European Radiology最新文献

Objectives: The purpose of this study was to establish a predictive deep learning (DL) model for clinically significant portal hypertension (CSPH) based on a single cross-sectional non-contrast CT image and to compare four representative positional images to determine the most suitable for the detection of CSPH.

Materials and methods: The study included 421 patients with chronic liver disease who underwent hepatic venous pressure gradient measurement at our institution between May 2007 and January 2024. Patients were randomly classified into training, validation, and test datasets at a ratio of 8:1:1. Non-contrast cross-sectional CT images from four target areas of interest were used to create four deep-learning-based models for predicting CSPH. The areas of interest were the umbilical portion of the portal vein (PV), the first right branch of the PV, the confluence of the splenic vein and PV, and the maximum cross-section of the spleen. The models were implemented using convolutional neural networks with a multilayer perceptron as the classifier. The model with the best predictive ability for CSPH was then compared to 13 conventional evaluation methods.

Results: Among the four areas, the umbilical portion of the PV had the highest predictive ability for CSPH (area under the curve [AUC]: 0.80). At the threshold maximizing the Youden index, sensitivity and specificity were 0.867 and 0.615, respectively. This DL model outperformed the ANTICIPATE model.

Conclusion: We developed an algorithm that can predict CSPH immediately from a single slice of non-contrast CT, using the most suitable image of the umbilical portion of the PV.

Key points: Question CSPH predicts complications but requires invasive hepatic venous pressure gradient measurement for diagnosis. Findings At the threshold maximizing the Youden index, sensitivity and specificity were 0.867 and 0.615, respectively. This DL model outperformed the ANTICIPATE model. Clinical relevance This study shows that a DL model can accurately predict CSPH from a single non-contrast CT image, providing a non-invasive alternative to invasive methods and aiding early detection and risk stratification in chronic liver disease without image manipulation.

Objectives: This study compared three-dimensional myocardial extracellular volume (ECV) quantification among single-phase and dual-phase photon-counting detector computed tomography (PCD-CT) and dual-phase energy-integrating detector computed tomography (EID-CT), using cardiac magnetic resonance (CMR) imaging as the reference.

Materials and methods: This retrospective study included 80 patients who underwent both CMR and cardiac CT (40 patients underwent PCD-CT and 40 underwent EID-CT). Pearson correlation coefficients and intraclass correlation coefficients (ICCs) were used to evaluate the correlation and reliability of CT-ECV with CMR-ECV. Subgroup analysis was performed based on heart rate.

Results: The single-phase and dual-phase ECV measurements on PCD-CT showed excellent correlation and consistency with CMR (correlation coefficient: single-phase, 0.84; dual-phase, 0.91; ICC: single-phase, 0.90; dual-phase, 0.94). Both methods significantly outperformed EID-CT (correlation coefficient: 0.62; ICC: 0.76; all p < 0.05). No significant differences were observed between the two PCD-CT methods (all p > 0.05). PCD-CT demonstrated high consistency across both HR subgroups (ICC: HR ≤ 60 bpm: 0.89 for single-phase, 0.85 for dual-phase; HR > 60 bpm: 0.87 for single-phase, 0.88 for dual-phase), whereas EID-CT exhibited reduced consistency in patients with HR > 60 bpm (ICC: HR ≤ 60 bpm: 0.78; HR > 60 bpm: 0.60).

Conclusion: Compared with EID-CT, PCD-CT showed superior correlation and reliability for myocardial ECV quantification, particularly in patients with elevated heart rates. The single-phase PCD-CT method demonstrated performance equivalent to the dual-phase approach.

Key points: Question How does myocardial ECV quantification using PCD-CT compare with conventional EID-CT? Findings Single-phase and dual-phase methods on PCD-CT significantly outperform EID-CT in myocardial ECV quantification. Clinical relevance These findings suggest PCD-CT significantly outperforms EID-CT as a promising alternative to CMR for myocardial ECV assessment, particularly under challenging conditions such as higher heart rates, offering improved reliability for clinical myocardial tissue characterization.

Objectives: Percutaneous ablation is a treatment modality used in the management of primary and metastatic lung malignancies. Factors affecting local control rate (LCR) following microwave ablation (MWA) are poorly understood due to inconsistent and limited reporting of histology and procedure-related outcomes.

Materials and methods: Single-center retrospective study of patients with primary and metastatic lung malignancies who underwent MWA between January 2010 and July 2021. Patient, tumor and ablation characteristics were recorded. Outcomes evaluated included technical success, local tumor control and adverse events (AEs).

