To evaluate three commercial AI software tools for pulmonary nodule detection and segmentation and to assess their impact on guideline-based management recommendations. A total of 740 CT and PET-CT studies from clinical routine were analyzed using three software tools (S1, S2, S3). We compared the total number of detected nodules and “actionable” nodules (per British Thoracic Society (BTS) definition). We further evaluated how measurement variations between tools affected hypothetical management according to Fleischner Society and BTS guidelines for incidental nodules. The tools differed significantly in the total number of detections (S1: 1336; S2: 1060; S3: 1536; p < 0.001) and wrong findings (S1: 965; S2: 720; S3: 1169; p < 0.001). However, the detection of actionable nodules was comparable across all tools (S1: 375; S2: 341; S3: 373; p = 0.73). While no statistically significant differences were found in mean diameter or volume measurements, small absolute variations led to significant differences in management. Specifically, S2 triggered significantly more 1-year follow-up recommendations than S3 under BTS guidelines (p < 0.001). No significant management differences were observed when applying Fleischner Society guidelines. While the three included AI tools show comparable performance in detecting actionable nodules, minor measurement variations significantly impact downstream management when using guidelines with narrow thresholds, such as the BTS criteria. Fleischner Society guidelines appear more robust to these inter-software variations. Question How do commercial software tools for pulmonary nodule detection perform in real-world settings and impact hypothetical management under BTS and Fleischner guidelines? Findings Detection of actionable nodules was comparable across all tools, but small absolute measurement variations triggered significantly more 1-year follow-up recommendations under BTS guidelines. Clinical relevance AI software can cause inconsistent BTS-based management due to narrow thresholds, while Fleischner criteria appear more stable. Frequent detection of benign lesions potentially poses a risk of overdiagnosis and overtreatment in standalone AI-based reporting.

Die Aktualisierung der American Thoracic Society/European Respiratory Society(ATS/ERS)-Klassifikation der interstitiellen Pneumonien von 2025 erweitert deren Einteilung über rein idiopathische Entitäten hinaus und umfasst nun auch Krankheitsbilder mit identifizierbaren Auslösern [1]. Zentraler Bestandteil dieser Aktualisierung ist die terminologische Neuausrichtung der Klassifikation zu einem deskriptiven, primär morphologisch orientierten Konzept mit dem Ziel der klaren Unterscheidung von radiologischen und histologischen Mustern auf der einen und von klinisch definierten Erkrankungen auf der anderen Seite. Die akute interstitielle Pneumonie (AIP) wird nun als idiopathische diffuse alveoläre Schädigung (DAD) bezeichnet, die desquamative interstitielle Pneumonie (DIP) als Alveolarmakrophagenpneumonie (AMP). Das Muster der bronchiolozentrischen interstitiellen Pneumonie (BIP) wird als eigenständiges morphologisches Muster eingeführt, der Begriff der Hypersensitivitätspneumonitis (HP) soll künftig ausschließlich der multidisziplinären Diagnose vorbehalten sein. Darüber hinaus bietet die aktualisierte Klassifikation ein biologisch begründetes und klinisch anwendbares System zur Kategorisierung in interstitielle und alveoläre Füllmuster und die auch prognostisch relevante Aufteilung in fibrotische und nichtfibrotische Verläufe. Hervorzuheben ist zudem eine stärkere Betonung der transparenten Angabe der diagnostischen Sicherheit im interdisziplinären Board für interstitielle Lungenerkrankungen (ILD-Board), inklusive der Vergabe von „provisorischen“ Diagnosen bzw. Verwendung des Begriffs der „unklassifizierbaren ILD“ bei niedriger und sehr niedriger diagnostischer Sicherheit. Die aktuelle Klassifikation kann helfen, den diagnostischen Ablauf und den Entscheidungsprozess im ILD-Board zu standardisieren und transparenter zu gestalten und dadurch auch die klinische Versorgung zu optimieren. Es ist zu erwarten, dass dieses Update die Grundlage für künftige Forschung und neue Erkenntnisse im Bereich der ILD sein wird. Daneben soll jedoch nicht unerwähnt bleiben, dass die neue Klassifikation hinsichtlich der Einführung des Musters der BIP und der provisorischen Entität der „idiopathischen BIP“ von einigen ExpertInnen kritisch beurteilt wird. Befürchtet wird konkret, dass durch die stärkere Betonung eines morphologischen Musters und die terminologische Abgrenzung zur HP die diagnostische Aufmerksamkeit für potenziell auslösende Expositionen abgeschwächt werden könnte.

