Interaction between clinicians and artificial intelligence to detect fetal atrioventricular septal defects on ultrasound: how can we optimize collaborative performance?

OBJECTIVES

Artificial intelligence (AI) has shown promise in improving the performance of fetal ultrasound screening in detecting congenital heart disease (CHD). The effect of giving AI advice to human operators has not been studied in this context. Giving additional information about AI model workings, such as confidence scores for AI predictions, may be a way of improving performance further. Our aims were to investigate whether AI advice improved overall diagnostic accuracy (using a single CHD lesion as an exemplar), and to see what, if any, additional information given to clinicians optimized the overall performance of the clinician-AI team.

METHODS

An AI model was trained to classify a single fetal CHD lesion (atrioventricular septal defect, AVSD), using a retrospective cohort of 121,130 cardiac four chamber images extracted from 173 ultrasound scan videos (98 with normal hearts, 75 with AVSD). A ResNet50 model architecture was used. Temperature scaling of model prediction probability was performed on a validation set, and gradient-weighted class activation maps (grad-CAMs) produced. Ten clinicians (two consultant fetal cardiologists, three trainees in pediatric cardiology, and five fetal cardiac sonographers) were recruited from a center of fetal cardiology to participate. Each participant was shown 2000 fetal four chamber images in a random order (1,000 normal and 1,000 AVSD). The dataset was comprised of 500 images, each shown in four conditions: 1) image alone without AI output; 2) image with binary AI classification; 3) image with AI model confidence; 4) image with gradient-weighted class activation map image overlays. The clinicians were asked to classify each image as normal or AVSD.

RESULTS

20,000 image classifications were recorded from 10 clinicians. The AI model alone achieved an accuracy of 0.798 (95% CI 0.760 - 0.832), sensitivity of 0.868 (95% CI 0.834 - 0.902) and specificity of 0.728 (95% CI 0.702 - 0.754, and the clinicians without AI achieved an accuracy of 0.844 (95% CI 0.834 - 0.854), sensitivity of 0.827 (95% CI 0.795 - 0.858) and specificity of 0.861 (95% CI 0.828 - 0.895). Showing a binary (normal or AVSD) AI model output resulted in significant improvement in accuracy to 0.865 (p <0.001). This effect was seen in both experienced and less experienced participants. Giving incorrect AI advice resulted in significant deterioration in overall accuracy from 0.761 to 0.693 (p <0.001), which was driven by an increase in both type I and type II error by the clinicians. This effect was worsened by showing model confidence (accuracy 0.649, p <0.001) or grad-CAM (accuracy 0.644, p <0.001).

CONCLUSIONS

AI has the potential to improve performance when used in collaboration with clinicians, even if the model performance does not reach expert level. Giving additional information about model workings such as model confidence and class activation map image overlays did not improve overall performance, and actually worsened performance for images where the AI model was incorrect. This article is protected by copyright. All rights reserved.

A universal deep-learning model for zinc finger design enables transcription factor reprogramming

Cys2His2 zinc finger (ZF) domains engineered to bind specific target sequences in the genome provide an effective strategy for programmable regulation of gene expression, with many potential therapeutic applications. However, the structurally intricate engagement of ZF domains with DNA has made their design challenging. Here we describe the screening of 49 billion protein-DNA interactions and the development of a deep-learning model, ZFDesign, that solves ZF design for any genomic target. ZFDesign is a modern machine learning method that models global and target-specific differences induced by a range of library environments and specifically takes into account compatibility of neighboring fingers using a novel hierarchical transformer architecture. We demonstrate the versatility of designed ZFs as nucleases as well as activators and repressors by seamless reprogramming of human transcription factors. These factors could be used to upregulate an allele of haploinsufficiency, downregulate a gain-of-function mutation or test the consequence of regulation of a single gene as opposed to the many genes that a transcription factor would normally influence.

Left ventricular myocardial tissue characteristics and function among healthy subjects with varying atherosclerotic cardiovascular disease risk

 

The Atherosclerotic Cardiovascular Disease Risk (ASCVD) score by pooled cohort equation is a reliable predictor for future ASCVD events and is used to guide primary prevention in asymptomatic aging subjects. ASCVD risk is associated with burden of coronary artery disease as measured from computed tomography angiography.

Purpose

We aim to investigate the association between ASCVD risk and cardiac magnetic resonance (CMR) derived left ventricular (LV) myocardial tissue characteristics (T1 values) and LV systolic and diastolic ventricular function in a large cohort of healthy subjects.

Methods

We selected all healthy subjects who underwent CMR from the UK-Biobank cohort study. We collected patient characteristics, cardiovascular risk factors, blood pressure at CMR, medication use and cholesterol levels. We used AI-CMRQC, our quality-controlled tool for analysis of cardiovascular function metrics from CMR using artificial intelligence [1,2], to automatically extract septal T1 values from native T1 maps and LV ejection fraction (EF), peak ejection and early filling rates (PER, PEFR), peak systolic longitudinal strain and diastolic strain-rate, mitral valve annular plane systolic excursion and diastolic peak velocity from cine long and short axis acquisitions. Subjects were stratified for low (&lt;7.5%), intermediate (7.5–21%) and high (&gt;21%) ASCVD risk. One-way repeated measures ANOVA was used to examine the association between cardiovascular metrics and ASCVD risk groups.

