Mitral valve (MV) assessment is key to diagnosing valvular disease and to addressing its serious downstream complications. Cardiac magnetic resonance (CMR) has become an essential diagnostic tool in MV disease, offering detailed views of the valve structure and function, and overcoming the limitations of other imaging modalities. Automated detection of the MV leaflets in CMR could enable rapid and precise assessments that enhance diagnostic accuracy. To address this gap, we introduce DeepValve, the first deep learning (DL) pipeline for MV detection using CMR. Within DeepValve, we tested three valve detection models: a keypoint-regression model (UNET-REG), a segmentation model (UNET-SEG) and a hybrid model based on keypoint detection (DSNT-REG). We also propose metrics for evaluating the quality of MV detection, including Procrustes-based metrics (UNET-REG, DSNT-REG) and customized Dice-based metrics (UNET-SEG). We developed and tested our models on a clinical dataset comprising 120 CMR images from patients with confirmed MV disease (mitral valve prolapse and mitral annular disjunction). Our results show that DSNT-REG delivered the best regression performance, accurately locating landmark locations. UNET-SEG achieved satisfactory Dice and customized Dice scores, also accurately predicting valve location and topology. Overall, our work represents a critical first step towards automated MV assessment using DL in CMR and paving the way for improved clinical assessment in MV disease.

Background: A subset of patients with mitral valve prolapse (MVP) are at risk of life-threatening ventricular arrhythmias (VAs), and a high burden of premature ventricular contractions (PVCs) is associated with increased mortality in patients with MVP. Previous studies indicate that PVC burden could either increase with increasing heart rate (HR), decrease with increasing HR, or with no relationship to HR. The association between PVC burden and HR in MVP is not known, nor if there is an association with VAs. Purpose: We aimed to explore the relationship between PVC burden and HR in patients with MVP. Furthermore, we aimed to explore whether the different PVC profiles were associated with severe VA. Methods: In this cross-sectional ambispective outcome study, we included patients with MVP and available Holter monitorings with >100 PVCs per 24-hours seen at our tertiary centre. The Holter monitoring with the highest PVC burden was used in case of multiple recordings. We investigated the relationship between hourly PVC burden and hourly mean HR by univariable mixed linear regression. We defined PVC profiles as (1) fast-HR-dependent-PVC (F-HR-PVC) when there was a significant positive correlation, (2) slow-HR-dependent-PVC (S-HR-PVC) when there was a significant negative correlation, and (3) independent-HR-PVC (I-HR-PVC) when no correlation between PVC and HR was found. The significance was set to 0.05. Severe VA was defined as sustained ventricular tachycardia, non-sustained ventricular tachycardia with haemodynamic instability, ventricular fibrillation, aborted cardiac arrest, or appropriate shock by a primary preventive implantable cardioverterefibrillator. Severe VA was recorded from retrospective medical history, prospective ICD interrogation or monitoring with implantable loop recorder. Results: We included 75 patients with >100 PVCs at Holter monitoring (median age 48 years [IQR 35-58], 58% female). The median PVC burden was 1.9% per 24-hours (IQR 0.4-7.2). We found F-HR-PVC in 49 patients (65%), S-HR-PVC in 1 (1%), and I-HR-PVC in 25 (33%). Twelve (16%) patients had severe VA and these had higher PVC burden compared to those without severe VA (7.3% per 24-hours [IQR 4.2-10.4] vs 1.1% per 24-hours [IQR 0.3-5.3], p=0.004). The median PVC burden was 1.5% (IQR 0.3-8.0) for F-HR-PVC, 5.7% for S-HR-PVC and 1.5% (IQR 0.3-8.0) for I-HR-PVC. We found no association between different PVC profiles and the presence of severe VA (F-HR-PVC 65% vs 67%, S-HR-PVC 2% vs 0%, I-HR-PVC 33% vs 33%, p>0.05 for all groups). Conclusion: Fast-HR-dependent-PVC was most common in MVP patients, and slow-HR-dependent-PVC were rare. A higher PVC burden was associated with severe VA. However, different PVC profiles did not infer higher risk of severe VA in our study. Further studies are needed to explore whether distinct PVC profiles can predict severe VA in larger patient populations.