Background/Objectives: Coronary artery disease (CAD) remains the leading cause of death worldwide. Traditional cardiovascular risk assessment is based on chronological age and other clinical factors, with inherent limitations and poor accuracy. Objective was to estimate the artificial intelligence (AI)-enhanced biological cardiovascular age calculation derived from coronary computed tomography angiography (CTA) reports using a large language model (LLM), in predicting major adverse cardiovascular events (MACE). Methods: Coronary CTA reports were analyzed using a LLM (ChatGPT-4.0v, OpenAI), from symptomatic patients with suspected CAD who underwent coronary CTA for clinical indications. Patients in which the LLM successfully analyzed the key metrics (1) coronary artery calcium (CAC) score and (2) coronary CTA reports (coronary stenosis severity (CAD-RADS), high-risk anatomy, non-calcified plaque, cardiac function (LVEF and others) were included. Results: 386 CTA reports were uploaded, and 346 (89.6%) included. The mean biological age (bioAGE) was 57.2 ± 10.9 and the chronological 58.5 ± 10.8 years. 137 (39.6%) were women. The intra-individual deviation in bioAGE was high (median: 8.8; IQR 9.98). BioAGE exceeded chronological age in 45.4% patient and was lower or equal in 54.6%) MACE rate was 8.7% comprising 2 deaths, 5 myocardial infarctions, and 22 late revascularizations. The accuracy for prediction of MACE was higher for bioAGE (c = 0.768; 95% CI: 0.681–0.855, p < 0.001) compared to chronological age (c = 0.590; 95% CI: 0.492–0.689, p = 0.102) Conclusions: Biological age calculation from coronary CTA reports using LLM is feasible, yet intra-individual deviations are high. The accuracy for prediction of MACE is improved by bioAGE compared to chronological.

Purpose To evaluate the coronary artery disease (CAD) profile and valvular and structural alterations seen at coronary CT angiography (CTA) after radiation therapy (RT) for breast cancer in a case-control study. Materials and Methods Patients who underwent clinically indicated coronary CTA were included in this retrospective study. The following parameters were evaluated: coronary artery calcium (CAC) score, Coronary Artery Disease Reporting and Data System (CAD-RADS) score, high-risk plaque (HRP) phenotypes, and extracoronary findings (valvular fibrosis, fibrous adherence). Patients with breast cancer who underwent RT were propensity score-matched with controls (level: P = .05) for age, body mass index, and major cardiovascular risk factors to reduce selection bias and confounding. Results Among 154 female patients (mean age, 65.23 years ± 10.4 [SD]; 77 patients in RT group, 77 patients in control group), there was no evidence of a difference in CAC score between the RT and control groups (201.1 vs 75.4 Agatston units [AU], P = .64). No difference was seen in coronary stenosis severity (CAD-RADS score) (P = .40) or obstructive disease (>50% stenosis) rate (29% [22 of 77] vs 23% [18 of 77], P = .46; odds ratio [OR], 1.31 [95% CI: 0.63, 2.73]). No between-group difference was observed in high-risk plaque phenotype rate (9% [seven of 77] vs 17% [13 of 77], P = .23) or CAC or CAD-RADS scores. There was no evidence of a difference between left versus right RT for CAC (309 vs 120 AU, P = .23), coronary stenosis severity (CAD-RADS score, P = .43), or HRP phenotype rate (left, 11% [four of 36]; right, 6% [two of 34]). The prevalences of valvular fibrosis and calcifications were low (5% [four of 77] vs 3% [two of 77] [P = .68] and 16% [12 of 77] vs 9% [seven of 77] [P = .37], respectively; OR, 1.84 [95% CI: 0.68, 5.25]). The prevalence of fibrous adherence of the left anterior descending coronary artery or right coronary artery to the chambers was 1.4-fold higher in the RT group (29% [22 of 77] vs 13% [17 of 77] [P = .46]; OR, 1.41 [95% CI: 0.68, 2.97]). Conclusion RT for breast cancer was not associated with more severe coronary stenosis, HRP phenotypes, or valvular fibrosis in this case-control population cohort. Keywords: Coronary Arteries, Radiation Effects, CT-Coronary Angiography, Cardiac, Breast, Breast Cancer, Radiation Therapy, Cardiovascular Risk, Computed Tomography, Coronary Artery Disease Supplemental material is available for this article. © RSNA, 2026.

