[Intl_DxMedPhys] Hybrid Canon IR/CT displayed CTDIvol accuracy

[Szczykutowicz, Timothy P]

The vendors are like used car salepeople sometimes. In my book, I give this advice (see below an excerpt from my book), to wade through this slice versus channel BS. Feel free to use, it should get to the bottom of these issues quickly.

I also heard some vendors with less than 16 cm of collimation put in specs saying they can do 16 cm volume because of some Frankenstein definition of helical scanning (Philips).

Sometimes I am so confused by what the vendors say, at acceptance testing I rely on the ratio of CTDIvol and DLP to figure out what the heck the thing is doing…

In the older days, some of this confusion was legit due to detector costs and DAS issues (like where you had systems that used thicker slices at wider cone angles). Now it is all marketing fluff and helps non one.

From the chapter in my book called “Buyer’s Guide”
I highlighted the text I refer to above. Sorry about the ? mark symbols, some copy paste thing from a pdf.

17.1.4 Slice Number
The “slice wars” in MDCT were a time of rapid growth in the number of detector modules
the major MDCT vendors provided. As MDCT scanners grew in slice number from
1–4 to 16–64/128, beam collimation coverages increased from being the multiple mm
range to 4 cm. This allowed for scan times to be drastically reduced. Exams that would
have taken minutes on single-slice scanners could be performed in seconds. It also opened
up new possibilities for MDCT scanning. With smaller scan times, angiographic exams
became possible; in order to capture the hemodynamics of the arterial blood vessels with
reasonable amounts of CT contrast agents, fast scanning is mandatory. Before 2- to 4-cm
detector coverages, many angiographic exams were not clinically feasible.
As shown in Figure 1.11, the “slice wars” occurred in the late 1990s early 2000s. Since
about 2005, we have not seen significant increases in the number of detector modules or
beam collimation. Multiple vendors now offer scanners with beam collimation at 16 cm.
As a buyer, one is now presented with the choice from many vendors to buy a low-end
new scanner with a 1-cm beam collimation to that vendor’s premium line having a collimation
of 8 to 16 cm. In general, all indications will benefit from a protocol using a faster
scanning acquisition, which is often afforded by using a scanner with a wider beam collimation
or dual-source configuration. The relative benefit, however, is not the same for all
indications. An outpatient imaging center performing routine head, chest, and abdomen
pelvis imaging would not be fully realizing the benefits of a wide axial or dual-source scanner.
On the other hand, a center doing a lot of cardiac, CTA work, perfusion imaging (i.e.,
brain perfusion or body perfusion), 4DCT gated imaging of the lung fields, pediatric
BUYER’S GUIDE OF OPTIONAL FEATURES IN CT 515
imaging, etc. would benefit relatively more from the purchase of a scanner with a wide
axial coverage or dual-source configuration. This is because scan times usually do not matter
for routine imaging where arterial contrast dynamics are not needed or where cardiac
or respiratory motion will not be a diagnostically limiting source of artifact.
The definition of slice is somewhat unclear. Most vendors will claim more slices than
they have detector rows. It is important to understand exactly what is meant by slice number,
beam collimation, and number of detector rows.
• Number of Slices. This is the most common number used to define how “good” an
MDCT scanner is. For example, a person may say, “All of my department’s CT scanners
have 32 or more slices,” implying their department has a fleet of relatively up-todate
MDCT scanners. The number of slices reported by vendors, however, cannot be
taken to actually refer to the number of detector modules along the z (i.e., patient
superior-inferior) axis of the scanner. Nor can this number be used to compute the
beam collimation in units of distance (i.e., mm or cm). The number of slices reported
by a vendor number represents the number of data samples collected per view angle in
the z-direction during a scan. In general, as the number of reported slices increases,
the beam collimation also increases. Also in general, as the number of slices increases,
the number of actual detector rows increases. Caveats here are that many vendor’s
reports of the number of slices is equal to double the number of actual detector rows.
This causes confusion when one vendor claims to have a 128-slice scanner compared
to another vendor’s 64 slices, when both scanners actually have the same detector
design. Vendors claim double the actual number of detector elements as the slice
number because they change the focal spot location during the acquisition of the projection
data. This is referred to as deflecting or wobbling the focal spot, and it produces
and extra set of projection measurements for each detector row. Hence, the vendor is
accurate in referring to their slice number as double the number of detector rows. The
majority of Siemens scanners report their slice number in the “double” way, and some
of the other CT vendors do as well. To cut through the vendor marketing claims on
slice number, ask your sales representative to provide you with the following, “What is
the widest beam collimation in mm this scanner has, and at that collimation, what is
the smallest slice thickness I can reconstruct.” The answers to that question will yield
the important information about a vendor’s detector capabilities.
• Beam Collimation. Beam collimation is simply the physical extent (i.e., distance in
mm or cm) of the x-ray beam at isocenter. This is also referred to as beam width.
Referring to this quantity as beam width can be confusing, as some people confuse
this with slice width. On old single-slice scanners, beam collimation and slice collimation
were equal. This is not the case for modern multi-slice MDCT scanners. For
example, a 128-slice Siemens scanner with a beam collimation of 38.4 mm can reconstruct
0.5 mm slices over the entire 38.4 mm coverage. Beam collimation is 10 mm
on 16-slice scanners, 4 cm on most of the major CT vendors premium 64/128-slice
scanners of the mid 2000s, and 8 to 16 cm on modern premium scanners.
• Number of Detector Rows. This is the actual number of discrete channels of data
coming from the detector array. Focal spot wobbling or deflection cannot change the
actual number of distinct detector channels. It should be noted that the detector element
size for each channel may not be equal across the entire detector in the z-direction.
Most modern scanners have equally sized detector elements across the entire
516 THE CT HANDBOOK
detector surface. On some older scanners, however, detector channels on the outside
edges of the detector in the z-direction may be larger than those in the center. Therefore,
be careful when reviewing scanner quotes for detector coverage and detector row
number. A vendor may cite a beam collimation and smallest detector size that cannot
be used together. Let us consider a hypothetical scanner to illustrate this point. Consider
a scanner with a maximum beam collimation of 2 cm. At 2 cm, the scanner may
only allow 1-mm slices to be reconstructed, but if the scanner is operated in 1-cm collimation,
it allows 0.5-mm slices to be reconstructed. In this case, our hypothetical
scanner had a detector configuration consisting of 20  0.5 mm detector elements in
the center with another 5  1 mm detector elements on each side of the central
0.5-mm detectors. This can be an issue if you want to use full detector coverage (i.e.,
to achieve fast scan times and avoid tube heating issues) but also realize the best spatial
resolution your scanner is capable of.
In general, scanner cost and option sets are closely tied to the number of slices the
scanner has. Price and default option packages will usually increase as the number of slices
increases. Some scanner options may only be available on scanners with a certain minimum
number of slices from some vendors. Some vendors will also sell the same scanner
make and model with different slice options. For example, if you desire to buy a cardiac
gating package option on some scanners, the vendor often will then include the option for
the highest slice mode available on that scanner.



