, and airway alterations. However, segmentation faces challenges in distinguishing pulmonary structures due to heterogeneity in image acquisition and pathological manifestations. Unlike existing tools, which usually target a single anatomical structure and rely predominantly on either deep learning or rule-based approaches, our hybrid pipeline pairs U-Net-based AI segmentation with tailored image processing refinements to produce a reliable and simultaneous segmentation of lungs, airways, pulmonary vessels, and parenchymal injury patterns, while enabling quantitative characterization across a spectrum of disease severities and types (inflammatory and infectious).

Methods: This retrospective observational study employed 19 chest CT scans from COVID-19 public datasets for deep learning, 8 annotated scans from the EXACT'09 challenge to validate airway segmentation, and 20 retrospective HRCT scans from COVID-19 and idiopathic pulmonary fibrosis patients for pipeline validation. The pipeline performs preliminary segmentation of lungs, airways, and pathological regions using U-Nets, followed by image processing to refine results, include vasculature, and classify injury patterns in ground-glass opacities, reticulations/consolidations, and air-filled pathological spaces. Three radiologists validated segmentations on a 1-5 scale, and the Kruskal-Wallis test was conducted to assess differences across raters, pathologies, and severities.

Results: The proposed pipeline visually outperformed established tools (LungCTAnalyzer, PTK, TotalSegmentator). Airway's segmentation achieved a Dice coefficient of 0.91 [0.89-0.92] on the EXACT'09 dataset. Radiologists assigned scores of 4 and 5 to segmentation completeness and accuracy, respectively, for both airways and vessels. Parenchymal injury patterns scored 4 for completeness, accuracy, and classification. Ratings were consistently high with no significant differences among raters, diseases, and severity levels.

Conclusion: The proposed pipeline introduces a novel, comprehensive, and hybrid approach for simultaneous, multi-structure lung segmentation, demonstrating reliable and potentially generalizable performance across inflammatory and infectious pulmonary diseases.

Aim: To evaluate the association of anatomical, clinical, and procedural factors with endovascular treatment failure, including both proximal and distal splenic artery embolization (SAE).

Material and methods: In 11 centers, all patients following blunt splenic injury (BSI) were retrospectively evaluated, and those who had received SAE were selected. Data collected included: patient demographics and characteristics, mechanism and grading of BSI, endovascular management, and outcomes. Technical and clinical success were defined as successful embolization of the bleeding artery and stabilization of the haemodynamic status and laboratory data in 1 or 2 sessions, respectively. Rebleeding during follow-up and subsequent splenectomy were considered as treatment failure. The rate of complications related to the endovascular procedure was evaluated.

Results: The management of two hundred and forty-seven participants was evaluated. Technical and clinical success were 100% and 91.9% (227/247), respectively. A second embolization was performed in 5 cases. Rescue splenectomy occurred in 20 (8.1%) patients. Statistically significant associations were identified between endovascular treatment failure and GCS and the presence of other lesions at CT at patient presentation. No anatomical or procedural factors were found to be statistically significant; in the surgical group, a larger diameter of the splenic artery was observed. The complication rate was 15.2% (26/171), all relating to the vascular access, e.g., hematoma or pseudoaneurysm, and all managed conservatively.

Conclusion: SAE is a safe and effective procedure; unsuccessful cases resulted statistically associated with some clinical factors, but no correlation with anatomical factors was observed.

Aims: The aim of this study is to analyze how radiotherapy (RT) is perceived and portrayed by Italian media and to determine whether there is any bias or misinformation about. The study will also assess the influence of these perceptions on patients' and medical students' decisions to specialize in radiotherapy.

Methods: A comprehensive review was conducted of 436 articles published in "Corriere della Sera" between 1895 and 2023, using keywords such as "radiotherapy" and "radiation." The articles were classified into positive, neutral, and negative categories, and the dominant themes and trends were analyzed.

Results: Articles on radiotherapy (RT) have significantly increased since year 2000, with a notable rise in negative publications focused on toxicities and alleged malpractice. Out of 436 articles, 74 were negative, with this trend growing in recent years, emphasizing risks over benefits in media coverage of RT.

Conclusions: The influence of media on public perception of RT is significant and influences clinical and therapeutic decisions. It is essential that the RT community continues working with media and communication professionals to promote accurate information about the benefits and advances of RT for the patients and for healthcare professionals.

Advances in knowledge: This study highlights the importance of accurate media portrayal of RT to improve public understanding of its benefits. Collaboration between radiation oncologists and media can help disseminate positive outcomes and dispel harmful myths to ensure a balanced and informed perception of RT.

Purpose: Magnetic resonance imaging-only (MRI-only) radiotherapy workflow capitalizes on the superior soft tissue contrast of MRI while eliminating computed tomography-MRI (CT-MRI) registration uncertainties. Most existing studies focus on optimizing individual components of this workflow, i.e., synthetic CT (sCT) generation, dose calculation accuracy, and the reliability of MRI-based patient positioning, while often neglecting comprehensive evaluation of the whole clinical process and its interdependent technical requirements.

Methods: A total of 55 pelvic cancer patients, who underwent the standard radiotherapy workflow based on CT-MRI co-registration and involving a variety of imaging scanners, treatment planning systems (TPSs), and linear accelerators (LINACs), were included. For each patient, a fully integrated MRI-only approach was developed through a step-by-step evaluation of the different radiotherapy phases, including MRI-alone delineation, sCT generation with dose calculation, and cone beam CT-MRI (CBCT-MRI)-guided positioning. Comparative evaluations with the standard workflow were performed for target segmentation (dice similarity coefficient [DSC] and Hausdorff distance [HD]), dose calculation (DVH parameters and gamma analysis), and daily patient positioning (CBCT-CT versus CBCT-MRI registrations).

