European Radiology最新文献

Objectives: To conduct a meta-analysis of the diagnostic performance of non-contrast magnetic resonance pulmonary angiography (NC-MRPA) and ventilation-perfusion (V/Q) scintigraphy for the detection of acute pulmonary embolism (PE).

Materials and methods: Systematic searches of electronic databases were conducted from 2000 to 2024. Primary outcomes were per-patient sensitivity and specificity of NC-MRPA and V/Q scintigraphy. The pooled sensitivities, specificities, and 95% confidence intervals (95% CI) were calculated using a random-effect analysis. Summary receiver-operating characteristic (SROC) curves and the area under the curve (AUC) were obtained.

Results: A total of 3709 studies (1941 NC-MRPA studies) were identified through systematic searches, with eight published MRI and nine published V/Q investigations meeting inclusion criteria. The results showed that NC-MRPA had a pooled sensitivity of 0.88 (95% CI: 0.83-0.91) and specificity of 0.97 (95% CI: 0.93-0.98), yielding an AUC of 0.92 (95% CI: 0.85-0.96). V/Q scanning had a pooled sensitivity of 0.81 (95% CI: 0.76-0.85) and specificity of 0.84 (95% CI: 0.74-0.91), yielding an AUC of 0.87 (95% CI: 0.75-0.91). The pooled proportion of non-diagnostic tests for V/Q scans (34.7%, 95% CI: 30.8-38.7) was greater than that of NC-MRPA studies (3.31%, 95% CI: 1.65-4.97).

Conclusion: This meta-analysis suggests that NC-MRPA is more specific than V/Q scintigraphy for the detection of PE, with comparable accuracy and sensitivity. NC-MRPA yielded fewer non-diagnostic scans than V/Q scintigraphy and is a feasible alternative imaging modality for diagnosing PE in patients for whom intravenous contrast administration poses a substantive risk.

Key points: Question V/Q lung scintigraphy has been used as a reserve, alternative modality for patients who cannot undergo CT pulmonary angiography. Findings Non-contrast MR angiography (MRA) is a feasible alternative for diagnosing PE in patients for whom intravenous iodinated contrast administration poses a substantial risk. Clinical relevance Non-contrast MRA provides similar sensitivity and superior specificity to V/Q scintigraphy for diagnosing PE, without ionizing radiation exposure.

Objectives: Adenomatous colorectal polyps require endoscopic resection, as opposed to non-adenomatous hyperplastic colorectal polyps. This study aims to evaluate the effect of artificial intelligence (AI)-assisted differentiation of adenomatous and non-adenomatous colorectal polyps at CT colonography on radiologists' therapy management.

Materials and methods: Five board-certified radiologists evaluated CT colonography images with colorectal polyps of all sizes and morphologies retrospectively and decided whether the depicted polyps required endoscopic resection. After a primary unassisted reading based on current guidelines, a second reading with access to the classification of a radiomics-based random-forest AI-model labelling each polyp as "non-adenomatous" or "adenomatous" was performed. Performance was evaluated using polyp histopathology as the reference standard.

Results: 77 polyps in 59 patients comprising 118 polyp image series (47% supine position, 53% prone position) were evaluated unassisted and AI-assisted by five independent board-certified radiologists, resulting in a total of 1180 readings (subsequent polypectomy: yes or no). AI-assisted readings had higher accuracy (76% +/- 1% vs. 84% +/- 1%), sensitivity (78% +/- 6% vs. 85% +/- 1%), and specificity (73% +/- 8% vs. 82% +/- 2%) in selecting polyps eligible for polypectomy (p < 0.001). Inter-reader agreement was improved in the AI-assisted readings (Fleiss' kappa 0.69 vs. 0.92).

Conclusion: AI-based characterisation of colorectal polyps at CT colonography as a second reader might enable a more precise selection of polyps eligible for subsequent endoscopic resection. However, further studies are needed to confirm this finding and histopathologic polyp evaluation is still mandatory.

Key points: Question This is the first study evaluating the impact of AI-based polyp classification in CT colonography on radiologists' therapy management. Findings Compared with unassisted reading, AI-assisted reading had higher accuracy, sensitivity, and specificity in selecting polyps eligible for polypectomy. Clinical relevance Integrating an AI tool for colorectal polyp classification in CT colonography could further improve radiologists' therapy recommendations.