Peripartum cardiomyopathy (PPCM) occurs in 0.02% to 0.1% of live births in European countries. Truncating variants in the Titin gene (TTNtv) are the most common genetic findings in women with PPCM. However, data on pregnancy-associated cardiomyopathy (PAC) and pregnancy tolerance among TTNtv-positive women are limited. The objectives of the study were to report maternal cardiac outcomes during pregnancy and postpartum and fetal outcomes in women with TTNtv and to explore the effect of parity on long-term cardiac phenotype. We conducted an ambispective cohort study including consecutive TTNtv genotype-positive women, regardless of phenotype, with ≥1 cardiological evaluation. Data on pregnancies and maternal and fetal outcomes were collected from questionnaires and medical records. Arrhythmic burden and device implantation were recorded. Left ventricular dimensions and function were assessed by echocardiography. A total of 107 TTNtv-positive women were included (age 43 [IQR: 32-55] years, 17% probands). Among 78 women with prior pregnancies, 206 pregnancies resulted in 167 (81%) live births. PAC occurred in 3 of 167 live births and in 3 of 76 women, corresponding to a cumulative incidence of 1.8% and 3.9%. All affected women recovered completely with medical therapy. Remaining pregnancies were largely uncomplicated, without increased maternal or fetal adverse events. At median of 18 years (IQR: 9-31) after pregnancy, no differences were observed between parous and nulliparous women in arrhythmic burden, cardiac function, or ventricular dimensions. In this TTNtv-positive cohort, PAC incidence appeared higher than reported PPCM rates in the general European population. However, pregnancy was well tolerated and was not associated with adverse long-term cardiac remodeling.

Keratoconus is a progressive eye disease that results in thinning of the cornea, leading to visual impairment. Mitral valve prolapse (MVP) is a common disorder affecting around 2–4% of the general population. Previous studies have found an overrepresentation of MVP in individuals with keratoconus, with a prevalence of 38–65%, suggesting a shared underlying mechanism. In this case-control study, patients with keratoconus were enrolled from a quality and research registry. They were examined by a 2D echocardiography to identify if they had MVP, billowing or normal mitral leaflets. Controls were matched from the population-based Trøndelag Health Study. Patients and controls underwent a detailed echocardiographic examination to detect abnormal mitral valves. We included 101 patients (age 33 [IQR 29–40], 75% male) with keratoconus and 101 matched individuals. MVP was found in 2 (2%), while billowing was found in 5 (5%) of keratoconus patients. No significant association was found between keratoconus and the prevalence of MVP or billowing compared to the control group. Moreover, no associations were found between severity of keratoconus with presence of MVP nor with billowing of the mitral valves. We could not confirm the previously reported association between keratoconus and MVP, suggesting that routine screening for MVP in keratoconus patients may not be warranted. However, we cannot rule out the possibility of an association in other gender, age and ethnic groups different than ours.

Background: Mitral valve prolapse (MVP) is a leading cause of mitral regurgitation in high-income countries. Although the course of MVP is often benign, a subset of patients progresses to significant valvular insufficiency, and a smaller subset develops ventricular arrhythmias or sudden cardiac death. This study aimed to identify genetic variants associated with MVP and/or mitral annulus disjunction (MAD) using whole-genome sequencing (WGS) in a Norwegian cohort. Methods: WGS was performed in 93 individuals diagnosed with MVP and/or MAD, of whom 92 passed genomic quality control criteria. Variants were called using Genome Analysis ToolKit (GATK) Best Practices and compared with non-Finnish European reference data from gnomAD (n = 34 029). Analyses were restricted to curated gene panels comprising: (i) non-syndromic MVP and cardiomyopathy genes, (ii) syndromic/connective tissue disorder genes, and (iii) arrhythmia-related genes. Enriched or underrepresented variants were identified using Fisher’s exact test with Bonferroni correction for multiple testing. Independent variants were derived through linkage disequilibrium pruning. A sensitivity analysis was conducted using low-coverage WGS data from an independent large Norwegian cohort – the Trøndelag Health Study (HUNT). Results: Across predefined gene panels, several coding and regulatory variants were significantly associated with MVP and/or MAD. A missense variant in FLNA and regulatory variants in TBX5 and SMAD4 were enriched in these patients. Additional significant associations involved regulatory variants in XYLT1, HS3ST4, SH3PXD2B , and a synonymous variant in COL1A2 . Furthermore, several regulatory variants in arrhythmia-related genes ( KCNK3, CACNA1D, NR2F1, KCNJ4 and others) were identified. Sensitivity analyses revealed population-level differences, underscoring the need of cautious interpretation of these findings. Conclusion: This targeted WGS study expands the genetic landscape of MVP and/or MAD, identifying several coding and regulatory variants across extracellular matrix, developmental, and electrophysiological pathways. These results support a multifactorial genetic architecture of MVP. Graphical Whole-genome sequencing of 92 Norwegian patients with mitral valve prolapse (MVP) and/or mitral annulus disjunction (MAD) identified coding and regulatory genetic variants across three curated gene categories: (i) non-syndromic MVP/cardiomyopathy genes, (ii) syndromic MVP/connective tissue disorder genes, and (iii) arrhythmia-related genes. Key associations included variants in FLNA, TBX5, COL1A2, SMAD4 , and several electrophysiological genes including KCNK3, CACNA1D, NR2F1 , and KCNJ4 . The findings implicate extracellular matrix (ECM) remodeling, transforming growth factor-β (TGF-β) signaling, cardiac developmental pathways, cytoskeletal integrity, and electrophysiological mechanisms, supporting a multifactorial genetic architecture of MVP/MAD.