Bringing particle physics to patient care

Cancer care might not be the first thing that comes to mind when you think of particle physics, but it’s all in a day’s work for doctoral candidate Matteo Maspero.

Maspero focused his Master’s study in medical physics, developing a neutron detector for hospital use with Boron neutron capture radiotherapy.

“I was really into the development of the detector part, then I discovered the application of particle physics in the medical world.”

Maspero’s team is working to create reliable synthetic CT scans that offer all the benefits and none of the radiation.

Now, his work in the department of radiotherapy at UMC Utrecht involves magnetic resonance imaging in the context of prostate cancer treatment. Maspero’s current research on prostate cancer seeks to eliminate the need for a CT scan.

“What you want to know is where is the prostate such that you increase the radiation to the prostate and diminish the rest,” Maspero said. “But that’s not exactly clear from CT.”

MRIs, then, are inevitably needed after a CT scan to delineate soft tissues with more precision. This imaging double feature is the current standard, allowing physicians to see the tissue and plan how far the radiation will enter a patient’s body.

“But, you are asking the patient to undergo two imaging processes,” Maspero noted. “And you are acquiring an image in two different moments.”

It’s a setup that can lead to errors. Eliminating the need for a CT, Maspero explained, would cut out not only the double imaging but also the radiation.

However, MRI images cannot illustrate how far radiation will penetrate tissue. In collaboration with a scanner vendor, Maspero’s team is working to create reliable synthetic CT scans that offer all the benefits and none of the radiation.

In the name of science, Maspero himself undergoes an MRI a week.

“We do the MRI, we do our delineation of the prostate, and then we send everything into the software,” Maspero said. “It should also cost less because you are only doing one image.”

In the name of science, Maspero himself undergoes an MRI a week.

“If some colleagues need to optimize some protocols, who do you think they ask?” he said, laughing. “That’s how it goes, that’s research!”

Maspero’s research outside of the MRI machine involves working with a mix of databases, image processing systems, and a host of software.

“Our research is generally on a Unix system, but I use a bit of everything depending on what we need,” Maspero said. “SQL to retrieve patient data, a lot of patch scripts or terminals from Linux, Unix when we need to process a high amount of data…”

The list goes on to include Python, C, C++, Java, and MATLAB. On any given project, Maspero can work with up to 10 different programs.

His most recent MATLAB code utilizes a synthetic CT scan program.

״What we’re working on currently is deep learning. It seems to work too easily to be true!״

“The fake CT that comes out doesn’t have an internal error,” Maspero said. Internal errors, in the case of prostate cancer imaging, are usually gas in the rectum or bowel that can slightly affect the image. “We show in a paper that you need to do that. That’s why we developed our method to insert the error as you see it from the MRI.”

For a particle physicist, working with patient data presents its own unique challenges. Facing clinical problems involving both software and real patients and solving problems in unorthodox ways is part of why Maspero enjoys his work.

“Since last year AI has started to enter the world of MRI,” Maspero added. “What we’re working on currently is deep learning. It seems to work too easily to be true! The big challenge is to really understand what we are doing with these new techniques.”

״[Science] is for society. If we share, we go a step further all together. That’s what we are supposed to do.״

In a field that often collaborates with companies developing new techniques, software, and systems, keeping code open source isn’t always a priority. When the steps to creating a new, marketable product are not so far away, many choose not to divulge the details.

Maspero, however, feels that science should always strive to be as open a community as possible.

“I am definitely for open source. If you manage to get to a certain point, the rest of the world is not stupid — they will also reach your point,” he said. “[Science] is for society. If we share, we go a step further all together. That’s what we are supposed to do. It’s what I would like to do.”

When it comes to his own field, Maspero thinks open source tools have an additional, tangible benefit: validation before clinical use.

‍Matteo Maspero preparing MRI test.

‍“When you share something with the community, you get checked,” he said. “If [the code] is going to be used clinically, you need validation by other groups. That’s quite essential.”

His advice for his fellow programmers, both in his field and beyond, is simple: just do it.

“Try to take the perspective of another person when you’re making your own code,” Maspero said. “They might not know the jargon you’re using. If you find a new technique and use [that code], keep the reference also.”

Part of Maspero’s open source inclinations stems from the real world applications of his own work. When it comes to cancer care, time is of the essence. If a new technique can help save lives, then that’s a strong motivation to share it with the rest of the field.

“Usually when there is cancer involved, it’s easy to have someone in the family that had an experience — that was the case for me,” he said. “So that motivates you.”

He had a similar answer for what drives him to face so many challenges on a daily basis.

“I like to think that my work is not only for myself,” Maspero said after a thoughtful pause. “I like to think I have the patient in the back of my mind.”


Originally published at codeocean.com.