Results

12,493 healthy subjects were included (females n=6,000). Mean age was 62.7±7.5 years, 3.2% had diabetes, 12.5% received treatment for hypertension and 23% smoked. ASCVD risk score could be calculated in 9,487 subjects. Mean ASCVD risk was 12.7±9%. 38% of subjects had low, 43% intermediate and 19% high ASCVD risk.

T1 values fell across the incremental ASCVD groups (low: 943±51 ms, intermediate: 921±47 ms, high: 918±50 ms, P&lt;0.001). Indexed LV PEFR (low: 188±45 ml/ms·m2, intermediate: 170±48 ml/ms·m2, high: 147±45 ml/ms·m2, P&lt;0.001), diastolic longitudinal strain rate (low: 1.25±0.36, intermediate: 1.20±0.37, high: 1.17±0.36, P&lt;0.001) also fell consistently with incremental risk. A statistically significant, but clinically less relevant decrease was seen for LVEF (low: 59±6%, intermediate 58±6%, high: 57±6% p&lt;0.001) and longitudinal systolic strain (low: 21±3.5%, intermediate: 21±3.4%, high: 21±3.5%).

Conclusion

Increasing ASCVD risk was associated with lower native T1 values and decreasing metrics of diastolic and systolic LV function. The fall in T1 might suggest fibrofatty replacement in the LV myocardium in patients with incremental ASCVD risk that could contribute to the observed deterioration of LV function.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): NIHR MedTech cooperation awarded to Guy's and ST Thomas NHS Foundation Trust

Wideband RCS reduction due to plasma generated by radioactive nuclei for cylindrical object

Radar cross section reduction has been one of the most important research topics in recent years. Plasma-based stealth is a method of reducing the radar cross section, which dampens the electromagnetic waves and reduces the amount of return waves. In this paper, a coating of the radioactive nucleus \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{241}{\mathrm{Am}}$$\end{document}241Am on the surface of the cylinder with a radius of 10 cm is considered and the range of the emitted alpha particles and the electron density generated in the air are obtained using the Geant4 code under standard temperature and pressure conditions. By finite element method solution, the radar cross section of the conductive cylindrical object has been simulated and extracted in the presence and absence of plasma created by alpha-particles. The obtained results show a reduction of 5–8 dB \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{m}^2$$\end{document}m2 in the radar cross section in the frequency range of 2–12 GHz for specific activity source of 1 Ci/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathrm{cm}^2$$\end{document}cm2.

Subthreshold delayed afterdepolarizations mediated by reduced tissue conductivity form a substrate for unidirectional block and reentry within the infarcted heart

Funding Acknowledgements

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): British Heart Foundation, Wellcome Trust

Background

Delayed afterdepolarizations (DADs) due to spontaneous calcium release (SCR) events at the subcellular scale have been associated with arrhythmia formation in the border zone (BZ) of infarcted hearts. DADs may not only summate to form ectopic focal sources but may also inactivate sodium channels forming a substrate for unidirectional conduction block and reentry. The role played by infarct anatomy and altered intracellular coupling in facilitating this phenomenon is not fully understood.

Purpose

To use computational modelling to investigate the role of anatomical properties of the infarct BZ in creating a substrate for DAD-mediated conduction block and reentry.

Methods

MRI data from a porcine post-infarction heart was used to build the computational model. A phenomenological model was used to simulate SCRs in the BZ. Arrhythmia susceptibility was quantified by pacing the model followed by a pause, to see whether DADs would occur, and an extra S2 beat with different coupling intervals (CIs). Tissue conductivity in the BZ was decreased to investigate the effect of uncoupling on DAD-mediated conduction block.

Results

Subthreshold DADs occurring within the infarct BZ inactivated the fast sodium channels which resulted in block of S2 beats. This occurred most readily in narrow isthmuses where electrotonic load was attenuated by the non-conducting scar. DADs rendered the entire isthmus area refractory establishing a substrate for unidirectional block and reentry (see Fig. A). Reduced tissue conductivity in the BZ reduced electrotonic load on cells undergoing DADs. This led to more local tissue depolarization (Vm) as uncoupling prevented current from flowing to neighboring cells at rest (Fig. B-C). Reduced tissue conductivity also enhanced DAD-mediated block by increasing the vulnerable window for reentry initiation (700ms &lt; S2 CI &lt; 900ms as shown in Fig. D).

Conclusion

Subthreshold DADs provide a substrate for arrhythmogenesis in the infarct BZ. Tissue uncoupling enhanced the arrhythmogenic risk by increasing the time window of unidirectional block.

Automated non-invasive pressure-volume loop analysis of cardiac aging in a large cohort of healthy community dwellers

Background/Introduction

Pressure-volume loops (PVloops) provide a wealth of information on cardiac function that is not readily available from cardiac imaging alone.

Methods

To estimate left ventricular (LV) PVloops non-invasively have been available, but have so far not been used to interrogate ventricular function in large patient cohorts, due to the complexity of estimating PVloops. A new method was recently validated that construct PVloops non-invasively from cine cardiac magnetic resonance (CMR), based on the time-varying elastance model [1]. At the same time, we have validated a framework for automated, quality controlled analysis of cine CMR in large cohorts of patients/subjects [2]. Combining these two methods could automated PVloop estimation, enabling analysis of ventricular pressure-volume relationships in large study populations.

