European Radiology Experimental最新文献

Objectives: To develop a nomogram based on low-dose one-stop dual-energy and perfusion computed tomography (LD-DE&PCT) for predicting the efficacy of transcatheter arterial chemoembolization (TACE) combined with lenvatinib and immune checkpoint inhibitors (TACE-LEN-ICIs) in unresectable hepatocellular carcinoma (uHCC) patients.

Materials and methods: This prospective, multicenter study included uHCC patients who underwent LD-DE&PCT scanning. The relationships between quantitative LD-DE&PCT-derived parameters and the efficacy of TACE-LEN-ICIs were analyzed using logistic regression analysis. A nomogram incorporating the independent predictors was constructed, and its predictive performance was evaluated by the area under the receiver operating characteristic curve (AUROC).

Results: A total of 125 lesions from 71 uHCC patients were enrolled, with 71 lesions (56.8%) classified as the objective response (ObR) group and 54 lesions (43.2%) as the non-response (NR) group. Univariate analysis revealed significant differences in tumor size, corona enhancement, tumor location, iodine concentration in the arterial phase (IC-AP), normalized iodine concentration in the arterial phase (NIC-AP), effective atomic number in the arterial phase (Z_eff-AP), slope of spectral HU curve in the arterial phase (λ_HU-AP), and permeability surface area product (PS) between ObR and NR groups. Among these, NIC-AP exhibited the highest predictive value (AUROC = 0.770; 95% confidence interval [CI]: 0.682‒0.858). Multivariate analysis identified tumor size, NIC-AP, and PS as independent predictors. The nomogram showed excellent performance (AUROC  = 0.913; 95% CI: 0.858-0.968). The total radiation dose was 19.02 ± 5.39 mSv.

Conclusion: The LD-DE&PCT-based nomogram can accurately predict the response to TACE-LEN-ICIs in uHCC patients.

Relevance statement: Low-dose one-stop dual-energy and perfusion CT provides a noninvasive method to predict response to TACE combined with lenvatinib and immune checkpoint inhibitors in unresectable HCC.

Key points: Predicting response to TACE-LEN-ICIs in uHCC helps treatment decision-making. NIC-AP and PS from LD-DE&PCT, and tumor size were independent predictive biomarkers. NIC-AP was the best parameter for predicting response to TACE-LEN-ICIs in uHCC.

Objectives: This study aims to evaluate the actual energy consumption of two generations of 1.5-T magnetic resonance imaging (MRI) scanners, quantify the benefits in terms of primary energy savings resulting from technological replacement, and compare field estimates of primary energy consumption with those reported in environmental product declarations (EPDs).

Materials and methods: Two 1.5-T MRI scanner models, the old model version and its new model replacement, were monitored using a power quality analyzer connected to the electrical cabinet. Electrical power consumption data were collected over 2-week periods, both before and after the scanner replacement. Primary energy consumption was projected over 10 years, and the resulting values were compared with those reported in the EPDs for the two scanners.

Results: Over 10 years, cumulative energy consumption is estimated to be 1,010.4 MWh for the new unit versus 1,206.7 MWh for the old unit, corresponding to a 16.3% reduction. Considering the range of European primary energy factors (PEFs), energy savings varied from 235.6 to 687.1 MWh. Comparison with EPDs revealed significant discrepancies (± 40%) depending on the national PEF used, demonstrating that EPDs can both overestimate and underestimate actual energy consumption.

Conclusion: Replacement of an old MRI model resulted in measurable energy savings, particularly in non-productive phases. However, EPDs do not always reflect clinical operation or the impact of national energy mixes. While energy efficiency is central to sustainable radiology, it should not be the sole driver for equipment replacement, which must remain primarily guided by clinical and diagnostic criteria.

Relevance statement: For a radiology department focused on more sustainable practices, it is essential to have accurate data on the environmental performance of medical imaging equipment, which should not be based solely on EPDs, but on real data based on usage patterns and national energy mixes.

Key points: Replacing the old MRI scanner reduced energy consumption by 16.3%, mainly due to lower use in non-productive modes. Over a 10-year operational period, the primary energy consumption savings varied from 235.6 to 687.1 MWh. A discrepancy emerged between EPD-reported and real-world measurements, highlighting the importance of on-site validation for sustainability assessments.

