1663 Home Los Alamos National Laboratory
An illustration of red cells against a light purple background.

Cancer in the Crosshairs

By Eleanor Hutterer|August 01, 2020

High-energy protons could provide better imaging, better targeting, and better treatment.

The relationship between cancer and radiation is… complicated. Radiation can definitely cause cancer. But it can also cure cancer. And when used in medical imaging, radiation can help diagnose cancer. 

In medicine, “radiation” is a catch-all term used for several different types of electromagnetic waves (e.g., x-rays, gamma rays) or subatomic particles (e.g., protons, positrons) that are used for imaging or therapy. Of the particles presently used for cancer treatment, protons are at the cutting edge, and researchers at Los Alamos have recently reported several discoveries that could make proton-based cancer diagnosis and treatment even better.

What they’ve demonstrated amounts to improved imaging accuracy, tighter tumor targeting, and even the potential for earlier diagnosis. Los Alamos medical physicist Matt Freeman puts it succinctly, saying, “Our goal is to apply imaging techniques to make therapy better. This is where our work lies, at the place where imaging and therapy meet.”

Better targeting

The Los Alamos Neutron Science Center, or LANSCE, is home to a particle accelerator that produces protons with energies up to 800 megaelectronvolts (MeV). LANSCE has a sister facility of sorts in the GSI Helmholtz Centre for Heavy Ion Research, in Darmstadt, Germany. Scientists from the two centers collaborate frequently, and it was during a 2009 visit to Darmstadt that Los Alamos physicist Frank Merrill and his German colleagues had an idea while brainstorming over a beer. Merrill was interested in new applications for LANSCE’s high-energy protons, and the Germans were interested in new ways of treating tumors while they are still very small. As the scientists conversed, they realized their two interests just may be the answers to each other’s questions. High-energy protons might be the way to treat very small tumors. Prost! (German for “cheers!”)

If an animal cell’s DNA is damaged in such a way that the cell loses its ability to control its own life cycle and begins to divide rampantly, the result is cancer. But if a cancer cell’s DNA is damaged such that the cell can’t successfully divide at all, then the cell dies. Modern medicine has devised multiple ways to target cancer cells with DNA-damaging radiation. Most common is the injection or implantation of a radioactive substance or the transmission of a beam of radiation into the body from an outside source. With each of these methods, the goal is to do more damage to the cancer cells than to neighboring, non-cancer cells. It is in this way that protons are just about perfect. 

Other types of therapeutic beam radiation pass all the way through a human body, potentially damaging everything on their path to and from the tumor. But protons only travel a certain distance, based on their starting energy; they gradually lose energy along the way, then stop and dump their remaining energy rapidly as they come to rest. The energy the protons lose on the way in does cause some unwanted damage, but most of the damage is concentrated onto the tumor. As long as the proton energy is suitably controlled, there is no exit path and therefore no exit-path damage.

Because they are positively charged, protons scatter as they pass through an object—they get pushed around a little bit by each atom they encounter, being drawn toward negative charges and repelled by positive ones. This scatter is largest at the end of the trajectory, causing the energetic protons to laterally spread into the immediately surrounding tissue, which, if the aim is true, should be tumor tissue. 

A graph showing how deep protons can penetrate tissue. Tissue depth in cenimeters, from 0 to 35 in intervals of 5, is on the X-axis, and lateral spread of the protons in cenimeters for 800 MeV and 200 MeV with a range of -2 to 2 cm in intervals of 2, is on the Y-axis.
Tissue penetration power and dose deposition of high-energy protons (800 MeV) compared to what is presently used in proton-beam therapy (230 MeV). At 30 cm, the 230-MeV protons stop and dump their energy into about 1.5 cm of surrounding tissue. By contrast, the 800-MeV beam doesn’t stop at all and spreads minimally (0.3 cm) as it travels through the tissue.

Different materials have different stopping power—that is, how much energy they absorb from a proton per unit volume of material. Increasing the proton beam’s energy increases the beam’s penetration range. When the stopping power is known, the penetration depth can be precisely tuned by controlling the energy of the beam. For example, 230-MeV protons have a penetration range of 30 centimeters (cm) in soft human tissue. Protons for cancer therapy typically range from 230 to 330 MeV and are commonly produced by a cyclotron—a particle accelerator that is compact enough for clinical spaces. 

Although it’s ideal for protecting tissue from exit-path radiation, current proton therapy can’t target tumors smaller than about a centimeter in diameter. This is because of the way the protons spread laterally into the tissue when they come to a stop. Protons that start at 230 MeV dump their energy into about 1.5 cm of surrounding tissue. If the tumor is much smaller than that, too much healthy tissue gets the dose. 

Merrill and his German colleagues realized that high-energy protons could theoretically target tumors as small as 1.0 millimeter (mm). High-energy protons, like the ones at LANSCE, actually do pass all the way through the patient with relatively little scatter, taking most of their energy with them. This means minimal lateral spread, which means tighter tumor margins, which means less risk to surrounding tissue, which means smaller tumor targets. It would be like painting with a 0.3-cm paintbrush, compared to the current standard 1.5-cm paintbrush.

