New technique uses AI, machine learning for image reconstruction

Researchers have developed a novel technique that reconstructs medical images using artificial intelligence (AI) and machine learning, according to a new study published in Nature. This saves radiologists valuable time and could potentially result in patients being exposed to lower radiation doses.

The technique, known as automated transform by manifold approximation (AUTOMAP), takes advantage of newly trained algorithms and large datasets. It also requires significant computing power.

“An essential part of the clinical imaging pipeline is image reconstruction, which transforms the raw data coming off the scanner into images for radiologists to evaluate,” author Bo Zhu, PhD, a research fellow at Massachusetts General Hospital’s Athinoula A. Martinos Center for Biomedical Imaging, said in a prepared statement. “The conventional approach to image reconstruction uses a chain of handcrafted signal processing modules that require expert manual parameter tuning and often are unable to handle imperfections of the raw data, such as noise. We introduce a new paradigm in which the correct image reconstruction algorithm is automatically determined by deep learning artificial intelligence.”

Zhu added that AUTOMAP uses AI to “teach” imaging systems to “see” in a specific way that helps radiologists work with the best possible images when making their evaluations. “This approach allows our imaging systems to automatically find the best computational strategies to produce clear, accurate images in a wide variety of imaging scenarios,” Zhu said in the statement.

Lead author Matt Rosen, PhD, director of the Low-field MRI and Hyperpolarized Media Laboratory and co-director of the Center for Machine Learning at the Athinoula A. Martinos Center for Biomedical Imaging, explained in the same statement how AUTOMAP leads to greater value-based care.

“Our AI approach is showing remarkable improvements in accuracy and noise reduction and thus can advance a wide range of applications,” he said. “We're incredibly excited to have the opportunity to roll this out into the clinical space where AUTOMAP can work together with inexpensive GPU-accelerated computers to improve clinical imaging and outcomes.”