Objectives: This study investigated the influence of hepatic vessels on the quantification of magnetic resonance imaging (MRI) proton density fat fraction (PDFF) and R2* using automated whole-liver segmentation.

Materials and methods: This prospective multicenter study included patients with chronic liver disease having paired liver biopsy and MR exams with a standardized multiecho chemical-shift gradient echo sequence. Automated whole-liver segmentation was performed, generating two masks per patient, one including and the other excluding the major hepatic vessels. PDFF and R2* were quantified and graded for both masks. Histological grading of hepatic steatosis and iron overload severity was used as a reference standard.

Results: A total of 377 patients were evaluated, of whom 54% had hepatic steatosis and 20% had iron overload on biopsy readings. Stratified by histological grades, there were no statistically significant differences in the distribution of PDFF or R2* between the two segmentation masks. Overall, PDFF and R2* values were minimally lower when vessels were included, with a bias of -0.06% for PDFF and -0.25 s^-1 for R2*. A lower coefficient of variation was obtained for both imaging parameters after excluding vessels. Patients were classified in the same PDFF grades despite the segmentation approach, and only 7 cases (1.9% of the study population) were reclassified for R2* grading, all being upgraded after vessel exclusion.

Conclusion: Excluding hepatic vessels entails nonsignificant differences in PDFF and R2* quantification. Although with limited impact, vessel exclusion improves biomarker precision in research settings demanding high accuracy and increases clinicians' confidence when using automatic tools in clinical practice.

Relevance statement: Fat and iron quantification on MRI are key imaging biomarkers for the accurate non-invasive assessment of patients with chronic liver disease. Proton density, fat fraction, and R2* quantification show minimal differences if hepatic vessels are included or excluded from the liver segmentation mask.

Key points: The effect of hepatic vessels on proton density, fat fraction, and R2* quantification was evaluated. No significant differences were found, excluding hepatic vessels, although their inclusion showed a small negative bias. Vessel exclusion may improve clinicians' confidence and precision in high-sensitivity applications.

The latest technological advancements in CT enable the exploration of unprecedented limits of spatial resolution in in vivo imaging. Nowadays, ultra-high-resolution imaging is available by using CT with detector elements at or smaller than 0.25 mm along the z-axis, like those used on photon-counting CT (PCCT) scanners. However, spatial resolution represents a complex criterion of imaging performance affected not only by detector elements, but also by other complex variables that can interact with each other. Knowledge of these variables and the metrics to evaluate spatial resolution is key to performing accurate cardiothoracic examinations with optimized CT protocols, which can eventually reduce acquisition times and radiation doses. This opens to a sustainable cardiothoracic radiology that permits accurate cardiac CT evaluations also in patients previously excluded, due to high calcium score, metallic stents or obesity, and allows to reduce radiation doses to never-seen levels. In this article, we review the technical advancements that allowed such an increase in spatial resolution in PCCT, along with all technical determinants of spatial resolution, the metrics to evaluate it, the clinical impact of UHR at PCCT and its challenges on cardiothoracic imaging. RELEVANCE STATEMENT: Knowledge of the ultra-high spatial resolution capabilities of new photon-counting CT technology is key to its best uses - performing accurate diagnostic examinations at unmatched low radiation doses and scanning patients previously excluded from cardiac CT examinations. KEY POINTS: Photon-counting CT scanners enable radiologists to evaluate cardiothoracic examinations with an exceptional spatial resolution, along with spectral information and dose reduction. The ultra-high spatial resolution in cardiovascular imaging enables accurate assessment in patients previously excluded, such as obese, those with extensive calcifications or metallic stents. Ultra-high spatial resolution empowers the visualization and follow-up of focal and diffuse lung diseases at unmatched accuracy and low radiation doses.

