Mihir M. Sanghvi

Towards the Semantic Enrichment of Free-Text Annotation of Image Quality Assessment for UK Biobank Cardiac Cine MRI Scans

Image quality assessment is fundamental as it affects the level of confidence in any output obtained from image analysis. Clinical research imaging scans do not often come with an explicit evaluation of their quality, however reports are written associated to the patient/volunteer scans. This rich free-text documentation has the potential to provide automatic image quality assessment if efficiently processed and structured. This paper aims at showing how the use of Semantic Web technology for structuring free-text documentation can provide means for automatic image quality assessment. We aim to design and implement a semantic layer for a special dataset, the annotations made in the context of the UK Biobank Cardiac Cine MRI pilot study. This semantic layer will be a powerful tool to automatically infer or validate quality scores for clinical images and efficiently query image databases based on quality information extracted from the annotations. In this paper we motivate the need for this semantic layer, present an initial version of our ontology as well as preliminary results. The presented approach has the potential to be extended to broader projects and ultimately employed in the clinical setting.

Automated cardiovascular magnetic resonance image analysis with fully convolutional networks

BackgroundCardiovascular resonance (CMR) imaging is a standard imaging modality for assessing cardiovascular diseases (CVDs), the leading cause of death globally. CMR enables accurate quantification of the cardiac chamber volume, ejection fraction and myocardial mass, providing information for diagnosis and monitoring of CVDs. However, for years, clinicians have been relying on manual approaches for CMR image analysis, which is time consuming and prone to subjective errors. It is a major clinical challenge to automatically derive quantitative and clinically relevant information from CMR images.MethodsDeep neural networks have shown a great potential in image pattern recognition and segmentation for a variety of tasks. Here we demonstrate an automated analysis method for CMR images, which is based on a fully convolutional network (FCN). The network is trained and evaluated on a large-scale dataset from the UK Biobank, consisting of 4,875 subjects with 93,500 pixelwise annotated images. The performance of the method has been evaluated using a number of technical metrics, including the Dice metric, mean contour distance and Hausdorff distance, as well as clinically relevant measures, including left ventricle (LV) end-diastolic volume (LVEDV) and end-systolic volume (LVESV), LV mass (LVM); right ventricle (RV) end-diastolic volume (RVEDV) and end-systolic volume (RVESV).ResultsBy combining FCN with a large-scale annotated dataset, the proposed automated method achieves a high performance in segmenting the LV and RV on short-axis CMR images and the left atrium (LA) and right atrium (RA) on long-axis CMR images. On a short-axis image test set of 600 subjects, it achieves an average Dice metric of 0.94 for the LV cavity, 0.88 for the LV myocardium and 0.90 for the RV cavity. The mean absolute difference between automated measurement and manual measurement is 6.1 mL for LVEDV, 5.3 mL for LVESV, 6.9 gram for LVM, 8.5 mL for RVEDV and 7.2 mL for RVESV. On long-axis image test sets, the average Dice metric is 0.93 for the LA cavity (2-chamber view), 0.95 for the LA cavity (4-chamber view) and 0.96 for the RA cavity (4-chamber view). The performance is comparable to human inter-observer variability.ConclusionsWe show that an automated method achieves a performance on par with human experts in analysing CMR images and deriving clinically relevant measures.

The impact of menopausal hormone therapy (MHT) on cardiac structure and function: Insights from the UK Biobank imaging enhancement study

Background The effect of menopausal hormone therapy (MHT)–previously known as hormone replacement therapy–on cardiovascular health remains unclear and controversial. This cross-sectional study examined the impact of MHT on left ventricular (LV) and left atrial (LA) structure and function, alterations in which are markers of subclinical cardiovascular disease, in a population-based cohort. Methods Post-menopausal women who had never used MHT and those who had used MHT ≥3 years participating in the UK Biobank who had undergone cardiovascular magnetic resonance (CMR) imaging and free of known cardiovascular disease were included. Multivariable linear regression was performed to examine the relationship between cardiac parameters and MHT use ≥3 years. To explore whether MHT use on each of the cardiac outcomes differed by age, multivariable regression models were constructed with a cross-product of age and MHT fitted as an interaction term. Results Of 1604 post-menopausal women, 513 (32%) had used MHT ≥3 years. In the MHT cohort, median age at menopause was 50 (IQR: 45–52) and median duration of MHT was 8 years. In the non-MHT cohort, median age at menopause was 51 (IQR: 48–53). MHT use was associated with significantly lower LV end-diastolic volume (122.8 ml vs 119.8 ml, effect size = -2.4%, 95% CI: -4.2% to -0.5%; p = 0.013) and LA maximal volume (60.2 ml vs 57.5 ml, effect size = -4.5%, 95% CI: -7.8% to -1.0%; p = 0.012). There was no significant difference in LV mass. MHT use significantly modified the effect between age and CMR parameters; MHT users had greater decrements in LV end-diastolic volume, LV end-systolic volume and LA maximal volume with advancing age. Conclusions MHT use was not associated with adverse, subclinical changes in cardiac structure and function. Indeed, significantly smaller LV and LA chamber volumes were observed which have been linked to favourable cardiovascular outcomes. These findings represent a novel approach to examining MHT’s effect on the cardiovascular system.