Purpose

Evaluate if CMR-based non-invasive PVloops can be used to interrogate the impact of cardiac ageing on LV function occurring in a large population of healthy community dwellers.

Methods

Non-invasive PVloops were calculated from a full cardiac cycle LV volume curve and brachial blood pressure data using a recently validated method based on the time-varying elastance model [1], in 7,650 healthy community dwellers from the UKBiobank population study. The LV volume curve was automatically obtained using our state-of-the-art, quality controlled deep learning (DL) based cine CMR analysis framework [2]. External Work, pressure-volume-area (PVA), end-systolic pressure (Pes), ventricular elastance (Ees, an estimate of contractility) and arterial elastance (Ea) and energy per ejected volume (EEV: PVA/ stroke volume) were calculated from the PVloops. We performed univariate regression between PVloop parameters and age. We also calculated the additional impact of cardiovascular risk-factors in a multivariate analysis.

Results

See results in table 1. With age, LV volumes fall (p&lt;0.001) in healthy subjects, while systolic blood pressure and Pes increases (both p&lt;0.001). As a result of the higher afterload, PVA (p=0.894) and EW (p=0.499) do not significantly change with age despite a lower SV. Arterial elastance (Ea) increased, and so did contractility, as measured by Ees (p&lt;0.001). Due to all these changes, EEV increased with age (p&lt;0.001). In multivariate analysis, cardiovascular risk factors hypercholesterolemia and hypertension negatively impacted Pes, PVA, Ees and EEV. Diabetes and smoking habits did not.

Conclusion

Non-invasive CMR-based PVloop analyses capture the impact of known changes occurring during cardiac ageing on cardiac work, contractility and energetic expenditure. Obtaining PVloops automatically using our AI analysis system in this large cohort of healthy subjects allows to formulate reference for assessment of cardiac disease.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): The authors acknowledge financial support (support) the National Institute for Health Research (NIHR) Cardiovascular MedTech Co-operative (previously existing as the Cardiovascular Healthcare Technology Co-operative 2012 - 2017) award to the Guy's and St Thomas' NHS Foundation Trust, in partnership with King's College London and the NIHR comprehensive Biomedical Research Centre of the Guy's & St Thomas' NHS Foundation Trust. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health Univariate regression analysisExample of estimated PV loop

Fairness in AI: are deep learning-based CMR segmentation algorithms biased?

Background/Introduction

Artificial intelligence (AI) is providing opportunities to transform cardiovascular medicine. A particular challenge in the application of AI technology is their potential for intrinsic and extrinsic biases, such as those based on gender and/or ethnicity. Unless satisfactorily addressed, these biases could lead to inequalities in early diagnosis, treatments and outcomes. Fairness in AI is a relatively new but fast-growing research field which deals with assessing and addressing potential bias in AI models.

Purpose

To perform the first analysis that assesses bias in AI-based cardiac MR segmentation models in a large-scale database.

Methods

A state-of-the-art deep learning (DL) based segmentation network, the “nnU-Net” framework [1], was used for automatic segmentation of both ventricles and the myocardium from cine short-axis cardiac MR over the full cardiac cycle. The dataset used consisted of end-diastole and end-systole short-axis cine cardiac MR images of 5,903 subjects (61.5±7.1 years). The nnU-Net network was trained and evaluated using a 5-fold cross validation (splits: train 60% / validation 20% / test 20%). Data on race and gender were obtained from the UK Biobank database and their distribution is summarized in Figure 1. To assess gender and racial bias in the segmentation network, we compared the Dice scores - which measure the overlap between manual and automatic segmentations – and the absolute error in measurements of biventricular volumes and function between patients grouped by ethnicity and gender.

Results

Figure 2 shows the Dice scores and the volumetric and functional measures for the full database, stratified by gender and by ethnicity. Results on the overall population showed an excellent agreement between the manual and automatic segmentations which is consistent with previous reported results [2–3]. However, we find statistically significant differences in Dice scores as well as volumetric measures between different ethnicities, showing that the segmentation network is biased against minority racial groups. No significant differences were found in Dice scores between genders. Similarly, for the end diastolic, end systolic volumes and ejection fraction, there were statistically significant differences in absolute error between the overall population and all racial groups except white.

Conclusion(s)

We have shown, for the first time, that racial bias exists in DL-based cardiac MR segmentation models. Our hypothesis is that this bias is a result of the unbalanced nature of the training data, and this is supported by the results which show that there is racial bias but not gender bias when trained using the UK Biobank database, which is gender-balanced but not race-balanced. In this work we want to highlight the potential issue of bias in DL-based image segmentation models when translating into a clinical environment.

Funding Acknowledgement

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): - EPSRC- Wellcome EPSRC Centre for Medical Engineering at the School of Biomedical Engineering and Imaging Sciences, King's College London Figure 1Figure 2

Large-scale, multi-vendor, multi-protocol, quality-controlled analysis of clinical cine CMR using artificial intelligence

Funding Acknowledgements

Type of funding sources: Public grant(s) – National budget only. Main funding source(s): Advancing Impact Award scheme of the EPSRC Impact Acceleration Account at King’s College London

Background

Artificial intelligence (AI) has the potential to facilitate the automation of CMR analysis for biomarker extraction. However, most AI algorithms are trained on a specific input domain (e.g., scanner vendor or hospital-tailored imaging protocol) and lack the robustness to perform optimally when applied to CMR data from other input domains.