Background:HER2 exon 20 insertions are rare oncogenic driver mutations in non-small cell lung cancer (NSCLC). Trastuzumab deruxtecan (T-DXd) has proven efficacy in pretreated metastatic HER2-mutant NSCLC. However, its use as first-line therapy in oligometastatic settings remains undocumented. Case Presentation:A 66-year-old female presented with a right hilar lung mass and a singular symptomatic left frontal brain metastasis. Biopsy confirmed poorly differentiated adenocarcinoma (TTF-1 positive, PD-L1 TPS 0%) with an ERBB2 exon 20 insertion (p.Y772_A775dup). The patient declined standard chemo- and immunotherapy. An individualized treatment plan was developed, initiating off-label first-line T-DXd after resection and consolidating radiotherapy of the affected cerebral region. After four cycles of T-DXd, imaging demonstrated significant tumor reduction. Subsequent surgical resection of the primary tumor revealed a complete pathological response (ypT0N0, R0). Adjuvant T-DXd therapy was continued thereafter. Conclusion:This case illustrates the high efficacy of T-DXd as the sole first-line treatment in oligometastatic HER2-mutant NSCLC. Prospective studies are needed to systematically explore the efficacy and safety of this approach.

PURPOSE:Preoperative differentiation between benign and malignant parotid tumors (PTs) remains challenging despite clinical examination, cross-sectional imaging, and biopsy. Accurate malignancy assessment is crucial to optimize surgical planning and minimize morbidity. METHODS:We retrospectively analyzed 66 patients who underwent parotidectomy between 2008 and 2024, including 33 with malignant and 33 with benign PTs matched for demographics. All patients had contrast-enhanced computed tomography (CE-CT), and segmented tumor volumes were evaluated using a three-dimensional convolutional neural network. Diagnostic performance was compared with standard clinical and radiologic workup. RESULTS:Standard clinical and radiologic workup achieved a sensitivity of 60.6 %, increasing to 69.7 % with fine needle aspiration cytology (FNAC) or core needle biopsy (CNB). The deep learning model achieved an area under the ROC curve of 0.94, with both sensitivity and specificity exceeding 90 % at optimized thresholds, outperforming conventional diagnostics. CONCLUSION:Deep learning applied to CE-CT demonstrated strong diagnostic performance for the preoperative classification of PTs in this cohort and may serve as a powerful non-invasive adjunct to standard diagnostic modalities without adding procedural burden to the diagnostic workup.

