Context and Rationale
In December 2019, a new strain of coronavirus, the coronavirus disease 2019 (COVID-19) was identified in Wuhan, China. Within a few months, the virus was able to spread over nearly the entire planet, through human-to-human contamination via aerosols from coughing and sneezing, but also via contaminated surfaces, eventually becoming a major pandemic [1,2]. The illness was first characterized by severe respiratory illness in humans leading to up to 20% to cases of acute respiratory distress
syndrome (ARDS). In most patients afflicted, non-contrast chest computed tomography (CT)-scans show bilateral ground glass opacities, nodules and “crazy paving” with a peripheral and basal distribution. Clinicians worldwide face these lungs infectious with no simple therapies to address them.
Our technical proposal
DIVA-CT scan without VR: optimization of transfer function and probability of clots
We have developed a software platform called DIVA [7-9] that provides visualization and analysis of medical images for pre-operative planning (https://diva.pasteur.fr/). The platform can integrate any tomographic image (MRI, CT-scan, 3D echography) and provide a full volumetric representation with an optional virtual reality (VR) interface. DIVA is already deployed in multiple hospitals with applications in breast cancer surgery, craniostenosis in children, facial trauma surgery and in heart surgery planning. DIVA is also being leveraged for medical students in the contexts of facial surgery (Université de Paris [10]) and in liver surgery (DU de pédagogie appliqué à l’enseignement medical & hospital Paul Brousse). Three-dimensional reconstructions in DIVA can be visualized on a 2D screen through an interface that has already deployed in the hospital setting. DIVA does not pretreat data, nor does it require segmentation for its reconstructions; it utilizes data in its native form (e.g. DICOM format).
At the end of this project, we will have a stand-alone software allowing CT-scan angiographies to be visualized and analyzed in seconds. We will develop a Bayesian learner allowing the probability of clots to be overlaid onto the full volumetric representation of the scan. This application may be found useful beyond the COVID-19 crisis and will offer new possibilities for analyzing vascular anomalies in children.
We conclude by emphasizing that while we aim at being useful during this crisis, we cannot be sure that we will be. Adopting new technology in the current context is challenging in hospitals and we point out that some hospital may adopt a strategy of directly adding blood thinners (anticoagulants, antiplatelets) to patients arriving with severe COVID-19 symptoms without performing angiography. Recent advances using the DIVA software ensures that we will advance with the full support of medical staff.
Références
[1] Cobey S., Science 10.1126/science.abb5659 (2020).
[2] Kissler et al., Science 10.1126/science.abb5793 (2020).
[3] J. B. Moore, C. H. June, Science 10.1126/science.abb8925 (2020)
[4] Grillet F. et al, Radiology https://doi.org/10.1148/radiol.2020201544 (2020)
[5] Oudkerk M., Radiology, https://doi.org/10.1148/radiol.2020201629 (2020)
[6] Couzin-Frankel J., Science 368 (6490), 455-456, https://doi.org/10.1126/science.368.6490.455
[7] M. El Beheiry, et al, bioRxiv 2020.04.09.019935 (2020)
[8] F. Ruiz, et al, J.P.R.A.S 2020, (In Press)
[9] Blanc et al., https://www.biorxiv.org/content/10.1101/2020.03.26.000448v1 (2020)
[10] Bouaoud J. et al, (In preparation)
[11] El Beheiry M., (In preparation)
fr