Supine breast MRI offers easier comparison with other screening methods than the prone MRI, but is often impaired by respiratory artifacts. This study aims to identify a suitable acceleration technique for motion-corrected supine breast MRI with a dedicated wearable coil (BraCoil), and to assess the quality of the resulting images. This prospective, single-center study included ten healthy female volunteers. Supine breast MRI was acquired using the BraCoil equipped with seven accelerometers. The acquisition protocol included unenhanced T1- and T2-weighted sequences; 2-, 3-, and 4-fold acceleration, as well as "elliptical scanning", were tested for T1-weighted images. Image reconstruction was performed using two methods: the manufacturer's standard algorithm and an advanced nonrigid motion correction algorithm. The images were compared to each other and to conventional prone MRI via radiological scoring. Additionally, comparisons were made with supine breast MRI obtained using a multi-purpose torso coil reported in previous studies. This study showed that using 2-fold acceleration was optimal for motion-corrected supine breast MRI using the BraCoil. It also demonstrated that enabling the "elliptical scanning" would shorten the measurement time by 33% without compromising image quality. Finally, for the data acquired, the BraCoil provides superior motion-corrected image quality compared to the torso coil. KEY POINTS: BraCoil image quality outperforms the body coil for both uncorrected and motion-corrected images in supine acquisition for T1w and T2w sequences. Enabling the "elliptical scanning" in T1-weighted sequences does not impair image quality while shortening the measurement time by 33% (compared to the same sequence without enabling this option). A 2-fold acceleration for the T1-weighted sequence provides sufficient scan time reduction, whereas higher acceleration would lead to even shorter scan time at the cost of significantly lower image quality.

Objective: Preoperative localization of nonpalpable breast lesions has traditionally been performed by wire-guided localization (WGL). Radiofrequency identification (RFID) tag localization provides a less invasive alternative. The aim of this systematic review and meta-analysis was to investigate outcomes of RFID in terms of clinical utility, efficacy, and safety, also in comparison with WGL.

Materials and methods: Following PRISMA guidelines, studies reporting on outcomes post-RFID tag localization, and comparing outcomes post-RFID tag localization and WGL were included. Positive margins and re-excision rates were estimated using meta-analyses of proportions. Further meta-analyses compared positive margins and re-excision rates between RFID tag localization and WGL. Random effects models were used.

Results: Nineteen studies involving 3,234 patients were included. Localization was performed for 497 benign and 2,741 malignant lesions. No study reported failure to retrieve an inserted RFID tag. Failed localization rates ranged from 0.0 to 60.7% across studies. After RFID tag localization, the pooled rate of positive margins was 12% (95% confidence interval (CI) 10-15%) and the pooled re-excision rate was 13% (95% CI 10-16%) in 14 and 16 studies, respectively; heterogeneity was high, I² = 54.6% and 54.9, respectively. In three comparative studies, RFID tag localization was associated with significantly lower rates of positive margins than WGL, odds ratio (OR) 0.71 (95% CI 0.54-0.95), p = 0.021; however, no difference was observed in re-excision rates, OR 1.13 (95% CI 0.88-1.45), p = 0.346. Heterogeneity was low in both analyses, I² = 0.0%. Moderate bias was reported in 16/19 studies, serious bias in 3/19.

Conclusion: RFID tag localization provides an effective alternative to WGL.

Relevance statement: This systematic review shows that RFID tag localization of nonpalpable breast lesions provides a less invasive, safe and effective alternative to WGL-guided localization for selected patients, considering its higher cost. Randomized trials are required to elucidate the benefit of RFID tag localization over WGL and other non-wire localization techniques.

Key points: The pooled rates of positive margins and re-excision after RFID tag localization were 12% and 13%, respectively. RFID localization had significantly lower positive margin rates than wire-guided localization (WGL); however, no difference was observed in re-excision rate. RFID localization provides an effective alternative to WGL and may be of benefit in selected patients. Randomized trials are required to better elucidate the benefit of RFIS tag localization.

Objective: Cryoablation, a minimally invasive, image-guided procedure, induces tumor necrosis through freezing/thawing cycles. This pilot study investigates the immunomodulatory effects of cryoablation in early breast cancer (BC) patients by analyzing blood and surgical samples, with a focus on T-cell subsets, regulatory T cells (Tregs), serum cytokines, and high-mobility group box 1 (HMGB1) levels.

