Non-invasive detection of severe PH in lung disease using magnetic resonance imaging

Introduction

Severe pulmonary hypertension (mean pulmonary artery pressure ≥35 mmHg) in chronic lung disease (PH-CLD) is associated with high mortality and morbidity. Data suggesting potential response to vasodilator therapy in patients with PH-CLD is emerging. The current diagnostic strategy utilises transthoracic Echocardiography (TTE), which can be technically challenging in some patients with advanced CLD. The aim of this study was to evaluate the diagnostic role of MRI models to diagnose severe PH in CLD.

Methods

167 patients with CLD referred for suspected PH who underwent baseline cardiac MRI, pulmonary function tests and right heart catheterisation were identified. In a derivation cohort (n = 67) a bi-logistic regression model was developed to identify severe PH and compared to a previously published multiparameter model (Whitfield model), which is based on interventricular septal angle, ventricular mass index and diastolic pulmonary artery area. The model was evaluated in a test cohort.

Results

The CLD-PH MRI model [= (-13.104) + (13.059 * VMI)-(0.237 * PA RAC) + (0.083 * Systolic Septal Angle)], had high accuracy in the test cohort (area under the ROC curve (0.91) (p < 0.0001), sensitivity 92.3%, specificity 70.2%, PPV 77.4%, and NPV 89.2%. The Whitfield model also had high accuracy in the test cohort (area under the ROC curve (0.92) (p < 0.0001), sensitivity 80.8%, specificity 87.2%, PPV 87.5%, and NPV 80.4%.

Conclusion

The CLD-PH MRI model and Whitfield model have high accuracy to detect severe PH in CLD, and have strong prognostic value.

Uncertainty Estimation for Heatmap-Based Landmark Localization

Automatic anatomical landmark localization has made great strides by leveraging deep learning methods in recent years. The ability to quantify the uncertainty of these predictions is a vital component needed for these methods to be adopted in clinical settings, where it is imperative that erroneous predictions are caught and corrected. We propose Quantile Binning, a data-driven method to categorize predictions by uncertainty with estimated error bounds. Our framework can be applied to any continuous uncertainty measure, allowing straightforward identification of the best subset of predictions with accompanying estimated error bounds. We facilitate easy comparison between uncertainty measures by constructing two evaluation metrics derived from Quantile Binning. We compare and contrast three epistemic uncertainty measures (two baselines, and a proposed method combining aspects of the two), derived from two heatmap-based landmark localization model paradigms (U-Net and patch-based). We show results across three datasets, including a publicly available Cephalometric dataset. We illustrate how filtering out gross mispredictions caught in our Quantile Bins significantly improves the proportion of predictions under an acceptable error threshold. Finally, we demonstrate that Quantile Binning remains effective on landmarks with high aleatoric uncertainty caused by inherent landmark ambiguity, and offer recommendations on which uncertainty measure to use and how to use it. The code and data are available at https://github.com/schobs/qbin.

A transparent artificial intelligence framework to assess lung disease in pulmonary hypertension

Recent studies have recognized the importance of characterizing the extent of lung disease in pulmonary hypertension patients by using Computed Tomography. The trustworthiness of an artificial intelligence system is linked with the depth of the evaluation in functional, operational, usability, safety and validation dimensions. The safety and validation of an artificial tool is linked to the uncertainty estimation of the model's prediction. On the other hand, the functionality, operation and usability can be achieved by explainable deep learning approaches which can verify the learning patterns and use of the network from a generalized point of view. We developed an artificial intelligence framework to map the 3D anatomical models of patients with lung disease in pulmonary hypertension. To verify the trustworthiness of the framework we studied the uncertainty estimation of the network's prediction, and we explained the learning patterns of the network. Therefore, a new generalized technique combining local explainable and interpretable dimensionality reduction approaches (PCA-GradCam, PCA-Shape) was developed. Our open-source software framework was evaluated in unbiased validation datasets achieving accurate, robust and generalized results.

How should studies using AI be reported? lessons from a systematic review in cardiac MRI

Recent years have seen a dramatic increase in studies presenting artificial intelligence (AI) tools for cardiac imaging. Amongst these are AI tools that undertake segmentation of structures on cardiac MRI (CMR), an essential step in obtaining clinically relevant functional information. The quality of reporting of these studies carries significant implications for advancement of the field and the translation of AI tools to clinical practice. We recently undertook a systematic review to evaluate the quality of reporting of studies presenting automated approaches to segmentation in cardiac MRI (Alabed et al. 2022 Quality of reporting in AI cardiac MRI segmentation studies-a systematic review and recommendations for future studies. Frontiers in Cardiovascular Medicine 9:956811). 209 studies were assessed for compliance with the Checklist for AI in Medical Imaging (CLAIM), a framework for reporting. We found variable-and sometimes poor-quality of reporting and identified significant and frequently missing information in publications. Compliance with CLAIM was high for descriptions of models (100%, IQR 80%-100%), but lower than expected for descriptions of study design (71%, IQR 63-86%), datasets used in training and testing (63%, IQR 50%-67%) and model performance (60%, IQR 50%-70%). Here, we present a summary of our key findings, aimed at general readers who may not be experts in AI, and use them as a framework to discuss the factors determining quality of reporting, making recommendations for improving the reporting of research in this field. We aim to assist researchers in presenting their work and readers in their appraisal of evidence. Finally, we emphasise the need for close scrutiny of studies presenting AI tools, even in the face of the excitement surrounding AI in cardiac imaging.

