Showing papers in "European Radiology in 2022"

Breast cancer screening in women with extremely dense breasts recommendations of the European Society of Breast Imaging (EUSOBI)

Abstract: Abstract Breast density is an independent risk factor for the development of breast cancer and also decreases the sensitivity of mammography for screening. Consequently, women with extremely dense breasts face an increased risk of late diagnosis of breast cancer. These women are, therefore, underserved with current mammographic screening programs. The results of recent studies reporting on contrast-enhanced breast MRI as a screening method in women with extremely dense breasts provide compelling evidence that this approach can enable an important reduction in breast cancer mortality for these women and is cost-effective. Because there is now a valid option to improve breast cancer screening, the European Society of Breast Imaging (EUSOBI) recommends that women should be informed about their breast density. EUSOBI thus calls on all providers of mammography screening to share density information with the women being screened. In light of the available evidence, in women aged 50 to 70 years with extremely dense breasts, the EUSOBI now recommends offering screening breast MRI every 2 to 4 years. The EUSOBI acknowledges that it may currently not be possible to offer breast MRI immediately and everywhere and underscores that quality assurance procedures need to be established, but urges radiological societies and policymakers to act on this now. Since the wishes and values of individual women differ, in screening the principles of shared decision-making should be embraced. In particular, women should be counselled on the benefits and risks of mammography and MRI-based screening, so that they are capable of making an informed choice about their preferred screening method. Key Points • The recommendations in Figure 1 summarize the key points of the manuscript

Radiology artificial intelligence: a systematic review and evaluation of methods (RAISE)

Abstract: There has been a large amount of research in the field of artificial intelligence (AI) as applied to clinical radiology. However, these studies vary in design and quality and systematic reviews of the entire field are lacking.This systematic review aimed to identify all papers that used deep learning in radiology to survey the literature and to evaluate their methods. We aimed to identify the key questions being addressed in the literature and to identify the most effective methods employed.We followed the PRISMA guidelines and performed a systematic review of studies of AI in radiology published from 2015 to 2019. Our published protocol was prospectively registered.Our search yielded 11,083 results. Seven hundred sixty-seven full texts were reviewed, and 535 articles were included. Ninety-eight percent were retrospective cohort studies. The median number of patients included was 460. Most studies involved MRI (37%). Neuroradiology was the most common subspecialty. Eighty-eight percent used supervised learning. The majority of studies undertook a segmentation task (39%). Performance comparison was with a state-of-the-art model in 37%. The most used established architecture was UNet (14%). The median performance for the most utilised evaluation metrics was Dice of 0.89 (range .49-.99), AUC of 0.903 (range 1.00-0.61) and Accuracy of 89.4 (range 70.2-100). Of the 77 studies that externally validated their results and allowed for direct comparison, performance on average decreased by 6% at external validation (range increase of 4% to decrease 44%).This systematic review has surveyed the major advances in AI as applied to clinical radiology.• While there are many papers reporting expert-level results by using deep learning in radiology, most apply only a narrow range of techniques to a narrow selection of use cases. • The literature is dominated by retrospective cohort studies with limited external validation with high potential for bias. • The recent advent of AI extensions to systematic reporting guidelines and prospective trial registration along with a focus on external validation and explanations show potential for translation of the hype surrounding AI from code to clinic.

Systematic review of the radiomics quality score applications: an EuSoMII Radiomics Auditing Group Initiative

Abstract: The main aim of the present systematic review was a comprehensive overview of the Radiomics Quality Score (RQS)-based systematic reviews to highlight common issues and challenges of radiomics research application and evaluate the relationship between RQS and review features.The literature search was performed on multiple medical literature archives according to PRISMA guidelines for systematic reviews that reported radiomic quality assessment through the RQS. Reported scores were converted to a 0-100% scale. The Mann-Whitney and Kruskal-Wallis tests were used to compare RQS scores and review features.The literature research yielded 345 articles, from which 44 systematic reviews were finally included in the analysis. Overall, the median of RQS was 21.00% (IQR = 11.50). No significant differences of RQS were observed in subgroup analyses according to targets (oncological/not oncological target, neuroradiology/body imaging focus and one imaging technique/more than one imaging technique, characterization/prognosis/detection/other).Our review did not reveal a significant difference of quality of radiomic articles reported in systematic reviews, divided in different subgroups. Furthermore, low overall methodological quality of radiomics research was found independent of specific application domains. While the RQS can serve as a reference tool to improve future study designs, future research should also be aimed at improving its reliability and developing new tools to meet an ever-evolving research space.• Radiomics is a promising high-throughput method that may generate novel imaging biomarkers to improve clinical decision-making process, but it is an inherently complex analysis and often lacks reproducibility and generalizability. • The Radiomics Quality Score serves a necessary role as the de facto reference tool for assessing radiomics studies. • External auditing of radiomics studies, in addition to the standard peer-review process, is valuable to highlight common limitations and provide insights to improve future study designs and practical applicability of the radiomics models.