Purpose

To develop and validate a robust CMR analysis tool for automatic segmentation and cardiac function analysis which achieves state-of-the-art performance for multi-vendor short-axis cine CMR images. 

Methods

The current work is an extension of our previously published quality-controlled AI-based tool for cine CMR analysis [1]. We deployed an AI algorithm that is equipped to handle different image sizes and domains automatically - the ‘nnU-Net’ framework [2] - and retrained our tool using the UK Biobank (UKBB) cohort population (n = 4,872) and a large database of clinical CMR studies obtained from two NHS hospitals (n = 3,406). The NHS hospital data came from three different scanner types: Siemens Aera 1.5T (n = 1,419), Philips Achieva 1.5T and 3T (n = 1,160), and Philips Ingenia 1.5T (n = 827). The ‘nnU-net’ was used to segment both ventricles and the myocardium. The proposed method was evaluated on randomly selected test sets from UKBB (n = 488) and NHS (n = 331) and on two external publicly available databases of clinical CMRs acquired on Philips, Siemens, General Electric (GE), and Canon CMR scanners – ACDC (n = 100) [3] and M&Ms (n = 321) [4]. We calculated the Dice scores - which measure the overlap between manual and automatic segmentations - and compared manual vs AI-based measures of biventricular volumes and function. 

Results

 Table 1 shows that the Dice scores for the NHS, ACDC, and M&Ms scans are similar to those obtained in the highly controlled, single vendor and single field strength UKBB scans. Although our AI-based tool was only trained on CMR scans from two vendors (Philips and Siemens), it performs similarly in unseen vendors (GE and Canon). Furthermore, it achieves state-of-the-art performance in online segmentation challenges, without being specifically trained on these databases. Table 1 also shows good agreement between manual and automated clinical measures of ejection fraction and ventricular volume and mass.

Conclusions

We show that our proposed AI-based tool, which combines training on a large-scale multi-domain CMR database with a state-of-the-art AI algorithm, allows us to robustly deal with routine clinical data from multiple centres, vendors, and field strengths. This is a fundamental step for the clinical translation of AI algorithms. Moreover, our method yields a range of additional metrics of cardiac function (filling and ejection rates, regional wall motion, and strain) at no extra computational cost.

Assessing survival and re-hospitalisation following transvenous lead extraction in cardiac resynchronisation therapy devices depending on reimplantation timing: a propensity score matched analysis

Funding Acknowledgements

Type of funding sources: None.

Background

 Among patients undergoing transvenous lead extraction (TLE), differences in complication rate and 1-year mortality has been explored in patients with cardiac resynchronisation therapy (CRT) devices.  Longer term outcomes and the influence of timing of reimplantation of device, with respect to rehospitalisation and longer-term mortality is poorly understood. 

Purpose

 The purpose of this study was to evaluate whether early reimplantation following TLE in patients with CRT devices influenced survival and rehospitalisation.

Methods

 Clinical data from consecutive patients undergoing TLE in the reference centre between the years 2000 to 2019 were prospectively collected.  Patients surviving to discharge who were re-implanted with the same device were included. The cohort was split depending on whether or not they had a CRT device at time of explant.  The association between TLE in CRT patients and all-cause mortality and re-hospitalisation was assessed by Kaplan Meier estimates in a 1:1 propensity-score matched cohort, with a calliper of 0.10.  Early reimplantation was defined as reimplantation within 7 days of TLE, and late reimplantation as reimplantation after greater than 7 days of TLE.

Results

 Of 1005 patients included in the analysis, 285 (25%) had a CRT device.  After matching, 192 CRT patients were compared with 192 non-CRT patients.  Propensity scores were calculated using 39 baseline characteristics, including age, gender, co-morbidities, TLE indication, left ventricular ejection fraction, baseline creatinine and technical extraction data.  Mean follow up was 53.5 ± 38.3 months, mean age at explant was 67.7 ± 12.1 years, 83.3% were male and 54.4% had an infective indication for TLE.  In the matched cohort, there was no significant difference between the CRT and non-CRT group with respect to long-term mortality (hazard ratio [HR] = 1.01, 95% confidence interval [CI] [0.74-1.39], p = 0.093) or rehospitalisation (HR = 1.2 [0.87-1.66], p = 0.265).  A similar proportion of patients were reimplanted within 7 days in the CRT and non-CRT groups (59.4% vs 61.5%, p = 0.754).  In the matched non-CRT group, late reimplantation was associated with similar mortality to early reimplantation (HR = 1.33 [0.86-2.05], p = 0.208) and rehospitalisation (HR = 0.88 [0.53-1.45], p = 0.603).  In the matched CRT group, late reimplantation was associated with higher mortality (HR = 1.64 [1.04-2.57], p = 0.032) and rehospitalisation (HR = 1.57 [1.00-2.46], p = 0.049] (see figure).