Real-Time Prediction of Segmentation Quality.

Recent advances in deep learning based image segmentation methods have enabled real-time performance with human-level accuracy. However, occasionally even the best method fails due to low image quality, artifacts or unexpected behaviour of black box algorithms. Being able to predict segmentation quality in the absence of ground truth is of paramount importance in clinical practice, but also in large-scale studies to avoid the inclusion of invalid data in subsequent analysis.

Cardiovascular magnetic resonance imaging for amyloidosis: The state-of-the-art

Amyloidosis results from insoluble precursor proteins being deposited in the extracellular compartment. The prognosis of the disease is predominantly determined by cardiac involvement due to amyloid accumulation that contributes to cardiac dysfunction and disturbed conduction of cardiac electrical signals. The clinical and radiological manifestations of amyloidosis are often non-specific, making amyloidosis a diagnostic challenge both for clinicians and radiologists. Cardiovascular magnetic resonance imaging, including conventional sequences, late gadolinium enhancement, T1 mapping and determination of extracellular volume fraction is a multi-dimensional modality for the assessment and diagnosis of cardiac amyloidosis and, in addition, is an excellent tool for risk stratification and disease tracking.

Variation in lung function and alterations in cardiac structure and function—Analysis of the UK Biobank cardiovascular magnetic resonance imaging substudy

Background Reduced lung function is common and associated with increased cardiovascular morbidity and mortality, even in asymptomatic individuals without diagnosed respiratory disease. Previous studies have identified relationships between lung function and cardiovascular structure in individuals with pulmonary disease, but the relationships in those free from diagnosed cardiorespiratory disease have not been fully explored. Methods UK Biobank is a prospective cohort study of community participants in the United Kingdom. Individuals self-reported demographics and co-morbidities, and a subset underwent cardiovascular magnetic resonance (CMR) imaging and spirometry. CMR images were analysed to derive ventricular volumes and mass. The relationships between CMR-derived measures and spirometry and age were modelled with multivariable linear regression, taking account of the effects of possible confounders. Results Data were available for 4,975 individuals, and after exclusion of those with pre-existing cardiorespiratory disease and unacceptable spirometry, 1,406 were included in the analyses. In fully-adjusted multivariable linear models lower FEV1 and FVC were associated with smaller left ventricular end-diastolic (−5.21ml per standard deviation (SD) change in FEV1, −5.69ml per SD change in FVC), end-systolic (−2.34ml, −2.56ml) and stroke volumes (−2.85ml, −3.11ml); right ventricular end-diastolic (−5.62ml, −5.84ml), end-systolic (−2.47ml, −2.46ml) and stroke volumes (−3.13ml, −3.36ml); and with lower left ventricular mass (−2.29g, −2.46g). Changes of comparable magnitude and direction were observed per decade increase in age. Conclusions This study shows that reduced FEV1 and FVC are associated with smaller ventricular volumes and reduced ventricular mass. The changes seen per standard deviation change in FEV1 and FVC are comparable to one decade of ageing.

Prospective association between handgrip strength and cardiac structure and function in UK adults.

