IMAT/VMAT basics

One of the newest and most interesting external beam delivery techniques today goes by many names: intensity modulated arc therapy (IMAT), volumetric modulated arc therapy (VMAT), etc. To make things more confusing, vendors have each given this technique their own proprietary names: RapidArc (Varian), SmartArc (Philips), and VMAT^TM [didn't I already say VMAT?] (Elekta). Maybe the most general and correct name for IMAT would be cone-beam dynamic angle fluence modulated x-ray therapy (CBDAFMXT), but you might confuse that acronym with a chemotherapy drug name... In this post I'll discuss some of the basics of this arc-based form of IMRT (and just call it IMAT to keep things simple).

At its most basic IMAT is essentially conventional IMRT, but with the gantry moving in one or more rotating arcs, rather than delivering from a small number of fixed angles. This means that most of the concepts and advantages and disadvantages of IMRT apply to IMAT (detailed below). IMAT was developed (and marketed!) as a conventional linac-based alternative to helical tomotherapy and as a more conformal / lower critical structure dose and faster version of static angle IMRT.

The hierarchy of IMRT techniques.

In the figure showing the IMRT hierarchy, IMAT is on the branch of cone-beam, dynamic gantry IMRT. In order to deliver IMAT, a linac must have some of the following capabilities: gantry motion with beam on, dynamic MLC (i.e. leaf motion with beam on and gantry rotating), and variable dose rate.

Planning of IMAT is very similar to conventional IMRT. The plan is determined by inverse planning methods. The degrees of freedom are increased by considering gantry rotation speed, dose rate, number of field shapes, number of arcs, etc. For planning, arcs are usually approximated with a finite number of angles (e.g. 36). Constraints can be more tightly matched with multiple arcs at the expense of delivery time. Another important aspect in IMAT optimization is that MLC leaf speed limits the beam shape "distance" from one angle to the next, i.e. the MLC leaf positions cannot vary greatly from one angle to the next and thus beam shape "interconnectedness" must be taken in to account.

Advantages of IMAT include:

• Highly conformal target volume dose with lower dose to critical structures than IMRT or 3DCRT, as dose is spread over more angles.
• Faster delivery times and lower MU's (especially single arc IMAT) when compared with IMRT.
• Non-co-planar arcs possible.
• Comparable plans to helical tomotherapy, but performed with a conventional linac.

Disadvantages of IMAT include:

• Higher cost of hardware and software licensing relative to IMRT.
• Increased complexity of plans makes QA a poor diagnostic tool (i.e. hard to determine source of QA failures).

IMAT delivery techniques are the obvious(?) next step following IMRT. In fact, it's hard to come up with a list of concrete disadvantages of IMAT over IMRT. (Please comment if you feel otherwise.) In our clinic it's one of the few new techniques that everyone seems to have adopted with open arms.

Further reading:

• Cedric X Yu and Grace Tang, Intensity-modulated arc therapy: principles, technologies and clinical implementation, 2011 Phys. Med. Biol. 56 R31 doi:10.1088/0031-9155/56/5/R01 (open access).
• David Shepard, Clinical Implementation of Intensity Modulated Arc Therapy, presentation, 2009, http://www.medicaldosimetry.org/meetings/2009handouts/Shepard_VMAT.pdf

The many faces of bolus: Part 2

Previously I discussed the role of bolus material in radiation therapy and some of the forms it takes. This post shows a couple of other examples.

Pink bolus molded into shape.

Super Stuff bolus, also known generically as pink bolus, is a moldable bolus material with the consistency of gelatin. The material is described by the manufacturer as a "hydophilic organic polymer" and is sold in individual powder packets. Pink bolus is supposed to have a density of 1.02 g/cm³. To use the bolus, you add the necessary amount of water, allow the material to set, i.e. coming to its gelatin-like consistency, and then knead it into the shape you want. Over time pink bolus will lose its shape and must be re-shaped. Eventually it will lose some consistency due to moisture loss and a new batch must be made. Care must also be taken to remove as many air bubbles as possible.

A packet of pink bolus powder.

Recently in our clinic we treated a patient with classic (i.e. non-HIV related) Kaposi sarcoma of the leg with photons. For this we decided to use rice grains as the bolus material. As with all bolus, the idea of using rice is to simulate tissue and modify the dose distribution as desired. In this case, increase of skin dose is desired.

