1 February 2019 in Radiology
Radiomics-based machine learning of MRI studies can be useful to predict the tumor type of brain metastases. PurposeTo 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 methodsThis 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 algo...
30 July 2019 in Frontiers in Oncology
Background and Purpose: To date, imaging studies quantifying the amount of vasogenic edema reduction (VE) in patients with brain metastases (BM) treated with glucocorticoids (GC) have included a very limited number of patients and showed ambiguous results. Here, we aim to determine the radiological effect of GC on VE in BM patients in a large cohort with multiple primary tumor entities in a cross-sectional approach. Materials and Methods: This monocentric retrospective study includes 299 patients first-ever diagnosed with 2,759 intra-axial BM on the respective MRI. 126/299 patients received GC prior to MRI due to mass effect of edema on cranial CT scan and clinical symptoms (GC-pos) and 173 patients did not (GC-neg). GC dose was documented in 85/126 patients. All BM and their respective VE were semi-automatically segmented on post-contrast T1-weighted images. Results: VE volumes were higher in GC-pos compared to GC-neg (p = 0.009) and did not correlate with GC dose. Multivariate linear regression analysis with interaction terms on the assumption that BM volume and BM number influence the probability of GC administration shows that large and higher numbers of BM under GC treatment generate less VE than without (p < 0.001 and p = 0.038, respectively). The primary tumor type and total BM volume did not influence VE volume. Conclusion: Use of GC is especially effective for treatment of VE formation in patients with larger and multiple BM regardless of primary tumor type and dosage. However, based on the present data a direct causative relationship between GC and VE cannot be proven.
3 July 2015 in European Radiology
1 October 2014 in American Journal of Neuroradiology
BACKGROUND AND PURPOSE: No previous study compares neuroradiology training programs and teaching schedules across the globe, to our knowledge. This study was conducted to better understand international program requisites. MATERIALS AND METHODS: Data from 43 countries were collected by an e-mail-based questionnaire (response rate, 84.0%). Radiologists across the world were surveyed regarding the neuroradiology training schemes in their institutions. Answers were verified by officers of the national neuroradiology societies. RESULTS: While many countries do not provide fellowship training in neuroradiology (n = 16), others have formal postresidency curricula (n = 27). Many programs have few fellows and didactic sessions, but the 1- or 2-year duration of fellowship training is relatively consistent (n = 23/27, 85%). CONCLUSIONS: There is a wide variety of fellowship offerings, lessons provided, and ratios of teachers to learners in neuroradiology training programs globally.
16 September 2016 in Clinical Neuroradiology