The Importance of Mitral Valve Prolapse Doming Volume in the Assessment of Left Ventricular Stroke Volume with Cardiac MRI

There remains a debate whether the ventricular volume within prolapsing mitral valve (MV) leaflets should be included in the left ventricular (LV) end-systolic volume, and therefore factored in LV stroke volume (SV), in cardiac magnetic resonance (CMR) assessments. This study aims to compare LV volumes during end-systolic phases, with and without the inclusion of the volume of blood on the left atrial aspect of the atrioventricular groove but still within the MV prolapsing leaflets, against the reference LV SV by four-dimensional flow (4DF). A total of 15 patients with MV prolapse (MVP) were retrospectively enrolled in this study. We compared LV SV with (LV SV_MVP) and without (LV SV_standard) MVP left ventricular doming volume, using 4D flow (LV SV_4DF) as the reference value. Significant differences were observed when comparing LV SV_standard and LV SV_MVP (p < 0.001), and between LV SV_standard and LV SV_4DF (p = 0.02). The Intraclass Correlation Coefficient (ICC) test demonstrated good repeatability between LV SV_MVP and LV SV_4DF (ICC = 0.86, p < 0.001) but only moderate repeatability between LV SV_standard and LV SV_4DF (ICC = 0.75, p < 0.01). Calculating LV SV by including the MVP left ventricular doming volume has a higher consistency with LV SV derived from the 4DF assessment. In conclusion, LV SV short-axis cine assessment incorporating MVP dooming volume can significantly improve the precision of LV SV assessment compared to the reference 4DF method. Hence, in cases with bi-leaflet MVP, we recommend factoring in MVP dooming into the left ventricular end-systolic volume to improve the accuracy and precision of quantifying mitral regurgitation.

Automated 4D flow cardiac MRI pipeline to derive peak mitral inflow diastolic velocities using short-axis cine stack: two centre validation study against echocardiographic pulse-wave doppler

BACKGROUND

Measurement of peak velocities is important in the evaluation of heart failure. This study compared the performance of automated 4D flow cardiac MRI (CMR) with traditional transthoracic Doppler echocardiography (TTE) for the measurement of mitral inflow peak diastolic velocities.

METHODS

Patients with Doppler echocardiography and 4D flow cardiac magnetic resonance data were included retrospectively. An established automated technique was used to segment the left ventricular transvalvular flow using short-axis cine stack of images. Peak mitral E-wave and peak mitral A-wave velocities were automatically derived using in-plane velocity maps of transvalvular flow. Additionally, we checked the agreement between peak mitral E-wave velocity derived by 4D flow CMR and Doppler echocardiography in patients with sinus rhythm and atrial fibrillation (AF) separately.

RESULTS

Forty-eight patients were included (median age 69 years, IQR 63 to 76; 46% female). Data were split into three groups according to heart rhythm. The median peak E-wave mitral inflow velocity by automated 4D flow CMR was comparable with Doppler echocardiography in all patients (0.90 ± 0.43 m/s vs 0.94 ± 0.48 m/s, P = 0.132), sinus rhythm-only group (0.88 ± 0.35 m/s vs 0.86 ± 0.38 m/s, P = 0.54) and in AF-only group (1.33 ± 0.56 m/s vs 1.18 ± 0.47 m/s, P = 0.06). Peak A-wave mitral inflow velocity results had no significant difference between Doppler TTE and automated 4D flow CMR (0.81 ± 0.44 m/s vs 0.81 ± 0.53 m/s, P = 0.09) in all patients and sinus rhythm-only groups. Automated 4D flow CMR showed a significant correlation with TTE for measurement of peak E-wave in all patients group (r = 0.73, P < 0.001) and peak A-wave velocities (r = 0.88, P < 0.001). Moreover, there was a significant correlation between automated 4D flow CMR and TTE for peak-E wave velocity in sinus rhythm-only patients (r = 0.68, P < 0.001) and AF-only patients (r = 0.81, P = 0.014). Excellent intra-and inter-observer variability was demonstrated for both parameters.

CONCLUSION

Automated dynamic peak mitral inflow diastolic velocity tracing using 4D flow CMR is comparable to Doppler echocardiography and has excellent repeatability for clinical use. However, 4D flow CMR can potentially underestimate peak velocity in patients with AF.

Semi-automatic thresholding of RV trabeculation improves repeatability and diagnostic value in suspected pulmonary hypertension

Objectives

Right ventricle (RV) mass is an imaging biomarker of mean pulmonary artery pressure (MPAP) and pulmonary vascular resistance (PVR). Some methods of RV mass measurement on cardiac MRI (CMR) exclude RV trabeculation. This study assessed the reproducibility of measurement methods and evaluated whether the inclusion of trabeculation in RV mass affects diagnostic accuracy in suspected pulmonary hypertension (PH).

Materials and methods

Two populations were enrolled prospectively. (i) A total of 144 patients with suspected PH who underwent CMR followed by right heart catheterization (RHC). Total RV mass (including trabeculation) and compacted RV mass (excluding trabeculation) were measured on the end-diastolic CMR images using both semi-automated pixel-intensity-based thresholding and manual contouring techniques. (ii) A total of 15 healthy volunteers and 15 patients with known PH. Interobserver agreement and scan-scan reproducibility were evaluated for RV mass measurements using the semi-automated thresholding and manual contouring techniques.

