CT Scans are Getting Faster, Safer and Smarter
CT scans are a crucial diagnostic tool in every hospital. A scan can quickly locate problems from a fracture to a tumor, resulting in massively improved medical outcomes.
But that doesn’t mean CT scans are entirely safe.Patients worry that a CT scan could lead to cancer, and they’re not wrong to worry about this. There is a risk in this process. But medical innovators are working to make the process safer.
How does a CT scan work?
A Computed Tomography scan, or CT scan, uses a sophisticated array of mobile x-ray tubes to obtain images from multiple angles. The CT software combines these images to create a “slice” – a cross-sectional representation of part of the patient’s body. This image reveals things that a traditional x-ray might miss, including tumors and lesions.
From the patient’s perspective, a CT scan means lying flat on a mechanized gurney that slides into the donut-shaped gantry of the CT machine. It’s reasonably comfortable, and it’s not particularly noisy – especially compared to the much louder MRI.
Sometimes, patients need to ingest a contrasting agent to help improve the quality of the CT scan. For example, gastric examinations require the patient to drink a barium solution. If it’s a cardiovascular scan, the patient might need an iodine-based agent delivered by IV.
CT scans are remarkably common. For every thousand people in the US, there are approximately 276 CT scans performed each year. That’s because this technique offers remarkable benefits, like eliminating the need for certain kinds of exploratory surgery.
But that’s not to say that CT scans couldn’t be better, or that current techniques don’t have any issues.
What are the current issues of CT scans?
There are some occasional difficulties in administering a CT scan. The patient has to lie still, which can be an issue for children or those with claustrophobia.
Some patients also have issues with the contrast required for imaging. Between 5% and 8% of patients have an allergic reaction to the dye, and these responses can be quite severe.
But the biggest concern is radiation. CT scans expose the patient to a certain amount of ionizing radiation that could potentially damage the patient’s DNA. With enough time and bad luck, that can lead to cancer.
It’s important to note that ionizing radiation is everywhere, from radon gas to cosmic rays left over from the Big Bang. This natural background radiation gives us a benchmark for measuring the dose in a typical CT scan.
A study by the Mayo Clinic looks at different types of CT scan and estimates how long it would take to accumulate an equivalent dose of typical background radiation:
- Sinus scan = 2 months of normal radiation
- Head scan = 8 months
- Chest scan = 2 years
- Abdomen and pelvis scan = 10 years
This means that if someone had four CT scans of their abdomen, they would receive 40 years’ worth of typical background radiation.
It’s hard to quantify the exact health risk of all this exposure. Ionizing radiation is measured in millisieverts (mSv), and most cancer research data looks at doses of 1000 mSv or more. A typical CT scan involves around 2-10 mSv.
But one thing is certain – the greater the dose of radiation, the bigger the risk. If we could reduce exposure, patients could have multiple CT scans over their lifetime without worrying about risk.
How are CT scans improving?
Innovators are working hard to reduce risk, either by minimizing radiation or by improving the CT scan process
Targeted beamlets to reduce radiation dosage
In our current technology, a CT scan uses an x-ray tube that emits a wide beam of ionizing radiation. The tube moves around and repeatedly emits the same energy until the radiographer has enough information.
A team at UCL has devised an innovative solution to scan using less radiation. Their model involves placing a radiation-blocking mask over the x-ray tube. This mask contains several small slits, which emit focused beams – or beamlets – of radiation. With careful control, the team can move these beamlets in a cyclonic motion pattern, ensuring that the entire area is irradiated.
Beamlets are far less radioactive than x-ray beams. And because they’re so thin, the controller can focus them on the precise area they want to scan. The result is a sharper image, with a much lower radiation dose for the patient.
It’s a technique that’s still at the development stages, although the UCL team have published a well-received paper in Physical Review Applied. In time, this technique may reduce radiation doses by several orders of magnitude, allowing for more frequent scans.
AI can do more with less
CT scans work by taking a number of individual x-rays and compositing them into a detailed image. The more x-rays you take, the better the image quality. This is why some scans involve a much higher dose of radiation – you have to take lots of images to get sufficient quality.
But what if there was a way of taking fewer images and using digital tools to fill in the blanks? This is exactly the approach taken by a research team at Rensselaer Polytechnic Institute. Their model uses a sophisticated machine learning framework to analyze CT images, filter out the noise, and create a high-res result.
Machine learning is a branch of artificial intelligence that uses statistical analysis to find deep-lying patterns. Essentially, a machine learning algorithm will run millions of tests on huge sets of data, and “learn” from the results.
In the past, any attempt to artificially sharpen a CT image has failed, because it’s hard to guess what’s noise and what’s an important anomaly. However, the Rensselaer team has reported encouraging results, and their AI-powered image enhancement shows a promising rate of positive results.
The result is a system that allows CT operators to do more with less. They can take a limited number of images and use AI to fill in the blanks. This makes the process quicker and cheaper, while reducing the patient’s radiation dose.
Better sharing and storage with blockchain
CT scans, like other high-tech imaging systems, can be quite data intensive. Some estimates suggest that the US will produce 35 Zettabytes of imaging data in 2020 alone. Storing this much information is a challenge. Sending it to another location is an even bigger challenge.
Without the right infrastructure in place, some clinicians may find that they don’t have access to older scans. This leaves them with one option: order a new scan, and expose the patient to more radiation.
One solution is blockchain, the technology associated with cryptocurrencies like Bitcoin. Blockchain offers encrypted, decentralized storage of any kind of data, including medical information. By hosting imaging data on blockchain, healthcare providers could overcome any issues with storing or sharing data.
There are several companies already offering blockchain solutions for CT scans. DeepRadiology is a radiology specialist that has developed an AI-powered blockchain solution to power collaboration. Medical Diagnostic Web, meanwhile, is using blockchain to create a distributed marketplace of imaging specialists. Other companies are working to make blockchain-based medical imaging a part of everyday life.
Nanoparticles offer better contrasts
Contrast is an essential part of the CT scan process, especially in gastrointestinal and cardiovascular scans. However, there are two big problems: allergic reactions; and the patient can pass the contrast as urine quite quickly.
One alternative is to use nanoparticles as contrast. These are infinitesimally small particles of heavy metals that react to x-rays. Nanoparticles rarely cause an allergic reaction, plus they tend to persist in the bloodstream for longer.
Some studies have shown exciting results for nanoparticles made of gold, iron oxide, and tiny semiconductor crystals known as quantum dots. The next challenge is to figure out how to produce nanoparticles at scale.
Another advantage of the nanoparticle approach is that you can perform multiple CT scans at once. Because each particle has a different density, they’re easy to distinguish in CT images. So, a radiographer could give the patient two doses of contrasts with different nanoparticles in each, then perform a single scan. That would eliminate the need to perform two separate scans.
What is the future of medical imaging?
CT scans are one of the most popular diagnostic tools in the world, and nothing is likely to challenge that title in the near future.
Instead, imaging is going to focus on working smarter and faster. AI and blockchain will play a huge role. In 2020, during the height of the pandemic, Chinese researchers developed an AI that could detect Covid-19 from a chest scan. This allowed them to move faster on positive results, administering faster treatment and isolating from other patients.
Ultimately, the goal is to build systems that can tell a lot about a patient with relatively little imaging. That means systems that are faster and more efficient. It also means that patients will worry a lot less about the risk of radiation.