European Radiology最新文献

Objective: To determine if macroscopic intralesional fat detected in bone lesions on CT by Hounsfield unit (HU) measurement and on MRI by macroscopic assessment excludes malignancy.

Materials and methods: All consecutive CT-guided core needle biopsies (CNB) of non-spinal bone lesions performed at a tertiary center between December 2005 and September 2021 were reviewed. Demographic and histopathology data were recorded. All cases with malignant histopathology were selected, and imaging studies were reviewed. Two independent readers performed CT HU measurements on all bone lesions using a circular region of interest (ROI) to quantitate intralesional fat density (mean HU < -30). MRI images were reviewed to qualitatively assess for macroscopic intralesional fat signal in a subset of patients. Inter-reader agreement was assessed with Cronbach's alpha and intraclass correlation coefficient.

Results: In 613 patients (mean age 62.9 years (range 19-95 years), 47.6% female), CT scans from the CNB of 613 malignant bone lesions were reviewed, and 212 cases had additional MRI images. Only 3 cases (0.5%) demonstrated macroscopic intralesional fat on either CT or MRI. One case demonstrated macroscopic intralesional fat density on CT in a case of metastatic prostate cancer. Two cases demonstrated macroscopic intralesional fat signal on MRI in cases of chondrosarcoma and osteosarcoma. Inter-reader agreement was excellent (Cronbach's alpha, 0.95-0.98; intraclass correlation coefficient, 0.90-0.97).

Conclusion: Malignant lesions rarely contain macroscopic intralesional fat on CT or MRI. While CT is effective in detecting macroscopic intralesional fat in primarily lytic lesions, MRI may be better for the assessment of heterogenous and infiltrative lesions with mixed lytic and sclerotic components.

Clinical relevance statement: Macroscopic intralesional fat is rarely seen in malignant bone tumors and its presence can help to guide the diagnostic workup of bone lesions.

Key points: • Presence of macroscopic intralesional fat in bone lesions has been widely theorized as a sign of benignity, but there is limited supporting evidence in the literature. • CT and MRI are effective in evaluating for macroscopic intralesional fat in malignant bone lesions with excellent inter-reader agreement. • Macroscopic intralesional fat is rarely seen in malignant bone lesions.

Objective: To test the ability of high-performance machine learning (ML) models employing clinical, radiological, and radiomic variables to improve non-invasive prediction of the pathological status of prostate cancer (PCa) in a large, single-institution cohort.

Methods: Patients who underwent multiparametric MRI and prostatectomy in our institution in 2015-2018 were considered; a total of 949 patients were included. Gradient-boosted decision tree models were separately trained using clinical features alone and in combination with radiological reporting and/or prostate radiomic features to predict pathological T, pathological N, ISUP score, and their change from preclinical assessment. Model behavior was analyzed in terms of performance, feature importance, Shapley additive explanation (SHAP) values, and mean absolute error (MAE). The best model was compared against a naïve model mimicking clinical workflow.

Results: The model including all variables was the best performing (AUC values ranging from 0.73 to 0.96 for the six endpoints). Radiomic features brought a small yet measurable boost in performance, with the SHAP values indicating that their contribution can be critical to successful prediction of endpoints for individual patients. MAEs were lower for low-risk patients, suggesting that the models find them easier to classify. The best model outperformed (p ≤ 0.0001) clinical baseline, resulting in significantly fewer false negative predictions and overall was less prone to under-staging.

Conclusions: Our results highlight the potential benefit of integrative ML models for pathological status prediction in PCa. Additional studies regarding clinical integration of such models can provide valuable information for personalizing therapy offering a tool to improve non-invasive prediction of pathological status.

Clinical relevance statement: The best machine learning model was less prone to under-staging of the disease. The improved accuracy of our pathological prediction models could constitute an asset to the clinical workflow by providing clinicians with accurate pathological predictions prior to treatment.

Key points: • Currently, the most common strategies for pre-surgical stratification of prostate cancer (PCa) patients have shown to have suboptimal performances. • The addition of radiological features to the clinical features gave a considerable boost in model performance. Our best model outperforms the naïve model, avoiding under-staging and resulting in a critical advantage in the clinic. •Machine learning models incorporating clinical, radiological, and radiomics features significantly improved accuracy of pathological prediction in prostate cancer, possibly constituting an asset to the clinical workflow.

Objectives: To evaluate the accuracy of combined imaging and blood test indices related to liver fibrosis (LF) compared to magnetic resonance elastography (MRE) for estimating severe LF (F3-4) in preoperative patients.

Methods: This retrospective study included patients who underwent MRE, gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced MRI, and dynamic CT before liver resection. Liver stiffness measurement (LSM) using MRE, liver-to-spleen signal intensity ratio (LSR) using Gd-EOB-DTPA-enhanced MRI, and spleen volume normalized to body surface area (SV/BSA) using CT volumetry were measured. Laboratory parameters, including levels of type IV collagen 7S and hyaluronic acid, were also measured. Logistic regression and receiver operating characteristic analyses were performed to identify parameters that could estimate severe LF more accurately than LSM alone.