Results: A total of 669 nodules were identified from 383 patients (197 male/186 female with a median age of 59 years (IQR = 21 years)) across 459 ablation sessions. This comprised 83 primary and 586 metastatic lung tumors. One-year and two-year local LCRs (±SD) across the cohort were 90.3% (±1.1%) and 84.7% (±1.4%), respectively. Multivariable analysis identified ablation power settings, age, concomitant chemotherapy and histology as significant covariates associated with time to local recurrence (TTLR). Standard power was associated with longer TTLR, compared to low power, with no significant difference in AE rates. A significant difference in TTLR was noted between different histological groups.

Conclusion: MWA of primary and metastatic lung tumors is safe and effective. LCRs differ across different histologies. Standard power MWA is associated with longer TTLR without increased AEs rates.

Key points: Question Reported local control rates for microwave ablation (MWA) vary widely, with limited and heterogeneous data on histology-specific and parameter-related outcomes, leaving a gap in understanding. Findings MWA achieved 90.3% and 84.7% one- and two-year local control rates, with longer time to recurrence for standard power MWA. Histology-specific differences noted in time to recurrence. Clinical relevance MWA is a safe and effective treatment for primary and secondary lung cancers, offering comparable local control rates to radiotherapy and surgical resection. Further refinement of treatment parameters may enhance outcomes and address histology-specific challenges.

Objectives: To address the unmet clinical need for validated risk stratification tools in salvage CT-guided percutaneous lung biopsy (PNLB) following nondiagnostic transbronchial lung biopsy (TBLB). We aimed to develop machine learning models predicting severe adverse events (SAEs) in PNLB (Model 1) and diagnostic success of salvage PNLB post-TBLB failure (Model 2).

Materials and methods: This multicenter predictive modeling study enrolled 2910 cases undergoing PNLB across two centers (Center 1: n = 2653 (2016-2020); Center 2: n = 257 (2017-2022)) with complete imaging and clinical documentation meeting predefined inclusion and exclusion criteria. Key variables were selected via LASSO regression, followed by development and validation of Model 1 (incorporating sex, smoking, pleural contact, lesion size, and puncture depth) and Model 2 (including age, lesion size, lesion characteristics, and post-bronchoscopic pathological categories (PBPCs)) using ten machine learning algorithms. Model performance was rigorously evaluated through discrimination metrics, calibration curves, and decision curve analysis to assess clinical applicability.

Results: A total of 2653 and 257 PNLB cases were included from two centers, where Model 1 achieved external validation ROC-AUC 0.717 (95% CI: 0.609-0.825) and PR-AUC 0.258 (95% CI: 0.0365-0.708), while Model 2 exhibited ROC-AUC 0.884 (95% CI: 0.784-0.984) and PR-AUC 0.852 (95% CI: 0.784-0.896), with XGBoost outperforming other algorithms.

Conclusion: The dual XGBoost system stratifies salvage PNLB candidates by quantifying SAE risks (AUC = 0.717) versus diagnostic yield (AUC = 0.884), addressing the unmet need for personalized biopsy pathway optimization.

Key points: Question Current tools cannot quantify severe adverse event (SAE) risks versus salvage diagnostic success for CT-guided lung biopsy (PNLB) after failed transbronchial biopsy (TBLB). Findings Dual XGBoost models successfully predicted the risks of PNLB SAEs (AUC = 0.717) and diagnostic success post-TBLB failure (AUC = 0.884) with validated clinical stratification benefits. Clinical relevance The dual XGBoost system guides clinical decision-making by integrating individual risk of SAEs with predictors of diagnostic success, enabling personalized salvage biopsy strategies that balance safety and diagnostic yield.

Objectives: Irreversible electroporation (IRE) is a relatively recent non-thermal ablation technique for hepatocellular carcinoma (HCC) in high-risk anatomical locations, but post-procedural recurrence limits long-term survival. This study aimed to investigate the risk factors, patterns, and outcomes of post-IRE recurrence.

Materials and methods: This study retrospectively analyzed 180 patients with solitary HCC (≤ 5 cm) who underwent IRE at five centers. Recurrence was categorized as local tumor progression (LTP), intrahepatic distant recurrence (IDR), aggressive intrasegmental recurrence (AIR), and extrahepatic distant recurrence (EDR). Competing-risk regression models and dynamic risk curves were utilized to assess risk factors and temporal trends.

Results: Over a mean follow-up of 40.9 ± 12.6 months, 50.6% of patients developed recurrence. IDR was the predominant pattern (41.7%), followed by LTP (10.0%), EDR (8.3%), and AIR (1.7%). Multivariate analysis identified cirrhosis and tumor size as predictors of LTP; recurrent HCC and cirrhosis as predictors of IDR; and the male sex, cirrhosis, and tumor size as predictors of EDR (all p < 0.05). Only EDR was an independent risk factor for worse overall survival (p < 0.001). Recurrence risk demonstrated a bimodal pattern, with peaks at 21 and 33 months.