Objectives: The Scrotal and Penile Imaging Working Group (SPIWG) of the European Society of Urogenital Radiology (ESUR) aimed to formulate recommendations on the imaging modalities and minimal technical requirements for abdominopelvic imaging in the follow-up of adult patients treated for testicular germ-cell tumors (TGCT).

Methods: The SPIWG members performed an extensive literature search, reviewed the current clinical practice, and reached a consensus based on the opinions of experts in the field.

Results: Recurrence in patients treated for TGCT mainly occurs in retroperitoneal lymph nodes (LNs). Abdominopelvic CT and MRI are equivalent assessing retroperitoneal LNs. MRI has the advantage of avoiding radiation exposure, and moreover, diffusion-weighted images (DWI) may increase the detection rates without the need for contrast administration. In patients treated for stage I TGCT, the ESUR-SPIWG recommends MRI over CT for the detection of retroperitoneal LNs during the follow-up after treatment. CT, however, remains the follow-up imaging of choice in patients with advanced disease. When MRI is used, the recommended minimal requirements are at least one high-quality anatomical sequence (T1-WI or T2-WI) in axial and coronal planes, and DWI in the same axial plane, ≤ 4 mm contiguous slices from the diaphragm to the perineum. When CT is used, the recommended minimal requirement is a standard-dose contrast-enhanced CT in the portal-venous phase, scanned from the diaphragm to the perineum.

Conclusions: In this paper, the ESUR-SPIWG provides recommendations on the imaging modalities and minimal technical requirements for abdominopelvic imaging in the follow-up of adult patients treated for TGCT.

Key points: Question There are no recommendations on the preferred imaging modality or scan sequences required for abdominopelvic imaging in the follow-up after treatment for testicular cancers. Findings The European Society of Urogenital Radiology Scrotal and Penile Imaging Working Group (ESUR-SPIWG) provides recommendations for abdominopelvic imaging in follow-up after treatment for testicular cancers. Clinical relevance Recurrence of testicular germ-cell tumors mainly occurs in retroperitoneal lymph nodes. Both CT and MRI provide similar morphological assessments, but radiation exposure can be avoided by using MRI instead of CT.

Objectives: Forensic age estimation from orthopantomograms (OPGs) can be performed more quickly and accurately using convolutional neural networks (CNNs), making them an ideal extension to standard forensic age estimation methods. This study evaluates improvements in forensic age prediction for children, adolescents, and young adults by training a custom CNN from a previous study, using a larger, diverse dataset with a focus on dental growth features.

Methods: 21,814 OPGs from 13,766 individuals aged 1 to under 25 years were utilized. The custom CNN underwent 1000 epochs of training and validation using 16,000 and 4000 OPGs, respectively. The best model was chosen by the least mean absolute error (MAE) and evaluated with an additional test dataset of 1814 independent OPGs. Furthermore, the CNN was applied to OPGs from 15 available forensic age estimations conducted by experts certified by the Study Group on Forensic Age Diagnostics (AGFAD), and the results were compared.

Results: A MAE of 0.93 ± 0.81 years and a mean-signed error (MSE) of -0.06 ± 1.23 years were achieved in the test dataset. 63% of predictions were accurate within 1 year, and 95% within 2.5 years. Results of the CNN were comparable to those obtained by experts, effectively highlighting discrepancies in the reported ages of individuals.

Conclusion: Using a large and diverse dataset along with custom deep learning techniques, forensic age estimation can be significantly improved, often providing predictions accurate to within 1 year. This approach offers a reliable, robust, and objective complement to standard forensic age estimation methods.

Key points: Question The potential of custom convolutional neural networks for forensic age estimation, along with a large, diverse dataset, warrants further investigation, offering valuable support to experts. Findings For 1814 test-orthopantomograms, 63% of predictions were accurate within 1 year and 95% within 2.5 years, similar to expert estimates in 15 forensic cases. Clinical relevance Many individuals' fates depend on accurate age estimation. Forensic age estimation can benefit from applying CNN-based methods to further enhance reliability and accuracy.

Objectives: To develop and validate an ultrasomics-based machine-learning (ML) model for non-invasive assessment of interstitial fibrosis and tubular atrophy (IF/TA) in patients with IgA nephropathy (IgAN).

