Precise IQ Engine (PIQE) for MRI

Valentin H. Prevost, Ph.D.

In magnetic resonance imaging, image sharpness is one of the main criteria to allow delineation of fine structures that are important for clinical diagnosis. One way to improve the sharpness is to increase the acquired matrix size, but this is generally associated with a lower signal to noise ratio (SNR) or a longer scan time. Historically, some methods have been proposed to artificially increase the reconstructed matrix size, but this often leads to additional Gibbs ringing artifacts immediately adjacent to high-contrast interfaces.
Following the strong success of its Deep Learning-based denoising solution (Advanced intelligent Clear-IQ Engine: AiCE), Canon has recently introduced Precise IQ Engine (PIQE), a solution that enables increased resolution reconstructions. With the use of multiple Deep Learning algorithms, PIQE reduces noise from the images, increases the matrix size, and can potentially remove any Gibbs ringing artifacts.

PIQE can then triple the in-plane matrix size leading to a higher image sharpness, while maintaining a high SNR, and without altering the image quality. PIQE can be used at both 1.5 and 3T, without requiring acquisition adjustments, and can benefit several scenarios, such as imaging small structures or even fast imaging when patient status requires.

Two leading radiology experts involved in the product evaluation of PIQE share their feedback.

The first is Dr. Benoit Sauer, a Radiologist and the General Director of MIM (Medical Imaging) Group (based in Strasbourg, in France).

“Super Resolution means gaining spatial resolution with the same SNR and less artifacts (e.g. Gibbs ringing). This is especially needed in small structures, like ear, cartilage, or coronary artery, to propose a better diagnosis. PIQE is a promising solution that fits perfectly with these specifications.”
Dr. Benoit Sauer
Radiologist and General Director of MIM (Medical Imaging) Group (based in Strasbourg, France).

Could you comment on the value of PIQE from your recent experience?

“One of the big challenges with MRI is to achieve high image quality in a limited time. Historically, the MR community and companies focused their improvements on the optimization of sequences, but in my opinion, working on image reconstruction is the new trend to follow! AiCE has ‘paved the way’ forwards for Deep Learning Reconstructions (DLR), and PIQE is the next step. We tested it in our two Canon 1.5T systems and we observed important image quality improvements, both in terms of SNR and sharpness.”

Would you consider using PIQE on all your patients?

“Yes, but in different ways. In our clinical practice, we have several different types of protocols, depending on patient status and indications, and PIQE has been found useful in all of these cases. For example, in MSK imaging, gaining spatial resolution with the same scan time is what I’m looking for, especially for the evaluation of cartilage lesions or meniscal tears. For brain imaging in elderly or patients with severe comorbidities, we created faster protocols to be combined with PIQE that save precious scan time and avoid motion artifacts. Our philosophy is to offer personalized care and answer the clinical question, while maintaining optimal patient comfort.”

What does the term ‘Super Resolution’ mean to you, and could it impact your clinical activity?

“Super Resolution means gaining spatial resolution with the same SNR and less artifacts (e.g. Gibbs ringing). This is especially needed in small structures, like ear, cartilage, or coronary artery, to propose a better diagnosis. PIQE is a promising solution that fits perfectly with these specifications.

A second expert involved in this evaluation was Dr. Daniel Chow, Co-Director of the Center for Artificial Intelligence in Diagnostic Medicine and Assistant Professor-in-Residence at the Department of Radiological Sciences at the University of California, Irvine, USA.

What is your opinion on artificial intelligence and its place in medical imaging?

“Artificial intelligence has the potential to improve equity. 3T systems are generally more expensive and have a smaller bore size. Getting higher resolution images with 1.5T, while also gaining a larger bore size AND going faster have the potential to help us reach more patients.”

What were your thoughts when you reviewed the PIQE images?

"I initially thought the 1.5T images reconstructed with PIQE were generated from a 3T scanner, and I am very impressed by how far advances in PIQE technology has matured! PIQE clearly improved spatial resolution, which can make spotting things like metastasis or subtle lesions easier."

“I initially thought the 1.5T images reconstructed with PIQE were generated from a 3T scanner, and I am very impressed by how far advances in PIQE technology has matured! PIQE clearly improved spatial resolution, which can make spotting things like metastasis or subtle lesions easier.”
Dr. Daniel Chow
Co-Director of the Center for Artificial Intelligence in Diagnostic Medicine and Assistant Professor-in-Residence at the Department of Radiological Sciences at the University of California, Irvine, USA.