Rice bolus box for treating a patient's leg/foot.

For this patient we built a polystyrene foam box and filled it with loose rice grains. It took approximately 10 kg of dry parboiled rice to fill the box with the patient's leg. The patient plus rice box was then scanned with the CT and planned as normal.

Some leftover rice (not) used as bolus material.

The open access article linked below from Ahn et al. shows some dosimetric comparisons between the use of rice as a bolus and a water bolus for irradiating extremities. I will warn you that both methods create a mess at best :)

Further reading:

• Ahn SK, Kim YB, Lee IJ, Song TS, Son DM, Jang YJ, Cho JH, Kim JH, Kim DW, Cho JH, Suh CO.   Evaluation of a Water-based Bolus Device for Radiotherapy to the Extremities in Kaposi's Sarcoma Patients.   J Korean Soc Ther Radiol Oncol. 2008 Sep;26(3):189-194.   http://dx.doi.org/10.3857/jkstro.2008.26.3.189 (Open access. In Korean with abstract and figure captions in English.)

Dose-volume histogram basics

A dose-volume histogram (DVH) is a mathematical tool to assess the appropriateness of a given radiation therapy plan. It can be used to assess whether a plan meets desired constraints for a voulme of interest, within certain limitations. DVH’s are widely used and understanding how they work is a basic skill for treatment plan assessment. In this post I’ll discuss some DVH basics.

A typical cumulative dose-volume histogram (cDVH).

A DVH is nothing more than a histogram, but it is important to understand where the data comes from and how the DVH is representing the data. Modern treatment plans are created based on 3D image sets created using CT, MRI, etc. These data sets consist of voxels (the 3D equivalent of pixels). A volume of interest, e.g. a PTV, consists of a subset of these voxels. The basic data in a DVH is generated by binning the dose values from each voxel in the volume. (Interpolation may be necessary if the bound of the volume intersects a voxel.) This binned dose frequency data comprises a differential dose-volume histogram, or dDVH, which I will discuss in more detail in a future post. The dDVH looks like a common histogram and gives you an idea of how many voxels receive a certain dose, e.g. the dDVH might show that 85% of the PTV voxels received 98% - 102% of the prescribed dose and 46% received exactly 100% of the prescribed dose.

The more familiar form of DVH is the cumulative dose-volume histogram, or cDVH. This DVH is calculated by summing the dDVH starting at the dose of interest, D, up to the max dose, D_max (Eq. 1).

Eq. 1

The cDVH displays the percent/number of voxels in a volume which receive at least a dose D, i.e. the cDVH of a volume irradiated perfectly uniformly to 100 cGy will show that 100% of the voxels received at least 30 cGy, 50 cGy, 80 cGy, etc, but 0% received 105 cGy. Thus for an ideal treatment plan, the cDVH’s of the target volumes will have a rectangular, step-down function appearance and the cDVH’s of critical volumes will drop immediately to zero.

In the real world treatment plans are not ideal (I know, it’s sad). Instead acceptable dose constraints are set for targets and critical structures. DVH’s can be used to determine if these constraints are mets. One caveat is that standard DVH’s do not directly provide spatial information about the dose distribution. One less than ideal method is to create sub-volumes, but creating useful/meaningful sub-volumes is a non-trivial exercise.

Top image from Vorwerk et al. Radiation Oncology 2008 3:31, doi:10.1186/1748-717X-3-31, used under CC License terms.

Compensator-based IMRT

Intensity modulated radiation therapy (IMRT) is almost always performed with the use of a multileaf collimator (MLC). This is, however, not the only way to deliver static angle IMRT. Another method is with the use of compensator blocks. In this post I will talk a little bit about this less common IMRT technique.

Brass IMRT field compensator from .decimal, Inc.

As discussed in a previous post, IMRT requires fluence modulation not possible with conventional poured / hand-cut blocks. This fluence modulation is necessary to achieve the desired target matching and critical structure sparing via inverse planning optimization. This fluence modulation is typically achieved using an MLC, which has many advantage as well as disadvantages. An alternative method, in use since at least the mid 1990's, is fluence modulation via solid compensator blocks designed for each individual field. The above image shows a sample compensator made of milled brass.