Results

Total RV mass correlated more strongly with MPAP and PVR (r = 0.59 and 0.63) than compacted RV mass (r = 0.25 and 0.38). Using a diagnostic threshold of MPAP ≥ 25 mmHg, ROC analysis showed better performance for total RV mass (AUC 0.77 and 0.81) compared to compacted RV mass (AUC 0.61 and 0.66) when both parameters were indexed for LV mass. Semi-automated thresholding was twice as fast as manual contouring (p < 0.001).

Conclusion

Using a semi-automated thresholding technique, inclusion of trabecular mass and indexing RV mass for LV mass (ventricular mass index), improves the diagnostic accuracy of CMR measurements in suspected PH.

Image Phenotyping of Preterm-Born Children Using Hyperpolarized 129Xe Lung Magnetic Resonance Imaging and Multiple-Breath Washout

Rationale: Preterm birth is associated with low lung function in childhood, but little is known about the lung microstructure in childhood. Objectives: We assessed the differential associations between the historical diagnosis of bronchopulmonary dysplasia (BPD) and current lung function phenotypes on lung ventilation and microstructure in preterm-born children using hyperpolarized 129Xe ventilation and diffusion-weighted magnetic resonance imaging (MRI) and multiple-breath washout (MBW). Methods: Data were available from 63 children (aged 9-13 yr), including 44 born preterm (⩽34 weeks' gestation) and 19 term-born control subjects (⩾37 weeks' gestation). Preterm-born children were classified, using spirometry, as prematurity-associated obstructive lung disease (POLD; FEV1 < lower limit of normal [LLN] and FEV1/FVC < LLN), prematurity-associated preserved ratio of impaired spirometry (FEV1 < LLN and FEV1/FVC ⩾ LLN), preterm-(FEV1 ⩾ LLN) and term-born control subjects, and those with and without BPD. Ventilation heterogeneity metrics were derived from 129Xe ventilation MRI and SF6 MBW. Alveolar microstructural dimensions were derived from 129Xe diffusion-weighted MRI. Measurements and Main Results: 129Xe ventilation defect percentage and ventilation heterogeneity index were significantly increased in preterm-born children with POLD. In contrast, mean 129Xe apparent diffusion coefficient, 129Xe apparent diffusion coefficient interquartile range, and 129Xe mean alveolar dimension interquartile range were significantly increased in preterm-born children with BPD, suggesting changes of alveolar dimensions. MBW metrics were all significantly increased in the POLD group compared with preterm- and term-born control subjects. Linear regression confirmed the differential effects of obstructive disease on ventilation defects and BPD on lung microstructure. Conclusion: We show that ventilation abnormalities are associated with POLD, and BPD in infancy is associated with abnormal lung microstructure.

First Genotype-Phenotype Study in TBX4 Syndrome: Gain-of-Function Mutations Causative for Lung Disease

Rationale: Despite the increased recognition of TBX4 (T-BOX transcription factor 4)-associated pulmonary arterial hypertension (PAH), genotype-phenotype associations are lacking and may provide important insights. Objectives: To compile and functionally characterize all TBX4 variants reported to date and undertake a comprehensive genotype-phenotype analysis. Methods: We assembled a multicenter cohort of 137 patients harboring monoallelic TBX4 variants and assessed the pathogenicity of missense variation (n = 42) using a novel luciferase reporter assay containing T-BOX binding motifs. We sought genotype-phenotype correlations and undertook a comparative analysis with patients with PAH with BMPR2 (Bone Morphogenetic Protein Receptor type 2) causal variants (n = 162) or no identified variants in PAH-associated genes (n = 741) genotyped via the National Institute for Health Research BioResource-Rare Diseases. Measurements and Main Results: Functional assessment of TBX4 missense variants led to the novel finding of gain-of-function effects associated with older age at diagnosis of lung disease compared with loss-of-function effects (P = 0.038). Variants located in the T-BOX and nuclear localization domains were associated with earlier presentation (P = 0.005) and increased incidence of interstitial lung disease (P = 0.003). Event-free survival (death or transplantation) was shorter in the T-BOX group (P = 0.022), although age had a significant effect in the hazard model (P = 0.0461). Carriers of TBX4 variants were diagnosed at a younger age (P < 0.001) and had worse baseline lung function (FEV1, FVC) (P = 0.009) than the BMPR2 and no identified causal variant groups. Conclusions: We demonstrated that TBX4 syndrome is not strictly the result of haploinsufficiency but can also be caused by gain of function. The pleiotropic effects of TBX4 in lung disease may be in part explained by the differential effect of pathogenic mutations located in critical protein domains.

The reproducibility of manual RV/LV ratio measurement on CT pulmonary angiography

Objectives

Right ventricular (RV) dysfunction carries elevated risk in acute pulmonary embolism (PE). An increased ratio between the size of the right and left ventricles (RV/LV ratio) is a biomarker of RV dysfunction. This study evaluated the reproducibility of RV/LV ratio measurement on CT pulmonary angiography (CTPA).