Conclusion

  In this closely matched population, TLE in CRT patients resulted in similar long-term outcomes compared with non-CRT patients.  Early reimplantation post CRT explant was associated with reduced long-term mortality and rehospitalisation.  This suggests a longer duration without biventricular pacing post TLE may induce negative reverse-remodelling and should be avoided in a CRT population. Abstract Figure. Kaplan-Meier survival curves

Long-term survival following transvenous lead extraction: importance of indication and comorbidities

Funding Acknowledgements

Type of funding sources: None.

Background/Introduction: 

The significant rise in cardiac implantable electronic devices (CIED) has been paralleled by an increase in the number of procedures required for the removal of such devices and their associated leads.  High procedural success rates with low rates of major in hospital complications is well recognised.  Longer term mortality following transvenous lead extraction (TLE) is less well characterised. Long term outcomes are important as they should inform the decision making and consent process, especially in non-infected cases where there may not be a class I indication for lead removal. 

Purpose

 The purpose of this study was to evaluate the factors influencing survival in patients undergoing TLE depending on extraction indication.

Methods

 Clinical data from consecutive patients undergoing TLE in the reference centre between the years 2000 to 2019 were prospectively collected.  Only patients surviving to discharge were included. The total cohort was divided into groups depending on whether there was an infective or non-infective indication for TLE.  We evaluated the association of demographic, clinical, device related and procedure-related factors on mortality.

Results

 A total of 1151 patients were included in the analysis.  632 (54.9%) and 519 patients (45.1%) were for infective and non-infective indications respectively. Analysis of long-term outcomes on the total cohort (mean 66-month follow-up) revealed a mortality of 34.1% (392 deaths).  A higher proportion of patients died in the infection vs the non-infection group (38.6% vs 28.5%, p &lt; 0.001).  Local infection (hazard ratio [HR] = 1.4, 95% confidence interval [CI] [1.12-1.75]) was associated with similar long-term mortality risk as systemic infection (HR = 1.3, CI[0.99-1.72]). Multivariate analysis demonstrated increased risk of mortality with higher age (HR = 1.05, CI[1.04-1.07]), eGFR &lt; 60ml/min/1.73m2 (HR = 1.55, CI[11.22-1.97]), higher cumulative co-morbidity burden (HR = 1.15, CI [1.06-1.23], and reduced risk per percentage increase in LVEF (HR = 0.98, CI[0.96-1.00]). Kaplan-Meier survival analysis demonstrated statistically worse prognosis in patients with a higher number of leads extracted and increasing co-morbidities.

Conclusion

  Long-term mortality for patients undergoing TLE remains high. Consensus guidelines recommend evaluating risk for major complications when determining whether to proceed with TLE. This study suggests assessing longer-term outcomes when considering TLE, particularly for non-infective indications. Abstract Figure.

First-phase ejection fraction predicts response to cardiac resynchronization therapy and adverse outcomes

Background

Cardiac Resynchronization Therapy (CRT) is an important therapeutic treatment for chronic heart failure. However, even in carefully selected cases up to 40% of patient fail to respond. First-phase ejection fraction (EF1), the ejection fraction up to the time of maximal ventricular contraction, is a novel and more sensitive echocardiographic measure of early systolic function.

Purpose

We examined the value of EF1, to predict response to CRT and clinical outcomes after CRT.

Methods

We analysed echocardiograms from 197 patients who underwent CRT between 2009 and 2018 and were followed to determine clinical outcomes.

Results

Volumetric response rate (reduction in end-systolic volume ≥15%) was 92.3% vs. 12.1%, for those with EF1 in the highest vs. lowest tertiles (p&lt;0.001). A cut-off value of 11.9% for EF1 had &gt;85% sensitivity and specificity for prediction of response to CRT; on multivariate logistic regression analysis incorporating previously defined predictors, EF1 was the strongest predictor of response (OR: 1.563, 95% CI: 1.371–1.782, p&lt;0.001) (table 1). EF1 was also the strongest predictor of a clinical composite score (OR: 1.115, 95% CI: 1.044–1.191, p=0.001). Improvement in EF1 at 6 months after CRT implantation (6.5±5.8% vs 1.8±4.3% in responders vs. non-responders, p&lt;0.001) was the best predictor of heart failure re-hospitalization and death after median follow-up of 29.2 months (HR: 0.819, 95% CI:0.765–0.876, p&lt;0.001).

Conclusion

EF1 is a promising marker to identify patients likely to respond to CRT and most likely provides a measure of myocardial viability that determines response.

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): British Heart Foundation, Wellcome/EPSRC Centre for Medical Engineering

Synergy in the heart: RV systolic function plays a key role in optimizing LV performance during exercise

