Acta radiologica最新文献

BackgroundPeripheral zone (PZ) PI-RADS category 4 includes lesions with diverse pathological outcomes, leading to varying prostate cancer (PCa) detection rates between pure category 4 (pCategory-4) and upgraded category 4 (Category-3+1) PZ lesions, as well as different lesion sizes.PurposeTo compare PCa detection rates for pCategory-4 and Category-3+1, considering lesion size.Material and MethodsThis retrospective study included 293 participants with PI-RADS V2.1 category-4 PZ lesions, who underwent MRI-targeted biopsy between 2012 and 2021. Overall and clinically significant PCa (csPCa) detection rates for pCategory-4 and Category-3+1 lesions were compared using Pearson's chi-square (χ²) test. In addition, PCa detection rates were analyzed by lesion size (1-5 mm, 5-10 mm, 10-15 mm, and >15 mm) using Spearman's test. Logistic regression analysis included age, PSA, PSA density, lesion volume, and size/scale for PZ lesions.ResultscsPCa detection rates were 60.4% for pCategory-4 and 25.8% for Category-3+1, while overall PCa detection rates were 69.4% and 36.2%, respectively. pCategory-4 showed higher cancer detection rates than Category-3+1 (overall PCa: χ² = 22.34; P <0.0001, csPCa: χ² = 21.88; P <0.001). Larger lesions (>5 mm) were more likely to harbor PCa, with significant differences in detection rates observed for pCategory-4 and Category-3+1 (overall PCa: χ² = 20.05; P <0.001).ConclusionpCategory-4 lesions have significantly higher PCa detection rates compared to Category-3+1. Larger lesion size is associated with increased PCa detection in pCategory-4 lesions but not in Category-3+1.

BackgroundHuman epidermal growth factor receptor 2 (HER-2) is a key biomarker in breast cancer, guiding therapeutic decisions and prognosis. Conventional assessment relies on tissue biopsy, an invasive procedure that may impose both physical and financial burdens on patients.PurposeTo develop an interpretable deep learning-based imaging framework capable of non-invasively predicting preoperative HER-2 expression.Material and MethodsWe retrospectively analyzed magnetic resonance imaging data and clinical records from 450 patients with pathologically confirmed HER-2 status across four medical centers. Several conventional machine learning algorithms were compared with a deep neural network model. A ResNet-based architecture was used to generate a probability score (D-score) reflecting the likelihood of HER-2 positivity. Independent clinical predictors were identified through logistic regression and integrated with the D-score to construct a combined predictive framework. Model performance was evaluated using receiver operating characteristic analysis, and interpretability techniques were applied to visualize the contribution of individual features.ResultsThe combined deep learning model achieved an area under the curve of 0.809 in the external validation cohort, outperforming the clinical model. Interpretability analysis identified the D-score, rim enhancement, and diameter of the largest axillary lymph node as the most influential predictors, consistent with established clinical knowledge.ConclusionThe proposed model enables accurate, non-invasive, and interpretable prediction of HER-2 expression in breast cancer. It may serve as a preoperative stratification tool, support individualized treatment planning, and reduce reliance on invasive diagnostic procedures.