BACKGROUND:High-resolution spectral photon counting computed tomography (PCCT) may allow characterization of structural alterations within the conduction system - for identifying patients with atrial fibrillation. OBJECTIVE:To perform a quantitative tissue analysis of the interatrial septum (IAS) along the anterior internodal bundle, the sinoatrial (SA) node and its vascular supply with PCCT in patients with atrial fibrillation (AF) compared to controls. METHODS:From 802 patients referred to PCCT coronary angiography, 130 (65 with AF and 65 controls in sinus rhythm)(age 70.3 ±12.2 years, 37.7% women) were included. SA-artery length/anatomy, IAS fibrosis (score1-4), IAS and SA-node density (HU) and iodine concentration; and left atrial wall thickening (LAWT) were quantified by PCCT. RESULTS:Patients with AF had shorter SA-arteries (p<0.001) but similar anatomy (left vs right). IAS fibrosis (Score 2-4) (87.7% vs 21.5%; p<0.001) and a fully fibrous IAS (78.5% vs 12.3%; p<0.001) were more prevalent in AF vs. controls. IAS density was higher in AF (55.3 vs -44.2 HU; p<0.001), and iodine concentration (1.45 vs -0.1mg/mL; p<0.001). The correlation between IAS density and iodine concentration was high (r = 0.749; p<0.001). Patients with positive left atrial remodeling (LAWT>2mm) had higher IAS fibrosis scores, higher IAS density and iodine concentration (p<0.001). SA-node density and iodine concentration was higher in AF (p<0.001). CONCLUSION:Patients with AF exhibit shorter SA-arteries and more IAS fibrosis- highlighting an interplay between vascular supply and fibrosis. Both features may serve as imaging biomarkers for identifying patients with occult asymptomatic AF undergoing coronary PCCT angiography for clinical indications - in whom reinforced screening for atrial fibrillation may be considered.

Artificial intelligence (AI) could facilitate and objectify quality assessment in the daily routine. The purpose was to explore the extent to which an AI prototype algorithm is able to replicate the perfect-good-moderate-inadequate (PGMI) system (perfect, good, moderate, inadequate). From a multicentre case collection, 200 standard mammograms (800 images) were selected. A deep learning-based prototype software was used to rate the images in analogy to the PGMI system. The AI results were compared with a reference standard obtained through consensus reading by three expert radiographers and one expert radiologist, using quadratically weighted Cohen’s kappa with confidence intervals (CI) and context-based interpretation. Frequency and reasons for disagreement were evaluated for challenging cases with a discrepancy of two or more grades and a discrepancy in assigning an inadequate. For overall PGMI per image, slight agreement between human consensus and AI was observed for CC views (κ = 0.14) and fair agreement for MLO views (κ = 0.25). The highest agreement was observed for the CC category “M. Pectoralis visibility” (substantial, κ = 0.75). Best category in MLO was “Pectoralis angle” (moderate, κ = 0.49). For other categories, fair, slight or poor agreement was observed. The work-up of disagreement gave insight into misinterpretations of anatomical landmarks and causality issues in the categorization. Transforming the PGMI system into a fully automated AI algorithm is challenging and may differ substantially between subcategories. Further research in computer science and quality assessment methodology is needed to pave the way for AI-based objective quality management in mammography. Profound evaluation of AI algorithms and their ability to replicate human interpretation, scoring, and classification are the basis and scientific framework toward AI-based objective quality management in mammography.

Background This study evaluated spectral computed tomography-based iodine density quantification for calculation of virtual non-contrast (VNC) images and extracellular volume (ECV) in the context of liver cirrhosis and sex. Methods A total of 156 consecutive patients (35 females) undergoing multiphasic Dual Energy CT (DECT) of the abdomen were retrospectively analyzed. Liver attenuation values and iodine density were quantified on true non-contrast (TNC), arterial-phase VNC (aVNC), and delayed phase VNC (dVNC) images. ECV was quantified using attenuation- (HU) and iodine density- (ID) based methods. Statistical analyses included comparison between TNC and VNC, ECV methods, and subgroup analyses for liver cirrhosis and sex. Logistic regression identified predictors of clinically relevant deviations of VNC. Results Both aVNC and dVNC showed significant, but not clinically relevant attenuation deviations (> 10 HU) from TNC with minor sex differences. BMI ≥30, sex, and ascites predicted aVNC deviations; BMI affected dVNC. No clinically relevant deviation was observed between cirrhotic and non-cirrhotic patients overall, but for high- and low-grade cirrhosis. ECV differentiated cirrhosis (AUC 0.74 [HU] and 0.67 [ID]) and varied with Child–Pugh scores. The ID-method showed significant sex-related differences in ECV Conclusion Iodine density quantification of the liver was influenced by sex and cirrhosis to a limited extent but remained within acceptable diagnostic thresholds. Sex-related findings are limited by male predominance. Elevated BMI may predict relevant deviation and should be considered. ID-ECV estimation yielded comparable results as HU-based evaluation and may indicate higher values in females.