Compensator-based IMRT is purported to have several advantages over MLC-based IMRT, including:

• Being static, each field is delivered more quickly (also lower MU's).
• Fluence patterns can be closer to the ideal, i.e. not limited by leaf size, speed, or leakage.
• Potentially cheaper.
• Avoids field splitting. (Did I ever mention I hate split fields?!?)

Along with these advantages come possible drawbacks, including:

• Long fabrication times, versus automated MLC patterens.
• Therapists must change compensator for each field.
• Potential for beam hardening.
• Large size / weight to achieve low dose regions.

Compensators can be fabricated from a range of materials, including brass, Wood's metal (Cerrobend), PMMA (Plexiglas), and tungsten powder composite. Milling. molding, or stacking and bolting are possible fabrication techniques. A handful of companies sell custom fabricated IMRT compensators on demand, delivering within one or two days of order.

Do you have any experience with this technique?

Further reading:

• Chang, S., Cullip, T., Deschesne, K., Miller, E., & Rosenman, J. Compensators: An alternative IMRT delivery technique. Journal Of Applied Clinical Medical Physics, 5(3), 2004. doi:10.1120/jacmp.v5i3.1965 (open access)
• P.C. Williams, IMRT: delivery techniques and quality assurance, British Journal of Radiology (2003) 76, 766-776, doi: 10.1259/bjr/12907222 (open access?)

Medical physics journals

If you want to keep up to date on the latest developments in medical physics, journals are one of the best resources. In this post I'm going to compile a list of medical physics journals and journals with medical physics related content. I will also mention the degree of open access for each journal (that I'm aware of) and the h-index as computed by Google Scholar.

Medical physics specific journals:

• Medical Physics, published by AAPM. Some article types made open access. h5-index: 45.
• Physics in Medicine and Biology, published by Institute of Physics. Open access option for authors with publishing fee. h5-index: not available.
• Journal of Applied Clinical Medical Physics, published by AAPM. Fully open access. h5-index: 15.
• Journal of Medical Physics, published by the Association of Medical Physicists of India. Fully open access. h5-index: 7.
• arXiv.org medical physics category. Fully open access. Overall for arXiv.org, h5-index: 256.
• Magnetic Resonance in Medicine, published by the International Society for Magnetic Resonance in Medicine. Paid access only. h5-index: 50.

Other journals with medical physics content:

• The Red Journal (International Journal of Radiation Oncology * Biology * Physics), published by ASTRO. Paid access only. h5-index: 68.
• The Green Journal (Radiotherapy and Oncology), published by ESTRO. Paid access only. h5-index: 48.
• Radiation Oncology, published by BioMed Central. Fully open access. h5-index: 23.
• Practical Radiation Onoclogy, published by ASTRO. Paid access only.  h5-index: N/A.
• Medical Dosimetry, published by the American Association of Medical Dosimetrists. Paid access only. h5-index: 15.

More info can be found on the state of open access and medical physics publications in my post about open access on Will Work for Science.

Any other additions to add?

Comparing dose distributions: The gamma test

In my last post I discussed dose distribution comparison with dose difference and distance-to-agreement (DTA) tests. Another widely used and closely related method for comparing dose distributions is the gamma test.

The gamma test was first introduced by Low et al. in 1998 as a single metric that combined features of both dose difference and DTA, while performing robustly in the regions where those are prone to failure. Conceptually, gamma is very similar to dose difference and DTA, but combines them into an abstract metric resembling a distance (Eq. 1). In this way both dose difference and DTA are taken into account for every point compared (rather than either-or as previously discussed).

Eq. 1

Eq. 2

In the above equations I have used somewhat different notation than Low et al. in an attempt to make things slightly clearer.

If we wish to compare two dose distributions, e.g. a measured versus a calculated distribution, we will have a dose, D_a(r_a), in the first distribution at point r_a, and a dose, D_b(r_b), at the corresponding point r_b in the second distribution. The DTA condition is fulfilled when D_a(r_a) = D_b(r_b+r), where r is an arbitrary point a distance |r| away from r_b. This condition defines an isodose contour in distribution b around point r_b. Away from this contour the DTA, d_DTA, is undefined. DTA is used with a threshold passing value, δ_DTA, e.g. 3mm. A DTA smaller than the threshold is considered passing for a simple DTA test. For gamma, δ_DTA is used to normalize the DTA value, such that a normal “passing” value would then be unity.