Comparison of chest CT severity scoring systems for COVID-19

Abstract: To compare the diagnostic performance and inter-observer agreement of five different CT chest severity scoring systems for COVID-19 to find the most precise one with the least interpretation time. This retrospective study included 85 patients (54 male and 31 female) with PCR-confirmed COVID-19. They underwent CT to assess the severity of pulmonary involvement. Three readers were asked to assess the pulmonary abnormalities and score the severity using five different systems, including chest CT severity score (CT-SS), chest CT score, total severity score (TSS), modified total severity score (m-TSS), and 3-level chest CT severity score. Time consumption on reporting of each system was calculated. Two hundred fifty-five observations were reported for each system. There was a statistically significant inter-observer agreement in assessing qualitative lung involvement using the m-TSS and the other four quantitative systems. The ROC curves revealed excellent and very good diagnostic accuracy for all systems when cutoff values for detection severe cases were > 22, > 17, > 12, and > 26 for CT-SS, chest CT score, TSS, and 3-level CT severity score. The AUC was very good (0.86), excellent (0.90), very good (0.89), and very good (0.86), respectively. Chest CT score showed the highest specificity (95.2%) in discrimination of severe cases. Time consumption on reporting was significantly different (< 0.001): CT-SS > 3L-CT-SS > chest CT score > TSS. All chest CT severity scoring systems in this study demonstrated excellent inter-observer agreement and reasonable performance to assess COVID-19 in relation to the clinical severity. CT-SS and TSS had the highest specificity and least time for interpretation. • All chest CT severity scoring systems discussed in this study revealed excellent inter-observer agreement and reasonable performance to assess COVID-19 in relation to the clinical severity. • Chest CT scoring system and TSS had the highest specificity. • Both TSS and m-TSS consumed the least time compared to the other three scoring systems.

First-generation clinical dual-source photon-counting CT: ultra-low-dose quantitative spectral imaging

Abstract: Evaluation of image characteristics at ultra-low radiation dose levels of a first-generation dual-source photon-counting computed tomography (PCCT) compared to a dual-source dual-energy CT (DECT) scanner.A multi-energy CT phantom was imaged with and without an extension ring on both scanners over a range of radiation dose levels (CTDIvol 0.4-15.0 mGy). Scans were performed in different modes of acquisition for PCCT with 120 kVp and DECT with 70/Sn150 kVp and 100/Sn150 kVp. Various tissue inserts were used to characterize the precision and repeatability of Hounsfield units (HUs) on virtual mono-energetic images between 40 and 190 keV. Image noise was additionally investigated at an ultra-low radiation dose to illustrate PCCT's ability to remove electronic background noise.Our results demonstrate the high precision of HU measurements for a wide range of inserts and radiation exposure levels with PCCT. We report high performance for both scanners across a wide range of radiation exposure levels, with PCCT outperforming at low exposures compared to DECT. PCCT scans at the lowest radiation exposures illustrate significant reduction in electronic background noise, with a mean percent reduction of 74% (p value ~ 10-8) compared to DECT 70/Sn150 kVp and 60% (p value ~ 10-6) compared to DECT 100/Sn150 kVp.This paper reports the first experiences with a clinical dual-source PCCT. PCCT provides reliable HUs without disruption from electronic background noise for a wide range of dose values. Diagnostic benefits are not only for quantification at an ultra-low dose but also for imaging of obese patients.PCCT scanners provide precise and reliable Hounsfield units at ultra-low dose levels. The influence of electronic background noise can be removed at ultra-low-dose acquisitions with PCCT. Both spectral platforms have high performance along a wide range of radiation exposure levels, with PCCT outperforming at low radiation exposures.

