To investigate the feasibility of tumor type prediction with MRI radiomic image features of different brain metastases in a multiclass machine learning approach for patients with unknown primary lesion at the time of diagnosis. Materials and methods: This single-center retrospective analysis included radiomic features of 658 brain metastases from T1-weighted contrast material-enhanced, T1-weighted nonenhanced, and fluid-attenuated inversion recovery (FLAIR) images in 189 patients (101 women, 88 men; mean age, 61 years; age range, 32-85 years). Images were acquired over a 9-year period (from September 2007 through December 2016) with different MRI units, reflecting heterogeneous image data. Included metastases originated from breast cancer (n = 143), small cell lung cancer (n = 151), non-small cell lung cancer (n = 225), gastrointestinal cancer (n = 50), and melanoma (n = 89). A total of 1423 quantitative image features and basic clinical data were evaluated by using random forest machine learning algorithms. Validation was performed with model-external fivefold cross validation. Comparative analysis of 10 randomly drawn cross-validation sets verified the stability of the results. The classifier performance was compared with predictions from a respective conventional reading by two radiologists. Results: Areas under the receiver operating characteristic curve of the five-class problem ranged between 0.64 (for non-small cell lung cancer) and 0.82 (for melanoma); all P values were less than .01. Prediction performance of the classifier was superior to the radiologists' readings. Highest differences were observed for melanoma, with a 17-percentage-point gain in sensitivity compared with the sensitivity of both readers; P values were less than .02. Conclusion: Quantitative features of routine brain MR images used in a machine learning classifier provided high discriminatory accuracy in predicting the tumor type of brain metastases.
BackgroundIn patients suffering from acute ischemic stroke from large vessel occlusion (LVO), mechanical thrombectomy (MT) often leads to successful reperfusion. Only approximately half of these patients have a favorable clinical outcome. Our aim was to determine the prognostic factors associated with poor clinical outcome following complete reperfusion.MethodsPatients treated with MT for LVO from a prospective single-center stroke registry between July 2015 and April 2019 were screened. Complete reperfusion was defined as Thrombolysis in Cerebral Infarction (TICI) grade 3. A modified Rankin scale at 90 days (mRS90) of 3–6 was defined as ‘poor outcome’. A logistic regression analysis was performed with poor outcome as a dependent variable, and baseline clinical data, comorbidities, stroke severity, collateral status, and treatment information as independent variables.Results123 patients with complete reperfusion (TICI 3) were included in this study. Poor clinical outcome was observed in 67 (54.5%) of these patients. Multivariable logistic regression analysis identified greater age (adjusted OR 1.10, 95% CI 1.04 to 1.17; p=0.001), higher admission National Institutes of Health Stroke Scale (NIHSS) (OR 1.14, 95% CI 1.02 to 1.28; p=0.024), and lower Alberta Stroke Program Early CT Score (ASPECTS) (OR 0.6, 95% CI 0.4 to 0.84; p=0.007) as independent predictors of poor outcome. Poor outcome was independent of collateral score.ConclusionPoor clinical outcome is observed in a large proportion of acute ischemic stroke patients treated with MT, despite complete reperfusion. In this study, futile recanalization was shown to occur independently of collateral status, but was associated with increasing age and stroke severity.
Objective: This work addresses two key problems of skin lesion classification. The first problem is the effective use of high-resolution images with pretrained standard architectures for image classification. The second problem is the high class imbalance encountered in real-world multi-class datasets. Methods: To use high-resolution images, we propose a novel patch-based attention architecture that provides global context between small, high-resolution patches. We modify three pretrained architectures and study the performance of patch-based attention. To counter class imbalance problems, we compare oversampling, balanced batch sampling, and class-specific loss weighting. Additionally, we propose a novel diagnosis-guided loss weighting method which takes the method used for groundtruth annotation into account. Results: Our patch-based attention mechanism outperforms previous methods and improves the mean sensitivity by 7 %. Class balancing significantly improves the mean sensitivity and we show that our diagnosis-guided loss weighting method improves the mean sensitivity by 3 % over normal loss balancing. Conclusion: The novel patch-based attention mechanism can be integrated into pretrained architectures and provides global context between local patches while outperforming other patch-based methods. Hence, pretrained architectures can be readily used with high-resolution images without downsampling. The new diagnosis-guided loss weighting method outperforms other methods and allows for effective training when facing class imbalance. Significance: The proposed methods improve automatic skin lesion classification. They can be extended to other clinical applications where high-resolution image data and class imbalance are relevant.