Materials and methods: Ten patients with early BC (cT1 cN0) underwent ultrasound-guided cryoablation using a cryoablation system followed by surgical resection. Peripheral blood mononuclear cells were isolated at four time points: pre-ablation (T0), day 2-3 (T1), 2-3 weeks post-ablation (T2), and post-surgery (T3). Immune cell populations, including Tregs and activated CD137⁺ T cells, were analyzed via flow cytometry. Serum HMGB1 and cytokines (e.g., IL-1β, IL-6, and TNF-α) were measured using Luminex assays. The histopathological analysis assessed the tumor response to ablation and immune infiltrates.

Results: Cryoablation significantly increased circulating HMGB1 levels at T1 (p = 0.047), with further elevation post-surgery (p = 0.023), suggesting immune activation. CD137⁺ T cells, predominantly the CD4⁺ subset, decreased significantly after surgery (p = 0.025), correlating with reduced interleukin-4 levels. Proliferating Tregs (Ki67⁺) also declined after the combined treatment (p = 0.046). Histopathology confirmed complete tumor ablation in 9 of 10 cases, with immune infiltrates, predominantly CD3⁺ lymphocytes (CD4⁺ and CD8⁺ equally represented).

Conclusion: Cryoablation induces significant immunological changes, including the release of HMGB1, modulation of CD137⁺ T cells, and decreased Treg proliferation, highlighting its potential as both local and systemic immunomodulatory therapy.

Relevance statement: Cryoablation triggers immune activation in early BC, as indicated by increased CD137⁺ T cells, reduced Tregs, elevated HMGB1, enhanced inflammatory cytokine release, and the presence of mild to intense inflammatory infiltrates in surgical samples. These findings suggest the potential efficacy of cryoablation in supporting immunotherapies in the treatment of BC.

Key points: Cryoablation is a promising nonsurgical treatment for early-stage BC. The procedure may induce immune activation by increasing HMGB1 and modulating T-cell populations. Tregs appear to decrease after cryoablation, suggesting immunomodulatory potential. Histopathology confirms effective tumor ablation in most treated patients. Cryoablation shows immunomodulatory effects and may provide a rationale for future combination with immunotherapy.

Background: The metabolic effects of combined metformin and aerobic exercise on skeletal muscle in prediabetes remain unclear. We evaluated individual and combined effects on skeletal muscle metabolism using multiparameter magnetic resonance imaging (MRI) and assessed prediabetes remission.

Materials and methods: In this 12-week randomized controlled trial, forty-two prediabetic adults aged 48.4 ± 13.1 years mean ± standard deviation) were randomized into control (n = 10), metformin (n = 10), exercise (n = 11), and combined therapy (n = 11) groups. Multiparameter MRI measured T2, apparent diffusion coefficient (ADC), fractional anisotropy (FA), intermuscular adipose tissue (IMAT), intramyocellular lipids, visceral adipose tissue (VAT), and muscle cross-sectional areas (MSCAs). Blood biomarkers were assessed by standard protocols. Analyses included ANOVA, Fisher's exact test, and correlation analysis.

Results: Normoglycemia occurred in 90% of the combined group, 80% of the exercise, 20% of the metformin, and 10% of the controls. Exercise alone significantly decreased IMAT%, T2, and ADC, and increased FA, whereas metformin alone had no significant effects. Combination therapy did not further improve blood biomarkers or change MRI parameters versus exercise, but uniquely reduced VAT and increased MSCAs. IMAT% reduction was borderline greater in subjects with high baseline IMAT after combination therapy (p = 0.051). MRI parameters, particularly IMAT%, T2, and FA, correlated with fasting glucose, incremental area under the curve of glucose and insulin, and hemoglobin A1c.

Conclusion: Aerobic exercise favorably altered skeletal muscle tissue characteristics and metabolic markers in prediabetes. Metformin did not attenuate these effects and may enhance IMAT reduction in subjects with high baseline IMAT.

Relevance statement: Multiparameter MRI provides a sensitive, noninvasive means to quantify skeletal muscle composition and function in prediabetes, enabling precise assessment of the metabolic effects of combined lifestyle and pharmacological interventions.

Trial registration: Chinese Clinical Trial Registry (ChiCTR2300072162), retrospectively registered on June 5, 2023.