Methods

20 inpatient CTPA scans performed to assess for acute PE were retrospectively identified from a tertiary UK centre. Each scan was evaluated by 14 radiologists who provided a qualitative overall opinion on the presence of RV dysfunction and measured the RV/LV ratio. Using a threshold of 1.0, the RV/LV ratio measurements were classified as positive (≥1.0) or negative (<1.0) for RV dysfunction. Interobserver agreement was quantified using the Fleiss κ and intraclass correlation coefficient (ICC).

Results

Qualitative opinion of RV dysfunction showed weak agreement (κ = 0.42, 95% CI 0.37-0.46). The mean RV/LV ratio measurement for all cases was 1.28 ± 0.68 with significant variation between reporters (p < 0.001). Although agreement for RV/LV measurement was good (ICC = 0.83, 95% CI 0.73-0.91), categorisation of RV dysfunction according to RV/LV ratio measurements showed weak agreement (κ = 0.46, 95% CI 0.41-0.50).

Conclusion

Both qualitative opinion and quantitative manual RV/LV ratio measurement show poor agreement for identifying RV dysfunction on CTPA.

Advances in knowledge

Caution should be exerted if using manual RV/LV ratio measurements to inform clinical risk stratification and management decisions.

Quantitative CT Evaluation of Small Pulmonary Vessels Has Functional and Prognostic Value in Pulmonary Hypertension

Background The in vivo relationship between peel pulmonary vessels, small pulmonary vessels, and pulmonary hypertension (PH) is not fully understood. Purpose To quantitatively assess peel pulmonary vessel volumes (PPVVs) and small pulmonary vessel volumes (SPVVs) as estimated from CT pulmonary angiography (CTPA) in different subtypes of PH compared with controls, their relationship to pulmonary function and right heart catheter metrics, and their prognostic value. Materials and Methods In this retrospective single-center study performed from January 2008 to February 2018, quantitative CTPA analysis of total SPVV (TSPVV) (0.4- to 2-mm vessel diameter) and PPVV (within 15, 30, and 45 mm from the lung surface) was performed. Results A total of 1823 patients (mean age, 69 years ± 13 [SD]; 1192 women [65%]) were retrospectively analyzed; 1593 patients with PH (mean pulmonary arterial pressure [mPAP], 43 mmHg ± 13 [SD]) were compared with 230 patient controls (mPAP, 19 mm Hg ± 3). The mean vessel volumes in pulmonary peels at 15-, 30-, and 45-mm depths were higher in pulmonary arterial hypertension (PAH) and PH secondary to lung disease compared with chronic thromboembolic PH (45-mm peel, mean difference: 6.4 mL [95% CI: 1, 11] [P < .001] vs 6.8 mL [95% CI: 1, 12] [P = .01]). Mean small vessel volumes at a diameter of less than 2 mm were lower in PAH and PH associated with left heart disease compared with controls (1.6-mm vessels, mean difference: -4.3 mL [95% CI: -8, -0.1] [P = .03] vs -6.8 mL [95% CI: -11, -2] [P < .001]). In patients with PH, the most significant positive correlation was noted with forced vital capacity percentage predicted (r = 0.30-0.40 [all P < .001] for TSPVVs and r = 0.21-0.25 [all P < .001] for PPVVs). Conclusion The volume of pulmonary small vessels is reduced in pulmonary arterial hypertension and pulmonary hypertension (PH) associated with left heart disease, with similar volume of peel vessels compared with controls. For chronic thromboembolic PH, the volume of peel vessels is reduced. In PH, small pulmonary vessel volume is associated with pulmonary function tests. Clinical trial registration no. NCT02565030 Published under a CC BY 4.0 license Online supplemental material is available for this article.

Unconventional Pathways to Carbide Phase Synthesis via Thermal Decomposition of UI4(1,4-dioxane)2

UI₄(1,4-dioxane)₂ was subjected to laser-based heating─a method that enables localized, fast heating (T > 2000 °C) and rapid cooling under controlled conditions (scan rate, power, atmosphere, etc.)─to understand its thermal decomposition. A predictive computational thermodynamic technique estimated the decomposition temperature of UI₄(1,4-dioxane)₂ to uranium (U) metal to be 2236 °C, a temperature achievable under laser irradiation. Dictated by the presence of reactive, gaseous byproducts, the thermal decomposition of UI₄(1,4-dioxane)₂ under furnace conditions up to 600 °C revealed the formation of UO₂, UI_x, and U(C_1-xO_x)_y, while under laser irradiation, UI₄(1,4-dioxane)₂ decomposed to UO₂, U(C_1-xO_x)_y, UC_2-zO_z, and UC. Despite the fast dynamics associated with laser irradiation, the central uranium atom reacted with the thermal decomposition products of the ligand (1,4-dioxane = C₄H₈O₂) instead of producing pure U metal. The results highlight the potential to co-develop uranium precursors with specific irradiation procedures to advance nuclear materials research by finding new pathways to produce uranium carbide.

Phenotyping of idiopathic pulmonary arterial hypertension: a registry analysis

BACKGROUND

Among patients meeting diagnostic criteria for idiopathic pulmonary arterial hypertension (IPAH), there is an emerging lung phenotype characterised by a low diffusion capacity for carbon monoxide (DLCO) and a smoking history. The present study aimed at a detailed characterisation of these patients.