Conclusion: IDR is the dominant post-IRE recurrence pattern, but only EDR predicts survival. The identified risk factors and bimodal recurrence peaks contribute to guiding stratified therapeutic strategies for HCC in high-risk locations.

Key points: Question Current knowledge of post-ablation recurrence patterns for hepatocellular carcinoma is primarily derived from thermal ablation, lacking data for non-thermal irreversible electroporation. Findings Intrahepatic distant recurrence was the most common recurrence pattern, while only extrahepatic distant recurrence predicted worse survival. Clinical relevance Irreversible electroporation provided excellent local control for high-risk hepatocellular carcinoma but struggled with recurrence. While extrahepatic distant recurrence predicted reduced survival, local or intrahepatic recurrences yielded favorable outcomes with salvage therapy.

Objectives: Biodegradable biliary stents are used to treat benign biliary strictures in adults. However, there is limited data regarding their use in pediatric patients. This study aims to assess the efficacy and safety of biodegradable biliary stents following pediatric liver transplantation (pLT).

Materials and methods: Consecutive pLT patients with benign biliary strictures were retrospectively evaluated in five tertiary centers between October 2014 and May 2024. All patients underwent percutaneous bilioplasty followed by the placement of self-expanding polydioxanone-based stents. Stricture features and treatment timing were assessed, as well as freedom from stricture recurrence and complications.

Results: A total of 102 patients (52 males, 50 females; median age at treatment = 5 years, interquartile range (IQR) = 2-11 years) were included. At baseline, 58/102 (57%) had a stricture length ≥ 10 mm, and 53/102 (52%) had intrahepatic duct involvement. Stenting was performed a median of 55 days (IQR = 15-128 days) following biliary drainage. Technically successful stent placement was achieved in 101/102 (99%) cases, and low-grade complications occurred in 19/102 (19%). During a median follow-up of 793 days (IQR = 341-1529 days), 24/102 (24%) patients had stricture recurrence with an estimated median time to recurrence of 2915 days. Eight patients were lost to follow-up before recurrence. Logistic regression did not identify any factors that were independent predictors of stricture recurrence, while Cox regression showed that recoil/residual stenosis was independently associated with earlier recurrence (p = 0.005; HR = 5.334; 95% CI = 1.674-17).

Conclusion: Biodegradable stents appear safe and effective for treating biliary strictures after pLT. Certain factors may contribute to failure and should inform patient selection and the optimal timing for stenting.

Key points: Question The management of biliary strictures after pediatric liver transplantation (pLT) poses several challenges due to the frailty of the population and refractoriness of the condition. Findings In this multicenter study self-expanding biodegradable biliary stents made of polydioxanone proved to be safe and effective for the treatment of biliary strictures after pLT. Clinical relevance Biodegradable biliary stents may improve the management of biliary strictures after pLT with a positive impact on outcomes and less invasiveness; some factors may help to predict the outcome and define the best timing for stenting.

Objectives: To evaluate the potential of LLMs to generate sequence-level brain MRI protocols.

Materials and methods: This retrospective study employed a dataset of 150 brain MRI cases derived from local imaging request forms. Reference protocols were established by two neuroradiologists. GPT-4o, o3-mini, DeepSeek-R1 and Qwen2.5-72B were employed to generate brain MRI protocols based on the case descriptions. Protocol generation was conducted (1) with additional in-context learning involving local standard protocols (enhanced) and (2) without additional information (base). Additionally, two radiology residents independently defined MRI protocols. The sum of redundant and missing sequences (accuracy index) was defined as performance metric. Accuracy indices were compared between groups using paired t-tests.

Results: The two neuroradiologists achieved substantial inter-rater agreement (Cohen's κ = 0.74). o3-mini demonstrated superior performance (base: 2.65 ± 1.61; enhanced: 1.94 ± 1.25), followed by GPT-4o (base: 3.11 ± 1.83; enhanced: 2.23 ± 1.48), DeepSeek-R1 (base: 3.42 ± 1.84; enhanced: 2.37 ± 1.42) and Qwen2.5-72B (base: 5.95 ± 2.78; enhanced: 2.75 ± 1.54). o3-mini consistently outperformed the other models with a significant margin. All four models showed highly significant performance improvements under the enhanced condition (adj. p < 0.001 for all models). The highest-performing LLM (o3-mini [enhanced]) yielded an accuracy index comparable to residents (o3-mini [enhanced]: 1.94 ± 1.25, resident 1: 1.77 ± 1.29, resident 2: 1.77 ± 1.28).

Conclusion: Our findings demonstrate the promising potential of LLMs in automating brain MRI protocoling, especially when augmented through in-context learning. o3-mini exhibited superior performance, followed by GPT-4o.