BackgroundPulmonary arteriovenous malformations (PAVMs) are rare vascular anomalies that can lead to serious complications, and accurate imaging is essential for detection and post-embolization follow-up.PurposeTo evaluate the diagnostic accuracy and clinical utility of time-resolved magnetic resonance angiography (TR-MRA) in detecting and monitoring PAVMs before and after embolization.Material and MethodsIn this prospective study, 29 patients were initially enrolled between January 2023 and December 2024. After exclusions and loss to follow-up, 27 patients (25 women; median age = 52 years; interquartile range [IQR] = 42-61 years) with 75 PAVMs confirmed on chest computed tomography (CT) were included in the final analysis. Pre-treatment TR-MRA was performed 1 day before embolization, and post-treatment TR-MRA and non-enhanced chest CT were conducted 6 months later. Conventional angiography and pre-procedure CT served as reference standards. TR-MRA was performed using a 3-T scanner with a temporal resolution of 1-1.2 s. Two independent readers evaluated TR-MRA findings.ResultsAmong 75 PAVMs, 11 were previously treated lesions, including two cases of recanalization. Pre-treatment TR-MRA detected 98% (62/63) of naïve PAVMs confirmed on CT and angiography. Post-treatment TR-MRA detected 98% (60/61) of treated lesions. Inter-observer agreement was substantial to excellent (κ = 0.74 for pre-procedure diagnosis of PAVMs and 1.00 for post-procedure follow-up; P <0.05).ConclusionTR-MRA demonstrated excellent diagnostic performance for both pre- and post-embolization evaluation of PAVMs, providing reliable, radiation-free surveillance with diagnostic performance comparable to conventional angiography.

BackgroundThe Lake Louise Criteria (LLC) were updated in 2018 to improve accuracy in evaluating myocarditis. However, the diagnostic value of combining conventional magnetic resonance imaging (MRI) with contrast-enhanced whole-heart MRI (CE WH-MRI) in the diagnosis of acute myocarditis (AM) has not been determined.PurposeTo assess the diagnostic accuracy of the updated LLC and test the incremental value of CE WH-MRI in diagnosis of AM.Material and MethodsBetween March 2020 and November 2023, a total of 37 patients with clinically suspected AM were prospectively recruited for this study. The cardiac MR (CMR) protocol for myocarditis and controls included T2-STIR, breath-hold steady-state free precession, native T1, T2, CE WH-MRI, late gadolinium enhancement (LGE), and post-contrast T1 mapping.ResultsFor global native T1, the ideal cutoff value was 1308.5 ms (area under the curve [AUC]=0.879, sensitivity=82%, specificity=79%); for global T2, 43.2 ms (AUC=0.889, sensitivity=96%, specificity=75%), for ECV, 30.5% (AUC=0.946, sensitivity=97%, specificity=93%). The CE WH-MRI sequence detected 268 myocardial involvement (MI) segments, whereas 2D-LGE images identified 181 MI segments. Among 37 patients, 34 (91.9%) met the updated LLC definition for diagnosis, the AUC of updated LLC was 0.946.ConclusionThe updated LLC, as a recommended criterion for the diagnosis of AM, had better diagnostic accuracy compared with CMR mapping imaging. Moreover, this study highlighted the additional diagnostic value of CE WH-MRI in the identification of AM. Then, multiparametric CMR imaging can provide a satisfactory diagnostic value to enhance the accuracy of diagnosing AM.

BackgroundHepatic venous malformations (VMs) are adult-onset vascular anomalies that continue to be inaccurately labeled as "hepatic hemangiomas." Despite the widespread adoption of the International Society for the Study of Vascular Anomalies (ISSVA) classification, which distinguishes VMs from vascular tumors, outdated terminology persists. This misclassification may hinder diagnostic accuracy and limit the application of appropriate management strategies, including sclerotherapy.PurposeTo analyze the clinical, radiographic, and histopathologic characteristics of hepatic VMs in adults and assess the accuracy of the existing diagnoses.Material and MethodsThis is a retrospective review of a large tertiary referral center with a statewide catchment area, analyzing adult patients with pathology-proven hepatic VMs referred between January 2000 and July 2021. The original diagnosis and data on clinical, radiographic, pathological, and treatment methods of pathology-proven lesions were collected and analyzed.ResultsA total of 24 adult patients (13 women; mean age = 53.5 years) met the inclusion criteria. In 20 (83.3%) cases, imaging labeled the lesion as "hemangioma" before pathology confirmed the same diagnosis; in 4 (16.7%) cases, imaging initially suggested metastases, but pathology labeled them as "hemangioma." Most lesions were solitary (71%) and asymptomatic (67%). Symptomatic lesions had a larger mean diameter (9.3 cm) compared with the overall cohort (4.42 cm). Careful re-review of imaging and histopathology confirmed all lesions to be VMs.ConclusionHepatic VMs are frequently misdiagnosed as hemangiomas. Accurate classification is essential for improving clinical understanding, guiding treatment, and aligning terminology with current vascular anomaly standards.