Background: Distinguishing vital from non-vital persistent cervical lymph nodes after chemoradiotherapy in HNSCC remains clinically challenging. We investigated whether image-level data augmentation improves CT-based radiomics classification for this task. Methods: We evaluated eight augmentation strategies and their 28 pairwise combinations in 55 patients, using Bayesian hyperparameter tuning with Optuna for parameter optimization. A radiomics pipeline comprising Radiomics features, five feature selectors, and seven classifiers was assessed using patient-level stratified 5-fold cross-validation. Configurations were ranked using a composite score defined as the mean of AUC, ACC and F1-score. Results: Feature selection improved the composite score from 0.659 to 0.742. The best augmented configuration, Window Contrast Variation, achieved a composite score of 0.803 and an AUC of 0.831, corresponding to an 8.2% relative point-estimate gain over feature selection alone and a 21.9% gain over the no-selection baseline when feature selection and augmentation were combined. Conclusions: These findings suggest that feature selection with optimized augmentation may enhance radiomics-based lymph node classification. However, individual augmentation-versus-baseline differences did not reach statistical significance in this limited sample, requiring confirmation in larger cohorts.

Background:Cardiovascular risk stratification is crucial in patients with end-stage liver disease (ESLD) yet the optimal noninvasive strategy remains debated. Our study aimed to assess the prognostic value of coronary computed tomography angiography (CTA) and coronary artery calcium (CAC) in patients undergoing orthotopic liver transplantation (LT). Methods:Patients with ESLD scheduled for LT referred to coronary CTA and the CACscore were included. The primary endpoint was all-cause mortality and the secondary endpoint was myocardial infarction (MI). Results:Four hundred fifty-eight patients for pre-LT risk stratification were enrolled with 270 LT recipients (79.3% male; mean age 61 ± 8.5 years) finally being included. The mean follow-up was 7.5 ± 3.1 years, range: 2-13. Among 248 patients undergoing CTA, the majority (n = 173, 69.8%) had coronary artery disease (CAD) by CTA (Coronary Artery Disease-Reporting and Data System[CAD-RADS] 1-5), and n = 75 (30.2%) had no CAD. Stenosis severity was minimal-to-mild (<50%) in 112 (45.1%), intermediate (50-70%) in 44 (17.7%), and severe (>70%) in 17 (6.5%) patients. The all-cause mortality rate was 46 (17.0%) (n = 3 cardiovascular). Stenosis severity (CAD-RADS) was associated with mortality (Kaplan-Meier analysis, P < 0.001). On multivariate Cox regression, total plaque burden were associated with all-cause mortality (hazard ratio [HR]: 1.1, P = 0.034; 95% confidence interval [CI]: 0.649-0.983 and HR: 1.1, P = 0.029; 95% CI: 1.0-1.6), while the CAC score was not. Six acute myocardial infarctions (MIs, 3 ST-elevation MI [STEMI] and 3 non-STEMI) occurred, none of them (0%) in patients with CAD-RADS 0-1 and 100% CAD-RADS 2-4. Conclusion:Coronary CTA is a valuable tool for pre-LT cardiovascular risk assessment. Patients with no or minimal CAD on CTA have an excellent prognosis regarding survival. Clinical relevance statement:Coronary CTA enables refined risk stratification in patients undergoing liver transplantation by assessing stenosis severity and plaque burden.