Results: The proposed MRI-only workflow is applicable throughout all radiotherapy phases and compatible across multiple imaging systems, treatment planning systems, and linear accelerators. MRI-only delineation showed excellent agreement for the prostate (DSC = 0.948 ± 0.028) and cervix (DSC = 0.940 ± 0.015), whereas contour agreement for the rectum and prostate bed (DSC = 0.696 ± 0.037 and 0.689 ± 0.047, respectively) fell within the range of inter-observer variability. Dosimetric comparisons revealed no significant differences between CT- and sCT-based plans (p > 0.05), with median gamma pass rates of 99.0% and 97.9% for the whole body using 3%/3 mm and 2%/2 mm criteria, respectively. CBCT-MRI registration indicated positioning errors comparable to CBCT-CT.

Conclusion: MRI-only pelvic radiotherapy workflows are clinically feasible through customized protocols for target and OAR segmentation, MRI scanner setup and immobilization, sequence selection, sCT-based dose calculation validation, and CBCT-MRI matching reliability.

Objective: To evaluate breast cancers missed by artificial intelligence-based computer-aided diagnosis (AI-CAD) in women newly diagnosed with breast cancer, identify factors associated with these missed cases, and assess the potential diagnostic value of standalone diffusion-weighted imaging (DWI) in detecting cancers overlooked by AI-CAD.

Materials and methods: This retrospective study included 414 women (mean age, 55.3 years) with pathologically confirmed breast cancer who underwent preoperative mammography, MRI with DWI, and surgery. Cancers were classified as AI-detected if the lesion had an abnormality score greater than 10 and was correctly localized by AI-CAD; otherwise, they were categorized as AI-missed. Clinicopathologic and imaging features were compared between groups. Two radiologists independently reviewed DWI of AI-missed cancers and assigned malignancy confidence scores using a 6-point Likert-type scale (≥3 considered positive). Interobserver agreement and diagnostic performance were analyzed.

Results: AI-CAD missed 127 of 414 breast cancers (30.7%). Multivariate regression analysis identified dense breasts (adjusted OR = 1.619; p = 0.049) and tumor size ≤ 2 cm (adjusted OR = 4.698; p < 0.001) as independent predictors of AI-missed cancer. Standalone DWI detected 83.5% and 79.5% of AI-missed cancers for Radiologists 1 and 2, respectively, with substantial agreement (κ = 0.61). DWI was effective in detecting mammographically occult or >1 cm tumors, but sensitivity declined for subcentimeter lesions.

Conclusion: Standalone DWI detects the majority of breast cancers missed by AI-CAD, supporting its potential role as a triage adjunct in AI-based screening, particularly for dense breasts and mammographically occult lesions. However, the retrospective, cancer-only design limits generalizability, highlighting the need for prospective multicenter screening trials for validation.

Purpose: To compare the diagnostic performance of radiomic features extracted from T2-weighted and arterial-phase T1-weighted MRI sequences using univariate, machine and deep learning analysis and to assess their effectiveness in predicting axillary lymph node (ALN) metastasis in breast cancer patients.

Methods: We retrospectively analyzed MRI data from 100 breast cancer patients, comprising 52 metastatic and 103 non-metastatic lymph nodes. Radiomic features were extracted from T2-weighted and subtracted arterial-phase T1-weighted images. Feature normalization and selection were performed. Various machine learning classifiers, including logistic regression, gradient boosting, random forest, and neural networks, were trained and evaluated. Diagnostic performance was assessed using metrics such as area under the curve (AUC), sensitivity, specificity, and accuracy.

Results: T2-weighted imaging provided strong performance in multivariate modeling, with the neural network achieving the highest AUC (0.978) and accuracy (91.1%), showing statistically significant differences over models. The stepwise logistic regression model also showed competitive results (AUC = 0.796; accuracy = 73.3%). In contrast, arterial-phase T1-weighted imaging features performed better when analyzed individually, with the best univariate AUC reaching 0.787. When multivariate modeling was applied to arterial-phase features, the best-performing logistic regression model achieved an AUC of 0.853 and accuracy of 77.8%.

Conclusion: Radiomic analysis of T2-weighted MRI, particularly through deep learning models like neural networks, demonstrated the highest overall diagnostic performance for predicting metastatic ALNs. In contrast, arterial-phase T1-weighted features showed better results in univariate analysis. These findings support the integration of radiomic features, especially from T2-weighted sequences, into multivariate models to enhance noninvasive preoperative assessment.

Thymomas are the most common primary tumors of the anterior mediastinum, frequently associated with paraneoplastic syndromes like myasthenia gravis. This preliminary study investigated the correlation between radiomic features extracted from venous-phase CT images, histological grading (WHO), and disease staging (Masaoka-Koga and TNM) in patients with thymomas. A total of 37 patients were analyzed, with 107 radiomic features extracted using PyRadiomics module. Statistical analysis revealed 11 significant radiomic features distinguishing early and advanced thymomas according to Masaoka-Koga/TNM staging (p < 0.05), with shape_Sphericity, shape_Maximum3DDiameter, and firstorder_Skewness being the most predictive. For WHO classification, 7 significant features differentiated low-risk and high-risk thymomas (p < 0.05), with shape_Sphericity, firstorder-Range, and firstorder_RootMeanSquared showing the highest performance. LASSO models demonstrated high accuracy, with an AUC of 0.9 for Masaoka-Koga/TNM staging and 0.82 for WHO classification. These findings suggest that radiomic features can effectively distinguish thymoma stages and risk levels, potentially aiding in treatment planning and prognosis. By enabling noninvasive tumor characterization, radiomic features could support more personalized treatment strategies and improve decision-making in clinical practice.