METHODS

We analysed data from two European pulmonary hypertension registries, COMPERA (launched in 2007) and ASPIRE (from 2001 onwards), to identify patients diagnosed with IPAH and a lung phenotype defined by a DLCO of less than 45% predicted and a smoking history. We compared patient characteristics, response to therapy, and survival of these patients to patients with classical IPAH (defined by the absence of cardiopulmonary comorbidities and a DLCO of 45% or more predicted) and patients with pulmonary hypertension due to lung disease (group 3 pulmonary hypertension).

FINDINGS

The analysis included 128 (COMPERA) and 185 (ASPIRE) patients with classical IPAH, 268 (COMPERA) and 139 (ASPIRE) patients with IPAH and a lung phenotype, and 910 (COMPERA) and 375 (ASPIRE) patients with pulmonary hypertension due to lung disease. Most patients with IPAH and a lung phenotype had normal or near normal spirometry, a severe reduction in DLCO, with the majority having no or a mild degree of parenchymal lung involvement on chest computed tomography. Patients with IPAH and a lung phenotype (median age, 72 years [IQR 65-78] in COMPERA and 71 years [65-76] in ASPIRE) and patients with group 3 pulmonary hypertension (median age 71 years [65-77] in COMPERA and 69 years [63-74] in ASPIRE) were older than those with classical IPAH (median age, 45 years [32-60] in COMPERA and 52 years [38-64] in ASPIRE; p<0·0001 for IPAH with a lung phenotype vs classical IPAH in both registries). While 99 (77%) patients in COMPERA and 133 (72%) patients in ASPIRE with classical IPAH were female, there was a lower proportion of female patients in the IPAH and a lung phenotype cohort (95 [35%] COMPERA; 75 [54%] ASPIRE), which was similar to group 3 pulmonary hypertension (336 [37%] COMPERA; 148 [39%] ASPIRE]). Response to pulmonary arterial hypertension therapies at first follow-up was available from COMPERA. Improvements in WHO functional class were observed in 54% of patients with classical IPAH, 26% of patients with IPAH with a lung phenotype, and 22% of patients with group 3 pulmonary hypertension (p<0·0001 for classical IPAH vs IPAH and a lung phenotype, and p=0·194 for IPAH and a lung phenotype vs group 3 pulmonary hypertension); median improvements in 6 min walking distance were 63 m, 25 m, and 23 m for these cohorts respectively (p=0·0015 for classical IPAH vs IPAH and a lung phenotype, and p=0·64 for IPAH and a lung phenotype vs group 3 pulmonary hypertension), and median reductions in N-terminal-pro-brain-natriuretic-peptide were 58%, 27%, and 16% respectively (p=0·0043 for classical IPAH vs IPAH and a lung phenotype, and p=0·14 for IPAH and a lung phenotype vs group 3 pulmonary hypertension). In both registries, survival of patients with IPAH and a lung phenotype (1 year, 89% in COMPERA and 79% in ASPIRE; 5 years, 31% in COMPERA and 21% in ASPIRE) and group 3 pulmonary hypertension (1 year, 78% in COMPERA and 64% in ASPIRE; 5 years, 26% in COMPERA and 18% in ASPIRE) was worse than survival of patients with classical IPAH (1 year, 95% in COMPERA and 98% in ASPIRE; 5 years, 84% in COMPERA and 80% in ASPIRE; p<0·0001 for IPAH with a lung phenotype vs classical IPAH in both registries).

INTERPRETATION

A cohort of patients meeting diagnostic criteria for IPAH with a distinct, presumably smoking-related form of pulmonary hypertension accompanied by a low DLCO, resemble patients with pulmonary hypertension due to lung disease rather than classical IPAH. These observations have pathogenetic, diagnostic, and therapeutic implications, which require further exploration.

FUNDING

COMPERA is funded by unrestricted grants from Acceleron, Bayer, GlaxoSmithKline, Janssen, and OMT. The ASPIRE Registry is supported by Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK.

Validation of time-resolved, automated peak trans-mitral velocity tracking: Two center four-dimensional flow cardiovascular magnetic resonance study

OBJECTIVE

We aim to validate four-dimensional flow cardiovascular magnetic resonance (4D flow CMR) peak velocity tracking methods for measuring the peak velocity of mitral inflow against Doppler echocardiography.

METHOD

Fifty patients were recruited who had 4D flow CMR and Doppler Echocardiography. After transvalvular flow segmentation using established valve tracking methods, peak velocity was automatically derived using three-dimensional streamlines of transvalvular flow. In addition, a static-planar method was used at the tip of mitral valve to mimic Doppler technique.

RESULTS

Peak E-wave mitral inflow velocity was comparable between TTE and the novel 4D flow automated dynamic method (0.9 ± 0.5 vs 0.94 ± 0.6 m/s; p = 0.29) however there was a statistically significant difference when compared with the static planar method (0.85 ± 0.5 m/s; p = 0.01). Median A-wave peak velocity was also comparable across TTE and the automated dynamic streamline (0.77 ± 0.4 vs 0.76 ± 0.4 m/s; p = 0.77). A significant difference was seen with the static planar method (0.68 ± 0.5 m/s; p = 0.04). E/A ratio was comparable between TTE and both the automated dynamic and static planar method (1.1 ± 0.7 vs 1.15 ± 0.5 m/s; p = 0.74 and 1.15 ± 0.5 m/s; p = 0.5 respectively). Both novel 4D flow methods showed good correlation with TTE for E-wave (dynamic method; r = 0.70; P < 0.001 and static-planar method; r = 0.67; P < 0.001) and A-wave velocity measurements (dynamic method; r = 0.83; P < 0.001 and static method; r = 0.71; P < 0.001). The automated dynamic method demonstrated excellent intra/inter-observer reproducibility for all parameters.

