• ARDS
• lung ultrasound
• segmentation

Detection and measurement of pulmonary edema with a fast, non expensive and non invasive technique would be a useful tool in the intensive care setting. Indeed the assessment of lung water in a critically ill patient is clinically used to guide fluid therapy and as a marker of severity in common diseases such as acute lung injury (ALI) and acute respiratory distress syndrome (ARDS).

The techniques more frequently employed for edema measurement are represented by dilution techniques, quantitative computed tomography and chest x-ray. They all present some drawbacks which reduce their clinical usability. CT scans are expensive and use ionizing radiations, chest x-rays are not very accurate especially when performed at the patient's bed, dilution techniques need arterial catheterization and underestimate edema in ARDS.

In the recent years a promising technique for lung exploration has been developed, namely lung ultrasound. It performs very well in the assessment of pulmonary edema, although in a semi-quantitative manner. For example, it can detect cardiogenic edema with a greater accuracy than x-rays. Lung ultrasound has also been compared to dilution techniques showing a good correlation between the two methods.

In this work we further study the accuracy of lung ultrasound in the assessment of pulmonary edema by comparing its findings to the lung weight calculated by means of quantitative computed tomography (the gold standard for non invasive evaluation of lung parenchyma).

In order to reach this goal we worked on two different fronts: on the one side we designed an experimental study (performed at the S. Chiara Hospital, Pisa) to gather lung ultrasound and CT data in patients with pulmonary edema. On the other side we wrote a software to calculate the lungs weight and volume from a chest CT scan by means of semi-automatic medical imaging segmentation techniques.