Cardiac magnetic resonance radiomics for prediction of incident heart failure: a feasibility study in the UK Biobank Imaging cohort

Background

Cardiac magnetic resonance (CMR) radiomics is a novel image quantification technique with the potential to improve image-based disease diagnosis and prediction.

Purpose

In this proof-of-concept study, we aimed to evaluate the utility of CMR radiomics in the prediction of incident heart failure (HF).

Methods

We studied 32,121 UK Biobank participants with CMR. Incident HF was defined from linked Hospital Episode Statistics. To create a balanced cohort, we identified as comparators an equal number of randomly selected subjects who did not develop the outcome of interest during this period. Radiomics shape, first-order and texture features were extracted from short-axis cine images (left and right ventricle, left ventricular myocardium) using the Pyradiomics toolbox. Vascular risk factors (VRFs) were considered as additional predictors. Feature selection was conducted using the sequential forward selection technique and modelling was performed using Support Vector Machine (SVM) methods with 5-fold cross-validation. Models were developed using 1) VRFs alone, 2) radiomics alone, and 3) VRFs and radiomics. We determined model performance using receiver operating characteristic (ROC) curve and area under the curve (AUC) scores.

Results

Over average follow-up time of 3.7 (±1.3) years, 209 participants experienced incident HF. Among vascular risk factors, age, body size, hypertension, diabetes, high cholesterol were chosen for the incident HF predictive model (Accuracy: 0.66, AUC: 0.73) by the SVM methods. The model based on radiomics features reached a marginal improvement compared to vascular risk factors alone (Accuracy: 0.71, AUC: 0.75). The combination of VRFs and radiomics features significantly improved the performance of the model to predict incident HF compared to VRFs alone (Accuracy: 0.77; AUC: 0.83; p<0.05)

Conclusion

We demonstrate the feasibility of CMR radomics features to predict incident HF and illustrate their added value over vascular risk factors.

Funding Acknowledgement

Type of funding sources: Public grant(s) – EU funding.

Successful primary use of VVDL+V ECMO with cephalic drain in neonatal respiratory failure

OBJECTIVE

To describe the use of double-lumen venovenous (VVDL) extracorporeal membrane oxygenation (ECMO) with cephalic draining cannula (VVDL+V) as a primary approach for all neonatal respiratory diagnoses and to compare our single-center experience with data as collected in the Extracorporeal Life Support Organization (ELSO) database.

STUDY DESIGN

We retrospectively reviewed all cases of ECMO for neonatal respiratory failure performed in the neonatal intensive-care unit at a large referral children's hospital, the Children's Healthcare of Atlanta at Egleston (CHOA-E). Comparisons were then made to neonatal respiratory ECMO data retrieved from the ELSO database.

RESULTS

At CHOA-E 162 of 189 cases were completed with the VVDL+V approach. Survival in the VVDL+V cohort was 89.1% versus 68.7% from ELSO, P<0.001. For those complications considered, the overall risk of complication favored the CHOA-E VVDL+V group as compared with ELSO (odds ratio (OR) 0.71 (0.52-0.7)) as did the risk of neurologic complications (OR 0.29, (0.15-0.58)), including intracranial hemorrhage (OR 0.39 (0.18-0.97), P=0.011).

CONCLUSION

The VVDL+V approach can be used successfully as the primary approach for ECMO for neonatal respiratory failure of various etiologies and in this single-center cohort this approach was associated with improved survival and lower rates of complication as compared with the ELSO database.

APOE and ACE polymorphisms and dementia risk in the older population over prolonged follow-up: 10 years of incidence in the MRC CFA Study

Background: dementia risk conferred by apolipoprotein-E (APOE) and angiotensin-1-converting enzyme (ACE) polymorphisms have been reported for the MRC Cognitive Function and Ageing Study (CFAS) at 6-year follow-up. We concentrate on incident dementia risk over 10 years.

Methods: participants come from MRC CFAS, a multi-centre longitudinal population-based study of ageing in England and Wales. Three follow-up waves of data collection were used: 2, 6 and 10 years. Logistic regressions were undertaken to investigate associations between APOE (n = 955) and ACE (n = 856) alleles/genotypes and incident dementia. Two types of control groups were used: non-demented and highly functioning non-demented. Results were back-weighted.

Results: compared to APOE ε3, ε2 conferred protection of odds ratio (OR) = 0.3 (95% confidence interval, CI = 0.1–0.6) and ε4 risk of OR = 2.9 (95% CI = 1.7–4.9) for incident dementia. Compared to ε3/ε3, the ε3/ε4 and ε4/ε4 genotypes conferred risks of OR = 3.6 (95% CI = 1.8–7.3) and OR = 7.9 (95% CI = 1.6–39.2), respectively. The ε3/ε2 genotype protected against dementia (OR = 0.2, 95% CI = 0.1–0.7), and ε2/ε2 had a similar protective effect but with wide CIs (OR = 0.3, 95% CI = 0.1–1.7). Restricting the control group accentuated these differentials. The effects of ACE alleles/genotypes on incident dementia risk were small.