CONCLUSION

Automated dynamic peak velocity tracing method using 4D flow CMR is comparable to Doppler echocardiography for mitral inflow assessment and has excellent reproducibility for clinical use.

Validation of Artificial Intelligence Cardiac MRI Measurements: Relationship to Heart Catheterization and Mortality Prediction

Background Cardiac MRI measurements have diagnostic and prognostic value in the evaluation of cardiopulmonary disease. Artificial intelligence approaches to automate cardiac MRI segmentation are emerging but require clinical testing. Purpose To develop and evaluate a deep learning tool for quantitative evaluation of cardiac MRI functional studies and assess its use for prognosis in patients suspected of having pulmonary hypertension. Materials and Methods A retrospective multicenter and multivendor data set was used to develop a deep learning-based cardiac MRI contouring model using a cohort of patients suspected of having cardiopulmonary disease from multiple pathologic causes. Correlation with same-day right heart catheterization (RHC) and scan-rescan repeatability was assessed in prospectively recruited participants. Prognostic impact was assessed using Cox proportional hazard regression analysis of 3487 patients from the ASPIRE (Assessing the Severity of Pulmonary Hypertension In a Pulmonary Hypertension Referral Centre) registry, including a subset of 920 patients with pulmonary arterial hypertension. The generalizability of the automatic assessment was evaluated in 40 multivendor studies from 32 centers. Results The training data set included 539 patients (mean age, 54 years ± 20 [SD]; 315 women). Automatic cardiac MRI measurements were better correlated with RHC parameters than were manual measurements, including left ventricular stroke volume (r = 0.72 vs 0.68; P = .03). Interstudy repeatability of cardiac MRI measurements was high for all automatic measurements (intraclass correlation coefficient range, 0.79-0.99) and similarly repeatable to manual measurements (all paired t test P > .05). Automated right ventricle and left ventricle cardiac MRI measurements were associated with mortality in patients suspected of having pulmonary hypertension. Conclusion An automatic cardiac MRI measurement approach was developed and tested in a large cohort of patients, including a broad spectrum of right ventricular and left ventricular conditions, with internal and external testing. Fully automatic cardiac MRI assessment correlated strongly with invasive hemodynamics, had prognostic value, were highly repeatable, and showed excellent generalizability. Clinical trial registration no. NCT03841344 Published under a CC BY 4.0 license. Online supplemental material is available for this article. See also the editorial by Ambale-Venkatesh and Lima in this issue. An earlier incorrect version appeared online. This article was corrected on June 27, 2022.

Kat-ARC accelerated 4D flow CMR: clinical validation for transvalvular flow and peak velocity assessment

BACKGROUND

To validate the k-adaptive-t autocalibrating reconstruction for Cartesian sampling (kat-ARC), an exclusive sparse reconstruction technique for four-dimensional (4D) flow cardiac magnetic resonance (CMR) using conservation of mass principle applied to transvalvular flow.

METHODS

This observational retrospective study (2020/21-075) was approved by the local ethics committee at the University of East Anglia. Consent was waived. Thirty-five patients who had a clinical CMR scan were included. CMR protocol included cine and 4D flow using Kat-ARC acceleration factor 6. No respiratory navigation was applied. For validation, the agreement between mitral net flow (MNF) and the aortic net flow (ANF) was investigated. Additionally, we checked the agreement between peak aortic valve velocity derived by 4D flow and that derived by continuous-wave Doppler echocardiography in 20 patients.

RESULTS

The median age of our patient population was 63 years (interquartile range [IQR] 54-73), and 18/35 (51%) were male. Seventeen (49%) patients had mitral regurgitation, and seven (20%) patients had aortic regurgitation. Mean acquisition time was 8 ± 4 min. MNF and ANF were comparable: 60 mL (51-78) versus 63 mL (57-77), p = 0.310). There was an association between MNF and ANF (rho = 0.58, p < 0.001). Peak aortic valve velocity by Doppler and 4D flow were comparable (1.40 m/s, [1.30-1.75] versus 1.46 m/s [1.25-2.11], p = 0.602) and also correlated with each other (rho = 0.77, p < 0.001).

CONCLUSIONS

Kat-ARC accelerated 4D flow CMR quantified transvalvular flow in accordance with the conservation of mass principle and is primed for clinical translation.

The quality of reporting in cardiac MRI artificial intelligence segmentation studies - a systematic review

Funding Acknowledgements

Type of funding sources: Public Institution(s). Main funding source(s): This work was supported by an NIHR AI Award, AI_AWARD01706. This research was also funded in part, by the Wellcome Trust [Grant number 205188/Z/16/Z ].

Background

There has been a rapid increase in the number of Artificial Intelligence (AI) studies of cardiac MRI (CMR) segmentation. AI has huge potential to improve image analysis assessments. However, advancement and clinical translation in this field depend on researchers presenting their work in a transparent and reproducible manner.

