Loading...

quantusCOVID

Study for the detection of COVID-19

X-Ray

COVID19 causes a wide variety of findings in this type of imaging.

Artificial Intelligence

Allows automatic image analysis optimizing the diagnostic role of X-Ray

COVID-19: What it is?

COVID-19 is an infectious disease caused by the SARS-CoV-2 virus, which can be transmitted before the appearance of symptoms (presymptomatic) or even if the person has no symptoms (asymptomatic). This together with a contagion rate, estimated by the WHO, of 1.4 to 2.5, is what has led to the COVID-19 being classified as a pandemic.

The virus mainly affects the airways starting at the throat and continuing through the bronchial tubes to the lungs. To fight the virus, the body has an inflammatory response that can damage the lungs. This is why medical images have played an important role in the absence of tests to detect the disease.

map of COVID19 detected cases in the world
+
19
M

Confirmed cases

+
11
M

Healed patients

+
712
mil

Deaths from SARS-CoV-2

Why use chest X-Ray?

Chest X-Ray is a common, non-invasive and rapid test, widely used by clinicians worldwide to diagnose all sorts of respiratory diseases and viral infections affecting the respiratory system. Given the lack of PCR tests and the saturation of health systems, X-rays have been used as a triage tool 1 for patients who come to health systems for health care and as a diagnostic tool to monitor the accelerated progression of lung abnormalities in infected patients

Imagen de Tomografía computerizada
IA aplicada a la medicina

What does Artificial Intelligence contribute?

Due to the exponential increase in patients, medical personnel have been overwhelmed, which is why we need to find ways to help them fight this disease. These days, medical imaging has been key to triage of patients in the emergency department, but it must be borne in mind that the evaluation of an image by a radiologist takes about 15 minutes. This is where the AI ​​intervenes reducing these times since the processing of an image takes a few seconds.

Therefore AI allows reducing the workload on doctors in a scenario like the current outbreak of COVID-19.

quantusCOVID

In a first study, tested on 3,528 patients,the results indicate that the texture analysis using Deep Learning techniques provides information associated with COVID-19 involvement in CT images. With 96.4% global success and a 1% false positive rate.
The study, "Computer-aided covid-19 patient screening using chest images (X-Ray and CT scans)" 3 was published in MedrXiv.
The next step is to have a larger database to corroborate that the results obtained are still valid and in this way be able to offer quantusCOVID19 to the world.

First results: ROC curve

Do you want to join the project?

Send us an email

References:

  1. TKim, Hyunjoong W et al. Emergency radiology, 1–5. 22 Jun. 2020. The role of initial chest X-ray in triaging patients with suspected COVID-19 during the pandemic.
  2. Borghesi A, Medica RM-L radiologia, 2020 U. COVID-19 outbreak in Italy: experimental chest X-ray scoring system for quantifying and monitoring disease progression. Springer.
  3. Xavier P Burgos-Artizzu, MedrXiv Jul. 2020. Computer-aided covid-19 patient screening using chest images (X-Ray and CT scans).