Background and Purpose: Intracranial hemorrhage has been observed in patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (COVID-19), but the clinical, imaging, and pathophysiological features of intracranial bleeding during COVID-19 infection remain poorly characterized. This study describes clinical and imaging characteristics of patients with COVID-19 infection who presented with intracranial bleeding in a European multicenter cohort. Methods: This is a multicenter retrospective, observational case series including 18 consecutive patients with COVID-19 infection and intracranial hemorrhage. Data were collected from February to May 2020 at five designated European special care centers for COVID-19. The diagnosis of COVID-19 was based on laboratory-confirmed diagnosis of SARS-CoV-2. Intracranial bleeding was diagnosed on computed tomography (CT) of the brain within one month of the date of COVID-19 diagnosis. The clinical, laboratory, radiologic, and pathologic findings, therapy and outcomes in COVID-19 patients presenting with intracranial bleeding were analyzed. Results: Eighteen patients had evidence of acute intracranial bleeding within 11 days (IQR 9–29) of admission. Six patients had parenchymal hemorrhage (33.3%), 11 had subarachnoid hemorrhage (SAH) (61.1%), and one patient had subdural hemorrhage (5.6%). Three patients presented with intraventricular hemorrhage (IVH) (16.7%). Conclusion: This study represents the largest case series of patients with intracranial hemorrhage diagnosed with COVID-19 based on key European countries with geospatial hotspots of SARS-CoV-2. Isolated SAH along the convexity may be a predominant bleeding manifestation and may occur in a late temporal course of severe COVID-19.
Background and Purpose: Intracranial hemorrhage (ICH) remains a major complication of endovascular treatment (ET) in acute stroke. The aim of this study was to identify clinical and imaging predictors for ICH in patients with acute ischemic stroke undergoing successful ET. Methods: We performed a retrospective analysis of patients with large vessel occlusion in the anterior circulation who underwent successful ET at our university medical center between 2015 and 2018. ICH was diagnosed on non-enhanced CT and a binary outcome was defined: ICH occurrence in the immediate post-interventional phase within 12–36 h (yes/no). The impacts of clinical, radiological, and interventional parameters on outcome were assessed in logistic regression models. Results: One hundred and seven patients fulfilled the inclusion criteria. 37 (34.6%) showed an ICH of which 7 (6.5%) patients were diagnosed as symptomatic and 30 (28.04%) as asymptomatic. Multivariable regression analyses identified a lower ASPECTS (adjusted odds ratio (OR) 1.95, 95%CI: 1.4–3.63, P = 0.037), low collateral score (adjusted OR 0.12, 95%CI: 0.03–0.49, P = 0.003) and high Net Water Uptake (NWU) (adjusted OR 1.56, 95%CI: 2.34–1.03, P = 0.007) as independent predictors of ICH after successful ET. Conclusions: CT-based quantitative NWU, ASPECTS, and collateral score mediate tissue vulnerability and are reliable independent predictors of a bleeding event after successful ET. This imaging-based prediction model might be useful for early stratification of patients at high risk of a bleeding event after ET, especially with low ASPECTS.
Background Ischemic water uptake in acute stroke is a reliable indicator of lesion age. Nevertheless, inter-individually varying edema progression has been observed and elevated water uptake has recently been described as predictor of malignant infarction. Aims We hypothesized that early-elevated lesion water uptake indicates accelerated “tissue clock” desynchronized with “time clock” and therefore predicts poor clinical outcome despite successful recanalization. Methods Acute middle cerebral artery stroke patients with multimodal admission-CT who received successful thrombectomy (TICI 2b/3) were analyzed. Net water uptake (NWU), a quantitative imaging biomarker of ischemic edema, was determined in admission-CT and tested as predictor of clinical outcome using modified Rankin Scale (mRS) after 90 days. A binary outcome was defined for mRS 0–4 and mRS 5–6. Results Seventy-two patients were included. The mean NWU (SD) in patients with mRS 0–4 was lower compared to patients with mRS 5–6 (5.0% vs. 12.1%; p < 0.001) with similar time from symptom onset to imaging (2.6 h vs. 2.4 h; p = 0.7). Based on receiver operating curve analysis, NWU above 10% identified patients with very poor outcome with high discriminative power (AUC 0.85), followed by Alberta Stroke Program Early CT Score (ASPECTS) (AUC: 0.72) and National Institutes of Health Stroke Scale (NIHSS) (AUC: 0.72). Conclusions Quantitative NWU may serve as an indicator of “tissue clock” and pronounced early brain edema with elevated NWU might suggest a desynchronized “tissue clock” with real “time clock” and therefore predict futile recanalization with poor clinical outcome.