Purpose

This systematic review aimed to evaluate the quality of reporting in AI studies involving CMR segmentation.

Methods

MEDLINE and EMBASE databases were searched for AI CMR segmentation studies on 18/11/2021. The flow of study inclusion is shown in Figure 1. Any AI method to segment any cardiac structure on CMR was eligible for inclusion. Each study was assessed for compliance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM).

Results

70 studies were included in the qualitative analysis. Studies were published between 2015 to 2021, with the majority (71%) published in 2020 and 2021. Most studies were performed in Europe (33%), China (27%) and the USA (26%). Short-axis sections were segmented in 70% of studies and most commonly included both ventricles (51%) or the left ventricle alone (30%). 20 different architecture implementations were represented. Figure 2 summarises the most relevant CLAIM domains to AI segmentation. The training sample eligibility criteria, demographics and clinical characteristics were not reported in 47% and 81% of studies, respectively. Ground truth annotations, source of the annotations and annotation tool were absent in 31%, 36% and 51% of studies respectively. Preprocessing steps and software libraries and packages used in training were not included in 27% and 24%. Details on the training approach including the number of models trained and method of selecting the final model were missing in 20% and 17% of the studies. Methods of validation or testing on external data, inter- and intra- rater variability and failure analysis were unreported in 57%, 63% and 74%, respectively.

Conclusion

This systematic review highlights important gaps in the AI literature of CMR studies. We identified key items missing in the dataset description, model development, validation and testing that limit the transparency, reproducibility and hence validity of published AI studies. This review may support closer adherence to established frameworks for reporting standards.

The Role of Artificial Intelligence in Predicting Outcomes by Cardiovascular Magnetic Resonance: A Comprehensive Systematic Review

Background and Objectives: Interest in artificial intelligence (AI) for outcome prediction has grown substantially in recent years. However, the prognostic role of AI using advanced cardiac magnetic resonance imaging (CMR) remains unclear. This systematic review assesses the existing literature on AI in CMR to predict outcomes in patients with cardiovascular disease. Materials and Methods: Medline and Embase were searched for studies published up to November 2021. Any study assessing outcome prediction using AI in CMR in patients with cardiovascular disease was eligible for inclusion. All studies were assessed for compliance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM). Results: A total of 5 studies were included, with a total of 3679 patients, with 225 deaths and 265 major adverse cardiovascular events. Three methods demonstrated high prognostic accuracy: (1) three-dimensional motion assessment model in pulmonary hypertension (hazard ratio (HR) 2.74, 95%CI 1.73−4.34, p < 0.001), (2) automated perfusion quantification in patients with coronary artery disease (HR 2.14, 95%CI 1.58−2.90, p < 0.001), and (3) automated volumetric, functional, and area assessment in patients with myocardial infarction (HR 0.94, 95%CI 0.92−0.96, p < 0.001). Conclusion: There is emerging evidence of the prognostic role of AI in predicting outcomes for three-dimensional motion assessment in pulmonary hypertension, ischaemia assessment by automated perfusion quantification, and automated functional assessment in myocardial infarction.

Right ventricular remodelling in pulmonary arterial hypertension predicts treatment response

OBJECTIVES

To determine the prognostic value of patterns of right ventricular adaptation in patients with pulmonary arterial hypertension (PAH), assessed using cardiac magnetic resonance (CMR) imaging at baseline and follow-up.

METHODS

Patients attending the Sheffield Pulmonary Vascular Disease Unit with suspected pulmonary hypertension were recruited into the ASPIRE (Assessing the Spectrum of Pulmonary hypertension Identified at a REferral Centre) Registry. With exclusion of congenital heart disease, consecutive patients with PAH were followed up until the date of census or death. Right ventricular end-systolic volume index adjusted for age and sex and ventricular mass index were used to categorise patients into four different volume/mass groups: low-volume-low-mass, low-volume-high-mass, high-volume-low-mass and high-volume-high-mass. The prognostic value of the groups was assessed with one-way analysis of variance and Kaplan-Meier plots. Transition of the groups was studied.

RESULTS

A total of 505 patients with PAH were identified, 239 (47.3%) of whom have died at follow-up (median 4.85 years, IQR 4.05). The mean age of the patients was 59±16 and 161 (32.7%) were male. Low-volume-low-mass was associated with CMR and right heart catheterisation metrics predictive of improved prognosis. There were 124 patients who underwent follow-up CMR (median 1.11 years, IQR 0.78). At both baseline and follow-up, the high-volume-low-mass group had worse prognosis than the low-volume-low-mass group (p<0.001). With PAH therapy, 73.5% of low-volume-low-mass patients remained in this group, whereas only 17.4% of high-volume-low-mass patients transitioned into low-volume-low-mass.

CONCLUSIONS

Right ventricular adaptation assessed using CMR has prognostic value in patients with PAH. Patients with maladaptive remodelling (high-volume-low-mass) are at high risk of treatment failure.

Quality of reporting in AI cardiac MRI segmentation studies - A systematic review and recommendations for future studies

Background

There has been a rapid increase in the number of Artificial Intelligence (AI) studies of cardiac MRI (CMR) segmentation aiming to automate image analysis. However, advancement and clinical translation in this field depend on researchers presenting their work in a transparent and reproducible manner. This systematic review aimed to evaluate the quality of reporting in AI studies involving CMR segmentation.

