in

Study presents the first automatic method to detect and segment intrauterine cavity

Twin-to-twin transfusion syndrome (TTTS) happens in round 10-15% of pregnancies with twins that share the identical placenta. Sometimes, this syndrome seems earlier than 24 weeks’ gestation as a result of irregular vascular communications positioned on the floor of the placenta. In consequence, blood circulation is just not balanced between the 2 twins, dramatically reducing their probabilities of survival.

Fetoscopic laser photocoagulation is the simplest therapy for this syndrome and it consists of closing irregular vascular connections positioned on the floor of the placenta to utterly separate the circulation of blood to the 2 twins, thus stopping problems associated to blood circulate imbalance, resembling dying by cardiac overload, untimely supply and miscarriage.

The maneuverability of the fetoscope inserted by the uterine wall of the mom and the power to burn all vessels that require sealing will depend on the correct choice of the fetoscope entry level on the floor of the intrauterine cavity. Planning the very best insertion level earlier than the operation requires a superb understanding of the affected person’s anatomy, which might be achieved utilizing a digital illustration of the mom’s uterus, through magnetic resonance imaging.

A examine not too long ago printed within the superior on-line version of the journal IEEE Transactions on Medical Imaging presents the primary computerized technique to detect and section the intrauterine cavity through three views (axial, sagittal and coronal) of the MRI by the use of synthetic intelligence and deep studying methods.

A examine performed by Miguel Ángel González Ballester, ICREA analysis professor with the Division of Info and Communication Applied sciences (DTIC) at UPF, with Jordina Torrents-Barrena, first creator of the examine, Gemma Piella and Mario Ceresa, members of the UPF BCN MedTech Unit. Eduard Gratacós and Elisenda Eixarch, members of the Fetal i+D Fetal Medication Analysis Middle, BCNatal-Barcelona Middle for Maternal-Fetal and Neonatal Medication (Hospital Clínic and Hospital Sant Joan de Déu), IDIBAPS, are co-authors of the examine and chargeable for the clinics.

The methodology introduced makes use of neural networks primarily based on the brand new paradigm of capsules to efficiently seize the interdependency of the anatomy current within the MRI, significantly for distinctive class situations (anatomies), such because the intrauterine cavity and/or placenta.”


Jordina Torrents-Barrena, first creator of the paper

“The tactic designed is predicated on a reinforcement studying framework that makes use of capsules to delimit the situation of the uterus. A capsule structure is subsequently designed to section (or refine) the entire intrauterine cavity”, Torrents-Barrena provides. The latter community encodes essentially the most discriminatory and strong options within the picture.

The proposed technique is evaluated by 13 efficiency measures and can also be in comparison with 15 neural networks which were beforehand printed within the literature. “Our synthetic intelligence technique has been educated utilizing magnetic resonance imaging from 71 pregnancies”, Torrents-Barrena affirms.

“Having a three-dimensional illustration permits us to guage totally different entry factors and select the one that gives the very best visibility of all placental vessels with the slightest motion”, feedback Elisenda Eixarch, co-author of the examine. “Undoubtedly, the appliance of this know-how will enable us to maneuver in direction of safer, extra exact surgical procedure”, she provides.

On common, the methodology introduced obtains a segmentation efficiency of over 91% for all checks and comparisons, highlighting the potential of this strategy to be used within the every day scientific apply as a surgical planning technique.





Source link

May 22, 2020 — A roundup of the latest news about COVID-19

ABLE LOCK WHEELCHAIR DOCKING STATION COMPATIBLE WITH EZ LOCK AND Q-STRAINT MOUNT