Background Handgrip strength, a measure of muscular fitness, is associated with cardiovascular (CV) events and CV mortality but its association with cardiac structure and function is unknown. The goal of this study was to determine if handgrip strength is associated with changes in cardiac structure and function in UK adults. Methods and results Left ventricular (LV) ejection fraction (EF), end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), mass (M), and mass-to-volume ratio (MVR) were measured in a sample of 4,654 participants of the UK Biobank Study 6.3 ± 1 years after baseline using cardiovascular magnetic resonance (CMR). Handgrip strength was measured at baseline and at the imaging follow-up examination. We determined the association between handgrip strength at baseline as well as its change over time and each of the cardiac outcome parameters. After adjustment, higher level of handgrip strength at baseline was associated with higher LVEDV (difference per SD increase in handgrip strength: 1.3ml, 95% CI 0.1–2.4; p = 0.034), higher LVSV (1.0ml, 0.3–1.8; p = 0.006), lower LVM (-1.0g, -1.8 –-0.3; p = 0.007), and lower LVMVR (-0.013g/ml, -0.018 –-0.007; p<0.001). The association between handgrip strength and LVEDV and LVSV was strongest among younger individuals, while the association with LVM and LVMVR was strongest among older individuals. Conclusions Better handgrip strength was associated with cardiac structure and function in a pattern indicative of less cardiac hypertrophy and remodeling. These characteristics are known to be associated with a lower risk of cardiovascular events.

The impact of cardiovascular risk factors on cardiac structure and function: Insights from the UK Biobank imaging enhancement study

Aims The UK Biobank is a large-scale population-based study utilising cardiovascular magnetic resonance (CMR) to generate measurements of atrial and ventricular structure and function. This study aimed to quantify the association between modifiable cardiovascular risk factors and cardiac morphology and function in individuals without known cardiovascular disease. Methods Age, sex, ethnicity (non-modifiable) and systolic blood pressure, diastolic blood pressure, smoking status, exercise, body mass index (BMI), high cholesterol, diabetes, alcohol intake (modifiable) were considered important cardiovascular risk factors. Multivariable regression models were built to ascertain the association of risk factors on left ventricular (LV), right ventricular (RV), left atrial (LA) and right atrial (RA) CMR parameters. Results 4,651 participants were included in the analysis. All modifiable risk factors had significant effects on differing atrial and ventricular parameters. BMI was the modifiable risk factor most consistently associated with subclinical changes to CMR parameters, particularly in relation to higher LV mass (+8.3% per SD [4.3 kg/m2], 95% CI: 7.6 to 8.9%), LV (EDV: +4.8% per SD, 95% CI: 4.2 to 5.4%); ESV: +4.4% per SD, 95% CI: 3.5 to 5.3%), RV (EDV: +5.3% per SD, 95% CI: 4.7 to 5.9%; ESV: +5.4% per SD, 95% CI: 4.5 to 6.4%) and LA maximal (+8.6% per SD, 95% CI: 7.4 to 9.7%) volumes. Increases in SBP were associated with higher LV mass (+6.8% per SD, 95% CI: 5.9 to 7.7%), LV (EDV: +4.5% per SD, 95% CI: 3.6 to 5.4%; ESV: +2.0% per SD, 95% CI: 0.8 to 3.3%) volumes. The presence of diabetes or high cholesterol resulted in smaller volumes and lower ejection fractions. Conclusions Modifiable risk factors are associated with subclinical alterations in structure and function in all four cardiac chambers. BMI and systolic blood pressure are the most important modifiable risk factors affecting CMR parameters known to be linked to adverse outcomes.

Automatic left ventricular analysis with Inline VF performs well compared to manual analysis: results from Barts Cardiovascular Registry

Background Manual left ventricular (LV) volumes and function analysis is time consuming and operator dependent. Automated and semi-automated LV analysis tools could be helpful, especially in high volume clinical and research centres. Inline VF (Siemens) is a fully-automated assessment tool performing LV volume analysis during scan acquisition. The aims of this study were: 1) to assess performance of Inline VF against manual analysis of LV volumes and function, 2) to derive conversion formulas from linear regression models and 3) to validate adjusted Inline VF parameters to ascertain whether this improves accuracy of the automated method.

Allergy Teaching in UK Medical Schools The Unmet Need

Aims. The inadequate provision of allergy services in the United Kingdom is attributed, in part, to poor undergraduate teaching of the subject. We ascertained the level of medical student knowledge about allergy, an extremely common disease, and congenital heart disease, an important yet rare condition, to see if the focus of medical student teaching was correctly matched to disease burden within the general population. Methods. An online quiz was designed consisting of 10 questions on common allergy problems and 10 questions on congenital heart disease to be taken by medical students. Results. Two hundred and fourteen respondents took the quiz, the majority (78.5%) of who were clinical medical students. There was a significant difference between the mean scores for the allergy (µ = 28.7%, standard deviation = 15.5) and congenital heart disease (µ = 40.1%, standard deviation = 21.3) sections; t(213) = −6.78, P < .0001. Discussion. Knowledge about basic allergic disease was very poor within a medical stude...