Around 80 million computed tomography (CT) scans are performed in the US each year. They allow health-care providers to noninvasively visualize internal structures such as bones, brains, and tumors with high resolution and in three dimensions. But building a three-dimensional image entails delivering an x-ray radiation dose that amounts to a few years’ worth of natural background radiation, typically 100–1000 times that from a standard 2D image. Exposure to that much additional ionizing radiation can increase the patient’s chance of eventually getting cancer.
A new technique developed by Charlotte Hagen and coworkers at University College London could lower the risk that CT scans pose to patients. Their cycloidal computed tomography (CCT) procedure uses novel scanning and data acquisition protocols to expose subjects to less radiation without compromising the scan’s resolution. It’s also simple enough that it could be integrated into existing hardware.
During a normal CT scan, an x-ray beam is rotated in angular steps around the subject, and the transmitted x rays are collected by a detector. The researchers placed a mask with 10 μm slits in front of the beam, which shielded the subject so it was only exposed to small x-ray beamlets. In theory, that shielding could reduce the delivered radiation dose. But less signal means less information. To obtain high-resolution images with beamlets, the scan had to advance in smaller lateral increments in addition to angular ones, a process known as dithering. All that starting and stopping also adds lag time during which the scan’s subject is still being exposed to radiation. In practice, dithering increases scan times and counteracts the radiation reduction from masking.
In their CCT procedure, Hagen and coworkers kept the mask but avoided dithering by translating and rotating the subject at the same time. (For scans of inanimate objects, the subject is moved; in a clinical setting, the beam is moved to avoid rotating the patient.) The first figure illustrates how that procedure changed their data acquisition. With a mask and dithering, the scan collects information at every point shown. Using a mask but scanning normally—rotation only—generates data at just a subset of those points, shown as black dots in the left plot, and produces a blurry picture. But by scanning normally while translating and rotating the subject simultaneously, the researchers distributed the sparse data more evenly. The distribution made it easier to interpolate into a clear image whose resolution rivaled that achieved with dithering. The second figure shows the image quality achieved by each technique, illustrated using a rabbit esophagus.
A comparison of signal-to-noise ratios in experimental scans with different x-ray doses indicates that CCT can achieve the same resolution as dithering with about 50% as much radiation; simulations suggest that the number could be as low as 20%. And with the full radiation dose, both experiments and simulations show that applying CCT can produce higher-resolution scans. (C. K. Hagen et al., Phys. Rev. Appl. 14, 014069, 2020.)