Methods

MEDLINE and EMBASE were searched for AI CMR segmentation studies in April 2022. Any fully automated AI method for segmentation of cardiac chambers, myocardium or scar on CMR was considered for inclusion. For each study, compliance with the Checklist for Artificial Intelligence in Medical Imaging (CLAIM) was assessed. The CLAIM criteria were grouped into study, dataset, model and performance description domains.

Results

209 studies published between 2012 and 2022 were included in the analysis. Studies were mainly published in technical journals (58%), with the majority (57%) published since 2019. Studies were from 37 different countries, with most from China (26%), the United States (18%) and the United Kingdom (11%). Short axis CMR images were most frequently used (70%), with the left ventricle the most commonly segmented cardiac structure (49%). Median compliance of studies with CLAIM was 67% (IQR 59-73%). Median compliance was highest for the model description domain (100%, IQR 80-100%) and lower for the study (71%, IQR 63-86%), dataset (63%, IQR 50-67%) and performance (60%, IQR 50-70%) description domains.

Conclusion

This systematic review highlights important gaps in the literature of CMR studies using AI. We identified key items missing-most strikingly poor description of patients included in the training and validation of AI models and inadequate model failure analysis-that limit the transparency, reproducibility and hence validity of published AI studies. This review may support closer adherence to established frameworks for reporting standards and presents recommendations for improving the quality of reporting in this field.

Systematic Review Registration

[www.crd.york.ac.uk/prospero/], identifier [CRD42022279214].

Cardiac magnetic resonance identifies raised left ventricular filling pressure: prognostic implications

AIMS

Non-invasive imaging is routinely used to estimate left ventricular (LV) filling pressure (LVFP) in heart failure (HF). Cardiovascular magnetic resonance (CMR) is emerging as an important imaging tool for sub-phenotyping HF. However, currently, LVFP cannot be estimated from CMR. This study sought to investigate (i) if CMR can estimate LVFP in patients with suspected HF and (ii) if CMR-modelled LVFP has prognostic power.

METHODS AND RESULTS

Suspected HF patients underwent right heart catheterization (RHC), CMR and transthoracic echocardiography (TTE) (validation cohort only) within 24 h of each other. Right heart catheterization measured pulmonary capillary wedge pressure (PCWP) was used as a reference for LVFP. At follow-up, death was considered as the primary endpoint. We enrolled 835 patients (mean age: 65 ± 13 years, 40% male). In the derivation cohort (n = 708, 85%), two CMR metrics were associated with RHC PCWP:LV mass and left atrial volume. When applied to the validation cohort (n = 127, 15%), the correlation coefficient between RHC PCWP and CMR-modelled PCWP was 0.55 (95% confidence interval: 0.41-0.66, P < 0.0001). Cardiovascular magnetic resonance-modelled PCWP was superior to TTE in classifying patients as normal or raised filling pressures (76 vs. 25%). Cardiovascular magnetic resonance-modelled PCWP was associated with an increased risk of death (hazard ratio: 1.77, P < 0.001). At Kaplan-Meier analysis, CMR-modelled PCWP was comparable to RHC PCWP (≥15 mmHg) to predict survival at 7-year follow-up (35 vs. 37%, χ2 = 0.41, P  = 0.52).

CONCLUSION

A physiological CMR model can estimate LVFP in patients with suspected HF. In addition, CMR-modelled LVFP has a prognostic role.

Cardiovascular medication in patients with raised NT-proBNP, but no heart failure in the SHEAF registry

OBJECTIVES

We aim to assess the association of cardiovascular medications with outcomes of patients referred to the diagnostic heart failure (HF) clinic with symptoms or signs of possible HF, raised N-terminal pro-brain-type natriuretic peptide (NT-proBNP) but no evidence of HF on transthoracic echocardiography (TTE).

METHODS

Data were collected prospectively into the Sheffield HEArt Failure (SHEAF) registry between April 2012 and January 2020. The inclusion criteria were symptoms or signs suggestive of HF, NT-proBNP >400 pg/mL, but no evidence of HF on TTE. Cox proportional-hazards regression model was used to investigate the association between the survival time of patients and different cardiovascular medications. The outcome was defined as all-cause mortality.

RESULTS

From the SHEAF registry, we identified 1766 patients with raised NT-proBNP with no evidence of HF on TTE. Survival was higher among the younger patients, and among those with hypertension or atrial fibrillation (AF). Mortality was increased with male gender, valvular heart disease and chronic kidney disease. Using univariate Cox proportional-hazards regression, the only cardiac therapeutic agent independently associated with all-cause mortality was beta-blocker (HR 0.86; 95% CI: 0.77 to 0.97; p=0.02). The use of beta-blockers was significantly higher in patients with AF (63% vs 39%, p<0.01) and hypertension (51% vs 42%, p<0.01). However, using multivariate Cox proportional-hazards regression to adjust for all variables associated with mortality, the influence of beta-blockers became non-significant (HR 0.96; 95% CI: 0.85 to 1.1, p=0.49).

CONCLUSION

When all variables associated with mortality are considered, none of the cardiovascular agents are associated with the improved survival of patients with suspected HF, raised NT-proBNP but no HF on echocardiography.