Results: A total of 81 patients (mean age, 67 years ± 9.9 [SD]; 58 men) were enrolled. Multivariable logistic regression analysis indicated that LSR (odds ratio [OR]: 0.14, 95% confidence interval [CI]: 0.05-0.37, p < 0.001), SV/BSA (OR: 1.25, 95% CI: 1.02-1.52, p = 0.03) and type IV collagen 7S (OR: 1.84, 95% CI: 1.12-3.00, p = 0.02) were associated with severe LF. Receiver operating characteristic analysis showed that for estimating severe LF, the area under the curve was significantly larger for the combination of LSR, SV/BSA, and type IV collagen 7S than for LSM alone (0.95 vs 0.85, p = 0.04).

Conclusion: The combined evaluation of LSR, SV/BSA, and type IV collagen 7S obtained by clinically common preoperative examinations was more accurate than MRE alone for estimating severe LF in preoperative patients.

Key points: Question What indicators among the imaging and blood tests commonly performed preoperatively can provide a more accurate estimate of severe LF compared to MRE? Findings The combination of LSR, SV/BSA, and type IV collagen 7S was more accurate than an LSM alone for estimating severe LF. Clinical relevance A combination of commonly performed non-invasive preoperative tests provides a more accurate estimation of severe LF than MR elastography, an examination with relatively limited.

Objective: First, to determine the frequency and spectrum of osteoid osteoma (OO)-mimicking lesions among presumed OO referred for radiofrequency ablation (RFA). Second, to compare patient sex and age, lesion location, and rates of primary treatment failure for OO based on histopathology results.

Materials and methods: A retrospective review was performed of all first-time combined CT-guided biopsy/RFA for presumed OO at a single academic center between January 1990 and August 2023. Lesions were characterized as "biopsy-confirmed OO", "OO-mimicking", or "non-diagnostic" based on pathology results. Treatment failure was defined as residual or recurrent symptoms requiring follow-up surgery or procedural intervention. Variables of interest were compared between pathology groups using Kruskal-Wallis, Fisher's exact, and Wilcoxon rank sum tests.

Results: Of 643 included patients (median 18 years old, IQR: 13-24 years, 458 male), there were 445 (69.1%) biopsy-confirmed OO, 184 (28.6%) non-diagnostic lesions, and 15 (2.3%) OO-mimicking lesions. OO-mimicking lesions included chondroblastoma (n = 4), chondroma (n = 3), enchondroma (n = 2), non-ossifying fibroma (n = 2), Brodie's abscess (n = 1), eosinophilic granuloma (n = 1), fibrous dysplasia (n = 1), and unspecified carcinoma (n = 1). OO-mimicking lesions did not show male predominance (46.7% male) like biopsy-proven OO (74.1% male) (p = 0.033). Treatment failure occurred in 24 (5.4%) biopsy-confirmed OO, 8 (4.4%) non-diagnostic lesions, and 2 (13.3%) OO-mimicking lesions without a significant difference by overall biopsy result (p = 0.24) or pairwise group comparison.

Conclusion: OO-mimicking pathology is infrequent, typically benign, but potentially malignant. OO-mimicking lesions do not exhibit male predominance. There was no significant difference in RFA treatment failure or lesion location among lesions with imaging appearances suggestive of OO.

Key points: Question What is the frequency and spectrum of OO-mimicking lesions among presumed OO and what, if any, differences exist between these pathologies? Finding The study cohort included 69.1% OO, 28.6% lesions with non-diagnostic histopathology, and 2.3% OO-mimicking lesions. There was no difference in treatment failure or location among lesions. Clinical relevance Routine biopsy of presumed OO at the time of RFA identifies OO-mimicking lesions, which are rare and likely benign.

Objectives: The leukodystrophy "vanishing white matter" (VWM) and "metachromatic leukodystrophy" (MLD) affect the brain's white matter, but have very different underlying pathology. We aim to determine whether quantitative MRI reflects known neuropathological differences and correlates with clinical scores in these leukodystrophies.

Methods: VWM and MLD patients and controls were prospectively included between 2020 and 2023. Clinical scores were recorded. MRI at 3 T included multi-compartment relaxometry diffusion-informed myelin water imaging (MCR-DIMWI) and multi-echo T2-relaxation imaging with compressed sensing (METRICS) to determine myelin water fractions (MWF). Multi-shell diffusion-weighted data were used for diffusion tensor imaging measures and neurite orientation dispersion and density imaging (NODDI) analysis, which estimates neurite density index, orientation dispersion index, and free water fraction. As quantitative MRI measures are age-dependent, ratios between actual and age-expected MRI measures were calculated. We performed the multilevel analysis with subsequent post-hoc and correlation tests to assess differences between groups and clinico-radiological correlations.

Results: Sixteen control (age range: 2.3-61.3 years, 8 male), 37 VWM (2.4-56.5 years, 20 male), and 14 MLD (2.2-41.7 years, 6 male) subjects were included. Neurite density index and MWF were lower in patients than in controls (p < 0.001). Free water fraction was highest in VWM (p = 0.01), but similar to controls in MLD (p = 0.99). Changes in diffusion tensor imaging measures relative to controls were generally more pronounced in VWM than in MLD. In both patient groups, MCR-DIMWI MWF correlated strongest with clinical measures.

