Search

Objectives To investigate the differentiation of premalignant from benign colorectal polyps detected by CT colonography using deep learning. Methods In this retrospective analysis of an average risk colorectal cancer screening sample, polyps of all size categories and morphologies were manually segmented on supine and prone CT colonography images and classified as premalignant (adenoma) or benign (hyperplastic polyp or regular mucosa) according to histopathology. Two deep learning models SEG and noSEG were trained on 3D CT colonography image subvolumes to predict polyp class, and model SEG was additionally trained with polyp segmentation masks. Diagnostic performance was validated in an independent external multicentre test sample. Predictions were analysed with the visualisation technique Grad-CAM++. Results The training set consisted of 107 colorectal polyps in 63 patients (mean age: 63 ± 8 years, 40 men) comprising 169 polyp segmentations. The external test set included 77 polyps in 59 patients comprising 118 polyp segmentations. Model SEG achieved a ROC-AUC of 0.83 and 80% sensitivity at 69% specificity for differentiating premalignant from benign polyps. Model noSEG yielded a ROC-AUC of 0.75, 80% sensitivity at 44% specificity, and an average Grad-CAM++ heatmap score of ≥ 0.25 in 90% of polyp tissue. Conclusions In this proof-of-concept study, deep learning enabled the differentiation of premalignant from benign colorectal polyps detected with CT colonography and the visualisation of image regions important for predictions. The approach did not require polyp segmentation and thus has the potential to facilitate the identification of high-risk polyps as an automated second reader. Key Points • Non-invasive deep learning image analysis may differentiate premalignant from benign colorectal polyps found in CT colonography scans. • Deep learning autonomously learned to focus on polyp tissue for predictions without the need for prior polyp segmentation by experts. • Deep learning potentially improves the diagnostic accuracy of CT colonography in colorectal cancer screening by allowing for a more precise selection of patients who would benefit from endoscopic polypectomy, especially for patients with polyps of 6–9 mm size.

Background and Purpose: Basilar artery occlusion is associated with high morbidity and mortality. Optimal imaging and treatment strategy are still controversial and prognosis estimation challenging. We, therefore, aimed to determine the predictive value of computed tomography perfusion (CTP) parameters for functional outcome in patients with basilar artery occlusion in the context of endovascular treatment. Methods: Patients with basilar artery occlusion who underwent endovascular treatment were selected from a prospectively acquired cohort. Ischemic changes were assessed with the posterior-circulation Acute Stroke Prognosis Early Computed Tomography Score on noncontrast computed tomography, computed tomography angiography (CTA) source images, and CTP maps. Basilar artery on CTA score, posterior-circulation CTA score, and posterior-circulation collateral score were evaluated on CTA. Perfusion deficit volumes were quantified on CTP maps. Good functional outcome was defined as modified Rankin Scale score ≤3 at 90 days. Statistical analysis included binary logistic regressions and receiver operating characteristics analyses. Results: Among 49 patients who matched the inclusion criteria, 24 (49.0%) achieved a good outcome. In univariate analysis, age, National Institutes of Health Stroke Scale score on admission, posterior cerebral artery involvement, absence of or hypoplastic posterior communicating arteries, basilar artery on CTA score, posterior-circulation Acute Stroke Prognosis Early Computed Tomography Score, and perfusion deficit volumes on all CTP parameter maps presented significant association with functional outcome ( P Conclusions: CTP imaging parameters contain prognostic information for functional outcome in patients with stroke due to basilar artery occlusion and may identify patients with higher risk of disability at an early stage of hospitalization.