The right ventricle (RV) is often overlooked in the evaluation of cardiac performance and treatment of left ventricular (LV) heart diseases. However, recent evidence suggests the RV may play an important role in maintaining systemic cardiac function and delivering stroke volume (SV). We used exercise cardiac magnetic resonance and biomechanical modeling to investigate the role of the RV in LV stroke volume regulation. We studied SV augmentation during exercise by pharmacologically inducing negative chronotropy (sHRi) in healthy volunteers and investigating training-induced SV augmentation in endurance athletes. SV augmentation during exercise after sHRi is achieved differently in the two ventricles. In the RV, the larger SV is driven by increasing contraction down to lower end-systolic volume (ESV; P < 0.001). In the LV, SV augmentation is achieved through an increase in end-diastolic volume (EDV; P < 0.001), avoiding contraction to a lower ESV. The same mechanism underlies the enhanced SV response observed in athletes. Changes in atrial area during SV augmentation suggest that the improved LV EDV response is sustained by the larger RV contractions. Using our biomechanical model, we explain this behavior by showing that the RV systolic function-driven regulation of LV SV optimizes the energetic cost of LV contraction and leads to minimization of the total costs of biventricular contraction. In conclusion, this work provides mechanistic understanding of the pivotal role of the RV in optimizing LV SV during exercise. It demonstrates why optimizing RV function needs to become a key part of therapeutic strategies in patients and training for athletes.NEW & NOTEWORTHY The right ventricle appears to have an important impact on maintaining systemic cardiac function and delivering stroke volume. However, its exact role in supporting left ventricular function has so far been unclear. This study demonstrates a new mechanism of ventricular interaction that provides mechanistic understanding of the key importance of the right ventricle in driving cardiac performance.

An observational study to compare the utilisation of computed tomography colonography with optical colonoscopy as the first diagnostic imaging tool in patients with suspected colorectal cancer

AIM

To evaluate the clinical and cost implications of using computed tomography colonography (CTC) compared to optical colonoscopy (OC) as the initial colonic investigation in patients with low-to-intermediate risk of colorectal cancer (CRC).

MATERIALS AND METHODS

A non-randomised, prospective single-centre study recruited 180 participants to compare the cost implications of two clinical pathways used in the diagnosis of low-to-intermediate risk of CRC that differ in the initial diagnostic test, either CTC or OC. Costs were compared using generalised linear models (GLM) and combined with quality-adjusted life years (QALYs, based on the EQ-5D-5L) to estimate cost-effectiveness at 6 months post-recruitment. Secondary outcomes assessed access to care and patient satisfaction.

RESULTS

Mean (SD, n) cost at 6 months post-recruitment per participant was £991 (£316, n=105) for the OC group and £645 (£607, n=68) for the CTC group, leading to an estimated cost difference of -£370 (95% CI: -£554, -£185, p<0.001). Assuming a £20,000 willingness-to-pay per QALY threshold, there was a 91.4% probability of CTC being cost-effective at month 6. The utilisation of CTC led to improved access to care, with a shorter mean time from referral from primary care to results (6.3 days difference, p=0.005). No differences in patient satisfaction were detected between both groups.

CONCLUSION

The utilisation of CTC as the first-line investigation for patients with low-to-intermediate risk of CRC has the potential to release OC capacity, of pivotal importance for patients more likely to benefit from an invasive diagnostic approach.

221Evaluating the ability of different substrate mapping techniques to identify scar-related ventricular tachycardia circuits using computational modelling

Funding Acknowledgements

National Institute for Health Research; British Heart Foundation; and The Wellcome Trust and Engineering and Physical Sciences Research Council.

Background

Accurate identification of targets for catheter ablation therapy of ventricular tachycardias (VTs) in the postinfarction heart remains a significant challenge. Identification of such targets often requires VT-induction to delineate the entry/exit points of the reentrant circuit sustaining the VT. However, inducibility may not be possible due to hemodynamic instability. In this scenario, substrate ablation strategies can still be performed to uncover the arrhythmogenic substrate during sinus or paced rhythm. However, substrate mapping may fail to accurately delineate the reentrant circuit resulting in VT recurrence after the procedure.

Purpose

To use computer simulations to compare the ability of different electroanatomical maps constructed following typical substrate ablation strategies to identify the VT exit site.

Methods

An image-based computational model of the porcine post-infarction left ventricle was constructed to simulate VT and paced rhythm. Electroanatomical maps were constructed based on the following features extracted from electrograms computed on the endocardial surface: activation time (AT), bipolar electrogram amplitude, signal fractionation and the reentry vulnerability index (RVI - a metric combining activation and repolarization timings to identify tissue susceptibility to reentry). Potential ablation targets during substrate mapping were compared for: highest 5% AT gradient; lowest 5% bipolar signal amplitudes; areas with fragmented signals (more than one peak); and lowest 5% RVI. The minimum distance, d, between the manually identified VT exit site and the targets was measured.

Results

The RVI performed better than the other metrics at detecting the VT exit site (see Figure). The minimum distance between sites of lowest RVI and the exit site was 3.2mm compared to 13.1mm and 15.9mm in traditional AT and voltage maps, respectively. As the scar was not transmural, parameters derived from all electrograms (including those located on dense scar regions) were used to construct the electroanatomical maps. This improved the performance of the RVI significantly, making it more specific than the other metrics as can be seen in the Figure.

Conclusions

Among all metrics investigated here, the RVI identified the vulnerable region closest to VT exit site. This finding suggests that activation-repolarization metrics may improve the detection of pro-arrhythmic regions without having to induce VT. Moreover, the RVI may be particularly well suited for detecting vulnerable regions within non-transmural scars.