Key points: Aerobic exercise improves muscle composition and glucose metabolism in prediabetes. Metformin adds benefits for reducing muscle fat in high-fat individuals. Multiparameter MRI enables noninvasive monitoring of skeletal muscle metabolic changes.

Background: Errors in radiology reports can result in inappropriate/harmful decisions. We investigated whether large language models can reduce the error rate.

Materials and methods: We developed the radiology-specific clinical language anomaly recognition engine (RadCLARE) network, an automated engine based on the bidirectional encoder representations from transformers (BERT)-base model, designed to detect semantic errors in Chinese radiology reports and trained using 1.4 million reports, including 615,920 digital radiography, 560,310 computed tomography reports, and 223,480 magnetic resonance reports. One thousand reports were randomly selected for expert manual annotation. Inter-reader agreement for error detection and classification was assessed using Cohen κ and Gwet AC1. The RadCLARE's detection was compared against the expert references. Changes in error rates before (baseline test dataset, BTD) and after (experimental test dataset, ETD) RadCLARE implementation were analyzed. Finally, radiologists were invited to complete questionnaires to evaluate satisfaction and rate the system across five dimensions.

Results: Among the 1,000 reports, a total of 506 errors were identified as the reference standard. Inter-reader agreement was substantial for error detection (κ = 0.77) and excellent for error classification (Gwet AC1 = 0.94). RadCLARE successfully detected 437/506 errors, with 87.3% accuracy, 88.3% precision, 86.4% recall, and 87.4% F1-score. The BTD comprised 571,264 reports, the ETD 873,030 reports. After RadCLARE implementation, the semantic error rate dropped significantly compared to the BTD (error rate, 0.85% [7408/873,030] versus 4.19% [23,909/571,264]; p < 0.001). The questionnaire results showed that 95.7% (44/46) of radiologists were satisfied with RadCLARE.

Conclusion: RadCLARE showed the capability for automatic detection of semantic errors in radiology reports.

Relevance statement: RadCLARE demonstrated high performance in detecting semantic errors in radiology reports. Future studies should aim to extend their applicability across multiple languages and institutions.

Key points: We developed the RadCLARE network, a BERT-based engine for detecting semantic errors in Chinese radiology reports. With the aid of RadCLARE, the semantic error rate in radiology reports dropped significantly from 4.19% to 0.85%. The large majority (96%) of radiologists who participated in the test were satisfied with the RadCLARE and felt that it reduced stress.

The integration of diagnostic imaging with radiation therapy (RT) is evolving into a continuous workflow, significantly advancing personalised oncology care. Recent technological innovations, particularly the incorporation of real-time magnetic resonance imaging (MRI) with linear accelerators, have markedly enhanced RT precision, improving target coverage and reducing radiation exposure to surrounding healthy tissues. Furthermore, real-time MRI enables the collection of quantitative imaging data during each treatment fraction, potentially leading to the identification of quantitative imaging biomarkers. These biomarkers can capture dynamic biological changes during RT, offering unprecedented insights into treatment response. The integration of these imaging biomarkers with clinical, genomic, and pathological data into artificial intelligence (AI)-supported clinical decision support systems promises to further refine therapeutic personalisation. In this context, AI plays a central role by automating labour-intensive tasks, extracting quantitative metrics, and integrating multidimensional data into clinically meaningful predictive models. This review outlines a vision for the future of RT, highlighting how the synergy of advanced imaging, AI, and multidomain data through three logical steps: (1) rethinking and reorganising the patient care journey; (2) from imaging "for" to imaging "with" RT; and (3) incorporation into clinical decision support systems. This integration will support the development of personalised, biologically driven treatment strategies. RELEVANCE STATEMENT: The longitudinal integration of diagnostic imaging and RT, facilitated by AI, could significantly enhance clinical workflow efficiency and therapeutic accuracy in oncology. KEY POINTS: Oncological care is transitioning from disease-centred to patient-centred, with tumour boards representing the junction for shared multidisciplinary decisions. Integrating advanced imaging with RT enables quantitative imaging biomarkers extraction that captures tumour changes throughout the course of treatment. Artificial intelligence plays a central role in automating resource-intensive processes and integrating large-scale multidomain data towards personalised medicine.