Over the last three decades, researchers from around the world have put in the work to understand the concept of transit dosimetry. That work has been translated into technological improvements and commercial solutions.
Has transit dosimetry come of age? I believe it has. Of course, it can be improved, but in its current form it provides access to information we've never had before—information that can enhance patient safety.
A Rich History of Research and Development
Transit dosimetry is not a brand new concept. Its development started
in the 80's to early 90's when electronic portal imaging devices were
developed to replace film based treatment field verification.
Camera-based systems, liquid-filled ion chamber matrix assemblies, and
early versions of silicon diode array systems were the early stages of
usage and development.
Even then, physicists were tuned in to the potential of EPID
technology, but we still had major hurdles to overcome. For example,
there was the fact that proper transit dosimetry requires photon fluence
to be converted to dose. Professor Rogers made significant progress in
overcoming this hurdle using the EGS3 Monte Carlo algorithm.
In 1990, we saw two interesting studies. In the first, the authors
predicted portal dose image calculated during the treatment planning,
and compared it with the image taken during the treatment to verify
geometric alignment and delivered dose.
The second study used an iterative approach to match the calculated
DRR with the measured portal dose image. CT data was modified to
represent the actual dose transient path and new dose calculations were
performed on the new CT dataset. This is interesting because the
described workflow is similar to modern cone beam CT-based dose
This work happened in 1990. The authors were not waiting for
commercial entities to develop the technology, or for money to be pumped
in. They laid the ground work, knowing the most important goal was
improving the quality of treatment, and improving the methodology to
measure treatment quality.
By 1998, we learned more about how EPIDs behaved, how they could be
used for dose verification, and how accurate the results could be;
EPIDs used as dosimeters show a linear response and good dynamic
rage with doses that can be as accurate as diodes or ion chambers.
EPIDs show good response to changes in field size and gantry position.
A point spread function successfully connects doses with grey scale values measured with CCD camera based EPID.
Accurate portal dose measurements from a CCD camera based EPID when compared with ion chamber measurements.
The liquid ion chamber matrix system shows potential for use in on-line radiotherapy dose verification.
Also during this time, Ben Mijneer and several of his colleagues
at the Netherlands Cancer Institute really set themselves apart with
their excellent work in this area. For example, they figured out the
constraints preventing good agreement of EPID-measured dose rate for
different geometries and inhomogeneities. They also discovered that it
is possible to measure the midplane dose from exit dosimetry and
evaluate the differences in patient anatomy.
At the turn of the century, physicists took what they learned about
EPIDS, and applied it to 3-dimensional dose reconstruction. This was
important work that helped Transit Dosimetry become a viable tool to
study in a clinical setting.
Transit Dosimetry vs. Pre-Treatment verification
When IMRT and VMAT techniques were introduced, quite a lot of
pre-treatment verification was carried out. To generate the data for
pre-treatment verification, you need access to the treatment unit, and
at most centers, there isn't enough time or manpower during work hours
to carry this out. If you have transit dosimetry, this problem can be
Transit dosimetry gives you a double advantage— you get extra
confidence that the treatment unit is performing adequately, but you
also have the added advantage of knowing the actual patient position and
This is supported by the findings of another study from the
Netherlands Cancer Institute. They found that combining information from
three fractions of EPID in vivo dosimetry was the best way to
distinguish systematic errors from non-clinically relevant
discrepancies, save hours of quality assurance time per patient plan and
enable verification of the actual patient treatment.
Gaining Confidence in Transit dosimetry
As with any new technology, especially in medical physics, it has to
be vetted and validated before we feel comfortable with it. This work
was done in two studies.
The first uses multiple 2D planes within the patient volume and
reconstructed a 3D dose grid. When compared with the TPS, the results
were within 2% at the dose prescription point. Within 50% isodose
surface of the prescribed dose, at least 97% percent were in agreement,
evaluated with a 3D gamma method with a 3% 3mm criteria. However, their
dose reconstruction model didn't include tissue inhomogeneities.
The second study addresses this gap. In evaluating in vivo PDI
measurements behind actual patients, researchers found that on average,
87% of the pixels inside the field fulfilled the 3% local dose and 3mm
DTA. That is wonderful!
These studies provide evidence for us to gain confidence in the
practice of transit dosimetry. For such a complex technology to meet
stringent QA criteria is further evidence that the technology has come
Sure, further improvements could be made, and that will happen over a
period of time. However the technology and knowledge exists today. If,
as a clinician, you can improve the quality of your patients'
treatments, even incrementally, it is worth taking advantage of that
Transit Dosimetry in a Busy Clinic? It is Possible
We’ve been doing Transit Dosimetry at the Edinburgh Cancer Center
since 2012. In 2013, we implemented a rigorous transit dosimetry program
for all conformal VMAT radical treatments. This excludes plans with a
treatment field larger than our EPID.
I have been collecting data on these patients who go through transit
dosimetry, and so far, that totals about 3,250 patients. This is only to
show that yes, we have implemented transit dosimetry on a large scale,
so it is possible.
Interpolation and results from this data are still forthcoming, but I
have been amazed at what we can see with transit dosimetry. Quite a lot
of change is happening during treatment, but we have not been truly
aware of these changes, even with some of the modern imaging
technologies available today.
Today, this information comes through with transit dosimetry. And if
there is an error, it forces you to look into different avenues of
treatment--not just the treatment delivery machine. Therefore, transit
dosimetry enhances our global understanding of treatment delivery, and
that is a boon for our patients.
Where Does Transit Dosimetry Stand?
Let's think of four parameters to gauge the status of transit
dosimetry. It can provide either true or wrong information, and it can
be either precise or imprecise in identifying errors in all clinical
From all the research that has been done, we know transit dosimetry
does not give false information. Researchers have developed and tested
3D dose distributions, and verified it in heterogeneous and homogeneous
phantoms, as well as various clinical sites. They’ve tried different
clinical techniques like conformal, VMAT and hypofractionated SABR.