Abstract Figure. VT and Substrate Mapping

563 Early systolic function is impaired in patients with heart failure and preserved ejection fraction and is a more important predictor of clinical outcomes than diastolic function

Funding Acknowledgements

British Heart Foundation

Aims

First-phase ejection fraction (EF1), the ejection fraction up to the time of maximal ventricular contraction may be impaired in heart failure with preserved ejection fraction (HFpEF), with homeostatic mechanisms preserving the overall ejection fraction. We examined whether EF1 is impaired and whether it predicts adverse outcomes in patients with HFpEF.

Methods and Results

We examined EF1 in patients with HFpEF diagnosed according to European Society of Cardiology guidelines, in subjects with pre-clinical diastolic dysfunction and in control subjects with no evidence of heart failure. The predictive value of EF1 for a combined end-point of re-hospitalisation for heart failure and death from any cause was compared to that of conventional echocardiographic and other indices. There was a progressive impairment of EF1 in patients with diastolic dysfunction and HFpEF compared to those without evidence of heart failure (P &lt; 0.001). In 180 HFpEF patients followed for a median of 14.7 months, 101 patients reached the primary endpoint (61 deaths and 40 hospitalisations). EF1 was the most powerful predictor of events (table 1). A cut-off value of 19.4% gave hazard ratios (for EF1 &lt; 19.4% compared to ≥19.4%) of 3.04 (figure 1), (95% confidence interval 2.014 – 4.604, P &lt; 0.001) unadjusted, and 2.976 (1.887 – 4.695, P &lt; 0.001) adjusted for age, gender, other echocardiographic indices (including EF, E/e’ ratio, stroke volume and left atrial volume index) and N-terminal pro-brain natriuretic peptide.

Conclusion

Early systolic function is impaired in HFpEF and is a powerful predictor of clinical outcomes. Therapies targeted at improving early systolic function may improve outcomes in HFpEF.

Table 1 HR CI (95%) P value Age 1.020 1.001 - 1.039 0.042 Female Gender 1.989 1.254 - 3.156 0.003 ln NT-proBNP 1.354 1.096 - 1.672 0.005 Haemoglobin 0.992 0.980 - 1.005 0.225 eGFR 0.997 0.988 - 1.006 0.568 Structural Heart Disease 0.748 0.439 - 1.274 0.285 Diastolic Dysfunction 0.694 0.445 - 1.082 0.107 Ejection Fraction 1.031 0.996 - 1.068 0.082 Stroke Volume index 0.963 0.940 - 0.986 0.002 TPAVF 0.995 0.980 - 1.005 0.230 EF1 0.900 0.869 - 0.932 &lt;0.001 Multivariate analysis of predictors of events

Abstract 563 Figure 1 Kaplan-Meier Curve

SiRNA/DOX lodeded chitosan based nanoparticles: Development, Characterization and in vitro evaluation on A549 lung cancer cell line

High-mobility group AT-hook2 (HMGA2), involved in epithelial mesenchymal transition (EMT) process, has a pivotal role in lung cancer metastasis. Lung cancer therapy with HMGA2 suppressing small interfering RNA (siRNA) has been introduced recently while doxorubicin (DOX) has been used as a frequent cancer chemotherapy agent. Both reagents have been faced with obstacles in clinic which make them ineffective. NanoParticles (NPs) provided a platform for efficient co delivery of the anticancer drugs. The aim of this study was production and in vitro characterization of different pharmacological groups (siRNA, DOX or siRNA-DOX) of carboxymethyl dextran thrimethyl chitosan nanoparticles (CMDTMChiNPs) on cytotoxicity, gene expression, apoptosis and migration of metastatic lung cancer cell line (A-549). CMDTMChiNPs were synthesized and encapsulated with siRNA, DOX or siRNA-DOX. Then the effects of HMGA2 siRNA and DOX co delivery was assessed in A549 viability and target genes (HMGA2, Ecadherin, vimentin and MMP9) by MTT and real time PCR, respectively. In addition capability of apoptosis induction and anti-migratory features of formulated NPs were analyzed by flowcytometry and wound healing assays. SiRNA-DOX-CMDTM ChiNPs approximate size were 207±5 with poly dispersity index (PDI) and zeta potential of 0.4 and 16.3±0.3, respectively. NPs loaded with DOX and siRNA were the most efficient drug formulations in A549 cell cytotoxicity, altering of EMT markers, apoptosis induction and migration inhibition. Generally our results showed that co delivery of HMGA2 siRNA and DOX by novel designed CMDTMChiNPs is a new therapeutic approach with great potential efficiency for lung cancer treatment.

A novel methodology for personalized simulations of ventricular hemodynamics from noninvasive imaging data

Current state-of-the-art imaging techniques can provide quantitative information to characterize ventricular function within the limits of the spatiotemporal resolution achievable in a realistic acquisition time. These imaging data can be used to personalize computer models, which in turn can help treatment planning by quantifying biomarkers that cannot be directly imaged, such as flow energy, shear stress and pressure gradients. To date, computer models have typically relied on invasive pressure measurements to be made patient-specific. When these data are not available, the scope and validity of the models are limited. To address this problem, we propose a new methodology for modeling patient-specific hemodynamics based exclusively on noninvasive velocity and anatomical data from 3D+t echocardiography or Magnetic Resonance Imaging (MRI). Numerical simulations of the cardiac cycle are driven by the image-derived velocities prescribed at the model boundaries using a penalty method that recovers a physical solution by minimizing the energy imparted to the system. This numerical approach circumvents the mathematical challenges due to the poor conditioning that arises from the imposition of boundary conditions on velocity only. We demonstrate that through this technique we are able to reconstruct given flow fields using Dirichlet only conditions. We also perform a sensitivity analysis to investigate the accuracy of this approach for different images with varying spatiotemporal resolution. Finally, we examine the influence of noise on the computed result, showing robustness to unbiased noise with an average error in the simulated velocity approximately 7% for a typical voxel size of 2mm(3) and temporal resolution of 30ms. The methodology is eventually applied to a patient case to highlight the potential for a direct clinical translation.