BackgroundAssessment of subtle hip fractures on radiographs can be difficult, especially among less experienced emergency physicians, which may prolong the diagnosis and ultimately time to surgery. Clinical artificial intelligence (AI) decision support tools have shown great potential in assisting the detection of fractures on radiographs.PurposeTo investigate how a CE-marked AI fracture detection tool affects junior doctors' diagnostic accuracy in detecting hip fractures on radiographs.Material and MethodsEight junior doctors with affiliation to the Accident and Emergency (A&E) department read 246 hip radiographic examinations with and without AI support. The reference standard was determined by two musculoskeletal radiologists, to measure sensitivity and specificity for readers without and with support from the AI tool as well as the AI tool's standalone performance.ResultsMean sensitivity in detecting hip fractures increased significantly from 0.89 (95% confidence interval [CI] = 0.85-0.93) without AI support to 0.94 (95% CI = 0.92-0.97) (χ² = 9.27; P = 0.002) with AI support and the false-negative cases was thereby reduced by 49%. There was no significant change in mean specificity 0.90 (95% CI = 0.86-0.93) to 0.91 (95% CI = 0.88-0.94) (χ² = 0.34; P = 0.56). The AI standalone performance was 0.99 (95% CI = 0.99-1.00) and 0.73 (95% CI = 0.67-0.80) in sensitivity and specificity, respectively.ConclusionOut of eight junior doctors, seven detected more fractures with AI assistance than without. The applied performance gain for readers highlights the value of the product.

BackgroundIn Sweden, approximately 1300 patients are diagnosed with renal cell carcinoma (RCC) every year. The use of a computed tomography (CT) scan of the thoracic cavity in the preoperative work up of kidney cancer has increased in Sweden, and current national guidelines recommend that all patients, regardless of tumor size, should be evaluated this wayPurposeTo investigate the need for the preoperative routine to include a CT scan of the thoracic cavity when investigating renal masses 4 cm or smaller.Material and MethodsBetween 2017 and 2022, 496 patients at a university hospital and a regional hospital received treatment with curative intent for T1a tumors. Patient data and pathological findings were registered from patient records.ResultsMedian follow-up was 38 months. A total of 260 patients were examined with a preoperative CT scan of the thoracic cavity without pathology: 46 had not been scanned, 118 had indeterminate lesions, and metastasis was suspected in two cases. During follow-up, six patients had local relapse and none was diagnosed with lung metastasis. In no case did the preoperative CT of the thoracic cavity contribute to an early discovery of lung metastases.ConclusionOur conclusion is that a chest CT scan is superfluous in the preoperative work-up. The cost, and the time to treatment, could be reduced by precluding the chest CT in the preoperative work up for small renal tumors.

BackgroundRecent advancements in medical imaging technology have significantly increased the prevalence of automatic and semi-automatic segmentation techniques for foot bones, offering promising potential for improving diagnostic accuracy and efficiency. However, a critical challenge remains the scarcity of literature on the reliability and validation of these automated systems, underscoring the need for comprehensive studies to ensure their trustworthiness in clinical practice.PurposeTo implement a fully automated foot bone segmentation method processed exclusively using convolutional neuronal networks (CNNs).Material and MethodsFoot bones of 50 computed tomography (CT) scans were manually segmented. Of them, 48 were used to train three CNNs of a customized and optimized three-dimensional (3D) U-Net structure for the segmentation process. The so trained networks were then applied on the remaining two CT scans. The Dice coefficient and the Intersection over Union (IoU) metric were calculated to evaluate the CNN's ability of proper foot bone segmentation.ResultsThe CNN accurately segmented 5,090,689/5,434,749 voxels in the test sets, achieving an overall Dice coefficient of 0.97 and IoU of 0.94. Excellent segmentation results were obtained for the hindfoot, midfoot, hallux, sesamoids, and proximal phalanges, while lower performance was noted for the intermediate and distal phalanges of the lesser toes.ConclusionThe CNN networks demonstrated excellent ability to recognize foot bone structures on CT. Our findings underscore the potential of deep learning models in providing reliable and accurate segmentation of foot bones, paving the way for more widespread clinical adoption.