BackgroundBenefit of thrombectomy in patients with a low initial Alberta Stroke Program Early CT Score (ASPECTS) is still uncertain. We hypothesized that, despite low ASPECTS, patients may benefit from endovascular recanalization if good collaterals are present.MethodsIschemic stroke patients with large vessel occlusion in the anterior circulation and an ASPECTS of ≤5 were analyzed. Collateral status (CS) was assessed using a 5-point-scoring system in CT angiography with poor CS defined as CS=0–1. Clinical outcome was determined using the modified Rankin Scale (mRS) score after 90 days. Edema formation was measured in admission and follow-up CT by net water uptake.Results27/100 (27%) patients exhibited a CS of 2–4. 50 patients underwent successful vessel recanalization and 50 patients had a persistent vessel occlusion. In multivariable logistic regression analysis, collateral status (OR 3.0; p=0.003) and vessel recanalization (OR 12.2; p=0.009) significantly increased the likelihood of a good outcome (mRS 0–3). A 1-point increase in CS was associated with 1.9% (95% CI 0.2% to 3.7%) lowered lesion water uptake in follow-up CT .ConclusionEndovascular recanalization in patients with ASPECTS of ≤5 but good collaterals was linked to improved clinical outcome and attenuated edema formation. Collateral status may serve as selection criterion for thrombectomy in low ASPECTS patients.
Background and Purpose: This study evaluates the benefit of endovascular treatment (EVT) for patients with extensive baseline stroke compared with best medical treatment. Methods: This retrospective, multicenter study compares EVT and best medical treatment for computed tomography (CT)–based selection of patients with extensive baseline infarcts (Alberta Stroke Program Early CT Score ≤5) attributed to anterior circulation stroke. Patients were selected from the German Stroke Registry and 3 tertiary stroke centers. Primary functional end points were rates of good (modified Rankin Scale score of ≤3) and very poor outcome (modified Rankin Scale score of ≥5) at 90 days. Secondary safety end point was the occurrence of symptomatic intracerebral hemorrhage. Angiographic outcome was evaluated with the modified Thrombolysis in Cerebral Infarction Scale. Results: After 1:1 pair matching, a total of 248 patients were compared by treatment arm. Good functional outcome was observed in 27.4% in the EVT group, and in 25% in the best medical treatment group ( P =0.665). Advanced age (adjusted odds ratio, 1.08 [95% CI, 1.05–1.10], P <0.001) and symptomatic intracerebral hemorrhage (adjusted odds ratio, 6.35 [95% CI, 2.08–19.35], P <0.001) were independently associated with very poor outcome. Mortality (43.5% versus 28.9%, P =0.025) and symptomatic intracerebral hemorrhage (16.1% versus 5.6%, P =0.008) were significantly higher in the EVT group. The lowest rates of good functional outcome (≈15%) were observed in groups of failed and partial recanalization (modified Thrombolysis in Cerebral Infarction Scale score of 0/1–2a), whereas patients with complete recanalization (modified Thrombolysis in Cerebral Infarction Scale score of 3) with recanalization attempts ≤2 benefitted the most (modified Rankin Scale score of ≤3:42.3%, P =0.074) compared with best medical treatment. Conclusions: In daily clinical practice, EVT for CT–based selected patients with low Alberta Stroke Program Early CT Score anterior circulation stroke may not be beneficial and is associated with increased risk for hemorrhage and mortality, especially in the elderly. However, first- or second-pass complete recanalization seems to reveal a clinical benefit of EVT highlighting the vulnerability of the low Alberta Stroke Program Early CT Score subgroup. REGISTRATION: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT03356392.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.