But the reduction in tumor size comes at a cost. Lower-energy protons don’t have exit paths, but high-energy protons do, which increases the collateral damage. Additionally, because the bulk of the protons’ energy remains with the protons through the tumor and out the other side, there needs to be some way of boosting the energy deposited in the tumor, relative to the entry and exit paths. The way to do this is to use a highly controlled beam of high-energy protons and rotate it around the patient during treatment, irradiating the tumor from 360 degrees. This dramatically dilutes the dose to entry- and exit-path tissues, while maximizing the dose to tumor tissue.

“Treating tumors with high-energy protons is a good idea, but it’s still theoretical, and Los Alamos isn’t exactly a cancer-treatment facility,” explains Laboratory physicist Michelle Espy, who works with Freeman and Merrill. “But we’re trying to think about what the clinics will want to be doing ten years from now. We’re developing the science now so that it’s ready when the clinicians need it.” 

When the accelerators in the clinical instruments are able to give more energy, the beams will be able to treat smaller tumors. And as the target sizes decrease, the need to see them precisely will increase. So, Freeman, Merrill, and Espy, along with their colleague Dale Tupa and students Ethan Aulwes, Rachel Sidebottom, and Brittany Broder have been working on that too.

Better imaging

Contemporary proton therapy is becoming more and more accessible, but concomitant improvements in imaging are needed. Treatment planning relies on careful imaging and measuring of the tumor target, to make a 3D map of intended hits and misses. But this is often done on one day, while the actual treatment is done over the course of weeks. Human bodies are squishy—they swell, shrink, and shift from one day to the next. If a target is very close to something crucial, and things have moved around a bit, the therapy could wind up doing more harm than good.

A digitally created image of a proton therapy machine showing the cyclotron, rotating gantry, magnetic lens, detector, and fixed beam line.
Proton therapy for treating cancer is typically done in one of two ways: either the patient lies immobile while the gantry-mounted proton beam rotates (center), or the proton beam is immobile while the patient is re-positioned (right). In either configuration, instantaneous beam’s-eye-view proton radiography could be enabled by adding magnetic lenses and detectors to existing clinical infrastructure.

The current state of the art for precise proton-beam targeting is to use x-ray imaging during, as well as before, the proton therapy. This gets around some of the squishy issues, but the x-rays are oriented at different angles to the proton beam, so even though it’s real-time, it’s still not a perfect view. The ideal would be to see the target as the therapy protons see it, from a “beam’s-eye view.” 

Freeman explains, “Beam’s-eye-view radiography can help prepare for a patient’s treatment because it provides a better map to more accurately constrain the dose, which helps spare nearby tissues. It can also be used to guide a dose in real time and correct for day-to-day or even minute-to-minute changes in patient anatomy.”

Conveniently, high-energy protons like the ones that could safely target mm-sized tumors can also be used to visualize the internal structures of things. The technology, called proton radiography, was discovered in the 1970s and has been developed and perfected at LANSCE for national security applications. Proton radiography, or pRad for short, produces an image of an object’s density by shooting protons through the object and then mapping the transmission of the protons to density variations in the material. 

Because the protons scatter as they pass through the object being imaged, without some kind of correction, the resulting image would be blurry. The thicker the object, the more scatter and the blurrier the image. And although the scatter is caused by the protons’ positive charge, so too is the solution.

“It’s just magnets!” exclaims Espy. “Because protons are charged, we can take all the scatter and bend it back using magnets to focus it, and we actually wind up with a very good image.” 

Magnetic focusing lenses cancel the would-be blur by focusing the protons onto a specific plane, similar to how optical lenses focus light onto a specific plane (e.g., a screen, a piece of film, or a retina). A series of magnetic quadrupoles surround the proton beam in a perpendicular orientation, each with two north and two south poles in alternating orientations. The magnetic field is zero at the center of the beam line and gets stronger towards the periphery, where more powerful steering is required to refocus the protons. The most highly scattered protons create blurring in the final image, so these get removed at an earlier collimation point, while the rest of the protons continue on to converge on the image plane.

An X-ray radiograph and a proton radiograph of an artifical head. Pink dots in the proton radiograph represent "noise."
An x-ray radiograph (top left) and a proton radiograph (bottom left) show a cross section through the neck of an artificial pediatric head, designed for medical imaging. (Right) A composite image. Although the proton radiograph portion is noisier, the anatomy is still easily identified with 800-MeV protons. With beam’s-eye-view imaging, the proton beam that produces the image would also deliver the treatment dose, improving accuracy and reducing radiologic exposure.

Proton radiography at Los Alamos is definitely not new—in fact, it’s a Laboratory specialty. But the idea of using pRad in tandem with high-energy proton-beam therapy for cancer treatment is new. And although the Laboratory isn’t the place to implement clinical beam’s-eye-view radiography, it is certainly the place to develop the science.

“We are evaluating lens-based pRad, of the sort done regularly at LANSCE, for guiding proton-beam therapy,” explains Freeman, “and we’re sharing our findings with the medical imaging and radiotherapy communities in the hope that they’ll invest in the needed infrastructure. When they do, they’ll simultaneously unlock new treatment and imaging capabilities.”

