International team led by the UOC has developed a new method that improves the quality of CT Scan Imagery

Talking about the diagnosis of the diseases or analyzing their effects on human body, CT or the computerized tomography scan is considered as one of the most promising medical tests. Keeping in mind the current scenario driven by COVID-19 pandemic, its role in examining the virus impact on the lungs has become more crucial. UOC guided international team has recently discovered a novel approach that promises to improvise the image quality obtained via CT scans.

This algorithm, tested against simulated data, allows to better differentiate between different types of body tissue and reduce the amount of radiation, the patient is otherwise exposed to, during this type of test. CT utilizes a computational process to combine many X-ray measurements taken from different angles of the body to produce a tomography image. This non-invasive procedure that reconstructs organs or tissues in three dimensions allows the doctor to see inside the subject without amputation.

In the words of Mohammad Mahdi Dehshibi, who was a postdoctoral researcher at the UOC’s Scene Understanding and Artificial Intelligence laboratory (SUNAI), in Teheran, Iran- “This technique helps experts to determine the presence of a tumor, its exact location, size and spread. It can also be used to diagnose muscle and bone disorders, infection or blood clots, heart disease, lung nodules, and liver masses.” However, CT risks exposing your body to high-dose X-rays, damaging your DNA structure and consequently cancer.

For example, in the course of a head CT scan, a person usually receives a radiation dose equivalent to the total amount of exposure to everyday life in 243 days. To reduce this radiation, the team led by Dehshibi has developed a new post-processing algorithm that improves the quality of reconstructed CT images. While conventional CT methods capture only a portion of the X-ray energy spectrum, the researchers tested a wider energy range divided by intervals to reach higher contrast. After testing with data generated using GATE / GEANT4 simulation software, the algorithm found that by reducing the noise while improving the quality of the image, it is possible to better identify different types of tissue with low-dose X-rays according to the results.