Improved coronary calcium detection and quantification with low-dose full field-of-view photon-counting CT: a phantom study

Abstract: The aim of the current study was to systematically assess coronary artery calcium (CAC) detection and quantification for spectral photon-counting CT (SPCCT) in comparison to conventional CT and, in addition, to evaluate the possibility of radiation dose reduction.Routine clinical CAC CT protocols were used for data acquisition and reconstruction of two CAC containing cylindrical inserts which were positioned within an anthropomorphic thorax phantom. In addition, data was acquired at 50% lower radiation dose by reducing tube current, and slice thickness was decreased. Calcifications were considered detectable when three adjacent voxels exceeded the CAC scoring threshold of 130 Hounsfield units (HU). Quantification of CAC (as volume and mass score) was assessed by comparison with known physical quantities.In comparison with CT, SPCCT detected 33% and 7% more calcifications for the small and large phantoms, respectively. At reduced radiation dose and reduced slice thickness, small phantom CAC detection increased by 108% and 150% for CT and SPCCT, respectively. For the large phantom size, noise levels interfered with CAC detection. Although comparable between CT and SPCCT, routine protocols CAC quantification showed large deviations (up to 134%) from physical CAC volume. At reduced radiation dose and slice thickness, physical volume overestimations decreased to 96% and 72% for CT and SPCCT, respectively. In comparison with volume scores, mass score deviations from physical quantities were smaller.CAC detection on SPCCT is superior to CT, and was even preserved at a reduced radiation dose. Furthermore, SPCCT allows for improved physical volume estimation.• In comparison with conventional CT, increased coronary artery calcium detection (up to 156%) for spectral photon-counting CT was found, even at 50% radiation dose reduction. • Spectral photon-counting CT can more accurately measure physical volumes than conventional CT, especially at reduced slice thickness and for high-density coronary artery calcium. • For both conventional and spectral photon-counting CT, reduced slice thickness reconstructions result in more accurate physical mass approximation.

COVID-19 mortality prediction in the intensive care unit with deep learning based on longitudinal chest X-rays and clinical data

Abstract: We aimed to develop deep learning models using longitudinal chest X-rays (CXRs) and clinical data to predict in-hospital mortality of COVID-19 patients in the intensive care unit (ICU). Six hundred fifty-four patients (212 deceased, 442 alive, 5645 total CXRs) were identified across two institutions. Imaging and clinical data from one institution were used to train five longitudinal transformer-based networks applying five-fold cross-validation. The models were tested on data from the other institution, and pairwise comparisons were used to determine the best-performing models. A higher proportion of deceased patients had elevated white blood cell count, decreased absolute lymphocyte count, elevated creatine concentration, and incidence of cardiovascular and chronic kidney disease. A model based on pre-ICU CXRs achieved an AUC of 0.632 and an accuracy of 0.593, and a model based on ICU CXRs achieved an AUC of 0.697 and an accuracy of 0.657. A model based on all longitudinal CXRs (both pre-ICU and ICU) achieved an AUC of 0.702 and an accuracy of 0.694. A model based on clinical data alone achieved an AUC of 0.653 and an accuracy of 0.657. The addition of longitudinal imaging to clinical data in a combined model significantly improved performance, reaching an AUC of 0.727 (p = 0.039) and an accuracy of 0.732. The addition of longitudinal CXRs to clinical data significantly improves mortality prediction with deep learning for COVID-19 patients in the ICU. • Deep learning was used to predict mortality in COVID-19 ICU patients. • Serial radiographs and clinical data were used. • The models could inform clinical decision-making and resource allocation.

Neuroimaging in patients with COVID-19: a neuroradiology expert group consensus

Abstract: Neurological and neuroradiological manifestations in patients with COVID-19 have been extensively reported. Available imaging data are, however, very heterogeneous. Hence, there is a growing need to standardise clinical indications for neuroimaging, MRI acquisition protocols, and necessity of follow-up examinations. A NeuroCovid working group with experts in the field of neuroimaging in COVID-19 has been constituted under the aegis of the Subspecialty Committee on Diagnostic Neuroradiology of the European Society of Neuroradiology (ESNR). The initial objectives of this NeuroCovid working group are to address the standardisation of the imaging in patients with neurological manifestations of COVID-19 and to give advice based on expert opinion with the aim of improving the quality of patient care and ensure high quality of any future clinical studies. • In patients with COVID-19 and neurological manifestations, neuroimaging should be performed in order to detect underlying causal pathology. • The basic MRI recommended protocol includes T2-weighted, FLAIR (preferably 3D), and diffusion-weighted images, as well as haemorrhage-sensitive sequence (preferably SWI), and at least for the initial investigation pre and post-contrast T1 weighted-images. • 3D FLAIR should be acquired after gadolinium administration in order to optimise the detection of leptomeningeal contrast enhancement.

