Point detection through multi-instance deep heatmap regression for sutures in endoscopy

Purpose: Mitral valve repair is a complex minimally invasive surgery of the heart valve. In this context, suture detection from endoscopic images is a highly relevant task that provides quantitative information to analyse suturing patterns, assess prosthetic configurations and produce augmented reality visualisations. Facial or anatomical landmark detection tasks typically contain a fixed number of landmarks, and use regression or fixed heatmap-based approaches to localize the landmarks. However in endoscopy, there are a varying number of sutures in every image, and the sutures may occur at any location in the annulus, as they are not semantically unique. Method: In this work, we formulate the suture detection task as a multi-instance deep heatmap regression problem, to identify entry and exit points of sutures. We extend our previous work, and introduce the novel use of a 2D Gaussian layer followed by a differentiable 2D spatial Soft-Argmax layer to function as a local non-maximum suppression. Results: We present extensive experiments with multiple heatmap distribution functions and two variants of the proposed model. In the intra-operative domain, Variant 1 showed a mean $$F_1$$ F 1 of $$+ 0.0422$$ + 0.0422 over the baseline. Similarly, in the simulator domain, Variant 1 showed a mean $$F_1$$ F 1 of $$+ 0.0865$$ + 0.0865 over the baseline. Conclusion: The proposed model shows an improvement over the baseline in the intra-operative and the simulator domains. The data is made publicly available within the scope of the MICCAI AdaptOR2021 Challenge https://adaptor2021.github.io/, and the code at https://github.com/Cardio-AI/suture-detection-pytorch/.

Long QT Syndrome 2 PAS Domain Variant Induces hERG1a/1b Subunit Imbalance in Patient-specific iPSC-cardiomyocytes

Background - Inherited long QT syndrome type 2 (LQT2) results from variants in the KCNH2 gene encoding the hERG1 potassium channel. Two main isoforms, hERG1a and hERG1b, assemble to form tetrameric channel. The N-terminal Per-Arnt-Sim (PAS) domain, present only on hERG1a subunits, is a hotspot for pathogenic variants, but it is unknown whether PAS domain variants impact hERG1b expression to contribute to the LQT2 phenotype. We aimed to use patient-specific induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) to investigate the pathogenesis of the hERG1a PAS domain variant hERG1-H70R. Methods - Human iPSCs were derived from a LQT2 patient carrying the PAS domain variant hERG1-H70R. CRISPR/Cas9 gene editing produced isogenic control iPSC lines. Differentiated iPSC-CMs were evaluated for their electrophysiology, hERG1a/1b mRNA expression, and hERG1a/1b protein expression. Results - Action potentials from single hERG1-H70R iPSC-CMs were prolonged relative to controls, and voltage clamp studies showed an underlying decrease in I Kr with accelerated deactivation. In hERG1-H70R iPSC-CMs, transcription of hERG1a and hERG1b mRNA was unchanged compared to controls based on nascent nuclear transcript analysis, but hERG1b mRNA was significantly increased as was the ratio of hERG1b/hERG1a in mRNA complexes, suggesting post-transcriptional changes. Expression of complex glycosylated hERG1a in hERG1-H70R iPSC-CMs was reduced due to impaired protein trafficking, whereas the expression of the complex glycosylated form of hERG1b was unchanged. Conclusions - Patient-specific hERG1-H70R iPSC-CMs reveal a newly appreciated mechanism of pathogenesis of the LQT2 phenotype due to both impaired trafficking of hERG1a and maintained expression of hERG1b that produces subunit imbalance and reduced I Kr with accelerated deactivation.

You Are Only as Good as You Are Behind Closed Doors

Virtues are standardly characterized as stable dispositions. A stable disposition implies that the virtuous actor must be disposed to act well in any domain required of them. For example, a politician is not virtuous if s/he is friendly in debate with an opponent, but hostile at home with a partner or children. Some recent virtue theoretic accounts focus on specific domains in which virtues can be exercised. I call these domain-variant accounts of virtue. This paper examines two such accounts: Randall Curren and Charles Dorn’s (2018) discussion of virtue in the civic sphere, and Michael Brady’s (2018) account of virtues of vulnerability. I argue that being consistent with the standard characterization of virtue requires generalizing beyond a domain. I suggest four actions the authors could take to preserve their accounts while remaining consistent with the standard characterization. I also discuss how virtue education could be enhanced by domain-variant accounts.

DomainScope: A disease network based on protein domain connections

Protein domains are highly conserved functional units of proteins. Because they carry functionally significant information, the majority of the coding disease variants are located on domains. Additionally, domains are specific units of the proteins that can be targeted for drug delivery purposes. Here, using information about variants sites associated with diseases, a disease network was built, based on their sharing the same domain and domain variation site. The result was 49,990 disease pairs linked by domain variant site and 533,687 disease pairs that share the same mutated domain. These pairs were compared to disease pairs made using previous methods such as gene identity and gene variant site identity, which revealed that over 8,000 of these pairs were not only missing from the gene pairings but also not found commonly together in literature. The disease network was analyzed from their disease subject categories, which when compared to the gene-based disease network revealed that the domain method results in higher number of connections across disease categories versus within a disease category. Further, a study into the drug repurposing possibilities of the disease network created using domain revealed that 16,902 of the disease pairs had a drug reported for one disease but not the other, highlighting the drug repurposing potential of this new methodology.