Background CT colonography does not enable definite differentiation between benign and premalignant colorectal polyps. Purpose To perform machine learning-based differentiation of benign and premalignant colorectal polyps detected with CT colonography in an average-risk asymptomatic colorectal cancer screening sample with external validation using radiomics. Materials and Methods In this secondary analysis of a prospective trial, colorectal polyps of all size categories and morphologies were manually segmented on CT colonographic images and were classified as benign (hyperplastic polyp or regular mucosa) or premalignant (adenoma) according to the histopathologic reference standard. Quantitative image features characterizing shape (n = 14), gray level histogram statistics (n = 18), and image texture (n = 68) were extracted from segmentations after applying 22 image filters, resulting in 1906 feature-filter combinations. Based on these features, a random forest classification algorithm was trained to predict the individual polyp character. Diagnostic performance was validated in an external test set. Results The random forest model was fitted using a training set consisting of 107 colorectal polyps in 63 patients (mean age, 63 years ± 8 [standard deviation]; 40 men) comprising 169 segmentations on CT colonographic images. The external test set included 77 polyps in 59 patients comprising 118 segmentations. Random forest analysis yielded an area under the receiver operating characteristic curve of 0.91 (95% CI: 0.85, 0.96), a sensitivity of 82% (65 of 79) (95% CI: 74%, 91%), and a specificity of 85% (33 of 39) (95% CI: 72%, 95%) in the external test set. In two subgroup analyses of the external test set, the area under the receiver operating characteristic curve was 0.87 in the size category of 6-9 mm and 0.90 in the size category of 10 mm or larger. The most important image feature for decision making (relative importance of 3.7%) was quantifying first-order gray level histogram statistics. Conclusion In this proof-of-concept study, machine learning-based image analysis enabled noninvasive differentiation of benign and premalignant colorectal polyps with CT colonography. © RSNA, 2021 Online supplemental material is available for this article.

Subclinical effects of coffee consumption (CC) with regard to metabolic, cardiac, and neurological complications were evaluated using a whole-body magnetic resonance imaging (MRI) protocol. A blended approach was used to estimate habitual CC in a population-based study cohort without a history of cardiovascular disease. Associations of CC with MRI markers of gray matter volume, white matter hyperintensities, cerebral microhemorrhages, total and visceral adipose tissue (VAT), hepatic proton density fat fraction, early/late diastolic filling rate, end-diastolic/-systolic and stroke volume, ejection fraction, peak ejection rate, and myocardial mass were evaluated by linear regression. In our analysis with 132 women and 168 men, CC was positively associated with MR-based cardiac function parameters including late diastolic filling rate, stroke volume (p &lt, 0.01 each), and ejection fraction (p &lt, 0.05) when adjusting for age, sex, smoking, hypertension, diabetes, Low-density lipoprotein (LDL), triglycerides, cholesterol, and alcohol consumption. CC was inversely associated with VAT independent of demographic variables and cardiovascular risk factors (p &lt, 0.05), but this association did not remain significant after additional adjustment for alcohol consumption. CC was not significantly associated with potential neurodegeneration. We found a significant positive and independent association between CC and MRI-based systolic and diastolic cardiac function. CC was also inversely associated with VAT but not independent of alcohol consumption.

IntroductionAs white matter hyperintensities (WMHs) of the brain are associated with an increased risk of stroke, cognitive decline, and depression, elucidating the associated risk factors is important. In addition to age and hypertension, pre-diabetes and diabetes may play important roles in the development of WMHs. Previous studies have, however, shown conflicting results. We aimed to investigate the effect of diabetes status and quantitative markers of glucose metabolism on WMH volume in a population-based cohort without prior cardiovascular disease.Research design and methods400 participants underwent 3 T MRI. WMHs were manually segmented on 3D fluid-attenuated inversion recovery images. An oral glucose tolerance test (OGTT) was administered to all participants not previously diagnosed with diabetes to assess 2-hour serum glucose concentrations. Fasting glucose concentrations and glycated hemoglobin (HbA1c) levels were measured. Zero-inflated negative binomial regression analyses of WMH volume and measures of glycemic status were performed while controlling for cardiovascular risk factors and multiple testing.ResultsThe final study population comprised 388 participants (57% male; age 56.3±9.2 years; n=98 with pre-diabetes, n=51 with diabetes). Higher WMH volume was associated with pre-diabetes (p=0.001) and diabetes (p=0.026) compared with normoglycemic control participants after adjustment for cardiovascular risk factors. 2-hour serum glucose (pConclusionOur results indicate that high 2-hour serum glucose concentration in OGTT, but not fasting glucose levels, may be an independent risk factor for the development of WMHs, with the potential to inform intensified prevention strategies in individuals at risk of WMH-associated morbidity.