BackgroundThe prognostic significance of incidental cardiovascular calcifications-including coronary artery calcification (CAC), thoracic aortic calcification (TAC), aortic valve calcification (AVC), and mitral annular calcification (MAC)-detected on non-gated, non-contrast low-dose chest computed tomography (LDCT) remains unclear.PurposeTo evaluate the long-term prognostic significance of incidental cardiovascular calcifications detected on screening LDCT.Material and MethodsThis retrospective cohort study included individuals who underwent LDCT at a single health promotion center between 2007 and 2013. A cardiovascular radiologist quantified CAC, TAC, AVC, and MAC using dedicated software. Multivariable Cox proportional hazards regression was used to assess associations with all-cause mortality (ACM) and major adverse cardiovascular events (MACE), defined as revascularization, myocardial infarction, stroke, or cardiovascular death. Incremental prognostic performance was evaluated using Harrell's concordance index (C-index).ResultsAmong the 2434 included individuals (1863 men; median age = 54.2 years), CAC, TAC, AVC, and MAC were identified in 506 (20.8%), 1215 (49.9%), 159 (6.5%), and 49 (2.0%), respectively. The highest TAC category (≥1000) showed the strongest association with ACM (hazard ratio [HR] = 3.11, 95% confidence interval [CI] = 1.57-6.16; P = 0.001). The highest CAC category (≥400) showed the strongest association with MACE (HR = 8.67, 95% CI = 4.46-16.88; P <0.001). However, a combined model incorporating CAC, TAC, AVC, and MAC did not provide significant incremental prognostic value beyond CAC alone for ACM or MACE.ConclusionIncidental TAC was associated with increased long-term risk of ACM, while CAC was associated with MACE. However, their combined incorporation did not provide significant incremental prognostic value.

BackgroundMalignant extrahepatic biliary obstruction is most commonly caused by pancreatic cancer or cholangiocarcinoma and is frequently diagnosed at an advanced stage. Palliation with percutaneous metallic stent placement is often required, but data on large-scale outcomes remain limited.PurposeTo investigate the technical and clinical outcomes of percutaneous metallic stent placement in patients with malignant extrahepatic biliary obstruction and to examine the factors that influence patient survival and stent patency duration.Material and MethodsBetween January 2007 and December 2020, 612 patients (370 men, 242 women, mean age = 63.6 years; age range = 25-90 years) with malignant extrahepatic biliary obstruction were included in this retrospective study.ResultsPercutaneous metallic stents were successfully placed in all 612 patients. A total of 76 (12%) patients had major (n = 50, 8%) or minor (n = 26, 4%) complications. The percutaneous drainage catheter was successfully removed in 553 (90%) patients. In total, 53 patients were lost to follow-up and seven patients underwent subsequent biliary operation after stent placement. Stent occlusion occurred in 158/493 (32%) patients and median stent patency time was 274 days. Multivariate Cox regression analyses revealed that primary malignancy (P <0.001), stent patency (P <0.001), chemotherapy (P <0.001), and isolated biliary obstruction (P = 0.001) were independently associated with longer survival.ConclusionPercutaneous metallic stent placement is safe and effective in patients with malignant extrahepatic biliary obstruction. In addition, primary malignancy, stent patency, chemotherapy, and isolated biliary obstruction are significantly associated with longer survival within this poor-prognosis cohort.