Format

Send to

Choose Destination
Magn Reson Med. 2017 Sep;78(3):897-908. doi: 10.1002/mrm.26467. Epub 2016 Oct 13.

Hybrid MRI-Ultrasound acquisitions, and scannerless real-time imaging.

Author information

1
Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA.
2
The Laboratory for Imagery, Vision and Artificial Intelligence, École de Technologie Supérieure, Montréal, QC, Canada.
3
Department of Physics, Soochow University, Taipei, Taiwan.
4
Google Inc, New York, New York, USA.

Abstract

PURPOSE:

To combine MRI, ultrasound, and computer science methodologies toward generating MRI contrast at the high frame rates of ultrasound, inside and even outside the MRI bore.

METHODS:

A small transducer, held onto the abdomen with an adhesive bandage, collected ultrasound signals during MRI. Based on these ultrasound signals and their correlations with MRI, a machine-learning algorithm created synthetic MR images at frame rates up to 100 per second. In one particular implementation, volunteers were taken out of the MRI bore with the ultrasound sensor still in place, and MR images were generated on the basis of ultrasound signal and learned correlations alone in a "scannerless" manner.

RESULTS:

Hybrid ultrasound-MRI data were acquired in eight separate imaging sessions. Locations of liver features, in synthetic images, were compared with those from acquired images: The mean error was 1.0 pixel (2.1 mm), with best case 0.4 and worst case 4.1 pixels (in the presence of heavy coughing). For results from outside the bore, qualitative validation involved optically tracked ultrasound imaging with/without coughing.

CONCLUSION:

The proposed setup can generate an accurate stream of high-speed MR images, up to 100 frames per second, inside or even outside the MR bore. Magn Reson Med 78:897-908, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

KEYWORDS:

MR-ultrasound imaging; hybrid imaging; image-guided therapy; machine learning; motion tracking

PMID:
27739101
PMCID:
PMC5391319
DOI:
10.1002/mrm.26467
[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Wiley Icon for PubMed Central
Loading ...
Support Center