BCNatal – Fetal imaging
A case study on
forms and images integration
Fetal and obstetrics disease treatment and diagnosis heavily relies in clinical imaging processing and analysis.
/ Scientific Manager – Fetal Medicine Research Center
BCNatal offers women comprehensive, state-of-the-art care in the field of obstetrics and fetal disease. It is also the number-one maternity ward in Catalonia in terms of number of births, with 6,300 births per year. BCNatal Fetal Medicine Research Center belongs to Hospital Sant Joan de Deu and Hospital Clínic of Barcelona, at the University of Barcelona. They are one of the groups with the largest scientific international production in their field. Over the last 10 years they have published over 450 research papers and have successfully developed large-budget research studies, including European and international projects.
|Tags||All, Clinical imaging|
There is the need to have platform to store all of BCNatal medical images (over 200,000) that allow to reuse them and create new clinical studies from a subset of these images.
SCollect is per se a multistudy solution, so organizations can have users with access to certain studies and with different roles for each study. We adapted this vision to allow BCNatal create studies by selecting a subset of images of their repository (i.e. filtering images with tag "Thorax" to create a thoracic related study).
Store on-premise all the digital clinical images related to all the patients of the institution.
All the imaging data is stored locally in BCNatal's servers, while the application is hosted in the cloud and access both the databases and the image repository through a secured ip tunnel.
Connection and communication with machine learning pipelines that process images and extract results.
BCNatal already has a team of data scientists designing and running machine learning algorithms over their patients' images, so we only had to automate their communication, creating a non-human user in SCollect that downloads images, process them to perform some calculations and uploads the results to the platform.
BCNatal ends up having a platform ready for long term research, where multiple studies can be created
Support for SCollect as a multistudy platform, and in the future we have planned to add features to the management website so that managers can cross data of different studies.
All the images are stored in BCNatal's servers.
Users can tag images. When opening an image for tagging within the platform, a DICOM viewer is presented to the user so he/she can tag the image accordingly. Tags are set up on a per study basis, so different studies can have different tags.
Automated image tagging and measurement is enabled in the platform so BCNatal's machine learning algorithms can execute over the platform's images and upload the results to SCollect.
The platform is configured for the use of two sites at the moment of writing this, but could be easily expanded to more sites in the future.
All the communications are secured end to end using SSL encryption.
Images of each study have forms associated with them so, depending on the tags of the image, users are requested to fill certain fields in the form. A DICOM viewer is presented to the user so he/she can perform the needed measures and the results are automatically written in the corresponding field.