Conclusion: Quantitative MRI correlates to clinical measures and yields differential profiles in VWM and MLD, in line with differences in neuropathology.

Key points: Question Can quantitative MRI reflect known neuropathological differences and correlate with clinical scores for these leukodystrophies? Finding Quantitative MRI measures, e.g., MWF, neurite density index, and free water fraction differ between leukodystrophies and controls, in correspondence to known histological differences. Clinical relevance MRI techniques producing quantitative, biologically-specific, measures regarding the health of myelin and axons deliver more comprehensive information regarding pathological changes in leukodystrophies than current approaches, and are thus viable tools for monitoring patients and providing clinical trial outcome measures.

Objective: Distinguishing post-COVID-19 residual abnormalities from interstitial lung abnormalities (ILA) on CT can be challenging if clinical information is limited. This study aimed to evaluate the diagnostic performance of radiologists in distinguishing post-COVID-19 residual abnormalities from ILA.

Methods: This multi-reader, multi-case study included 60 age- and sex-matched subjects with chest CT scans. There were 40 cases of ILA (20 fibrotic and 20 non-fibrotic) and 20 cases of post-COVID-19 residual abnormalities. Fifteen radiologists from multiple nations with varying levels of experience independently rated suspicion scores on a 5-point scale to distinguish post-COVID-19 residual abnormalities from fibrotic ILA or non-fibrotic ILA. Interobserver agreement was assessed using the weighted κ value, and the scores of individual readers were compared with the consensus of all readers. Receiver operating characteristic curve analysis was conducted to evaluate the diagnostic performance of suspicion scores for distinguishing post-COVID-19 residual abnormalities from ILA and for differentiating post-COVID-19 residual abnormalities from both fibrotic and non-fibrotic ILA.

Results: Radiologists' diagnostic performance for distinguishing post-COVID-19 residual abnormalities from ILA was good (area under the receiver operating characteristic curve (AUC) range, 0.67-0.92; median AUC, 0.85) with moderate agreement (κ = 0.56). The diagnostic performance for distinguishing post-COVID-19 residual abnormalities from non-fibrotic ILA was lower than that from fibrotic ILA (median AUC = 0.89 vs. AUC = 0.80, p = 0.003).

Conclusion: Radiologists demonstrated good diagnostic performance and moderate agreement in distinguishing post-COVID-19 residual abnormalities from ILA, but careful attention is needed to avoid misdiagnosing them as non-fibrotic ILA.

Key points: Question How good are radiologists at differentiating interstitial lung abnormalities (ILA) from changes related to COVID-19 infection? Findings Radiologists had a median AUC of 0.85 in distinguishing post-COVID-19 abnormalities from ILA with moderate agreement (κ = 0.56). Clinical relevance Radiologists showed good diagnostic performance and moderate agreement in distinguishing post-COVID-19 residual abnormalities from ILA; nonetheless, caution is needed in distinguishing residual abnormalities from non-fibrotic ILA.

Objectives: The assessment of lumbar central canal stenosis (LCCS) is crucial for diagnosing and planning treatment for patients with low back pain and neurogenic pain. However, manual assessment methods are time-consuming, variable, and require axial MRIs. The aim of this study is to develop and validate an AI-based model that automatically classifies LCCS using sagittal T2-weighted MRIs.

Methods: A pre-existing 3D AI algorithm was utilized to segment the spinal canal and intervertebral discs (IVDs), enabling quantitative measurements at each IVD level. Four musculoskeletal radiologists graded 683 IVD levels from 186 LCCS patients using the 4-class Lee grading system. A second consensus reading was conducted by readers 1 and 2, which, along with automatic measurements, formed the training dataset for a multiclass (grade 0-3) and binary (grade 0-1 vs. 2-3) random forest classifier with tenfold cross-validation.

Results: The multiclass model achieved a Cohen's weighted kappa of 0.86 (95% CI: 0.82-0.90), comparable to readers 3 and 4 with 0.85 (95% CI: 0.80-0.89) and 0.73 (95% CI: 0.68-0.79) respectively. The binary model demonstrated an AUC of 0.98 (95% CI: 0.97-0.99), sensitivity of 93% (95% CI: 91-96%), and specificity of 91% (95% CI: 87-95%). In comparison, readers 3 and 4 achieved a specificity of 98 and 99% and sensitivity of 74 and 54%, respectively.

Conclusion: Both the multiclass and binary models, while only using sagittal MR images, perform on par with experienced radiologists who also had access to axial sequences. This underscores the potential of this novel algorithm in enhancing diagnostic accuracy and efficiency in medical imaging.

Key points: Question How can the classification of lumbar central canal stenosis (LCCS) be made more efficient? Findings Multiclass and binary AI models, using only sagittal MR images, performed on par with experienced radiologists who also had access to axial sequences. Clinical relevance Our AI algorithm accurately classifies LCCS from sagittal MRI, matching experienced radiologists. This study offers a promising tool for automated LCCS assessment from sagittal T2 MRI, potentially reducing the reliance on additional axial imaging.