Talk:Hounsfield scale

Rationale not quite correct
The section on rationale is not quite correct.

"The above standards were chosen as they are universally available references and suited to the key application for which computed axial tomography was developed: imaging the internal anatomy of living creatures based on organized water structures and mostly living in air, e.g. humans"

The original CT scanner had a fixed path length in water for each ray. The patient's head was effectively put into a bucket of water. Having all rays with almost the same attenuation reduced the dynamic range requirements of the detector system. So HU being relative to water is a basic component of the original design.

Not the real rationale so can it either be edited or removed.

Dave

Math4fun - can you please justify reverting this page back?

Dave 2nd March 2009 —Preceding unsigned comment added by DJMott-MCR (talk • contribs) 14:11, 2 March 2009 (UTC)

Can be applied to cone beam CT if use scatter correction
"The Hounsfield scale applies to medical grade CT scans but not to cone beam computed tomography (CBCT) scans."

That is not really correct. It "applies" to CBCT equally it just is not normally used. The reason for this is that CBCT has inherent X-ray scatter which contaminates the CBCT reconstruction and makes the reconstruction values non quantitative. However if you search research journals you'll find countless people doing scatter corrected CBCT and thus getting images in Hounsfield scale. --DB

Spiral CT not needed to view in slices from multiple perspectives
"Since the early 1990s, with advances in computer technology and scanners using spiral CT technology, internal three-dimensional anatomy is viewable by three-dimensional software reconstructions, from multiple perspectives, on computer monitors."

The computer part may be correct but the ability to do a spiral acquisition is not related to the available viewing angles of the reconstruction. After the data cube has been reconstructed you can view it at whatever angle you may like. It is merely interpolation error and rendering (for 3D) ability that is the issue. So a small slice thickness axial acquistion is just as suitable for 3D volumetric rendering as a spiral scan. In fact this sentence is a touch irrelevant as CT data is almost always primarily viewed as slices as this is what most doctor's experience and current clinical guidelines are based on. --DB — Preceding unsigned comment added by Draco-bob (talk • contribs) 15:58, 4 July 2011 (UTC)

Units depend on X-ray voltages (or energy spectrum)
The typical values cited for higher absorbing materials such as bone depend very much on the x-ray spectrum used. For example, typical medical CT scanners operate at energies around 120kV while bio-medical microCT scanner operate more in the 60-90kV range. Synchrotron CT energies can be even lower. The absorption spectra of air, water, and bone are very different over these ranges. Hounsfield units being defined via air and water are therefore very inconstant when extrapolated beyond the absorption ranges they are defined in, eg to bone. In mircoCT, cortical bone values are near 6000 HU rather than the 3000 HU found in typical medical CTs. — Preceding unsigned comment added by 80.254.165.167 (talk) 07:47, 1 June 2012 (UTC)

The HU of common substances
A row for "Liver" appears twice in this table. If not an error it should be explained (the footnote for one entry does not explain anything). Arbalest Mike (talk) 01:21, 17 July 2017 (UTC)
 * Removed older (not cited) entry. Arbalest Mike (talk) 19:22, 17 July 2017 (UTC)