Deep learning reconstruction improves radiomics feature stability and discriminative power in abdominal CT imaging: a phantom study

Abstract: To compare image quality of deep learning reconstruction (AiCE) for radiomics feature extraction with filtered back projection (FBP), hybrid iterative reconstruction (AIDR 3D), and model-based iterative reconstruction (FIRST).Effects of image reconstruction on radiomics features were investigated using a phantom that realistically mimicked a 65-year-old patient's abdomen with hepatic metastases. The phantom was scanned at 18 doses from 0.2 to 4 mGy, with 20 repeated scans per dose. Images were reconstructed with FBP, AIDR 3D, FIRST, and AiCE. Ninety-three radiomics features were extracted from 24 regions of interest, which were evenly distributed across three tissue classes: normal liver, metastatic core, and metastatic rim. Features were analyzed in terms of their consistent characterization of tissues within the same image (intraclass correlation coefficient ≥ 0.75), discriminative power (Kruskal-Wallis test p value < 0.05), and repeatability (overall concordance correlation coefficient ≥ 0.75).The median fraction of consistent features across all doses was 6%, 8%, 6%, and 22% with FBP, AIDR 3D, FIRST, and AiCE, respectively. Adequate discriminative power was achieved by 48%, 82%, 84%, and 92% of features, and 52%, 20%, 17%, and 39% of features were repeatable, respectively. Only 5% of features combined consistency, discriminative power, and repeatability with FBP, AIDR 3D, and FIRST versus 13% with AiCE at doses above 1 mGy and 17% at doses ≥ 3 mGy. AiCE was the only reconstruction technique that enabled extraction of higher-order features.AiCE more than doubled the yield of radiomics features at doses typically used clinically. Inconsistent tissue characterization within CT images contributes significantly to the poor stability of radiomics features.• Image quality of CT images reconstructed with filtered back projection and iterative methods is inadequate for the majority of radiomics features due to inconsistent tissue characterization, low discriminative power, or low repeatability. • Deep learning reconstruction enhances image quality for radiomics and more than doubled the feature yield at doses that are typically used in clinical CT imaging. • Image reconstruction algorithms can optimize image quality for more reliable quantification of tissues in CT images.

Initial comparative analysis of pulmonary involvement on HRCT between vaccinated and non-vaccinated subjects of COVID-19

Abstract: To compare the high-resolution computed tomography (HRCT)–derived severity score in COVID-19 patients between those who had earlier received the vaccine against the SARS-CoV-2 and those who did not. A retrospective cross-sectional analysis of HRCT of the chest was done in correlation with the vaccination status of clinically diagnosed COVID-19 patients. The variable under evaluation was the CT severity score, whereby differential analysis of the variability on this parameter between incompletely (single dose) vaccinated, completely (both doses) vaccinated, and non-vaccinated individuals was the outcome. The analysis included 826 patients of which 581 did not receive any vaccination whereas 196 patients received incomplete (single dose) vaccination and 49 received complete vaccination. Mean CT severity score was lower in completely vaccinated patients (3.5 ± 6.3) vis-à-vis incompletely vaccinated (10.1 ± 10.5) and non-vaccinated (10.1 ± 11.4) individuals. The mean CT score was significantly lower in completely vaccinated patients of lower ages (≤ 60 years) compared to patients above that age. The incidence of severe disease (CT score ≥ 20) was significantly higher in the incompletely vaccinated and non-vaccinated patients compared to that in the completely vaccinated group. CT severity scores in individuals receiving both doses of SARS-CoV-2 vaccination were less severe in comparison to those receiving a single dose of vaccine or no vaccine at all. • Patients who received complete two doses of vaccination had significantly low mean CT scores compared to the partially vaccinated patients and non-vaccinated patients. • The mean CT scores were significantly lower in completely vaccinated patients of lower ages (< 60 years) while patients > 60 years did not show significantly different CT scores between the vaccinated and non-vaccinated groups. • Consolidations and ground-glass opacities were significantly lower in the group receiving complete vaccination as compared to the unvaccinated and incompletely vaccinated patients.

Air trapping in COVID-19 patients following hospital discharge: retrospective evaluation with paired inspiratory/expiratory thin-section CT