Materials and methods: In this multi-center retrospective study, 471 patients with primary IgA nephropathy from four institutions were included (training, n = 275; internal testing, n = 69; external testing, n = 127; respectively). The least absolute shrinkage and selection operator logistic regression with tenfold cross-validation was used to identify the most relevant features. The ML models were constructed based on ultrasomics. The Shapley Additive Explanation (SHAP) was used to explore the interpretability of the models. Logistic regression analysis was employed to combine ultrasomics, clinical data, and ultrasound imaging characteristics, creating a comprehensive model. A receiver operating characteristic curve, calibration, decision curve, and clinical impact curve were used to evaluate prediction performance.

Results: To differentiate between mild and moderate-to-severe IF/TA, three prediction models were developed: the Rad_SVM_Model, Clinic_LR_Model, and Rad_Clinic_Model. The area under curves of these three models were 0.861, 0.884, and 0.913 in the training cohort, and 0.760, 0.860, and 0.894 in the internal validation cohort, as well as 0.794, 0.865, and 0.904 in the external validation cohort. SHAP identified the contribution of radiomics features. Difference analysis showed that there were significant differences between radiomics features and fibrosis. The comprehensive model was superior to that of individual indicators and performed well.

Conclusions: We developed and validated a model that combined ultrasomics, clinical data, and clinical ultrasonic characteristics based on ML to assess the extent of fibrosis in IgAN.

Key points: Question Currently, there is a lack of a comprehensive ultrasomics-based machine-learning model for non-invasive assessment of the extent of Immunoglobulin A nephropathy (IgAN) fibrosis. Findings We have developed and validated a robust and interpretable machine-learning model based on ultrasomics for assessing the degree of fibrosis in IgAN. Clinical relevance The machine-learning model developed in this study has significant interpretable clinical relevance. The ultrasomics-based comprehensive model had the potential for non-invasive assessment of fibrosis in IgAN, which helped evaluate disease progress.

Objectives: To assess the diagnostic accuracy, in a validation cohort, of a score based on three CT items, which has shown good performance for predicting ischaemia complicating acute adhesive small-bowel obstruction (SBO).

Methods: This retrospective single-centre study of diagnostic accuracy included consecutive patients admitted for acute adhesive SBO in 2015-2022, who were treated conservatively or underwent surgery within 24 h after CT. The gold standard for ischaemia was an intraoperative diagnosis for operated patients, while the absence of ischaemia was confirmed either by its absence during surgery or by clinical follow-up in patients who did not undergo surgery. Three radiologists independently assessed the three score items, namely, decreased bowel-wall enhancement, diffuse mesenteric haziness, and closed-loop mechanism. Inter-observer agreement was evaluated by computing Fleiss' kappa. The diagnostic performance characteristics of the score were computed.

Results: Of the 164 patients analysed (median age, 70 [57-80] years; 88 [54%] males), 57 (34.8%) had surgery, including 41 (71.9%) with intra-operative evidence of bowel ischaemia, whereas 107 (65.2%) were treated conservatively. A score ≥ 2/3 had a sensitivity of 78% (95% CI: 62-89%), a specificity of 97% (95% CI: 92-99%), a positive predictive value of 89% (95% CI: 74-97%), and a positive likelihood ratio of 24 (95% CI: 9.03-63.79). Adding increased unenhanced bowel-wall attenuation and requiring ≥ 2/4 items did not improve score performance. Fleiss' kappa values indicated moderate to substantial agreement between observers: 0.64 [0.56-0.73] for decreased bowel-wall enhancement, 0.57 [0.48-0.66] for diffuse mesenteric haziness, and 0.68 [0.59-0.76] for closed-loop mechanism.

Conclusions: The results of this external validation study support the reproducibility and good diagnostic performance of the score based on three CT items for predicting bowel ischaemia complicating acute adhesive SBO.

Key points: Question The Millet score with three enhanced CT items for predicting bowel ischaemia complicating acute adhesive SBO has not been assessed in an external validation cohort. Findings Adding "increased unenhanced bowel-wall attenuation" to the "decreased bowel-wall enhancement", "diffuse mesenteric haziness", and "closed-loop mechanism" items did not improve score performance. Clinical relevance In an external validation cohort, a score based on three CT items performed well for predicting ischaemia in patients with acute adhesive SBO and showed acceptable inter-observer agreement. This score may help identify patients for surgery.

