CT image characteristics

A CT image consists of an array of picture elements called pixels. Each pixel represents the mean attenuation of an axial column of tissue about 1 cm in length. The attenuation is expressed as a Hounsfield number. Typical Hounsfield numbers for various tissues are given below:

• bone: +200 to +500
• soft tissue: +30
• water: 0 (by definition)
• fat: -40
• lung: -200 to -300

A CT scan image is displayed as if the observer is looking at the patient from the feet. Cranial and cervical scans may be displayed as if looking from above.

Because only a limited number of grey-scale steps can be perceived, the observer may select the range of Hounsfield numbers over which the grey-scale should be applied. This range is called the window. Different windows are appropriate for different tissues:

• a general abdominal scan may have a window of +20 to +200
• if soft tissue variation is important, e.g. metastases in the liver, a window of +20 to +40 might be selected
• if the architecture of the lung is important then a window of -200 to -300 might be selected