Through this work, it’s been validated that transit dosimetry is a
The precision in transit dosimetry still needs to be improved so that
it can detect all errors all the time. As it stands, in lung and breast
cases, transit dosimetry may not show actual differences depending on
what algorithm is used—some algorithms need a better solution for
heterogeneity corrections. The ideal would be a Monte Carlo based
calculation. This would shift transit dosimetry from being true and
imprecise, to being true and precise.
However, in order to get to that point, the medical physics community
needs to embrace transit dosimetry in its current state. Only by using
this technique can we identify specific problems and think of solutions.
More interest also shows vendors that transit dosimetry is a
sustainable market in which they can invest research and development
In the meantime, transit dosimetry provides insight that we've never
had before—valuable insight that can improve the quality of treatment
No, Transit Dosimetry Has Not Come of Age
By Geoff Budgell
When I started in medical physics, back in 1994, the new technology
was MLC and EPID. At my clinic back then, we were very proud because we
had two EPIDs! Even then, people were already thinking about how to use
EPID for doing dosimetry. But still, 21 years later, the development of
transit dosimetry has lagged behind.
Therefore, I argue: no, transit dosimetry has not (yet) come of age.
Today, still very few people are using transit dosimetry, especially
when compared to all the technologies we now have in routine clinical
use—FFF, IMRT, and VMAT among others. In addition, manufacturers have
been quite slow to get on board. So, why is that?
Why Has it Taken So Long to Develop?
For the first ten years of the transit dosimetry lifespan, we didn't
really have the right technology. We were using fluoroscopic devices or
liquid-filled ionization devices. It wasn't until amorphous silicon
devices came along that we got a stable, undistorted dosimeter.
But during that period, transit dosimetry was outshone by seemingly
more important technologies. IMRT, cone beam CT, VMAT, and SABR had much
more immediately appealing benefits for the patient.
I think also that there were doubts about commercial viability of
transit dosimetry. Some equipment manufacturers wondered whether it was
worth putting money into developing it, because they didn't know whether
they could sell it. Beyond that, they were already invested in
developing technologies like VMAT and CBCT.
Transit Dosimetry has a lot of growing up to do
So, has transit dosimetry come of age? Well, what do we actually mean by come of age?
There are some dictionary definitions. "Something that's come of age
has reached full, successful development," or we might say it's "reached
maturity; attained prominence, respectability, recognition or maturity;
or developed completely."
How do we apply those kinds of definitions to technology? I would
suggest there are 4 ways we can judge the maturity of technology:
Is the technology reliable and robust? If the
comments from different speakers at the IPEM Transit Dosimetry meeting
this year are any indication, transit dosimetry applications have
several problems that still need to be ironed out. Comments on problems
found with some of the currently available applications were very
wide-ranging, varying from the software failing to recognized images,
plans or patients to “the VMAT method is clunky” amongst many others.
Is it a highly automated process that operates smoothly?
More comments from the IPEM meeting indicate transit dosimetry is not
yet automated enough: for example presenters told us that there is
“Almost no automation!”, “Long processing time” “Time consuming, need to
automate the internal steps used in the software”
Has it been widely adopted? Most people today are interested in transit dosimetry but we know that it has not yet been widely adopted.
Is it well-understood? People aren't quite sure what they'd do with transit dosimetry, and there are a few myths that cause confusion.
3 Myths about Transit Dosimetry
Transit dosimetry replaces diodes. With transit
dosimetry, you're measuring something different than you are with
diodes. Transit dosimetry measures exit dose. It also measures 2D or 3D
dose distributions, as opposed to 1D with diodes. Transit dosimetry is
potentially very useful, whereas a diode measurement is almost useless,
telling you very little about the plan that you're delivering to the
Transit dosimetry is going to save you time and money
With transit dosimetry, you're inevitably going to see more problems
because we are starting to look into 2D and 3D. Investigating those
issues will take more time, and physics time costs money.
Transit dosimetry might save radiographer (therapist) time vs diodes,
but not necessarily if you think the purpose is to check the dose every
day. Positioning the panel for each field could potentially take longer
than putting diodes on the patient for each fraction.
However, transit dosimetry may save time since it could help you
reduce the number of pre-treatment verifications you are doing.
Transit dosimetry makes life easier Radiotherapy is
getting more complex. We can now see what we're treating. We are now
able to treat the right volumes. We can now account for movement, and
now with things like cone beam and transit dosimetry we can actually see
and measure what we're treating.
We need automated tools to do these complex things, and in that sense,
transit dosimetry is going to be important in the future. I don't think
we're there yet, but I think transit dosimetry is necessary. For one
thing, we're going to eventually need it for adaptive planning on the
fly. If you're taking images and re-planning live, then we're going to
need tools that can quickly measure the impact of those actions.
Where Does Transit Dosimetry Stand?
When technology is first introduced, there's a lot of excitement, and
you get what's called "the peak of inflated expectations." Then reality
starts to set in when people find out what the problems are. They then
fall into the trough of disillusionment, thinking perhaps it's not worth
it after all.
But then, once people have realized what the technology is capable of
and we start to sort out the problems, we move up the slope of
enlightenment. Then is becomes widely used and very productive.
Transit dosimetry, I would say, is somewhere on the slope of
enlightenment, we're moving towards where we can use it, but we're not
I don’t believe transit dosimetry will come of age until; first of all, we have tools that are robust, fully automated, and widely adopted; secondly, until the radiotherapy community actually knows what it's for and what it's going to be for in the future; and thirdly, when it's become a standard purchase with every linac that we buy.