Objectives: To compare an MRI-based radiomics signature with the programmed cell death ligand 1 (PD-L1) expression score for predicting immunotherapy response in nasopharyngeal carcinoma (NPC).

Methods: Consecutive patients with NPC who received immunotherapy between January 2019 and June 2022 were divided into training (n = 111) and validation (n = 66) sets. Tumor radiomics features were extracted from pretreatment MR images. PD-L1 combined positive score (CPS) was calculated using immunohistochemistry. The least absolute shrinkage and selection operator (LASSO) algorithm was used for feature selection and radiomics signature construction. Receiver operating characteristic (ROC) curve analysis was performed to assess prediction performance.

Results: A total of eleven radiomics features with the greatest discrimination capability were identified by the LASSO algorithm to construct the radiomics signature. In predicting patients with objective response to immunotherapy, radiomics score (Rd-score) yielded a significantly higher area under the ROC curve than that of CPS in both the training (0.790 vs. 0.645, p = 0.025) and the validation (0.735 vs. 0.608, p = 0.038) sets. Multivariate analysis identified the Rd-score as an independent influencing factor in predicting immunotherapy response (odds ratio = 19.963, p < 0.001). Kaplan-Meier analysis indicated that patients with Rd-score ≥ 0.5 showed longer progression-free survival than patients with Rd-score < 0.5 (log-rank p < 0.01).

Conclusion: An MRI-based radiomics signature demonstrated greater efficacy than the PD-L1 expression score in predicting immunotherapy response in patients with NPC.

Key points: Question How does an MRI-based radiomics signature compare with the programmed cell death ligand 1 expression score for predicting immunotherapy response in nasopharyngeal carcinoma? Findings The MRI-based radiomics signature demonstrated superior predictive value compared with programmed cell death ligand 1 expression score in identifying immunotherapy responders. Clinical relevance MRI-based radiomics are a promising novel noninvasive tool for predicting immunotherapy outcomes in nasopharyngeal carcinoma.

Objective: To investigate the prognostic value of baseline European Association of Nuclear Medicine Research Ltd. (EARL) standardized [¹⁸F]fluorodeoxyglucose positron emission tomography-computed tomography ([¹⁸F]FDG PET-CT) quantitative values for survival and to evaluate cutoff values identified in other studies.

Materials and methods: Pediatric and adolescent patients with high-grade osteosarcoma were included. Baseline [¹⁸F]FDG PET-CT, with EARL-accredited reconstructions, was the standard diagnostic staging procedure. Cox proportional hazard analysis for event-free survival (EFS) and overall survival (OS) was performed with clinical prognostic factors. Kaplan-Meier analysis and log-rank tests were applied to investigate the prognostic performance of the [¹⁸F]FDG PET indices.

Results: In total, 66 patients were included in this study. In the univariable Cox regression analysis, peak lean-body mass corrected SUV (SUL_peak) (hazard ratio (HR): 1.04), total lesion glycolysis (TLG) (HR: 1.0), and metabolic tumor volume (MTV) (HR: 1.0) were not associated with EFS or OS. Log-rank analysis showed a significant difference in EFS for all SUL_max and SUL_peak cutoffs. For MTV_total the maximum Youden, and for TLG_total, the maximum Youden and maximally selected rank cutoff resulted in a significant EFS difference. No cutoff for any measure showed a significant difference in OS between the groups. ROC curves for event status had an AUC of 0.67, 0.66, 0.64, and 0.64 for SUL_max, SUL_peak, MTV_total, and TLG_total, respectively.

Conclusion: In this study, the baseline EARL-standardized [¹⁸F]FDG PET indices of children and adolescents with osteosarcoma were not prognostic of EFS or OS. The proposed prognostic cutoffs from earlier studies suffer from important technical and statistical issues.

Key points: Question Prognostic value of baseline [¹⁸F]FDG PET-CT imaging markers have been reported for histologic response and survival in high-grade osteosarcoma but have not been validated for clinical practice. Findings Baseline SUV_peak, TLG_total, and MTV_total measured on EARL-accredited reconstructions were not prognostic factors for survival in pediatric and adolescent patients with high-grade osteosarcoma. Clinical relevance A wide range of values for SUV_peak and SUV_max cutoffs with similar prognostic value were identified, questioning the value of a single proposed cutoff. Lack of validation, with important technical and statistical issues of proposed prognostic cutoffs, limits clinical implementation.