Giralt‐Steinhauer, Eva

Background: Fundamental advances in stroke care will require pooling imaging phenotype data from multiple centers, to complement the current aggregation of genomic, environmental, and clinical information. Sharing clinically acquired MRI data from multiple hospitals is challenging due to inherent heterogeneity of clinical data, where the same MRI series may be labeled differently depending on vendor and hospital. Furthermore, the de-identification process may remove data describing the MRI series, requiring human review. However, manually annotating the MRI series is not only laborious and slow but prone to human error. In this work, we present a recurrent convolutional neural network (RCNN) for automated classification of the MRI series. Methods: We randomly selected 1000 subjects from the MRI-GENetics Interface Exploration study and partitioned them into 800 training, 100 validation and 100 testing subjects. We categorized the MRI series into 24 groups (see Table). The RCNN used a modified AlexNet to extract features from 2D slices. AlexNet was pretrained on ImageNet photographs. Since clinical MRI are 3D and 4D, a gated recurrent unit neural network was used to aggregate information from multiple 2D slices to make the final prediction. Results: We achieved a classification accuracy (correct/total cases) of 99.8%, 98.5% and 97.5% on the training, validation and testing set, respectively. The averaged F1-score (percent overlap between predicted cases and actual cases) over all categories were 99.8% 98.2% and 94.4% on the training, validation and testing set. Conclusion: We showed that automated annotation of MRI series by repurposing deep-learning techniques used for photographic image recognition tasks is feasible. Such methods can be used to facilitate high throughput curation of MRI data acquired across multiple centers and enable scientifically productive collaboration by researchers and, ultimately enhancing big data stroke research.

Background and Purpose— Hematoma volume is an important determinant of clinical outcome in spontaneous intracerebral hemorrhage (ICH). We performed a genome-wide association study (GWAS) of hematoma volume with the aim of identifying novel biological pathways involved in the pathophysiology of primary brain injury in ICH. Methods— We conducted a 2-stage (discovery and replication) case-only genome-wide association study in patients with ICH of European ancestry. We utilized the admission head computed tomography to calculate hematoma volume via semiautomated computer-assisted technique. After quality control and imputation, 7 million genetic variants were available for association testing with ICH volume, which was performed separately in lobar and nonlobar ICH cases using linear regression. Signals with P <5×10 − 8 were pursued in replication and tested for association with admission Glasgow coma scale and 3-month post-ICH dichotomized (0–2 versus 3–6) modified Rankin Scale using ordinal and logistic regression, respectively. Results— The discovery phase included 394 ICH cases (228 lobar and 166 nonlobar) and identified 2 susceptibility loci: a genomic region on 22q13 encompassing PARVB (top single-nucleotide polymorphism rs9614326: β, 1.84; SE, 0.32; P =4.4×10 −8 ) for lobar ICH volume and an intergenic region overlying numerous copy number variants on 17p12 (top single-nucleotide polymorphism rs11655160: β, 0.95; SE, 0.17; P =4.3×10 −8 ) for nonlobar ICH volume. The replication included 240 ICH cases (71 lobar and 169 nonlobar) and corroborated the association for 17p12 ( P =0.04; meta-analysis P =2.5×10 −9 ; heterogeneity, P =0.16) but not for 22q13 ( P =0.49). In multivariable analysis, rs11655160 was also associated with lower admission Glasgow coma scale (odds ratio, 0.17; P =0.004) and increased risk of poor 3-month modified Rankin Scale (odds ratio, 1.94; P =0.045). Conclusions— We identified 17p12 as a novel susceptibility risk locus for hematoma volume, clinical severity, and functional outcome in nonlobar ICH. Replication in other ethnicities and follow-up translational studies are needed to elucidate the mechanism mediating the observed association.

ObjectiveIn observational epidemiologic studies, higher plasma high‐density lipoprotein cholesterol (HDL‐C) has been associated with increased risk of intracerebral hemorrhage (ICH). DNA sequence variants that decrease cholesteryl ester transfer protein (CETP) gene activity increase plasma HDL‐C; as such, medicines that inhibit CETP and raise HDL‐C are in clinical development. Here, we test the hypothesis that CETP DNA sequence variants associated with higher HDL‐C also increase risk for ICH.MethodsWe performed 2 candidate‐gene analyses of CETP. First, we tested individual CETP variants in a discovery cohort of 1,149 ICH cases and 1,238 controls from 3 studies, followed by replication in 1,625 cases and 1,845 controls from 5 studies. Second, we constructed a genetic risk score comprised of 7 independent variants at the CETP locus and tested this score for association with HDL‐C as well as ICH risk.ResultsTwelve variants within CETP demonstrated nominal association with ICH, with the strongest association at the rs173539 locus (odds ratio [OR] = 1.25, standard error [SE] = 0.06, p = 6.0 × 10−4) with no heterogeneity across studies (I2 = 0%). This association was replicated in patients of European ancestry (p = 0.03). A genetic score of CETP variants found to increase HDL‐C by ∼2.85mg/dl in the Global Lipids Genetics Consortium was strongly associated with ICH risk (OR = 1.86, SE = 0.13, p = 1.39 × 10−6).InterpretationGenetic variants in CETP associated with increased HDL‐C raise the risk of ICH. Given ongoing therapeutic development in CETP inhibition and other HDL‐raising strategies, further exploration of potential adverse cerebrovascular outcomes may be warranted. Ann Neurol 2016;80:730–740

Background and Purpose— NINDS (National Institute of Neurological Disorders and Stroke)-SiGN (Stroke Genetics Network) is an international consortium of ischemic stroke studies that aims to generate high-quality phenotype data to identify the genetic basis of pathogenic stroke subtypes. This analysis characterizes the etiopathogenetic basis of ischemic stroke and reliability of stroke classification in the consortium. Methods— Fifty-two trained and certified adjudicators determined both phenotypic (abnormal test findings categorized in major pathogenic groups without weighting toward the most likely cause) and causative ischemic stroke subtypes in 16 954 subjects with imaging-confirmed ischemic stroke from 12 US studies and 11 studies from 8 European countries using the web-based Causative Classification of Stroke System. Classification reliability was assessed with blinded readjudication of 1509 randomly selected cases. Results— The distribution of pathogenic categories varied by study, age, sex, and race ( P <0.001 for each). Overall, only 40% to 54% of cases with a given major ischemic stroke pathogenesis (phenotypic subtype) were classified into the same final causative category with high confidence. There was good agreement for both causative (κ 0.72; 95% confidence interval, 0.69–0.75) and phenotypic classifications (κ 0.73; 95% confidence interval, 0.70–0.75). Conclusions— This study demonstrates that pathogenic subtypes can be determined with good reliability in studies that include investigators with different expertise and background, institutions with different stroke evaluation protocols and geographic location, and patient populations with different epidemiological characteristics. The discordance between phenotypic and causative stroke subtypes highlights the fact that the presence of an abnormality in a patient with stroke does not necessarily mean that it is the cause of stroke.