If combined with high-energy proton therapy, beam’s-eye-view pRad would not increase the overall radiation that a patient receives, and might even decrease it, because imaging and treatment could both be achieved with a single beam. But it doesn’t have to be combined with new therapy technology—beam’s-eye-view pRad could benefit proton therapy now. It’s actually better at pinpointing the stopping power than x-ray radiography is, so it would improve dose calculation and mapping.

 “This is certainly something clinics could implement,” says Espy. “Once you have the protons—which any proton-therapy clinic already does—you can add the optics to the existing system. Imaging would be nearly instantaneous.”

Seeing tiny tumors from a beams-eye view helps with accuracy, but the densities of tumors and healthy tissue are similar, and it can be hard to distinguish the two using a density-based imaging technique. So the team has been working on that too.

A dark-field proton radiography image that shows how this technique is useful for seeing the difference between tumors and healthy tissue.
Even state-of-the-art collimation for proton radiography of thin systems (top) cannot resolve, to any useful degree, a basket weave pattern in gold leaf. But dark-field proton radiography (bottom) clearly shows the basket-weave pattern by enhancing contrast between the layers. (Perpendicular black bars at left and right are for image orientation.)

Even better imaging, and a capability too

“The challenge of seeing a tumor is just that,” says Merrill. “It’s something that will keep scientists busy for a while.”

Because tumors and healthy tissue look similar to pRad, there has to be a way to tell them apart. There are two general approaches for this: either magnify their differences, or increase the sensitivity of the tool. In pursuit of beam’s-eye-view pRad, the Los Alamos team has done both.

Initially the scientists went down the path of magnifying differences. One difference between healthy cells and cancerous cells is the number of certain receptor proteins on their surfaces—often cancer cells have more. Gold nanoparticles can be tethered to molecules that specifically bind these proteins, so that when those molecules are introduced into the body, the tumor winds up effectively covered in gold. Gold is a good contrast agent because it’s not very toxic to cells and it decreases transmission relative to unlabeled tissue. Target cancer cells can take up about 100,000 nanoparticles per cell, which makes 1-mm tumors plainly visible to pRad.

But nanoparticle enhancement of contrast is difficult to do, and it depends on the tissue type and tumor location, so it isn’t always feasible. The team also wanted to increase the sensitivity of the imaging tool. “Dark-field microscopy” is an optical method of enhancing the contrast between two difficult-to-distinguish things by imaging with scattered particles (photons for light microscopy, electrons for electron microscopy), rather than unscattered particles. Dark-field imaging thus yields a plain black image when nothing is there. Dark-field pRad had previously been theorized but never demonstrated. The method would be very sensitive to minute differences, which is why the team wanted to try it.

The scientists were testing different collimation schemes for beam’s-eye-view pRad when they realized they were perfectly positioned to try and prove dark-field pRad. Prior attempts to demonstrate dark-field pRad had only used one collimator, but the trick, it turned out, was to use two collimators: the standard one to exclude the most highly scattered protons, which create blur, and another to exclude the least scattered protons. These bring noise to the image and carry comparatively little radiographic information, so excluding them increased the signal-to-noise ratio, which improved the image dramatically. 

The week before Christmas, 2019, the team was working to squeeze in one more run on LANSCE’s proton beam before leaving for the holidays. They were attempting to do dark-field pRad on a few pieces of gold leaf—less than a millionth of a meter thick and designed by Tupa to simulate a gold-labeled tumor. And lo and behold, it worked. They could see the millionth-of-a-meter gold leaves, arranged in a basket-weave pattern, way better than they had dared to hope. Prost again!

To the scientists, this is the most exciting of their recent discoveries. Whereas high-energy proton-beam therapy and beam’s-eye-view pRad are important proofs of principle, they are trees that will bear their fruit primarily outside of the Laboratory. But dark-field pRad represents a new institutional capability for Los Alamos. And it’s done, it’s here, it’s ready to use.

Because it’s so new, only just having been proven for the first time, the team is confident that the technique can be made to work even better. It will mainly be useful for visualizing very thin things and very subtle differences. Mission-centric applications include visualizing the evolution of high-explosive detonation products, the breakup of ejecta material, or the mixing of certain gases after being shocked. 

LANSCE is home to a pRad user facility, where scientists from around the world can come do pRad experiments. Now dark-field pRad will be available to those users too. Dark-field pRad is a lasting legacy for pRad, which itself is a lasting legacy for Los Alamos. 

These scientists have moved science forward on two fronts. First, they’ve fleshed out a three-pronged improvement to cancer treatment: High-energy proton-beam therapy could improve cancer prognoses by targeting tumors when they are still only millimeters across; beam’s-eye-view pRad can bring higher accuracy to treatment dose calculation and dose delivery; and dark-field pRad can aid in both imaging and treatment by better distinguishing tumors from healthy tissues. Second, and not to be outdone, they’ve developed an entirely new capability for the Laboratory and the proton radiography community at large. Prost, indeed! LDRD