Conclusions:APOE but not ACE is associated with late-onset incident dementia in the population. Using longer term follow-up with proper adjustment for attrition and incident cases increases estimates of risk.

The capsid gene of feline calicivirus contains linear B-cell epitopes in both variable and conserved regions

In order to map linear B-cell (LBC) epitopes in the major capsid protein of feline calicivirus (FCV), an expression library containing random, short (100- to 200-bp) fragments of the FCV F9 capsid gene was constructed. Analysis of this library showed it to be representative of the region of the capsid gene that encodes the mature capsid protein. The library was screened by using polyclonal antisera from a cat that had been challenged experimentally with F9 to identify immunoreactive clones containing LBC epitopes. Twenty-six clones that reacted positively to feline antisera in immunoblots were identified. FCV-derived sequence from these clones mapped to a region of the capsid that spanned 126 amino acids and included variable regions C and E. An overlapping set of biotinylated peptides corresponding to this region was used to further map LBC epitopes by using F9 antisera. Four principal regions of reactivity were identified. Two fell within the hypervariable region at the 5' end of region E (amino acids [aa] 445 to 451 [antigenic site (ags) 2] and aa 451 to 457 [ags 3]). However, the other two were in conserved regions (aa 415 to 421 [ags 1; region D] and aa 475 to 479 [ags 4; central region E]). The reactivity of the peptide set with antisera from 11 other cats infected with a range of FCV isolates was also determined. Ten of 11 antisera reacted to conserved ags 4, suggesting that this region may be useful for future recombinant vaccine design.

Do experienced diagnosticians agree about the diagnosis of dementia from survey data? The effects of informants' reports and interviewers' vignettes

Dementia in community settings is often diagnosed by computerized algorithms. This study examines the extent to which independent diagnosticians agreed among themselves in diagnosing dementia, severity and type when presented with data obtained during a population-based incidence study of cognitive decline and dementia. Secondly, it examines how judgements, based initially on respondents' self-reports and cognitive performance, were affected first by informants' reports and then by short case-vignettes written by trained lay interviewers. Thirdly, it compares diagnosticians' diagnosis of dementia with the algorithmic diagnosis (AGECAT). The items presented were selected from two screening interviews at wave 1 and wave 2 separated by an interval of 2 years and from wave 2 assessment and informant interviews, and included medical, psychiatric and ADL items and interviewers' own observations. The sample (N = 42) was derived from the first year of the wave 2 assessments, potential dementia cases entering consecutively while presumed normals were selected randomly. Informants were available in 30. Agreement on diagnosis and type of dementia improved with increasing information, particularly from informants, but remained poor regarding severity. The number of cases of dementia, defined operationally, increased from 10 to 12 and uncertain cases fell from eight to six, but no respondent initially diagnosed as a dementia case was rediagnosed as a non-case, or vice versa. Dementia type changed from agreement about Alzheimer's disease to agreement about vascular dementia in one case. Operational and algorithmic diagnoses showed good agreement. Causes of disagreement, the role of vignettes and the relevance of the results for population surveys are discussed.

Sequences of the ribonucleotide reductase-encoding genes of felid herpesvirus 1 and molecular phylogenetic analysis

The felid herpesvirus 1 (FHV-1) genes encoding the two ribonucleotide reductase (RR) subunits (RR1, large subunit and RR2, small subunit) were cloned and their nucleotide (nt) sequence determined. The RR1 open reading frame (ORF) is 2358 nts long and is predicted to encode a protein of 786 amino acids (aa). In common with herpesviruses in the Varicellovirus genus of the alphaherpesvirus subfamily, FHV-1 RR1 lacks the N-terminal serine threonine protein kinase region present in herpes simplex virus (HSV)-1 and -2. FHV-1 RR1 has a predicted aa identity of 47-64% with other alphaherpesvirus RR1 peptides, falling to 26-29% for gammaherpesviruses. The RR2 ORF is 996 nts long, predicted to encode a protein of 332 aa and has aa identities of 64-70% with alphaherpesviruses and 38-39% with gammaherpesviruses. Molecular phylogenetic analysis groups FHV-1 with equid herpesviruses 1 and 4 (EHV 1 and 4), pseudorabies virus (PRV) and bovid herpesvirus 1 (BHV 1) within the genus Varicellovirus.

Susceptibility in cell culture of feline immunodeficiency virus to eighteen antiviral agents

The in-vitro susceptibilities of two strains of feline immunodeficiency virus to 18 antiviral agents were determined in two cell lines. In terms of inhibiting p24 antigen production, the nucleoside-analogue reverse transcriptase inhibitors were the most effective compounds. Inhibition was also observed with aurintricarboxylic acid, phosphonoformate and butyldeoxynorjirimycin, but not with the other agents tested.