Dose difference is simply the difference of the two doses at the corresponding points: |D_a(r_a) = D_b(r_b)|. As with DTA, a pass/fail threshold, δ_DD, is used in the simple dose difference test, but is used to normalize the result in the gamma equation, such that the normal "passing" value would be unity.

We now have two components: normalized DTA and normalized dose difference. By squaring these values, adding, and taking the square root, we have a distance-like metric, Γ, shown in Eq. 1. Because DTA is only defined for values of r, such that D_a(r_a) = D_b(r_b+r), Γ is only defined when that condition is met (geometrically located along the DTA isodose contour).

Finally, the actual gamma index, γ, is determined by finding the minimum value of Γ by varying r. This essentially means traveling along the isodose contour and finding the point at which DTA is smallest.

The convention is for passing γ to be ≤ 1 and failing to be > 1. You will notice that a point yielding normalized DTA = 1 and normalized dose difference = 1 would now fail, since the corresponding γ would be √2.

What γ provides is a single value to evaluate, versus using separate tests and then considering both. As with DTA, γ presents challenges in efficient implementation (clearly Eq.'s 1 and 2 are not hand solvable).

Your comments (especially corrections) are appreciated.

Roy

Further reading:

• D. A. Low, W. B. Harms, S. Mutic, and J. A. Purdy, A technique for the quantitative evaluation of dose distributions, Med. Phys. 25, 656 (1998); http://dx.doi.org/10.1118/1.598248

Comparing dose distributions: DTA and dose-difference

Radiation therapy plan quality assurance often hinges on comparing calculated dose distributions with measured dose distributions. One of the most common techniques to compare dose distributions is the combined use of distance-to-agreement (DTA) and dose difference. In this post I will give an overview of these concepts.

A planar dose distribution from IMRT.

Dose difference is a very straight forward comparison of dose at corresponding points in two distributions. Given a point a_p in the planned distribution and the corresponding point a_m in the measured distribution, the dose difference is simply D(a_m) - D(a_p). A passing criterion is used, e.g. 3% of planned dose, such that if the measured dose difference is <= 3% the measured distribution "passes" at that point. The drawback of the dose difference test is that it is not robust in high gradient regions, as small misalignments can cause large dose differences.

Distance-to-agreement (DTA) is also very straight forward conceptually. Given a point a_p in the planned distribution and the corresponding point a_m in the measured distribution, the distance-to-agreement is the nearest point in the measured distribution from a_m, such that D(a_m + r) = D(a_p). As with dose difference, a passing criterion is chosen, e.g. 3 mm. If the matching dose level is found within a radius of <= 3 mm, the measured distribution "passes" at that point, . This technique is quite robust against misalignments in high gradient regions, as the matching dose level will still be nearby. However, this technique is prone to failure in low gradient regions, where even small misalignments can require a large radius to find the matching dose level. To avoid this to some extent, a dose threshold value can be used, such that dose below of the x% isodose line is not considered, where x% is usually a low dose, penumbra region.

In order to overcome the lack of  robustness in high and low gradient regions for dose difference and DTA respectively, the two tests are often used in conjunction.  This is done by evaluating the tests independently and then defining a point as passing if it passes either test. Thus a distribution might pass 70% for DTA and 70% for dose difference, but 90% for the combined test.

What we've seen is that both of these tests are conceptually quite straight forward, though each has its limitations. The subtleties and challenges arise in the efficient implementation of the algorithms (DTA in particular).

I discuss the gamma test in this post.


Further reading:

• I. J. Yeo and J. O. Kim, A Procedural Guide to Film Dosimetry, Medical Physics Publishing, 2004, ISBN: 9781930524194
• W. B. Harms, Sr., D. A. Low, J. W. Wong, and J. A. Purdy, A software tool for the quantitative evaluation of 3D dose calculation algorithms, Med. Phys. 25, p.c1830 (1998); http://dx.doi.org/10.1118/1.598363
• J. Van Dyk, R. B. Barnett, J. E. Cygler, and P. C. Shragge, "Commissioning and quality assurance of treatment planning computers," Med. Phys. 26, 261–273 (1993). http://dx.doi.org/10.1118/1.598801

Image by MBq and licensed under CC license terms.

