SubtleSYNTH™

AI-powered synthesis for faster MRI

SubtleSYNTH™ uses deep learning (DL) to generate synthetic STIR images (SynthSTIR’s) from already-acquired T1 and T2 weighted contrasts. It is complementary to Subtle’s FDA-cleared SubtleMR™ software, which enables accelerated image acquisition of up to 60% for MRI procedures on any vendor, make or model of scanner.  SubtleSYNTH creates synthetic STIR images with zero acquisition time that are interchangeable with conventionally acquired STIR images.

It was validated for use in the spine using a demographically diverse study recently published in AJNR.  The first commercial release will support spine imaging, and future releases will support other anatomies.

See images below.

Improve efficiency gains with 100% acceleration

Add an additional boost in scan time reduction on top of SubtleMR™

Eliminate one common sequence from the protocol

Improve image quality of STIR

  • GE
  • Siemens
  • Hitachi
  • Toshiba
  • Philips

Normal - Philips 3T - C-spine

Non-cord lesion - Philips 1.5T - C-spine

Trauma - Toshiba 1.5T - C-spine

Non-cord lesion - Hitachi 0.3T - C-spine

Degenerative - Hitachi 0.3T - C-spine

Degenerative - Siemens 3T - C-spine

Degenerative - Siemens 1.5T - C-spine

Infection - Siemens 1.5T - C-spine

Trauma - Siemens 1.5T - C-spine

Trauma - Siemens 1.5T - C-spine

Degenerative - Siemens 1.5T - C-spine

Normal - GE 1.5T - C-spine

Vascular - GE 1.5T - C-spine

Hemorrhage - GE 1.5T - C-spine

Trauma - GE 1.5T - C-spine

Trauma - GE 1.5T - C-spine

Vascular - GE 1.5T - C-spine

Cord lesion - GE 3T - C-spine

Trauma - GE 1.5T - C-spine

Normal - GE 3T - C-spine

“The diagnostic quality seen with SubtleSYNTH’s state-of-the-art technology has shown to be clinically equivalent to the normally acquired STIR images in our research. By adding novel products like SubtleSYNTH to their AI portfolio, centers will be able to boost the efficiency of their MRI exams, resulting in less time for patients in the scanner and ultimately better patient care.”

Lawrence N. Tanenbaum, MD, FACR Chief Technology Officer, RadNet