BACKGROUND/AIM Intraarterial Technetium-99m-Macroaggregated Albumin (99mTc-MAA) administration is an established method to predict particle distribution prior to radioembolization. This study aimed to analyse the impact of intraarterial administration of 99mTc-MAA on changes in liver-specific laboratory parameters and to assess whether such changes are associated with post-radioembolization hepatotoxicity. PATIENTS AND METHODS A total of 202 patients treated with radioembolization received prior mapping angiography with 99mTc-MAA administration. All patients underwent clinical and laboratory examinations, including liver-specific parameters at certain times before and after mapping angiography/99mTc-MAA administration, as well as before radioembolization and during follow-up. RESULTS Bilirubin increased temporarily after 99mTc-MAA administration (p

To identify the most important factors that impact brain volume, while accounting for potential collinearity, we used a data-driven machine-learning approach. Gray Matter Volume (GMV) was derived from magnetic resonance imaging (3T, FLAIR) and adjusted for intracranial volume (ICV). 93 potential determinants of GMV from the categories sociodemographics, anthropometric measurements, cardio-metabolic variables, lifestyle factors, medication, sleep, and nutrition were obtained from 293 participants from a population-based cohort from Southern Germany. Elastic net regression was used to identify the most important determinants of ICV-adjusted GMV. The four variables age (selected in each of the 1000 splits), glomerular filtration rate (794 splits), diabetes (323 splits) and diabetes duration (122 splits) were identified to be most relevant predictors of GMV adjusted for intracranial volume. The elastic net model showed better performance compared to a constant linear regression (mean squared error = 1.10 vs. 1.59, p

Objectives To determine a cut-off apparent diffusion coefficient (ADC) value distinguishing local recurrence from scar tissue in patients with rectal cancer treated with complete surgical tumour removal. Methods 72 patients were retrospectively included. Patients underwent 1.5T MRI including multiplanar T2-weighted turbo-spin-echo sequences (TSE) and axial single-shot epi-diffusion-weighted sequences (EPSE). Two independent observers measured mean tumour and scar tissue ADCs by manually drawing regions of interest (ROIs). The t -test and ROC analysis were used for comparison and determining an optimal discrimination threshold. As reference standard histopathological results were used in 23 patients (32%) and clinical follow-up in 49 patients (68%). Results Recurrent rectal cancer was found in 30 patients (4 female, 26 male, median age 63.13 years) and treatment related changes such as scar tissue in 42 patients (11 female, 31 male, median age 63.67 years). The mean ADC value of tumour recurrence was 1.02×10 −3 mm 2 /s (0.63–1.44×10 −3 mm 2 /s) and of scar tissue 1.77×10 −3 mm 2 /s (1.11–2.41×10 −3 mm 2 /s) showing a statistically significant difference (p −3 mm 2 /s with a sensitivity, specificity, and accuracy of 93%, 91%, and 92% respectively. Conclusions Diffusion weighted MRI allows for the differentiation of tumour recurrence from scar tissue after surgical resection of rectal cancer.

Objective Hepatocellular carcinoma (HCC) is the most common cause of primary liver cancer. A major part of diagnostic HCC work-up is based on imaging findings from sonography, computed tomography (CT), or magnetic resonance imaging (MRI) scans. Contrast-enhanced ultrasound (CEUS) allows for the dynamic assessment of the microperfusion pattern of suspicious liver lesions. This study aimed to evaluate the diagnostic value of CEUS compared with CT scans for assessing HCC. Methods We performed a retrospective, single-center study between 2004 and 2018 on 234 patients with suspicious liver lesions who underwent CEUS and CT examinations. All patients underwent native B-mode, color Doppler and CEUS after providing informed consent. Every CEUS examination was performed and interpreted by a single experienced radiologist (European Federation of Societies for Ultrasound in Medicine and Biology level 3). Results CEUS was performed on all included patients without occurrence of any adverse effects. CEUS showed a sensitivity of 94%, a specificity of 70%, a positive predictive value of 93% and a negative predictive value of 72% for analyzing HCC compared with CT as the diagnostic gold standard. Conclusions CEUS has an excellent safety profile and shows a high diagnostic accuracy in assessing HCC compared with corresponding results from CT scans.