Background/Objectives: Large language models (LLMs), such as ChatGPT, have emerged as potential clinical support tools to enhance precision in personalized patient care, but their reliability in radiological image interpretation remains uncertain. The primary aim of our study was to evaluate the diagnostic accuracy of ChatGPT-4o in interpreting chest X-rays (CXRs) and abdominal X-rays (AXRs) by comparing its performance to expert radiology findings, whilst secondary aims were diagnostic confidence and patient safety. Methods: A total of 500 X-rays, including 257 CXR (51.4%) and 243 AXR (48.5%), were analyzed. Diagnoses made by ChatGPT-4o were compared to expert interpretations. Confidence scores (1-4) were assigned and responses were evaluated for patient safety. Results: ChatGPT-4o correctly identified 345 of 500 (69%) pathologies (95% CI: 64.81-72.9). For AXRs 175 of 243 (72.02%) pathologies were correctly diagnosed (95% CI: 66.06-77.28), while for CXRs 170 of 257 (66.15%) were accurate (95% CI: 60.16-71.66). The highest detection rates among CXRs were observed for pulmonary edema, tumor, pneumonia, pleural effusion, cardiomegaly, and emphysema, and lower rates were observed for pneumothorax, rib fractures, and enlarged mediastinum. AXR performance was highest for intestinal obstruction and foreign bodies, and weaker for pneumoperitoneum, renal calculi, and diverticulitis. Confidence scores were higher for AXRs (mean 3.45 ± 1.1) than CXRs (mean 2.48 ± 1.45). All responses (100%) were considered to be safe for the patient. Interobserver agreement was high (kappa = 0.920), and reliability (second prompt) was moderate (kappa = 0.750). Conclusions: ChatGPT-4o demonstrated moderate accuracy for the interpretation of X-rays, being higher for AXRs compared to CXRs. Improvements are required for its use as efficient clinical support tool.

Background: While epicardial adipose tissue (EAT) is a known predictor of adverse cardiovascular outcomes, lipomatous hypertrophy of the interatrial septum (LHIS) is composed of metabolically active fat such as brown adipose tissue, which may exert a different effect. This study investigates the coronary atherosclerosis profile in patients with LHIS using CTA, compared with a propensity score-matched control group. Methods: A total of 142 patients were included (n = 71 with LHIS and n = 71 controls) and propensity score-matched for age, gender, BMI, and the major CV risk factors (matching level, <0.05). CTA imaging parameters included HRP, coronary stenosis severity (CADRADS), and CAC score. Results: The mean age was 60.9 years +/− 10.6, there were nine (6.3%) women, and the mean BMI is 28.04 kg/m2 +/− 4.99. HRP prevalence was significantly lower in LHIS patients vs. controls (21.1% vs. 40.8%; p < 0.011), while CAC (p = 0.827) and CADRADS (p = 0.329) were not different, and there was no difference in the obstructive disease rate. There was no difference in lipid panels (cholesterol, LDL, HDL, TG) and statin intake rate. Conclusions: HRP prevalence is lower in patients with LHIS than controls, while coronary stenosis severity and CAC score are not different. Clinical relevance: LHIS may serve as imaging biomarker for reversed CV risk.

Background:Anomalous left coronary artery from the pulmonary artery (ALCAPA) is a very rare, congenital condition. Patients typically exhibit symptoms within the first few weeks of life. 'Adult-type ALCAPA', which presents later in life, is even rarer and is mostly detected due to symptoms of heart failure or after life-threatening arrhythmias. Case summary:We present the case of a 53-year-old asymptomatic male, who was incidentally diagnosed with adult-type ALCAPA during a coronary computed tomography angiography (CCTA) screening performed for cardiac risk stratification. Further cardiac investigations included echocardiography, myocardial perfusion scintigraphy, invasive coronary angiography (ICA), and cardiac magnetic resonance imaging (CMRI). As he had never experienced any cardiac symptoms at rest or during exercise throughout his life, the patient has not yet decided to undergo the recommended cardiac surgery. The patient is currently seeking a second opinion. Discussion:Anomalous left coronary artery from the pulmonary artery is rare, but the increased use of CCTA for risk stratification in asymptomatic patients might lead to more frequent diagnosis in older individuals. As with our patient, a dilemma may arise when deciding on its treatment, because the guideline-recommended treatment is surgical correction in all patients. However, there is still little evidence to support this recommendation for older, asymptomatic patients, making advice and decisions difficult for such individuals.