Key points: QuestionBrain MRI protocoling is a time-consuming, non-interpretative task, exacerbating radiologist workload. Findingso3-mini demonstrated superior brain MRI protocoling performance. All models showed notable improvements when augmented with local standard protocols. Clinical relevanceMRI protocoling is a time-intensive, non-interpretative task that adds to radiologist workload; large language models offer potential for (semi-)automation of this process.

Objectives: To evaluate diagnostic delays in lung cancer among patients with interstitial lung disease (ILD), assess tumor characteristics and outcomes, and identify radiologic factors contributing to missed or delayed detection.

Materials and methods: In this retrospective study, 54 ILD patients with concurrent lung cancer from two tertiary ILD referral centers were reviewed. A panel of four thoracic radiologists determined the earliest callable date-when the majority would have raised concern about a nodule-compared to the date of first clinical suspicion. Tumor size, morphology, staging, growth rate (volume doubling time), treatment regimens, and survival were analyzed. For delayed cases, radiologic causes of missed detection were categorized by consensus.

Results: Diagnostic delays occurred in 50% of cases (27/54), with a mean delay of 3.18 years (IQR: 0.94-4.51). Delayed cases showed significantly slower tumor growth (mean volume doubling time: 1.42 vs. 0.22 years, p < 0.0001). The most common radiologic factors contributing to missed detection were obscuration by fibrosis, ground-glass opacity, or bronchovascular bundles (63%), misinterpretation as infection or exacerbation (15%), minimal interval growth (11%), and invisibility on CT (15%). Tumor size, morphology, stage, and histologic subtype-most commonly adenocarcinoma (56%)-were similar between groups. No significant difference in survival was observed between delayed and non-delayed cases (χ² = 0.4, p = 0.5).

Conclusion: Half of the lung cancer cases in ILD patients were diagnosed with significant delay, primarily due to parenchymal masking and subtle imaging findings, rather than biologically aggressive disease. Despite delays, survival outcomes and tumor characteristics did not differ significantly, suggesting a less aggressive tumor phenotype in delayed cases.

Key points: Question To evaluate diagnostic delays in lung cancer among patients with interstitial lung disease (ILD), assess tumor characteristics and outcomes, and identify radiologic factors contributing to missed or delayed detection. Findings Half of lung cancer cases in ILD patients had diagnostic delays, mainly due to parenchymal masking. Delayed cases showed slower growth but no survival difference; tumor size and histology were indifferent. Clinical relevance This study shows that lung cancer diagnosis is often delayed in ILD patients (avg. > 3 years). Clinicians should stay alert for malignancy in peripheral fibrotic regions with slow-growing nodules. Despite no clear survival impact, future work should assess the risks and benefits of earlier detection.

Objectives: To assess the effects of a bisacodyl micro-enema on rectal physio-morphology, image quality, and patient acceptance when used as preparation before MRI of rectal tumors.

Materials and methods: In this prospective single-center study, we made an intra-individual comparison of patients with suspected early rectal cancer who all completed MRI without (MRex) and with (MRin) a bisacodyl micro-enema. The width and the anal-oral extension of the submucosal edema in the healthy rectal wall, anal-oral extension of luminal fluid, and rectal distension were registered. Image artifacts were recorded with a figurative visual analog scale (fVAS), and patient acceptance with questionnaires at baseline, after MRex, and MRin. Significance levels were calculated with a t-test, the Wilcoxon signed-rank test, or McNemar's test for paired samples and inter-reader agreement with Gwets AC1 statistics.

Results: Consecutively, 50 patients (mean age, 65 years ± 10, 26 men) were included. The median (iqr) summed width of submucosa (MRex, 26 (12) mm, MRin 39 (13) mm, p < 0.01) and anal-oral fluid extension (MRex 0 (0) mm, MRin 134 (91) mm, p < 0.001) increased. Susceptibility artifacts on diffusion weighted imaging (axial plane p ≤ 0.04 and parallel plane p < 0.001) were reduced. No increase in rectal distension, motion artifacts, or side effects was registered, and the patient-reported acceptance remained unchanged.

Conclusion: Bisacodyl significantly increased the submucosal width and extension of intraluminal fluid, did not lead to distension of the rectum, improved image quality, and was well tolerated by the patients.

Key points: Question Bisacodyl micro-enema can improve MRI of early stage-rectal tumors, but the impact on rectal physio-morphology, image quality, and patient acceptance is unknown. Result Bisacodyl micro-enema increased the submucosal width and intraluminal fluid, did not distend the rectum, improved image quality, and was well tolerated by the patients. Clinical relevance Bisacodyl micro-enema represents a reliable preparation before MRI of rectal tumors and can improve imaging of early-stage cancer. This is increasingly important because of the shift toward earlier stages due to screening programs and the growing use of organ-saving treatment.