Radiation therapy availability around the world

As a medical physicist, you often have to explain to people what it is exactly that you do. A couple of years ago I met a diplomat from Togo who asked me what I was studying (at the time). My attempt to explain medical physics to him was made all the more difficult, when it became clear that radiation therapy was not something that was available in his country and he had no familiarity with it.

That discussion stuck in my mind and I began to wonder, what exactly is the availability of radiation therapy around the world. Fortunately, the International Atomic Energy Agency is much more concerned about that question than I am and has been keeping stats on that very topic since 1959 via their DIRAC project. Besides providing statistics on the number of clinics and machines in each country, a project named after a prominent physicist also lists stats on the number of medical physicists in each country.

IAEA statistics on radiation therapy machines per capital in 2010.

The above map color codes each country as a function of radiation therapy machines (linacs, teletherapy, or HDR) per capita as of 2010. A more up-to-date, interactive version is found here. The DIRAC site also provides the raw data and even information on individual clinics in each country.

While there are certainly non-negligible error bars on their data, the numbers are revealing, though largely what you'd expect. The highest GDP per capita (or likely highest health spending per capita) countries show up in green on the above map (5 or more machines per million) and the poorest countries show up in dark orange or red ( < 1 machine per million). Togo is red because it has zero machines.

I think the main implications on the medical physics end is with regards to education and dissemination of current knowledge to countries with few physicists. The US, population approximately 3.1x10^8, is listed as having 1728 therapy physicists (probably a low estimate) versus India, population approximately 1.2x10^9, listed as having 144 therapy physicists. Clearly the man-years of experience are highly concentrated in the green countries. I'll claim that it's our duty in the green countries to help educate our colleagues in the countries with less local access to their professional and academic peers.

New medical physicists in the US: Crunching the numbers

If you are entering the field of medical physics or have been around for a while, you might be wondering "how many new medical physicists are joining the ranks each year?" In the US this is an especially important question in light of the 2014 ABR residency mandate and possible effects of an aging population on cancer incidence.

As a recent graduate, one of the topics fresh on my mind is the number of jobs available to graduating students. This is of course a supply and demand game (or possibly a supply and supply game when you consider number of residency spots). Since potential medical physicists in the US can come from many different "sources" (i.e. accredited medical physics grad programs, non-accredited medical physics grad programs, non-medical physics grad programs), it would be somewhat difficult to directly count the number of new grads. I think it is therefore instructive to look at the raw stats of the number of people taking the ABR medical physics board exams, which are (recently?) available on the ABR website.

What we see in the data is a marked increase in the number of people taking all three parts of the ABR certification exam over the period of 2006 - 2010, with a slight downtick in 2008. The increase in the number taking Part 3 (Oral) in all specialties is +45% (220 in 2006 to 319 in 2010). The increase in the number taking Part 1 over the same time period is approximately +35%.

This presents the obvious question of what will happen to this trend when the ABR residency requirement takes full effect in 2014.


For more info on this topic:

• Mills MD, Thornewill J, Esterhay RJ. Future trends in the supply and demand for radiation oncology physicists. J Appl Clin Med Phy. 2010 Apr;11(2) (open access!)
• Jean Moore, Medical Physics Workforce Study: Overview, AAPM presentation, 2010
• Final report - AAPM Workforce Study Report (AAPM login required)

The many faces of bolus

Bolus is a simple, yet important technology used in radiation therapy. The most basic function of bolus material is to shape the dose distribution in a desired way. This generally falls into two categories: compensating for "missing" tissue and enhancing the build-up effect of MeV energy photon beams.

The bolus itself can be made of a huge variety of materials depending on the application. Below are some materials used as bolus, all of which are applied directly to the skin surface.

"Superflab" vinyl plastic bolus in 5mm and 3mm thicknesses.

Edge view of Superflab sheets.

Wet towels can provide bolus with tissue like properties.

Petroleum jelly infused gauze pads are sometimes used as bolus used over open wounds.

Brass chainmail bolus for photons. The bolus of choice for C-3PO and Michael Jackson.

What's the strangest material you've seen used as bolus?