Dual-energy computed tomography (DECT) detects monosodium urate (MSU) deposits in joints. However, the correlation between coronary atherosclerosis phenotypes and MSU-positive lesions in the cardiovascular system remains unclear. We investigated the correlation between coronary MSU-positive plaques on unenhanced DECT with the coronary atherosclerosis profile at coronary CT angiography. One hundred fifty rheumatologic patients were prospectively enrolled. Sixty of them underwent unenhanced DECT and 128-row DECT coronary angiography. Analysis included CAD-RADS stenosis severity, high-risk plaque (HRP) phenotypes, and coronary artery calcium (CAC) score. Of 60 patients, with a mean age of 63.7 years, including 7 females (11.7

Background/Objectives: Hepatocellular carcinoma (HCC) is the most common primary malignant tumour of the liver. In a cirrhotic liver, each nodule larger than 10 mm demands further work-up using CT or MRI. The Liver Imaging Reporting and Data System (LI-RADS) is still based on visual assessment and measurements. The purpose of this study was to evaluate whether semi-automated quantification of visual LR-5 lesions is appropriate and can objectify HCC classification for personalized radiomic research. Methods: A total of 52 HCC patients (median age 67 years, 17% females, 83% males) from a retrospective data collection were evaluated visually and compared by the results using an oncology software with features of LI-RADS-based structured tumour evaluation and documentation, semi-automated tumour segmentation, and texture analysis. Results: Software-based evaluation of non-rim arterial-phase hyperenhancement (APHE) and non-peripheral washout, as well as the LI-RADS-score, showed no statistically significant differences compared with visual assessment (p = 0.2, 0.7, 0.17), with a consensus between a human reader and the software approach in 98% (APHE), 89% (washout), and 93% (threshold growth) of cases, respectively. The software provided automated LI-RADS classification, structured reporting, and quantitative features for HCC registries and radiomic research. Conclusions: The presented work may serve as an outlook for LI-RADS-based automated qualitative and quantitative evaluation. Future research may show if texture analysis can be used to foster personalized medical approaches in HCC.

Background Stroke is a feared complication after TAVI. The objective was to assess whether left atrial appendage (LAA) filling-defect (FD) patterns from early and late-phase computed tomography (CT), predict stroke/TIA in patients with severe aortic stenosis. Methods 124 patients with severe aortic stenosis (79.5y, 46.8% females) who underwent CT-Angiography for TAVI-planning were included (66.1% underwent TAVI, 18.6% surgical, 15.3% conservative treatment).CT-image-analysis included: CT-density (HU) from LAA tip-to-base and HU-gradients (I-III), the HU-ratio LAA/aorta, left-atrial-wall-thickness (LAWT) and the periatrial fat attenuation index (FAI). Results Stroke/TIA rate was 9.6 %. LAA-HU-gradient was slightly higher in non-stroke patients (p = 0.087). Persisting FDs during the late-phase were associated with stroke (p = 0.047) but not early-phase FDs. Early-phase FDs with HU < 245 (n = 15) were correlated with stroke (p = 0.05). A LAA-HU-gradient > 10HU had 91 % sensitivity and 68 % specificity for prediction of stroke. LAA-HU gradient I had a moderate accuracy (c = 0.592; 95 %CI:0.472-0.711; p = 0.317) for discrimination of stroke during the early phase, which enhanced during the late phase (c = 0.686;95 %CI:0.503-0.868; p = 0.046). Patients with stroke had a higher rate of FDs with HU-progression from early to late phase (>10HU)(p = 0.013), while the ratios LAA/aorta, LAWT, and periatrial-FAI were not different. Among clinical parameters, only age predicted stroke but not CHA2DS2-VASc-score. In multivariate analysis, late-phase FDs (p = 0.059)(OR 5.66: 95 %CI:0.936-34.28) but not early-phase FD were associated with stroke, and none of the major conventional risk factors. Conclusion Persisting LAA-filling defects on CT during the late-phase, and early-phase FD with <245HU predict stroke, and a CT-density progression >10HU from early-to-late phase. LAA-FD may improve stroke risk stratification.