Successfully challenging the limits of conventional body and pelvic MR imaging

Successfully challenging the limitations of conventional body and pelvic MR imaging with advanced imaging techniques and deep learning reconstruction tools.

Choosing magnetic resonance imaging (MRI) to evaluate patients who need cross sectional abdominal and pelvic imaging presents unique challenges and opportunities. Compared to computed tomography (CT), which is also commonly used, MRI offers important advantages, including non-ionizing radiation, superior soft-tissue contrast, and a truer anatomic presentation of abdominal and pelvic masses using a variety of available contrast agents. The ubiquitous use of MR body and pelvic imaging, however, is hindered by its lower spatial resolution and motion artifacts which can affect image quality.  

Utaroh Motosugi, MD, PhD, Radiologist at Kofu Kyoritsu Hospital in Yamanashi, Japan, specializes in cases of hepatocellular carcinoma and pancreatic diseases. He utilizes MRI because of its many advantages over CT, despite its inherent limitations, which he has overcome by using new, deep learning reconstruction technology and iterative denoising processes. In a recent webinar hosted by GE Healthcare, Dr. Motosugi addresses specific advances in MRI techniques that challenge the limitations of image quality and motion artifacts and speaks about his experiences, illustrating clinical cases where using these techniques enabled comprehensive and accurate diagnoses for his patients.

Seeing the benefits of MRI and deep learning-based reconstruction

According to Dr. Motosugi, some of the advantages of MRI include: high contrast to noise ratio, the availability of many different contrasts, the ability to quantify fat and measure stiffness with MR elastography, as well as the ability to use a hepatobiliary contrast agent for diffusion weighted MRI to more clearly see liver metastases.

Dr. Motosugi acknowledges that two major challenges of using MRI have been limitations in signal-to-noise (SNR), and vulnerability of motion artifacts, which are usually due to patients’ inability to maintain breath hold for the duration of the image acquisition. Historically, there has been an inherent compromise between image quality and scan time. Better image quality could be achieved through higher signal-to-noise ratio (SNR), but the spatial resolution needed to show anatomical detail necessitated long scan times. These trade-offs made incremental improvements to address one or the other issue, but not both simultaneously, and not without sacrificing image quality or increasing acquisition time, which can be uncomfortable for the patient.

To challenge the issues in image quality, Dr. Motosugi is using a deep learning-based reconstruction engine, AIR™ Recon DL*, that makes full use of the acquisition’s raw data for maximum image quality. “As we know, typical spatial resolution for body MRI is around 10 cubic millimeters. This is not enough,” says Dr. Motosugi. “We are now getting clear images with smaller voxel size, more detailed structures and with much less noise compared to conventional images.”

In addition to a user-selectable SNR improvement, this advanced reconstruction technology employs an intelligent ringing suppression that preserves fine image details and achieves image sharpness and high resolution.

GE_RSNA2020_Content Graphcis_11.6.20.jpg

Case study with MRI deep learning-based reconstruction algorithm

To illustrate how he overcomes motion artifacts, Dr. Motosugi highlights a case of metastatic pancreatic cancer from renal cell carcinoma. Using compressed sensing for the 3D dynamic imaging, it allows him to acquire multiple arterial phases in a single breath hold while also improving the resolution of the image.  In the case mentioned, Dr. Motosugi was able to find an additional metastasis that he would not have seen using  a single phase or lower resolution technique. Dr. Motosugi also utilizes with a variety of  free breathing protocols that can be used for patients who are unable to hold their breath for a variety of reason (age, patient condition or pathology, cultural language), where the sequences track the patient’s breathing pattern without relying on the need for external trackers driving consistency in timing for arterial phases without the need to hold their breath creating a better patient experience overall.

Click HERE for on-demand access to Dr. Motosugi’s complete presentation and view his clinical case examples.

Click HERE to read about Dr. Motosugi’s first hand experience in our Spring edition of SIGNA™ Pulse of MR magazine. 


*Not yet CE marked for 1.5T. Not available for sale in all regions.