Diabetes enhances the proliferation of adult pancreatic multipotent progenitor cells and biases their differentiation to more β-cell production

Endogenous pancreatic multipotent progenitors (PMPs) are ideal candidates for regenerative approaches to compensate for β-cell loss since their β-cell-producing capacities as well as strategic location would eliminate unnecessary invasive manipulations. However, little is known about the status and potentials of PMPs under diabetic conditions. Here we show that β-cell metabolic stress and hyperglycemia enhance the proliferation capacities of adult PMP cells and bias their production of progeny toward β-cells in mouse and human. These effects are dynamic and correlate with functional β-cell regeneration when conditions allow.

Benchmarking framework for myocardial tracking and deformation algorithms: an open access database

In this paper we present a benchmarking framework for the validation of cardiac motion analysis algorithms. The reported methods are the response to an open challenge that was issued to the medical imaging community through a MICCAI workshop. The database included magnetic resonance (MR) and 3D ultrasound (3DUS) datasets from a dynamic phantom and 15 healthy volunteers. Participants processed 3D tagged MR datasets (3DTAG), cine steady state free precession MR datasets (SSFP) and 3DUS datasets, amounting to 1158 image volumes. Ground-truth for motion tracking was based on 12 landmarks (4 walls at 3 ventricular levels). They were manually tracked by two observers in the 3DTAG data over the whole cardiac cycle, using an in-house application with 4D visualization capabilities. The median of the inter-observer variability was computed for the phantom dataset (0.77 mm) and for the volunteer datasets (0.84 mm). The ground-truth was registered to 3DUS coordinates using a point based similarity transform. Four institutions responded to the challenge by providing motion estimates for the data: Fraunhofer MEVIS (MEVIS), Bremen, Germany; Imperial College London - University College London (IUCL), UK; Universitat Pompeu Fabra (UPF), Barcelona, Spain; Inria-Asclepios project (INRIA), France. Details on the implementation and evaluation of the four methodologies are presented in this manuscript. The manually tracked landmarks were used to evaluate tracking accuracy of all methodologies. For 3DTAG, median values were computed over all time frames for the phantom dataset (MEVIS=1.20mm, IUCL=0.73 mm, UPF=1.10mm, INRIA=1.09 mm) and for the volunteer datasets (MEVIS=1.33 mm, IUCL=1.52 mm, UPF=1.09 mm, INRIA=1.32 mm). For 3DUS, median values were computed at end diastole and end systole for the phantom dataset (MEVIS=4.40 mm, UPF=3.48 mm, INRIA=4.78 mm) and for the volunteer datasets (MEVIS=3.51 mm, UPF=3.71 mm, INRIA=4.07 mm). For SSFP, median values were computed at end diastole and end systole for the phantom dataset(UPF=6.18 mm, INRIA=3.93 mm) and for the volunteer datasets (UPF=3.09 mm, INRIA=4.78 mm). Finally, strain curves were generated and qualitatively compared. Good agreement was found between the different modalities and methodologies, except for radial strain that showed a high variability in cases of lower image quality.

Biophysical modeling to simulate the response to multisite left ventricular stimulation using a quadripolar pacing lead

BACKGROUND

Response to cardiac resynchronization therapy (CRT) is reduced in patients with posterolateral scar. Multipolar pacing leads offer the ability to select desirable pacing sites and/or stimulate from multiple pacing sites concurrently using a single lead position. Despite this potential, the clinical evaluation and identification of metrics for optimization of multisite CRT (MCRT) has not been performed.

METHODS

The efficacy of MCRT via a quadripolar lead with two left ventricular (LV) pacing sites in conjunction with right ventricular pacing was compared with single-site LV pacing using a coupled electromechanical biophysical model of the human heart with no, mild, or severe scar in the LV posterolateral wall.

RESULT

The maximum dP/dt(max) improvement from baseline was 21%, 23%, and 21% for standard CRT versus 22%, 24%, and 25% for MCRT for no, mild, and severe scar, respectively. In the presence of severe scar, there was an incremental benefit of multisite versus standard CRT (25% vs 21%, 19% relative improvement in response). Minimizing total activation time (analogous to QRS duration) or minimizing the activation time of short-axis slices of the heart did not correlate with CRT response. The peak electrical activation wave area in the LV corresponded with CRT response with an R(2) value between 0.42 and 0.75.

CONCLUSION

Biophysical modeling predicts that in the presence of posterolateral scar MCRT offers an improved response over conventional CRT. Maximizing the activation wave area in the LV had the most consistent correlation with CRT response, independent of pacing protocol, scar size, or lead location.