Timothy P. Szczykutowicz, Ph.D., DABR
Associate Professor
Departments of Radiology, Medical Physics and BME
University of Wisconsin Madison
Cell# 1-716-560-7751<tel:(716)%20560-7751>
Office# 1-608-263-5729
he/him/his

From: Intl_dxmedphys_wd_osu_list <intl_dxmedphys_wd_osu_list-bounces at lists.osu.edu> On Behalf Of Michael Masiar via Intl_dxmedphys_wd_osu_list
Sent: Friday, February 21, 2025 4:47 PM
To: Daniel Vergara <xvergarax at gmail.com>
Cc: intl_dxmedphys_wd_osu_list at lists.osu.edu
Subject: Re: [Intl_DxMedPhys] Hybrid Canon IR/CT displayed CTDIvol accuracy


WARNING: This email appears to have originated outside of the UW Health email system.
DO NOT CLICK on links or attachments unless you recognize the sender and know the content is safe.
I have a client with this system. If they buy the upgraded 320 slice detector its truly 320 x 0. 5 with a 16 cm detector. Whereas the other option is a "160" slice system with a 4cm detector. . . so its really 80 x 0. 5 and the 160 is
I have a client with this system. If they buy the upgraded 320 slice detector its truly 320 x 0.5 with a 16 cm detector.  Whereas the other option is a "160" slice system with a 4cm detector... so its really 80 x 0.5 and the 160 is some flying focal spot nonsense.  If your system doesn't have 320 as an option and the helical protocols say 160 x 0.5 ... they are really 80 x 0.5.  With that adjustment everything comes out ok, measuring CTDIvol between 5 and 20% below reported depending on the protocol.


Michael Masiar MS DABR
Certified Medical Physicist
Therapy Physics, Inc.
2501 Cherry Ave., Suite 270
Signal Hill, CA 90755
email: mmasiar at gmail.com<mailto:mmasiar at gmail.com>
cell: 310-625-9906
office: 562-317-0650



On Fri, Feb 21, 2025 at 2:02 PM Daniel Vergara via Intl_dxmedphys_wd_osu_list <intl_dxmedphys_wd_osu_list at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list at lists.osu.edu>> wrote:
For those of you who have worked with Canon CT scanners, in particular the hybrid IR/CT Alphenix CT scanner, what has been your experience with the displayed CTDIvol accuracy? Do you often get CTDI accuracy results with large % differences?
For those of you who have worked with Canon CT scanners, in particular the hybrid IR/CT Alphenix CT scanner, what has been your experience with the displayed CTDIvol accuracy? Do you often get CTDI accuracy results with large % differences?

We have measured repeated CTDIvol differences exceeding 20% for several protocols of different collimation (nT), tube voltage, phantom size, pitch, axial vs helical, scanned FOV and rotation time. And these high % differences have repeated over the course of a year. We have reached out to Canon reps for assistance, but have not gotten satisfactory help. We also tried measuring the CTDI using the Canon CTDI instructions, in which 10 peripheral measurements are made with defined time delays for certain rotation times and then averaged.

I'm fine admitting if we've performed the measurements incorrectly, somehow, but I need to know how testing CTDI on a Canon Alphenix differs from other scanners.


TIA
Daniel

--
Daniel Vergara
xvergarax at gmail.com<mailto:xvergarax at gmail.com>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.osu.edu/mailman/private/intl_dxmedphys_wd_osu_list/attachments/20250224/fccd5d75/attachment.html>