Individual lung models could save lives

With the help of artificial intelligence, the system uses CT data to calculate the condition of the patient’s lung. The marked areas show damage from a Covid-19 infection. (Image: Jakob Richter / TUM)

Acute lung failure! If Covid-19 is severe, artificial respiration is often the last resort. But this measure is risky: there is a risk of life-threatening overloading of the organ. As reported by the Technical University of Munich, digital models of the lungs of patients could enable individually adapted and therefore gentle ventilation in the future.

The availability of ventilators is known to be a major issue in the corona pandemic. They have to be ready for the particularly bad cases: The disease Covid-19 triggered by the coronavirus Sars-CoV-2 can lead to severe pneumonia. Similar to some other diseases, this can lead to acute respiratory distress syndrome (ARDS). In these cases, only artificial ventilation can save the patient’s life.

In this procedure, the patient is put under anesthesia and receives a tube into the trachea through which a device presses air into the organ to further enable the exchange of oxygen and carbon dioxide. But this process is not as simple as it sounds. Because the lungs are a very sensitive organ: If the artificial pressure causes permanent overstretching, it can lead to further life-threatening inflammatory reactions. Therefore, the ventilation pressure must be adjusted individually and depending on the patient’s condition.

Delicate emergency measure

To date, however, doctors have only had inaccurate information to set the optimal ventilation. “Until now, the treatment providers had no way of recognizing overstretching. From the trachea to the finest branches, the lungs have more than 20 stages of branching, and there is no measurement method to determine what happens at the micro-level of the lungs during ventilation, ”says Wolfgang Wall from the Technical University of Munich. The mechanical interactions between the different types of tissue, the flowing air and the liquid film on the tissue are extremely complex, says the scientist.

For this reason, he and his team have been working on the lung in more and more detail for years. They are based on simulations of the behavior of tissue and air flow as well as on micromechanical tests on real tissue samples. Her work has now led to a highly complex digital lung model that enables conclusions to be drawn about individual characteristics. To do this, a computer tomogram of the chest must first be performed and a breath of the patient analyzed. The computer model then uses artificial intelligence to calculate the lung volume from the data. It can also identify the condition of individual areas of the lung that are already damaged by the disease.

A digital twin of the patient’s lung is created

The system can thus show in detail which settings of the ventilator lead to which loads on the micro level in the lungs. Treatments can therefore adjust the ventilation pressure accordingly. So far, however, the clinical standard has been to determine the settings for ventilation using a rule of thumb based on body weight.

“Over 80 percent of Covid-19 deaths are due to acute lung failure. With long-term artificial ventilation of patients, the survival rate currently drops to around 50 percent, ”says Wall. “The aim of our work is that in the future, a digital lung model will help each ventilation site to optimally set the ventilation and so that we can significantly increase the chance of survival,” said the scientist.

As the TU Munich reports, the success of the scientists has already led to a spin-off: Wall, with three former employees and clinical partners, has the company “Ebenbuild“Founded. It should now bring the research results into clinical practice as quickly as possible.

Source: Technical University of Munich

Recent Articles

Related Stories