Abstract: The study reports our experience with paired inspiration/expiration thin-section computed tomographic (CT) scans in the follow-up of COVID-19 patients with persistent respiratory symptoms.From August 13, 2020, to May 31, 2021, 48 long-COVID patients with respiratory symptoms (27 men and 21 women; median age, 62.0 years; interquartile range: 54.0-69.0 years) underwent follow-up paired inspiration-expiration thin-section CT scans. Patient demographics, length of hospital stay, intensive care unit admission rate, and clinical and laboratory features of acute infection were also included. The scans were obtained on a median of 72.5 days after onset of symptoms (interquartile range: 58.5-86.5) and at least 30 days after hospital discharge. Thin-section CT findings included ground-glass opacity, mosaic attenuation pattern, consolidation, traction bronchiectasis, reticulation, parenchymal bands, bronchial wall thickening, and air trapping. We used a quantitative score to determine the degree of air trapping in the expiratory scans.Parenchymal abnormality was found in 50% (24/48) of patients and included air trapping (37/48, 77%), ground-glass opacities (19/48, 40%), reticulation (18/48, 38%), parenchymal bands (15/48, 31%), traction bronchiectasis (9/48, 19%), mosaic attenuation pattern (9/48, 19%), bronchial wall thickening (6/48, 13%), and consolidation (2/48, 4%). The absence of air trapping was observed in 11/48 (23%), mild air trapping in 20/48 (42%), moderate in 13/48 (27%), and severe in 4/48 (8%). Independent predictors of air trapping were, in decreasing order of importance, gender (p = 0.0085), and age (p = 0.0182).Our results, in a limited number of patients, suggest that follow-up with paired inspiratory/expiratory CT in long-COVID patients with persistent respiratory symptoms commonly displays air trapping.• Our experience indicates that paired inspiratory/expiratory CT in long-COVID patients with persistent respiratory symptoms commonly displays air trapping. • Iterative reconstruction and dose-reduction options are recommended for demonstrating air trapping in long-COVID patients.

Cardiac magnetic resonance imaging of myocarditis and pericarditis following COVID-19 vaccination: a multicenter collection of 27 cases

Abstract: To assess clinical and cardiac magnetic resonance (CMR) imaging features of patients with peri-myocarditis following Coronavirus Disease 2019 (COVID-19) vaccination. We retrospectively collected a case series of 27 patients who underwent CMR in the clinical suspect of heart inflammation following COVID-19 vaccination, from 16 large tertiary centers. Our patient’s cohort was relatively young (36.6 ± 16.8 years), predominately included males (n = 25/27) with few comorbidities and covered a catchment area of approximately 8 million vaccinated patients. CMR revealed typical mid-subepicardial non-ischemic late gadolinium enhancement (LGE) in 23 cases and matched positively with CMR T2 criteria of myocarditis. In 7 cases, typical hallmarks of acute pericarditis were present. Short-term follow-up (median = 20 days) from presentation was uneventful for 25/27 patients and unavailable in two cases. While establishing a causal relationship between peri-myocardial inflammation and vaccine administration can be challenging, our clinical experience suggests that CMR should be performed for diagnosis confirmation and to drive clinical decision-making and follow-up. • Acute onset of dyspnea, palpitations, or acute and persisting chest pain after COVID-19 vaccination should raise the suspicion of possible myocarditis or pericarditis, and patients should seek immediate medical attention and treatment to help recovery and avoid complications. • In case of elevated troponin levels and/or relevant ECG changes, cardiac magnetic resonance should be considered as the best non-invasive diagnostic option to confirm the diagnosis of myocarditis or pericarditis and to drive clinical decision-making and follow-up.

The emerging role of atrial strain assessed by cardiac MRI in different cardiovascular settings: an up-to-date review

Abstract: Abstract The left atrium (LA) has a crucial function in maintaining left ventricular filling, which is responsible for about one-third of all cardiac filling. A growing body of evidence shows that LA is involved in several cardiovascular diseases from a clinical and prognostic standpoint. LA enlargement has been recognized as a predictor of the outcomes of many diseases. However, LA enlargement itself does not explain the whole LA’s function during the cardiac cycle. For this reason, the recently proposed assessment of atrial strain at advanced cardiac magnetic resonance (CMR) enables the usual limitations of the sole LA volumetric measurement to be overcome. Moreover, the left atrial strain impairment might allow several cardiovascular diseases to be detected at an earlier stage. While traditional CMR has a central role in assessing LA volume and, through cine sequences, a marginal role in evaluating LA function, feature tracking at advanced CMR (CMR-FT) has been increasingly confirmed as a feasible and reproducible technique for assessing LA function through strain. In comparison to atrial function evaluations via speckle tracking echocardiography, CMR-FT has a higher spatial resolution, larger field of view, and better reproducibility. In this literature review on atrial strain analysis, we describe the strengths, limitations, recent applications, and promising developments of studying atrial function using CMR-FT in clinical practice. Key Points • The left atrium has a crucial function in maintaining left ventricular filling; left atrial size has been recognized as a predictor of the outcomes of many diseases . • Left atrial strain has been confirmed as a marker of atrial functional status and demonstrated to be a sensitive tool in the subclinical phase of a disease . • A comprehensive evaluation of the three phases of atrial function by CMR-FT demonstrates an impairment before the onset of atrial enlargement, thus helping clinicians in their decision-making and improving patient outcomes .